ABSTRACT

The effect of the preparation method on the properties of LaMnO₃ and La_0.8Sr_0.2MnO₃ perovskite was studied. Materials were prepared by four methods: sol-gel, chemical combustion, solvothermal and spray pyrolysis and characterized. The effect of the synthesis method on the texture, acid-base character of the surface, reducibility with hydrogen, oxygen desorption, surface composition and catalytic activity for combustion of lean methane was studied. It was found that synthesis method affects physicochemical properties of obtained materials-solvothermally produced materials exhibit well-developed surface area, presence of reactive oxygen species on surface and high catalytic activity for CH₄ combustion. Generally, LaMnO₃ and La_0.8Sr_0.2MnO₃ perovskites show catalytic activity for lean CH₄ combustion comparable or higher than the activity of 0.5 wt.% Pt/Al₂O₃ but lower than 1 wt.% Pd/Al₂O₃.

Keywords:

Perovskite, Different Synthesis Method, Lean Methane Combustion, Combustion Rate, Ventilation Air Methane

1. Introduction

Huge amounts of methane are emitted into atmosphere every year from numerous sources like coal mines, animal waste or landfill. Concentration of methane in the atmosphere has been increasing strongly since beginning of XX century. Methane is a greenhouse gas showing 23 times higher global-warming potential than CO₂ [1] . Emission of methane creates not only threat to the environment but also it is a waste of energy. Ventilation air methane (VAM) represents the largest part of methane emissions from coal mines; in addition concentration of methane in VAM is low and variable. Annually, up to 350 bln m³ of methane is released from Polish coal mines and only a part of this gas is used, the majority goes into the atmosphere as VAM with a methane concentration of 0.1% - 1.0% [2] . Conventional flame burners possess insufficient efficiency to oxidize diluted methane in VAM and catalytic oxidation is often proposed for this purpose. However methane is more difficult to oxidize than most of the other hydrocarbons. Catalytic oxidation of methane needs higher temperature than in the case of VOC oxidation therefore not only catalyst activity but also its thermal stability is a crucial problem. Various catalytic systems were proposed for methane oxidation. Noble metals are the most active catalysts however they are susceptible to sintering and expensive as well [3] [4] . Metal oxides seem to be more suitable for such applications than noble metals because they combine low volatility and reactivity with relatively high activity in oxidation reactions and low price [5] .

Crystal structure of ABO₃ mixed metal oxides with the perovskite like structure is very flexible towards substitutions and thus allows modifying their physicochemical properties including catalytic activity. Perovskites have been the subject of many studies regarding use their as high-temperature oxidation catalysts [5] [6] . Among different perovskite compositions studied as catalysts of methane oxidation, LaMnO₃ appeared to be especially effective in this reaction [7] . Catalytic performances of perovskites are strongly affected by the mobility of oxygen species [8] [9] [10] . Simultaneously, it is commonly known that the method of synthesis strongly affects catalytic performance of the perovskite-type oxides [11] [12] . The amorphous citrate method belongs to the most commonly used in the perovskites synthesis; it is simple and enables one to obtain perovskites at relatively low temperatures. L. Marchetti and L. Forni [6] have found that specific activity of La-Ma based perovskites prepared by amorphous citrate complex method decreases monotonically with increasing temperature of maximum of TPD-O₂ desorption peak. They also stated that two kinds of perovskite related oxygen are active in methane combustion: adsorbed oxygen species active at low temperatures and lattice oxygen species active at high temperatures [6] . In comparison, flame-spray pyrolysis leads to perovskites with small crystal size and good thermal stability, but the apparatus needed for synthesis is rather complicated. Catalytic activity of perovskites prepared by this method increase with increasing surface area of material [13] .

The most important advantages of the solvothermal method of perovskites synthesis are high quality and purity of the products, short reaction time, low dispersion of grain size and no need to final high temperature calcination of product. The main disadvantage is a relatively high cost. F. Teng et al. [14] prepared La_0.5Sr_0.5MnO₃ single-crystal cubes and nanoparticles by hydrothermal and citrate routes, respectively. They have found that after running at 600˚C for 48 h under reaction conditions, surface area of nanoparticles decreased significantly while cubes nearly maintained their surface area and latter show higher activity for methane combustion than nanoparticles [14] .

H. Najjar et al. [15] using LaMnO₃ prepared by chemical combustion for methane combustion found that the highest catalytic activity shows perovskite obtained with the use of fuel described by glycine-to-nitrate ratio of 0.8. This material exhibits the highest superficial concentration of manganese and the highest amount of active oxygen [15] .

A number of other methods of perovskites synthesis has been elaborated, these ones mentioned above belongs to the most commonly used. Ch. Zhang et al. [16] studied the effect of the method of LaMnO₃ synthesis on its catalytic performance for toluene oxidation. They have prepared this material by citrate sol-gel, glycine combustion and co-precipitation. Perovskite prepared by sol-gel method showed the best catalytic activity and durability. This superior catalytic performance was attributed to its higher specific surface area, better low temperature reducibility and more available surface adsorbed oxygen species. The lowest apparent activation energy of this material was in good agreement with its high catalytic activity for toluene oxidation [16] .

This paper reports results of comparative investigation of the physicochemical properties of the LaMnO₃ and La_0.8Sr_0.2MnO₃ perovskites prepared by different synthesis routes (spray pyrolysis, sol-gel with citric acid, solvothermal and chemical combustion). The goal of the work was to determine effect of the synthesis procedure on the properties and catalytic performances for lean methane oxidation of the LaMnO₃ and La_0.8Sr_0.2MnO₃ perovskites.

2. Experimental

2.1. Samples Preparation

Mixed oxides with general formula LaMnO₃ (LM) and La_0.8Sr_0.2MnO₃ (LSM) were prepared using following methods of synthesis: sol-gel (SG), chemical combustion (CC), microwave-assisted solvothermal (MS) and spray-pyrolysis (SP) method. The details of preparation are as follows:

Sol-gel method: Nitrates of corresponding metals (molar ratio 1:1 for LM and 0.8:0.2:1 for LSM) were dissolved in demineralised water acidified by HNO₃. To the homogeneous solution citric acid was added (molar ratio metal/citric acid = 1:2.5). The solution was then evaporated at 80˚C to produce viscous syrup, which was dried for 24 h at 110˚C and then grinded and calcined by using 3 steps during calcination: 4 h/180˚C, 1 h/300˚C and 4 h/750˚C.

Chemical combustion method: This method is based on the highly exothermic reaction between the metal salts and organic fuel―here glucose was used. It is self-sufficient energy synthesis because the heat required to the synthesis are provided by the reaction itself. Nitrates of corresponding metals (in molar ratio 1:1 for LM and 0.8:0.2:1 for LSM) were dissolved in a small amount of demineralized water. Glucose was then added to the obtained solution, a mass of glucose was equal to the mass of all salts. The resulting solution was concentrated in a rotary evaporator at 80˚C to produce viscous syrup, which was dried for 24 h at 110˚C and then calcined for 6 h at 750˚C.

Microwave-assisted solvothermal method: nitrates of corresponding metals were dissolved in ethylene glycol and reacted in a teflon melting pot in an autoclave (90 min, 200˚C, 40 bar) under microwave heating. The resulting suspensions were centrifuged, washed with acetone, dried (24 h/110˚C) and then calcined (4 h/750˚C, 10˚C/min).

Spray-pyrolysis method: The hydrated nitrates of corresponding metal were dissolved in distilled water (0.45 mol/dm³). The resulting solution was filtered and then sequentially pumped into the chamber system by a peristaltic pump. Powder synthesis reaction was carried out under the following conditions: solution flow rate 30 ml/min with performance of 0.1137 dm³/min, pressure 13.8 bar, the temperature 500˚C. The resulting powder was filtered through a cotton filter, dried 24 h/110˚C and then calcined 4 h at 750˚C (10˚C/min).

2.2. Characterization Techniques

The prepared mixed oxides were characterized by means of: X-ray diffraction (XRD), specific surface area determination, scanning electron microscopy (SEM), temperature programmed reduction with hydrogen (TPR H₂), decomposition of cyclohexanol (CHOL), temperature programmed desorption of CO₂ (TPD CO₂), temperature programmed desorption of oxygen (TPD O₂), X-ray photoelectron spectroscopy (XPS) and tests of catalytic activity for lean methane combustion.

By using an X'Pert Powder Diffractometer (PANalytical), working in the reflection geometry and using CuKα radiation (λ = 1.54056 Ǻ) X-ray diffraction (XRD) patterns were recorded. The data were collected in the 2Θ range 10˚ - 80˚ (step of 0.026˚). The fitting of the diffraction peaks was done by WinPLOTR programme [17] . The lattice parameters were calculated by DICVOL programme implemented in FullProf Suite package. The average size of crystallites was calculated based on Scherrer method (for line at about 2Θ ≈ 46˚ and with parameter K = 1), as described in [18] [19] .

The specific surface area was determined by the BET method (Brunauer-Emmett-Teller) using the Quantachrome Autosorb-1C. The BET surface was calculated from the N₂ adsorption isotherm at −196˚C (77 K), in the range of relative pressures p/p_o from 0.05 to 0.35. Prior to the measurements, samples were degassed at 200˚C for 2 h to remove adsorbed contaminants such as water.

The morphology of the samples was determined using a scanning electron microscope SEM/Hitachi S-3400N while elemental composition of selected sites on their surface were determined using energy-dispersive X-ray spectroscopy (EDS).

The susceptibility of the prepared materials to reduction was determined by temperature programmed reduction with hydrogen (TPR H₂). TPR H₂ measurements were made using Pulse ChemiSorb 2705 apparatus. The 50 mg of sample was placed in a U-tube quartz reactor and then heated to 900˚C (heating rate of 10˚C/min). Through the catalyst bed was passed the mixture of 5% vol. of H₂ in Ar (30 ml/min). Hydrogen consumption was measured by thermal conductivity detector (TCD).

The share of acid-base centers on the surface of tested materials were determined by decomposition of cyclohexanol (CHOL), determining the selectivity of the conversion of this compound to dehydration and dehydrogenation products. The dehydration of CHOL occurs on acid sites and leads to cyclohexene (CHEN), while the dehydrogenation of CHOL, leading to cyclohexanone (CHON), requires the interaction both acid and basic centers. It is accepted, that selectivity of CHOL decomposition to CHEN (S_CHEN) is a direct measure of acidic character of material surface, while the CHON/CHEN selectivities ratio (S_CHON/S_CHEN) gives information about basic nature of surface [1] . The 0.1 g of sample was diluted within silicon carbide to a volume of 1 cm³ and then placed in a quartz reactor between two layers of quartz wool. The reactor was moved into the furnace chamber and heated to 300˚C with a rate of 15˚C/min in a carrier gas stream (N₂, 10 dm³/h). Then the reactor was purged with nitrogen saturated by cyclohexanol (10 dm³/h). Products of CHOL decomposition were analyzed by gas chromatograph equipped with flame ionization detector (FID).

Measurements of temperature programmed desorption of oxygen (TPD O₂) was performed by using a Micromeritics AutoChem II 2920 analyzer equipped with a TCD detector. The 100 mg of samples were placed in a quartz reactor, then moved into the furnace chamber and heated in a stream of O₂ from 25˚C to 500˚C for 1 h. The reactor was cooled to 40˚C and a stream of inert gas was passed through the catalyst bed (He, 30 ml/min) for 0.5h. Next the samples were heated, in the stream of He, from 40˚C to 950˚C (10˚C/min.). The content of O₂ in the effluent gas was determined by a mass spectrometer (OmniStar QMS 200, Pfeiffer Vacuum) while monitoring the m/e signal 32 (O₂).

The surface composition of the prepared materials was determined by X-ray photoelectron spectroscopy (XPS), using a SPECS XPS/UHV system equipped with a PHOIBOS 100 spectrometer. The X-ray source was an Mg anode operating at 100 W (survey scan) and 250 W (high resolution spectra), Ar(+) sputtering (90'', 7 μA/cm², 3 keV). The analyzer mode was set at constant analyzer energy mode. Sample charging was compensated using an electron flood at 0.5 mA current and 0.1 eV energy. The C1s peak of the contamination carbon, at 284.6 eV, was taken as reference. The detection angle was normal to the surface. The spectra were collected and processed using SPECLAB software. Non-linear least-squares fitting algorithm was performed using peaks with a mix of Gaussian-Lorenzian shape and a Shirley baseline.

Catalytic activity in lean CH₄ combustion was tested in a continuous-flow fixed bed quartz reactor, placed in a tube furnace with a single heating zone. The temperature was measured with a thermocouple positioned a few millimeters below the top of the catalyst bed. Temperature of reactor was raised (5˚C/min) from 25˚C to 550˚C. The 200 mg of catalyst diluted in 200 mg of SiC was loaded into a tubular quartz reactor on a thin layer of quartz wool. The reagents mixture (0.6 vol% of CH₄ and 21 vol% of O₂ in argon) was fed to the reactor with flow rate corresponding to GHSV = 40,000 h⁻¹. The gas mixture flow rate was 8 dm³/h and was adjusted by mass flow controllers. The reaction products were analyzed by GC (PERKIN ELMER Autosystem XL with capillary column (30 m × 0.25 mm × 0.25 µm), equipped with FID detector. The reaction rate per m² ( $r_{SSA}$ ) of the catalyst was calculated using formula:

$r_{SSA}=\frac{F_{{\text{CH}}_4}}{SSA}\times \alpha \times {10}^6$

where $F_{{\text{CH}}_4}$ is the flow rate of methane ( $\text{mol}\cdot {\text{s}}^{-1}$ ), a is methane conversion and SSA is specific surface area.

3. Results and Discussion

Regardless of the method of synthesis used, the X-ray diffraction patterns of all materials are similar to each other and characteristic for perovskite-like structure (PDF card No. 00-032-0484 for LM and PDF card No. 00-040-1100 for LSM), as can be seen from Figure 1. All LM prepared materials crystallized in a trigonal system in crystallographic space group R-3c (the cell parameters of a = b ~ 5.5 Å) (Table 1). The LSM samples crystallized in regular system in the space group Pm-3m (cell parameters a = b = c ~ 3.9 Å), except of LSM-CC, which crystallized in a trigonal system (Table 1). Diffractograms of synthesized materials show no phase contaminants, only the XRD pattern of LSM-SP, besides reflections coming from La_0.8Sr_0.2MnO₃, shows a weak lines characteristic for La₂O₃ (2Θ = 30, 46,

* - For trigonal system a = b ≠ c ** - For regular system a = b = c *** - The angle between the axes.

67˚, PDF card No. 00-005-0602) (Figure 1(b)). X-ray diffraction patterns of LM-MS differs from XRD patterns of LSM-MS. The presence of strontium results in a less intense peaks, which translates into smaller particle size. An average size of crystallites are gathered in the Table 1 and were determined from XRD and SSA. The LM and LSM prepared by sol-gel, chemical combustion and solvothermal method show a similar crystallites size, taking into account the crystallites from XRD, comprising in the range of 22 - 37 nm. The average size of crystallites of materials prepared by spray pyrolysis do not lie within this range and amounts to 16 nm for LSM and 44 nm for LM (Table 1). An average crystal sizes were also calculated based on the values of specific surface area and density of LM (LSM) equal to 6.57 g/cm³ (Table 1). Estimated in this way corresponding values of crystal size are bigger than those calculated with Scherrer equation. It means that part of the crystallites surface is inaccessible for adsorption/reaction.

All studied LM and LSM mixed oxides show relatively low specific surface area (Table 2) that depends on the method of their synthesis. Among the LM samples, the largest SSA shows the one synthesized via solvothermal route, while

among the LSM materials, this one obtained by the sol-gel method. Both samples prepared by spray pyrolysis are characterized by the smallest value of SSA. The partial substitution of lanthanum by strontium, for all of used methods of synthesis, leads to a material with a higher specific surface area.

Acid-base properties of the studied materials were determined by the method of cyclohexanol (CHOL) decomposition. On the surface of the mixed metal oxides with perovskite like structure nucleophilic oxide ions O²⁻ are present (basic sites of a Lewis type) [20] and the electrophilic ions O⁻, ${\text{O}}_2^{2-}$ , ${\text{O}}_2^{-}$ [21] . The ratio (selectivity to CHON)/(selectivity to CHEN) denoted here as a Y is considered as a measure of the share of basic centers [22] . Samples LM-CC, LM-MS and LSM-SG show the highest share of basic sites (Table 2), respectively 0.6 (LM-CC and LM-MS) and 1.0 (LSM-SG). Acidic sites prevail on the surface of both materials prepared by SP as it is evidenced by the highest value of selectivity to CHEN: 84% for LM and 88% for LSM (Table 2).

The morphology of the studied perovskites was characterized by scanning electron microscopy (SEM) technique.

Typical SEM pictures of LM surface are presented in Figure 2. According to literature substitution of La by Sr doesn’t significantly affects the morphology of the material [23] . The morphology of LM perovskites depends on the method of their synthesis. LM-SG and LM-CC create agglomerates. Material synthesized by CC is more porous and particles are terminated with sharp edges. In contrast, LM-MS consists of fine, solid crystallites with regular shapes, between which free spaces are visible. The morphology of the LM-SP differs markedly from the other materials, because it looks like sintered scales with low porosity without visible crystallites. As it was pointed above, calculated from SSA crystal size of both materials prepared by MS method are bigger than in case of corresponding materials prepared by other methods.

The TPR H₂ profiles of studied materials are shown in Figure 3 and the temperatures of maximum of hydrogen consumption are summarized in Table 3.

The TPR H₂ profiles of Mn-containing materials show two intensive peaks of hydrogen consumption (Figure 3). First peak, with maximum of hydrogen consumption at the temperature range 343˚C - 459˚C (T_I), represents the reduction of Mn⁴⁺ to Mn³⁺ while the second one (maximum intensity in the range 716˚C - 849˚C (T_II)) is ascribed to the reduction of Mn³⁺ to Mn²⁺ [5] [16] . The presence of Mn⁴⁺ is related to the fact that the Mn³⁺ has a free electron, which making Mn³⁺ ion capable to adsorption of molecular oxygen and its transformation into an electrophilic forms, which are very active in the oxidation reactions [24] :

${\text{Mn}}^{\text{3}+}+{\text{O}}_{\text{2}}\to {\text{O}}_2^{-}/{\text{O}}^{-}+{\text{Mn}}^{\text{4}+}$

The shapes of TPR H₂ profiles of all LM and LSM materials are similar, regardless of the method of synthesis. Nevertheless, the influence of the synthesis method on the temperature of maximum of H₂ consumption is observed. The temperatures of reduction T_I and T_II of LM perovskites lower in the following order of synthesis method: SP > SG > CC > MS. LM-SP exhibits the highest value of T_I (397˚C) and T_II (849˚C) (Table 3)―spray pyrolysis leads to a material characterized by a low susceptibility to reduction, what probably results from a significant sintering of this material. On the other hand LM-MS, that exhibits the smallest size of crystallites, reacts with hydrogen at the lowest temperatures: T_I - 343˚C and T_II - 792˚C.

The TPD O₂ profiles of studied materials are shown in Figure 4. Amount of oxygen released from them are summarized in Table 4. The investigated materials release oxygen in the temperature range 200˚C - 900˚C. TPD O₂ profiles of all samples possess similar shape and show three regions of oxygen desorption. Based on the data presented in Table 4 and Figure 4 one can conclude that synthesis method affects the amount of oxygen desorbed from the surface of corresponding material. The highest amount of oxygen (per gram) release materials prepared by MS method, while the lowest these ones prepared by SP; these differences are smaller when amount of desorbed oxygen is calculated per m² of SSA. The amount of oxygen released from the surface of studied here materials are close to the values reported in the literature. M. Alifanti et al. [25] reported 267.3 µmol O₂/g while Rosso et al. found 263 µmol O₂/g [26] desorbed from the surface of LaMnO₃ prepared respectively by citrate and the modified citrate method. Other researchers reported the 262.3 µmol O₂/g and 437.8 µmol O₂/g (total amount) desorbed from LaMnO₃ and La_0.6Sr_0.4MnO₃ prepared by citric acid complexing coupled with hydrothermal method [27] .

Depending on the desorption temperature, three forms of oxygen are distinguished on the perovskite surface [28] . Desorption at temperatures up to ca

400˚C is associated with the surface oxygens, described as α₁-oxygen, which are commonly attributed to weakly chemisorbed oxygen molecules upon surface-oxygen vacancies. The oxygen described as α₂ is a near-surface oxygen associated with lattice defects such as dislocations and grains frontiers. This form of oxygen desorbs at temperatures from the range 400˚C - 700˚C and is completely associated with the oxidation of diluted methane, together with α₁-oxygen [29] . These both kinds of oxygen are designed as α-O₂ that generally desorbs below 700˚C. It is known, that the desorption temperature of α-oxygen strongly depends on the calcination temperature and it is supposed that at the temperature of desorption, α-oxygen species (eg. ${\text{O}}_2^{-}$ ) dissociate into atomic species [29] . Oxygen that desorbs above ca 750˚C is denoted as β and generally is linked with the lattice oxygen. This oxygen species can be directly associated with the reduction of B-site cation - here reduction of Mn⁴⁺ [28] .

All studied LM and LSM perovskites desorb α₁, α₂ and β oxygen (Figure 4). One can state the effect of synthesis method on the amount of various forms of desorbed oxygen. As it is seen on Figure 4, α-oxygen desorbs at wide range of temperatures. The LM perovskites evolve a small amount of oxygen at temperature range 180˚C - 350˚C (maximum of desorption around 250˚C); much higher amount of O₂ is desorbed from these materials at temperatures above 400˚C with maximum of desorption depending on the sample: at 555 and 862˚C (LM-MS); 632˚C (LM-SG) 634˚C (LM-SP) and 645˚C (LM-CC). Only for LM-MS two peaks of desorption are observed >400˚C. Partial substitution of La by Sr greatly lowers amount of oxygen desorbed from corresponding sample with the exception of material prepared by solvothermal method (LSM-MS) for which depletion of the amount of desorbing oxygen is negligible. Similar effect of Sr incorporation was reported by S. Ponce et al. [27] and D. M. A. Melo et al. [30] who investigated the La_1-xSr_xMnO₃ perovskites synthesized by: amorphous citrate decomposition [27] and by Pechini method [30] and showed that the amount of O₂ desorbed from its surface decreases with increasing Sr content. The LSM perovskites desorb oxygen in a wider temperature range than LM ones. The materials prepared by solvothermal method desorb the most of oxygen per gram, while the samples synthesized using spray pyrolysis method the least of all. The amount of α-oxygen (per gram) released by the LM perovskites decreases in the order: CC > SG > MS > SP. The sequence for the amount of α-oxygen per unit surface area is as follows: SG > CC > SP > MS. The samples prepared by MS method desorb much more β-oxygen (per gram and per m² of sample) than these ones prepared by other methods.

The onset temperature of α₁-oxygen desorption from LM perovskites increased with the order MS < CC < SG < SP while in the case of materials substituted by strontium the order is following: MS < SG < CC < SP (Figure 5).

Oxides of lanthanum and strontium are very reactive and in contact with the atmospheric air form surface carbonates, hence on the surface of materials large amount of carbon are present. The surface of all the studied materials is enriched with La (Table 5) with respect to the theoretical value. It can be explained by the fact that the surface of LaMnO₃ may exhibit one of the two ends with opposite polarity and charge: +1e (terminated LaO) and −1e (terminated MnO₂). Due to the enrichment of surface of materials with La, it is likely that they have LaO

T―theoretical ratio; E―experimental ratio.

termination. The surface characterized by such termination adsorbs oxygen atoms in the place of bridge La-La, which results in the more energetically favorable configuration of the material [31] [32] . Partial substitution of La by Sr changes composition of the surface: the content of La decreases while Mn increases (Table 5). The effect of the synthesis method on the surface composition of materials is observed, as evidenced by the differences in experimental ratio of La:Mn, La:Sr. The content of strontium on the surface of the material increases in the following order of synthesis methods: MS < CC < SG < SP. The theoretical content of Sr in the LSM is 7.42% by weight. LSM perovskites were dissolved in nitric acid and the content of strontium was determined by ICP method. It turned out that LSM prepared by both SG and CC methods show close to expected content of strontium while samples prepared by MS and SP contain less strontium than it comes from stoichiometry, respectively (wt.%): 6.12 and 6.95. Strontium nitrate used as one of the precursors in the LSM synthesis is relatively easily volatile (melting point/boiling point respectively 570/645˚C) [33] , what may explain lower content of Sr in the sample synthesized at high temperature (SP method). Good solubility of Sr(NO₃)₂ in ethylene glycol probably is a reason of lower content of Sr in the LSM-MS sample. Surface of LSM-MS shows the lowest Sr content what is consistent with the ICP analysis results. In turn, the highest content of Sr on the surface of LSM-SP probably results from the specificity of spray pyrolysis causing the accumulation of the more volatile component on the surface of the grain.

S. Ponce et al. [27] reported that the stability of Mn⁴⁺ ions is an important parameter which increases the catalytic activity of the materials. A measure of stability is the fact that, despite the thermal treatment at high temperatures and under high vacuum, Mn⁴⁺ ions are still present on the surface of material. They have shown that with increasing share of Mn⁴⁺ on the La_1-xSr_xMnO₃ (x = 0 - 0.5) surface its activity in the methane combustion also increases. As a result of the partial substitution of La by Sr, nonstoichiometric mixed oxides La_1-xSr_xMnO_3-δ is formed and oxidation of Mn³⁺ to Mn⁴⁺ takes place in order to compensate the charge [27] [34] . Similarly B.M. Nagabhushana [35] reported, that Mn⁴⁺ content increases with increase of Sr content. LSM materials studied in this work are characterized by higher share of Mn⁴⁺ on the surface than corresponding LM ones. Among LSM materials the largest share of Mn⁴⁺ on the surface was found for LSM-MS and LSM-SG, while the lowest values could be found in LSM-SP. When La³⁺ is partially substituted by Sr²⁺, the charge is compensated by the oxidation of Mn³⁺ to Mn⁴⁺ and/or formation of vacancies [36] . N. Yamazoe and Y. Teraoka observed decreasing of the high temperature oxygen desorption (β-oxygen) with increasing degree of Sr substitution. They explained this observation by the intrinsic nature of Mn ions: in order to mitigate the static Jahn-Teller distortion, LaMnO₃ has to have a certain number of cation vacancies, accompanied with formation of Mn⁴⁺ ions and oxygen-excess nonstoichiometry (LaMnO_3+δ) [37] . In our study we observed that substitution of La by Sr only slightly increase the share of Mn⁴⁺ on the surface of materials prepared by SG and CC methods while this effect is greater for LSM samples prepared by MS and SP methods. Since the samples prepared by both latter methods contain less Sr than it is required by the stoichiometry, for charge compensation part of Mn³⁺ passes to Mn⁴⁺ oxidation state. For these reasons there is only little space for introducing oxygen vacancies in the lattice of these materials because electroneutrality can be ensured by change Mn³⁺ to Mn⁴⁺. Oxygen species desorbing at high temperatures (β-O₂) may be formulated according following reaction [38] : 4Mn⁴⁺ + 2O²⁻⁽□) → 4Mn³⁺ + O₂ where (□) is a cation vacancies. Taking into account the high value of Mn⁴⁺/Mn ratio (Table 6) on the surface of LSM-MS, it is clear why concentration of β-O₂ on the surface of this material is distinctly higher than in case of other studied here perovskites.

In addition, as it was shown above, the morphological and textural properties of studied perovskites are strongly affected by the method of synthesis. S. Royer et al. have shown that amounts of desorbed different α-oxygens is mainly dependent on the specific surface area of the material [39] . All these factors make the TPD-O₂ profiles of the investigated materials differ significantly.

Three forms of oxygen are present (Table 6, Figure 5) on the surface of studied perovskites: lattice oxygen O²⁻ (BE about 528.8 eV), adsorbed oxygen species like hydroxyls and carbonate ${\text{OH}}^{-}/{\text{CO}}_3^{2-}$ (BE about 530.9 eV) and H₂O (BE about 533.0 eV) [40] [41] . Generally, the shifts of O 1s BE resulted from the synthesis method or substitution by Sr are insignificant.

However, both the method of synthesis and substitution by Sr exert a significant influence on the share of oxygen forms on the surface of these materials (Table 6). Lattice oxygen is the main form of oxygen species on the surface and it is confirmed by value of O_lat/O_ads ratio, which amounts in case of LM: 0.93; 1.85; 1.96; 2.45 respectively for MS, SG, CC and SP synthesis method and in case of LSM: 1.33; 1.61; 2.03; 2.23 respectively for SG, CC, MS and SP method. The high share of O²⁻ is observed on the surface of both samples prepared by SP method. Partial substitution of La by Sr in samples prepared by SG and CC method, lowers the share of O²⁻, while in the case of MS and SP method the effect is opposite. Higher share of lattice oxygen on the surface of the latter two materials, probably results from the presence of cationic vacancies (lower than stoichiometric content of Sr), what in turn limits the possibility of creating anionic vacancies on the surface of both latter materials.

The results of lean methane combustion on studied LM and LSM perovskites are reported in Figure 6, Table 7 and Table 8. Activity (temperature of 10%, 50% and 90% conversion of methane) of studied catalysts was compared with activity of 0.5% Pt/Al₂O₃ and 1% Pd/AlO₃ and presented in Table 7.

All prepared perovskites exhibit catalytic activity in lean methane combustion. Depending on the method of synthesis, the temperature of 10%, 50% and 90% conversion of CH₄ differs significantly. The highest activity exhibits LM-MS, it enables the total methane conversion at 451˚C. Slightly less active is LSM-CC (T_100% = 455˚C). The catalytic activity of 0.5% Pt/Al₂O₃ is low and comparable with the activity of both samples prepared by spray pyrolysis method while 1% Pd/Al₂O₃ is significantly more active than all other studied materials. The partial

substitution of 20% at. of La by Sr in the structure of LM perovskite decreases its activity for combustion of lean methane―only LSM-SG has higher activity than LM-SG.

It seems, that the high catalytic activity of LM samples prepared by MS and CC methods may be explained by their ability to release large amounts of α oxygen at relatively low temperatures (Table 4, Figure 4). It is suggested, that these

form of oxygen species are responsible for the oxidation of methane [30] . LM-SP released comparable amount of α₁ and α₂ oxygen (per surface unit) however but at higher temperatures than perovskites prepared by others methods. Apart of this, LM-SP exhibits the value of SSA among LM materials. Both factors make that among studied LM perovskites this one prepared by SP shows the lowest activity in the combustion of diluted methane. The rate of oxidation of methane (per m²) on the LM perovskites (Table 8) increases in the same order (SP < SG < CC < MS) as availability (temperature onset) of α₁-oxygen and does not correlate with either α₁-oxygen concentration (per m², Table 5) nor O_lat/O_ads ratio (share of adsorbed oxygen, Table 6). Taking into account low concentration of methane in the reaction mixture one can consider, that availability of active oxygen on the surface is at least as important as its amount for the activity of these materials. The confirmation of this conclusion is a linear relationship between rate of methane combustion and temperature of the first maximum of hydrogen consumption for LM perovskites prepared by studied methods (Figure 7).

Another factor contributing to higher catalytic activity of LM-MS and LM-CC, is relatively better developed specific surface area of this material, composed of fine crystal particles. In addition, LM-MS reaction with hydrogen starts at the lowest temperature, which suggests the presence of the most reactive oxygen species on the surface.

For the samples containing Sr there are no simple correlations between their properties and the catalytic activity. The highest rate of methane combustion shows material prepared by CC method while the lowest this one prepared by SP method (Table 8). At low temperature (300˚C), the rate of methane combustion on the LSM prepared by either by CC or SG is higher than on the corresponding LM. However the rate of combustion on the latter materials increases faster with temperature than on the LSM. The biggest amount (and at the lowest temperature) of α-oxygen desorbs from the surface of LSM-MS, however the reaction rate on this material is lower than on the LSM-CC and comparable to this one on the LSM-SG. Perovskites LSM-MS and LSM-CC show a very similar profile

TPR-H₂, possess close temperatures T_I and T_II and textures. In addition, the LSM-MS desorbs substantially more oxygen (α and β) counterpart prepared by CC. However, the latter perovskite shows a significantly higher catalytic activity than the previous one. The only one which differs these two materials is strontium content―its insufficiency in respect to the stoichiometric in the LSM-MS. Presently we can not explain why the abundance and availability of α-oxygen on the LSM-MS surface is not sufficiently reflected in its activity. Probably β-oxygen present in large amounts on the surface LSM-MS, but desorbing at relatively higher temperatures than in case of other LSM play more important role in methane combustion on this material. In this way one can suppose that unusual catalytic behavior of LSM-MS might be associated with lower than stoichiometric content of Sr resulting in a high share of Mn⁴⁺, whereby results of neither TPD-O₂ nor TPR-H₂ can be simply related to the rate of methane oxidation. Relatively low activity of LSM-SP should be attributed to its low SSA and a small amount of adsorbed oxygen species on its surface. High catalytic activity of materials prepared by CC method can be explained on the basis of H. Najjar, H. Batis conclusions who stated that this method of synthesis resulted in a decrease in superficial La/Mn atomic ratio and an increase of relative content of the surface oxygen species what leading to high catalytic activity of corresponding material in the combustion reactions [42] . Our results confirm these findings (Table 5 and 5): both materials prepared by CC show the low value of experimental La:Mn ratio and high concentration of surface oxygen species.

5. Conclusions

LaMnO₃ and La_0.8Sr_0.2MnO₃ perovskites were synthesized by four different preparation routes. It appeared that synthesis method affects physicochemical properties of obtained materials including catalytic activity in lean methane combustion. There is no simple answer to the question on the physicochemical reason for differences of catalytic behavior of studied perovskites. The use of solvothermal method allows one to obtain materials characterized by relatively well-developed surface area, the presence of reactive oxygen species on surface as well as high catalytic activity for lean methane combustion. Properties of the materials prepared by chemical combustion are very similar to that of solvothermally prepared. Spray pyrolysis leads to materials possessing the lowest specific surface area and porosity, characterized by the acidic character of the surface. LM and LSM perovskites prepared by the spray pyrolysis method show low susceptibility to reduction with H₂ and the highest share of lattice oxygen on the surface which can be correlated with the lowest activity in lean methane combustion. LM and LSM prepared by sol-gel method exhibit an intermediate properties between that ones prepared by solvothermal (or chemical combustion) and spray pyrolysis. LSM perovskites prepared by solvothermal and spray pyrolysis have a less than stoichiometric content of Sr.

All prepared LM and LSM materials show catalytic activity for lean CH₄, comparable to (or higher) than the activity of 0.5 wt.% Pt/Al₂O₃ but lower than 1 wt.% Pd/Al₂O₃. Qualitatively catalytic activity of lean methane combustion on studied perovskites can be correlated with the presence of α and β oxygen forms on their surface as well as availability of these species and the degree of surface development. The rate of methane combustion on LM perovskites quite well correlates with temperature of first maximum of hydrogen consumption. In case of LSM perovskites neither results of TPR-O₂ nor TPR-H₂ could be simply correlated with rates of methane combustion. Catalytic properties of LSM prepared by MS are affected by lower than assumed content of strontium leading to high content of β-oxygen on its surface. These species desorb from the LSM-MS surface at relatively higher temperatures than in the case of other LSM materials what may results in lower rate of methane combustion at low temperatures range.

Acknowledgements

The work was financed by a statutory activity subsidy from the Polish Ministry of Science and Higher Education for the Faculty of Chemistry of Wrocław University of Technology. We would like to thank dr W. Miśta for the TPD O₂ measurements.

Cite this paper

Miniajluk, N., Trawczyński, J., Zawadzki, M. and Tylus, W. (2018) LaMnO₃ (La_0.8Sr_0.2MnO₃) Perovskites for Lean Methane Combustion: Effect of Synthesis Method. Advances in Materials Physics and Chemistry, 8, 193-215. https://doi.org/10.4236/ampc.2018.84013

References