Influence of the Thermal Processing and Doping on LaMnO₃ and La_0.8A_0.2MnO₃ (A = Ca, Sr, Ba) Perovskites Prepared by Auto-Combustion for Removal of VOCs

Abstract

Single-phase oxygen stoichiometric LaMnO₃ and doped La_0.8A_0.2MnO₃ (A = Ca, Sr, Ba) perovskites have been prepared by a simple one-step auto-combustion method. Cation-deficient LaMnO_3+δ and La_0.8A_0.2MnO_3+δ were obtained by calcination of the former samples in air at 750 °C. The samples were characterized by X-ray powder diffraction, X-ray photoelectron spectroscopy, temperature-programmed reduction, temperature-programmed oxygen desorption, and N₂ physisorption in order to apply them as catalysts in the complete catalytic oxidation of acetone as a model volatile organic compound. The studied phases show the expected orthorhombic and rhombohedral perovskite crystal structures. Catalytic experiments performed with all the samples show measurable activity already at 100 °C. At 200 °C, doped La_0.8A_0.2MnO₃ samples show higher activity than undoped LaMnO_3, with increasing conversion with larger A-cation size. Calcined samples also show higher activity than as-prepared ones making La_0.8Ba_0.2MnO_3+δ the best catalyst at this temperature. All doped samples show >95% acetone conversion at T ≥ 250 °C with a weak dependence on the sample processing or A cation doping. The collected evidence confirms that the most important factors for the catalytic activity of these oxides are the Mn⁴⁺/Mn³⁺ molar ratio on the surface of the samples and the cation-deficiency of the bulk perovskite structure. In addition, increasing the symmetry of the bulk crystal structure appears to have an additional favourable effect. Despite the observation of the presence of surface carbonates, we show that it is possible to use the as-prepared samples without further thermal treatment with good results in the oxidation of acetone.

1. Introduction

Noble metal-based catalysts have a high activity for the oxidation of most volatile organic compounds (VOCs). However, they are relatively expensive and tend to be deactivated by coking, poisoning, thermal sintering, etc. [1]. Therefore, the design and development of new materials with high performance, low cost, chemical and thermal stability at the operation temperature and high tolerance to catalytic poisons, is increasingly pursued [2,3,4,5,6,7,8]. Transition metal oxides, especially perovskite manganites, are among the best catalysts tested for VOCs oxidation at low temperatures [6]. Catalytic activity of manganites can be generally attributed to the redox properties of manganese, showing mixed-valence states, the high mobility of oxygen within the crystal structure and the presence of adsorbed oxygen species [9,10,11,12,13,14,15,16,17]. LaMnO_3+δ perovskite has been evaluated for total oxidation of toluene [18,19,20], methanol [19], benzene [21], propane [22], methane [23], hexane [24], acetone [8] and chlorinated compounds [25,26], among others. The modification of LaMnO₃ through doping at the A-site by different species has been used to increase the activity of the catalyst [16,27].

When prepared in air at high temperatures, LaMnO₃ forms the metal-vacant phase La_1−εMn_1−εO₃ of rhombohedral symmetry (R$\overline{\mathit{3}}c$ space group), usually described as LaMnO_3+δ [28,29,30,31]. A similar effect has been described for low-doped (La_0.8A_0.2)MnO₃ (A = Ca, Sr, Ba) forming (La_0.8A_0.2)_1−εMn_1−εO₃ also usually described as La_0.8A_0.2MnO_3+δ [31] where δ represents excess oxygen respect to the metals, but not interstitial oxygen. These metal-vacant phases are stabilized by the oxidation of Mn³⁺ into Mn⁴⁺ in the presence of oxygen at moderate-to-high temperatures achieving equilibrium with maximum δ = 0.26 for LaMnO₃ annealed at 1000 °C in O₂ [29]. Preparation of oxygen-stoichiometric LaMnO₃ containing pure Mn³⁺ and La_0.8A_0.2MnO₃ containing up to 20% Mn⁴⁺ was usually achieved by firing non-stoichiometric phases at high temperature and low oxygen-partial-pressure atmosphere [28,29,30,31], observing a change in the symmetry of the perovskite structures from the metal-vacant rhombohedral R$\overline{\mathit{3}}c$ structure to the stoichiometric orthorhombic Pbnm structure. Reports that have shown rhombohedral LaMnO_3+δ and A-doped LaMnO₃ catalytic activity for oxidation of VOCs generally make no mention of the oxygen/metal stoichiometry calling them LaMnO₃ and La_1−xA_xMnO₃ [5,7,10,18,19,20,21,22,24,25,32,33,34]. Recent reviews on the use of perovskites in the catalytic oxidation of different VOCs have not investigated the relative effect of changing the oxygen stoichiometry or A-site dopant in the catalytic activity [4,5,27].

In this work, stoichiometric LaMnO₃ and alkaline-earth substituted La_0.8A_0.2MnO₃ (A = Ca, Sr, Ba) perovskites were prepared by a modified auto-combustion method. They were characterized and tested as catalysts in the complete catalytic combustion of acetone (chosen as model VOC). The cation-deficient rhombohedral LaMnO_3+δ and alkaline-earth substituted La_0.8A_0.2MnO_3+δ phases were also tested in identical conditions for comparison purposes. The effects of different metals substitution (A) and initial oxygen stoichiometry (δ) in the catalytic activity over several catalytic cycles are discussed.

2. Results

2.1. Structural Characterization

The auto-combustion synthesis method used in this work yields perovskite phases in one synthesis step. The as-prepared samples obtained were orthorhombic LaMnO₃ (LM) and La_0.8Ca_0.2MnO₃ (LCM), and mixed-phase orthorhombic and rhombohedral La_0.8Sr_0.2MnO₃ (LSM) and La_0.8Ba_0.2MnO₃ (LBM) (Figure 1) corresponding to oxygen-stoichiometric samples as described in [28,29,30,31]. ICP-AES spectroscopic results obtained for these samples (Table S1) indicate that La, Mn and A cations fall in the expected composition range within experimental uncertainty, considering the presence of ~1% weight from carbonated species (see discussion in Section 2.7 and Section 3 and Supplementary Materials), confirming the adequacy of the preparation method to obtain the desired perovskite phases. These phases correspond to bulk stoichiometric perovskites, showing no significant amount of cation vacancies. The change in the symmetry observed in doping is caused by the change of ionic radii of the A cation as predicted by the Goldschmidt tolerance factor in AMnO₃ perovskites [35,36]. The calcined samples show different symmetry relative to the as-prepared compounds at room temperature. They are orthorhombic La_0.8Ca_0.2MnO_3+δ (LCM750), rhombohedral LaMnO_3+δ (LM750), La_0.8Sr_0.2MnO_3+δ (LSM750) and biphasic rhombohedral and cubic La_0.8Ba_0.2MnO_3+δ (LBM750) (Figure 1). For all the samples studied, no additional unexplained diffraction peaks could be detected, (except for one peak assigned to CaCO₃ in the as-prepared LCM sample), confirming that the samples are >99% pure with respect to crystalline phases for LM, LSM and LBM and >95% pure for LCM. Note that amorphous phases in small proportions and crystalline phases below 1% weight show, in general, no detectable signal in a conventional XRD diffractogram. Rietveld plots, final refined unit cell parameters of the observed phases and fit agreement parameters are shown in Figures S1 and S2 as well as in Table S2. The results of the thermo-diffraction experiments are discussed in the following section.

2.2. Catalytic Performance

Acetone conversion (%) vs. temperature (°C) curves for the as-prepared and the calcined manganites are shown in Figure 2. At 100 °C, both groups of samples present similar conversion but at 150 °C the effect of A doping in the samples can be observed. Conversion values of LCM, LSM and LBM are at least 10% larger than that of LM between 150 °C and 250 °C, with a maximum difference of 20% for LCM at 150 °C. Similarly, conversion values of LCM750, LSM750 and LBM750 are at least 20% larger than those of LM750, with a maximum difference of 40% for LBM750 at 200 °C. Samples doped with Sr and Ca show similar catalytic activity both in the as-prepared or calcined form in the whole temperature range, while barium doping produces higher conversions above 150 °C. Conversion of acetone is larger than 95% for all doped samples below 250 °C. By observing the light-off curves in more detail one can see that all the calcined samples show a higher conversion of acetone between 150 °C and 250 °C with respect to the as-prepared ones.

Figure 3 shows the light-off curves corresponding to five catalytic cycles carried out on the as-prepared samples in comparison with the respective calcined samples. For the LM sample, as the catalytic cycles are carried out, there is no noticeable difference in the percentage of acetone conversion, and they are similar to the calcined sample. LM750 sample was also cycled four times and neither shows a significant change in catalytic activity over cycling (as shown in Figure S3). The LCM sample shows no clear trend of conversion over cycling, reaching values similar to those obtained for LCM750 in the best cycle. When the LSM sample is studied, it is observed that, as the cycles are carried out, the percent of acetone conversion increases, reaching the corresponding value of LSM750 conversion after four cycles. When five cycles are carried out for LBM, it is also observed that there is no variation in the acetone conversion, but these values differ from those obtained for the LBM750 sample with a maximum of 30% difference in acetone conversion at 200 °C.

Between 150 °C and 200 °C, the activation of the oxidation reaction occurs as CO₂ starts to be detected (Figure S4). At 150 °C, the formation of acetic acid is observed for all the samples tested. We define r as the ratio of [CO₂]_experimental/[CO₂]_theoretical. For most of the doped samples already at 250 °C, an r of 0.9 is reached, except for LCM and LCM750, which show an r of 0.87 and 0.75 respectively, detecting acetaldehyde as a by-product.

It is important to note that any change in the surface of the as-prepared samples will make them similar to the calcined ones because it will introduce the formation of metal vacancies at catalytic conditions. This surface modification may not be visible by X-ray powder diffraction, since the technique is more sensitive to the bulk structure. However, all the as-prepared samples start showing indications of the transformation from orthorhombic to rhombohedral phase between 200 °C and 300 °C on heating. Orthorhombic perovskite oxides usually transform reversibly to rhombohedral ones on heating [35] due to changes in the Goldschmidt tolerance factor. However, in this case, this effect is irreversible as it is mainly caused by the gain of oxygen, as depicted in Figure 4 for LBM (and Figure S5 for LM, LCM and LSM). We hypothesize that the as-prepared, stoichiometric samples show similar or slightly less activity than the calcined catalysts, but become quickly oxidized at the surface above 200 °C in the presence of O₂. This makes the activity of all the samples above 200 °C similar to the activity of the rhombohedral/cubic phases showing cation vacancies. Note that the LBM750 sample shows rhombohedral and cubic perovskite phases at room temperature (Figure 1). The proportion of cubic La_0.8Ba_0.2MnO_3+δ increases at higher temperatures, which improves oxygen mobility [37].

Table 1. Estimated temperature values (°C) for acetone conversions of 50 and 80% for the as-prepared and calcined samples.

Catalyst	T50 [°C]	T80 [°C]
LM	218	244
LM 750	208	238
LCM	208	235
LCM 750	190	227
LSM	209	234
LSM 750	193	228
LBM	200	231
LBM 750	168	198

shows the temperature values for conversions of 50% (T₅₀) and 80% (T₈₀) of acetone. For all the compositions tested, there are two general trends. There is a reduction in both temperatures with the size of the A cation, and the same is observed with calcination.

2.3. TPR

TPR measurements are shown in Figure 5 and the calculated amounts of consumed hydrogen from TPR profiles are summarized in

Table 2. Reducibility of perovskite samples.

Catalyst	H2 Uptake (mol H2 mol−1 cat.)
	T ≤ 600 °C	T > 600 °C	Total
LM	0.04	0.35	0.39
LM 750	0.23	0.42	0.65
LCM	0.18	0.34	0.52
LCM 750	0.25	0.32	0.57
LSM	0.25	0.28	0.53
LSM 750	0.29	0.28	0.57
LBM	0.30	0.24	0.54
LBM 750	0.32	0.25	0.57

. For all the studied samples, two reduction stages are observed in the TPR profile, one below 600 °C and the other at around 800 °C. Owing to the fact that La³⁺, Sr²⁺, Ca²⁺ and Ba²⁺ are not reducible under the measurement conditions, the water production is due to the reduction of manganese. According to the literature [38,39,40], the H₂ consumption below 600 °C corresponds to the reduction of Mn⁴⁺ to Mn³⁺ while that observed at high temperature it is attributed to the reduction of Mn³⁺ to Mn²⁺ (which will not be discussed because it is beyond the temperature range considered for the reaction conducted under oxidizing conditions in this work).

Then, based on the first reduction process (Table 2), it is possible to obtain the Mn⁴⁺ content of the perovskite being the 100% of Mn⁴⁺ equal to 1/2 mol H₂ mol⁻¹ cat.

The doping of the perovskite leads to an increase in the Mn⁴⁺ content in the perovskite structure, rising from 4% Mn⁴⁺ in LM to 30% in the case of the Ba-doped sample. Moreover, the calcination treatment also increases the Mn⁴⁺ content for all the samples.

2.4. O₂-TPD

The results of O₂-TPD analysis are presented in Figure 6. Evolved oxygen can be categorized as β-oxygen, which is ascribed to the bulk crystal structure oxygen in the perovskite framework [41,42]. In LM perovskite containing more than 95% of Mn³⁺, no oxygen removal is observed. For LCM, LSM and LBM, containing only ~20% of Mn⁴⁺ a peak starting around ~300 °C is detected for all the as-prepared samples (characterized by two broad features in LCM), probably related to the reduction of all Mn⁴⁺ to Mn³⁺ with the formation of oxygen-deficient perovskites (δ < 0 in La_0.8A_0.2MnO_3+δ). After the calcination treatment, which introduces cation vacancies in the perovskite, the four compounds show release of oxygen starting above 450 °C. The semi-quantitative nature of this technique prevents the correlation between peak areas and the amount of oxygen removed, as observed in LSM and LSM750 where the opposite area relation is observed with respect to the rest of the samples. The temperature of the processes is clearly the same for the three as-prepared A-doped samples and the four calcined samples. Bulk studies indicate that the heating treatment performed on the as-prepared perovskites generates a net oxygen gain and consequently an increase in the Mn⁴⁺/Mn³⁺ ratio, which can be correlated with the improvement in the catalytic response of the calcined samples.

2.5. Textural Properties by Means of N₂ Physisorption

Nitrogen physisorption isotherms for LBM and LBM750 are shown in Figure 7 (and Figure S6 for LM, LCM and LSM). According to the IUPAC classification, the curves in the figure correspond to type II. This is characteristic of low-porous or macroporous solids. In LBM, a very light hysteresis (H3 type, Boer) is perceived, which could indicate the presence of lamellar solids. For LBM750, the hysteresis effect disappears, denoting a slight modification in the porosity of the sample during the calcination process used to transform LBM into LBM750. This behaviour is similar in all the doped samples. The quantitative results of textural properties obtained for the as-prepared and calcined manganites are shown in

Table 3. Textural properties for as-prepared and calcined manganites.

Catalyst	SBET [m2g−1]	Vtotal [m3g−1]
LM	5.7	0.025
LM750	4.7	0.015
LCM	5.0	0.030
LCM750	6.9	0.030
LSM	7.0	0.030
LSM750	5.5	0.030
LBM	7.1	0.030
LBM750	6.7	0.030

. It is observed that both types of samples have similar specific surface areas (S_BET), with low values of area per gram, but comparable to those reported in the literature for other perovskites prepared by other methods [11,13,32,41]. A small decrease in the specific area can be observed when samples are subjected to calcination except for LCM750, where the area seems to slightly increase despite having a similar total pore volume (V_total).

2.6. SEM-EDS Characterization

SEM micrographs show an apparently homogeneous surface of all the samples (Figure 8). In order to determine the presence of significant impurities, the X-ray emission spectra (EDS) of different points on the surface of each sample, showing different aspects from the average, were collected and qualitative analysis of metals was performed. This can be observed in Figures S7–S14 where the micrograph and spectra of each marked point are shown. The variations in metal composition (as extracted from the area of peaks of each phase) suggest that there is no phase segregation. Notice that no region shows a significant excess/defect of one metal, as would correspond to the presence of separated La, Ca, Sr, Ba or Mn carbonates (as suggested in the following section) and all the metals La, Mn and A (when adequate) could be observed at every point selected as shown in

Table 4. Atom % obtained by EDS corresponding to the points indicated in the SEM images shown in Figure 8.

Sample	Atom %			Sample	Atom %
	O	Mn	La		O	Mn	La
LM	60.4	18.4	20.1	LM750	61.5	17.1	19.3
	44.3	23.6	30.4		37.9	25.0	36.7
	63.2	17.0	18.1		58.3	18.4	21.3
LCM	69.2	14.0	12.5	LCM750	57.5	19.2	18.2
	60.0	19.3	15.7		62.0	17.4	14.1
	48.6	24.4	19.8		61.5	18.3	14.7
LSM	62.6	18.0	13.6	LSM750	71.2	12.7	11.8
	66.9	14.6	12.9		67.8	14.8	11.4
	59.9	20.0	14.0		65.3	14.9	13.9
LBM	58.7	18.2	16.2	LBM750	67.6	14.2	13.4
	66.0	15.2	12.0		61.9	16.6	14.7
	66.1	15.6	12.7		69.4	14.3	11.0

, Tables S3 and S4. In addition, the SEM-EDS study allows ruling out the presence of large granular impurities on the surface of all the samples.

2.7. XPS Characterization

XPS measurements provide the chemical environment of the atoms present on the outmost layers of the samples.

Table 5. Surface atomic concentrations (%) determined from the XPS survey spectra and Mn cations ratios for each sample, being A the dopant element.

Sample	A	Element (Atom %)
		O	Mn (Mn4+/Mn3+ Ratio)	La	C	A
LM	-	48.9	6.9 (0.6)	4.4	39.8	-
LM750	-	46.3	8.1 (0.8)	5.1	40.5	-
LCM	Ca	49.7	8.5 (1.0)	4.4	32.7	4.7
LCM750	Ca	49.4	10.2 (1.1)	5.7	31.2	3.5
LSM	Sr	50.0	8.6 (0.7)	4.2	35.5	1.7
LSM750	Sr	47.8	8.9 (0.8)	4.2	37.6	1.5
LBM	Ba	45.9	7.4 (1.1)	3.8	41.2	1.7
LBM750	Ba	47.6	9.3 (1.1)	4.4	36.0	2.7

displays the surface composition (atom %) of all samples as calculated from the XPS survey scans. It is clear that, besides the expected elements present in the expected ratios, the samples also have considerable surface carbon content. This was expected based on the synthesis method and sample storage in air, as will be discussed in the next section. The Mn 2p XPS spectra (see Figure 9) are especially relevant for this study and presented similar features for all samples. They show the presence of both Mn³⁺ and Mn⁴⁺ oxidation states on the surface [42], although identification of the specific chemical bonds associated with them is difficult from the Mn 2p signal. This difficulty is due to the very similar binding energy values for several Mn bonds, such as oxides. In the LM sample, the Mn⁴⁺/Mn³⁺ ratio is the lowest and Mn³⁺ cations dominate, as expected for a sample with bulk Mn³⁺ oxidation state. The Mn⁴⁺/Mn³⁺ ratio increases with the A doping and with calcination for the same sample, also expected from the formation of bulk metal-deficient samples with the shift of Mn average charge closer to Mn⁴⁺. The presence of carbon in the form of carbonates justifies the observation of CaCO₃ in the LCM sample and of carbonated species in the La 3d_5/2, Ca 2p, Sr 3d, and Ba 3d_5/2 region fits (Figures S15 and S16) [43]. Since La, Ca, Sr and Ba are generally considered spectators in the catalytic activity of these manganites, the detailed analysis of the spectra for each of the elements in each sample is given in the Supplementary Materials (Figures S15 and S16).

3. Discussion

All results obtained from the bulk and surface characterization techniques applied to the as-prepared and calcined materials are consistent with the presence of perovskite manganites containing bulk Mn³⁺ in the case of as-prepared LaMnO₃ (LM) and Mn³⁺/Mn⁴⁺ mixtures for the rest of the doped (LCM, LSM, LBM) and calcined (LM750, LCM750, LSM750 and LBM750) samples. All the samples show the expected crystalline structures consistent with thermal treatment and oxygen content, with no indication of the presence of crystalline impurities, except for a small amount of CaCO₃ in LCM. The hydrogen consumption and oxygen evolution behaviour in programmed reduction/desorption studies are also consistent with the aforementioned characteristics. XPS spectra show the expected chemical complexity of the grain surface, also observed by SEM, with the presence of similar amounts of carbon species in all the samples. The preparation procedure of LM, LCM, LSM and LBM implies the combustion of acetates and amine species in an open atmosphere and is expected to produce carbonates of the metals. All the samples were stored in 30 cm³ Falcon tubes, therefore also exposed to atmospheric carbon during handling between the preparation and characterization steps. The similar carbon content in all the samples is consistent with carbonates or atmospheric carbon since La, Ca, Sr and Ba carbonates have been reported to be stable beyond the calcination temperature [44,45,46,47]. This is also confirmed by XPS results shown in Table 5.

EDS measurements provide additional evidence that there is no significant La/A/Mn segregation in the surface of the sample (see Figures S7–S14) expected for carbonate crystallization. In any case, these surface carbon species seem to have no detrimental effect on the catalytic properties of the material. The bulk stability of the materials has already been demonstrated. The comparable catalytic activity of as-prepared and calcined materials in this study shows they are good candidates either for systems that may use the raw catalysts without processing or require further processing (such as pre-formation of the catalyst, deposition to prepare supported catalysts and/or thermal treatments) since we found that calcination in air maintains or slightly increases, in general, the catalytic activity of the samples.

The doping of LaMnO₃ (LM) with larger Ba²⁺ and Sr²⁺ cations generates higher symmetry structures with larger Mn⁴⁺ content due to the decrease of the charge of the A site and the reduction of the tilt of MnO₆ octahedra. This change has a beneficial effect on the catalytic response. La_0.8Ba_0.2MnO₃ sample shows the best conversion of acetone among the as-prepared samples. This may be a consequence of the change in the proportion of rhombohedral and cubic perovskite phases. At higher temperatures, the proportion of cubic La_0.8Ba_0.2MnO_3+δ increases, improving the oxygen mobility and thus the catalytic response. Calcination of the samples introduces cation vacancies in the structure, which also increase Mn⁴⁺ and oxygen content respect to the stoichiometric samples. This also improves the catalytic conversion of acetone for all the sample compositions studied. It is reported that the oxidation of hydrocarbons on perovskites occurs by means of a mechanism in which the VOC interacts with the oxygen present in the catalyst. The active oxygen content is directly related to the amount of cation vacancies [48,49]. The slight variability of the catalytic response of as-prepared samples with cycling may be attributed to the incorporation of oxygen (consequently more Mn⁴⁺) in the surface of the sample during the catalytic process, consistent with the effect of doping, calcination, and thermal evolution of the samples at the working temperature. This makes the calcined La_0.8Ba_0.2MnO_3+δ (LBM750) sample the best catalyst of the series. The Mn⁴⁺ content is, however, similar for all calcined samples on the surface, and the catalytic activity of all the materials (above 90% conversion at 300 °C), confirms their similarity regarding catalytic performance.

4. Materials and Methods

4.1. Catalysts Preparation

LaMnO₃ and La_0.8A_0.2MnO₃ (A = Ca, Sr, Ba), perovskites were synthesized via an auto-combustion route, using ethylenediaminetetraacetic acid (EDTA) as organic fuel and chelating agent, and ammonium nitrate as combustion promoter. The selected method, already used for the preparation of other perovskite oxides [37,50,51,52] has been modified to obtain the desired material in one step directly from the initial dissolution of metals in stoichiometric proportion. This method is cheap, fast (3–5 h total from weighing the reactants to the obtainment of the final sample), and scalable, since it allows us to obtain several grams (5–20 g) of the final material in one beaker.

For this purpose, aqueous solutions containing stoichiometric amounts of La(NO₃)₃·6H₂O, Sr(NO₃)₂, Ba(NO₃)₂, Ca(CH₃CO₂)₂·H₂O and Mn(CH₃CO₂)₂·4H₂O (>99.9%, Sigma-Aldrich, San Luis, MO, USA) were prepared according to the desired composition. An alkaline solution (pH = 10) containing 1.1 moles of EDTA (>99.4%, Sigma-Aldrich, San Luis, MO, USA) per mole of cation and NH₄NO₃ (>99.0%, Sigma-Aldrich, San Luis, MO, USA) was also prepared and the aforementioned solutions were mixed. EDTA molar excess ensures the total complexation of the cations. A 10:1 NH₄NO₃:EDTA molar ratio was chosen to promote the auto-combustion process. The resulting solution was heated over a hot plate at 130 °C, constantly stirring and keeping the pH constant (monitored every 30 min and restored to pH = 10 using small aliquots of NH₄OH), until the formation of a gel. Then, the magnetic stirrer was removed, and the temperature of the hot plate was raised to 300 °C, where the gel self-ignited [Safety note: Depending on the size of the beaker, the mass of reactants and the EDTA/NH₄NO₃ ratio, the auto-combustion process may produce ejected ashes or sparks falling outside the beaker. This procedure must be performed inside a hood and the surface should be free of flammable products around the hot plate]. The combustion process resulted in a sponge-like black powder, corresponding to the desired perovskite compounds, as confirmed by X-ray powder diffraction. The samples were ground in an agate mortar until a fine and homogeneous powder was obtained. After that, a portion of each sample was subjected to a heating treatment in a muffle furnace at 750 °C in air for 8 h to obtain the calcined samples. The samples obtained directly from the synthesis process, LaMnO_3, La_0.8Ca_0.2MnO₃, La_0.8Sr_0.2MnO₃ and La_0.8Ba_0.2MnO₃ (LM, LCM, LSM and LBM, respectively) denote the as-prepared samples, while those obtained after the calcination process, LaMnO_3+δ, La_0.8Ca_0.2MnO_3+δ, La_0.8Sr_0.2MnO_3+δ and La_0.8Ba_0.2MnO_3+δ, (LM750, LCM750, LSM750 and LBM750, respectively) are described as calcined.

4.2. Catalysts Characterization

4.2.1. Inductively Coupled Plasma Atomic Emission Spectroscopy

The metal contents in the as-prepared materials were determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES) in an Iris Intrepid Thermo Scientific equipment. Three takes of each sample were prepared by acid digestion of a solution containing (2 to 3) mg of the material and Milli-Q water up to 0.05 L, in a Digestor Digiprep Jr. SCP-Science equipment. The results are consistent within uncertainty with the expected nominal formula including up to 1% weight of impurities such as carbonates, as discussed in Section 2.1.

4.2.2. X-ray Powder Diffraction

Conventional X-ray powder diffraction (XRD) measurements were performed over as-prepared and calcined samples in a Rigaku ULTIMA IV diffractometer of 285 mm radius in θ-θ geometry, using CuKα sealed-tube radiation (λ_ave = 1.5418 Å) operating at 40 kV and 30 mA. A diffracted-beam curved Ge monochromator, and NaI scintillation detector were used for data collection. The data, used for confirmation of phase purity and preliminary structural characterization of as-prepared samples, were collected at room temperature in the 2θ = 10° to 70° range in steps of 0.04° for 3 s. Data for Rietveld analysis at room temperature for all samples were collected in the 2θ = 10° to 120° range in steps of 0.02° for 10 s.

The as-prepared samples were also characterized through X-ray thermo-diffraction at the XPD-D10B beamline of the UVX ring of the Brazilian Synchrotron Light Laboratory (LNLS). The beamline was equipped with a θ–2θ reflection-geometry diffractometer with a Mythen-1000 linear position sensitive detector (PSD). Measurements were performed from RT to 500 °C in 100 °C steps in synthetic air atmosphere, similar to the catalytic conditions in the reactor (see below), using the in-house Canario furnace. X-rays with an energy of 10 keV (λ = 1.2372 Å) were used for data collection using a constant-monitor-counts strategy to determine the time of each step. The linear detector was scanned in 0.5° steps providing 6 measurements of each data point in one complete scan. Data reduction and averaging were carried out with the in-house software to obtain powder patterns in the 8° to 120° range with 0.005° 2θ-steps. The Rietveld method (as implemented in GSAS-II software suite [53]) was used to fit all X-ray powder diffraction data to confirm the desired phase formation, determine the phase composition and the structural parameters of the perovskite manganites.

4.2.3. Temperature-Programmed Reduction

Temperature Programmed Reduction was performed in a quartz reactor coupled to a TCD detector Autochem Micromeritics apparatus (TPR-TCD). 50 mg of catalysts were fed with a 5% H₂/Ar mixture at a flow rate of 60 cm³·min⁻¹, with a heating rate of 10 °C·min⁻¹. Prior to the temperature-programmed reduction (TPR) analysis, the sample was cleaned by heating up to 300 °C with a flow rate of 60 cm³·min⁻¹ containing Ar, using a heating rate of 10 °C·min⁻¹; then, the catalyst was kept for 1 h at 300 °C constant temperature and further cooled down to 25 °C under a continuous flow of Ar. The data obtained allow hydrogen consumption estimate for sample reduction in the range of temperature selected.

4.2.4. Temperature-Programmed Desorption

Temperature Programmed Desorption (TPD) was performed in a quartz reactor coupled to a quadrupole mass spectrometer (Pfeiffer, model Thermostar GSD301T1). A quantity of 200 mg of catalysts was fed with He, at a flow rate of 60 cm³·min⁻¹, with a heating rate of 10 °C·min⁻¹. Prior to the temperature-programmed desorption (TPD) analysis, the sample was cleaned by heating up to 500 °C with a flow rate of 60 cm³·min⁻¹ containing 100% O₂, using a heating rate of 10 °C·min⁻¹; then, the catalyst was kept for 1 h at 500 °C constant temperature, and further cooled down to 150 °C under a continuous flow of oxygen, and finally to 25 °C in He, except for LaMnO₃, which gains oxygen to form LaMnO_3+δ (similar to the calcined LM750 sample) at such conditions. This provides a semi-quantitative measure of the oxygen released from the sample.

4.2.5. Nitrogen Physisorption

Specific surface area (S_BET) measurements were performed applying the Brunauer-Emmett-Teller (BET) method to the nitrogen (Praxair, >99.95%) adsorption-desorption isotherms at −196 °C on outgassed samples for 10 h at 100 °C. The isotherms were recorded until relative pressures equal to 0.995 P_o using a Beckman Coulter SA3100 apparatus.

4.2.6. Scanning Electron Microscopy

The as-prepared and calcined samples were inspected at the micron scale by Scanning Electron Microscopy (SEM). Digital micrographs were acquired at 20 kV in a JEOL JS M-5900LV microscope. Elemental composition of selected areas of the samples was determined by means of an Energy Dispersive Spectrometry (EDS) probe (NORAN Instruments EDS-vantage probe).

4.2.7. X-ray Photoelectron Spectroscopy

X-ray photoelectron spectroscopy (XPS) measurements were performed using a SPECS system equipped with a PHOIBOS 150 1D-DLD hemispherical electron analyser. An Al-Kα (1486.6 eV, 150 W) X-ray source was used, and survey spectra were acquired with 50 eV pass energy, energy step size of 0.5 eV, and acquisition time of 0.1 s. For the high-resolution spectra, the analyser was adjusted to pass energy of 30 eV, an energy step size of 0.1 eV, and an acquisition time of 0.5 s. The Au 4f_7/2 XPS peak obtained from a clean Au foil was applied for the analyser’s energy calibration. An acceptance angle of ±5° was used and the base pressure of the analysis chamber was 4 × 10⁻¹⁰ mbar. The spectra were analysed using the CasaXPS software and the charge was corrected considering the adventitious C 1 s chemical component (C-C, C-H) set to 284.5 eV binding energy. The fitting procedure also addressed the peaks line shape asymmetry considering a 20% Gauss-Lorentz ratio as well as Shirley-type background [54].

4.3. Catalytic Evaluation

Catalytic experiments were carried out in a quartz tubular reactor (diameter = 8 mm) with a “u” shape. For the tests, 0.5 g of catalyst was used. All catalysts were pressed into pellets at 5 MPa, crushed and sieved in order to obtain particle size between 425 µm and 600 µm, and prevent a pressure drop in the reactor. Additionally, the pre- and post-catalytic zones were filled with ground quartz to improve the homogeneity of the gaseous blend (avoiding the development of a laminar flux), both at the inlet and outlet of the catalytic fixed bed. A pre-treatment of the catalyst was performed at 300 °C for 1800 s in Ar flow. The reactivity of the samples was determined at atmospheric pressure, between 100 °C and 300 °C, in 50 °C steps with an Ar/O₂/Acetone reaction mixture (60 cm³·min⁻¹, 3.42 cm³·min⁻¹ and 1500 ppm_v, respectively). The O₂ gas flow goes through a saturator containing acetone (>99.5%, Dorwil), submerged in a water/ice bath (at 0.0 ± 0.1 °C) and the resulting mixture is diluted in Ar. Quantification of reagents and reaction products was accomplished by gas chromatography on a Shimadzu GC 2014, using a HAYESEP R 60/80 mesh column (diameter 1/8 mm SS, length 4.5 m) and TCD and FID detectors in series.

Consecutive catalytic cycles were carried out for all the as-prepared and LM750 samples under the same conditions without any sample treatment between cycles to assess the deactivation of the catalyst.

The acetone conversion was defined as:

$$Acetone Conversion \left(\%\right)=\frac{\left[ Acet\right]{ }_{in}-\left[Acet\right]{ }_{out}}{\left[Acet\right]{ }_{in}}\times 100$$

5. Conclusions

Samples of stoichiometric LaMnO₃, La_0.8Ca_0.2MnO₃, La_0.8Sr_0.2MnO₃ and La_0.8Ba_0.2MnO₃ perovskite manganites were prepared by an auto-combustion method. Additionally, cation-deficient LaMnO_3+δ, La_0.8Ca_0.2MnO_3+δ, La_0.8Sr_0.2MnO_3+δ and La_0.8Ba_0.2MnO_3+δ were obtained from them by calcination at 750 °C in air. The preparation method provides active samples in multi-gram batches with a simple procedure in a few hours of work. Structural characterization, behaviour against reduction, oxygen desorption, bulk and surface composition and textural measurements were conducted and tests for catalytic oxidation of acetone as a model VOC compound were performed. The results reveal that these single-phase perovskite manganites show good catalytic activity and remain active throughout the cycling both in the as-prepared and calcined forms, even though they present low BET areas and the presence of carbonates on the surface. The doping with larger alkaline earth cations such as Ba²⁺ and incorporation of extra oxygen in the structure by calcination seem to improve their activity. Both cation-size of the dopant and calcination seem to have synergistic effects, making calcined La_0.8Ba_0.2MnO_3+δ (LBM750) material the one with the largest acetone conversion at 200 °C. However, this improvement is insignificant at 300 °C where all the doped samples show acetone conversion above 95%. The change of alkaline earth metal in the perovskite changes the structure’s bulk symmetry, increasing it from orthorhombic (Ca) to rhombohedral/cubic (Ba). This correlates with the improvement of activity with dopant size and with the equalization of the catalytic activity as the structural differences among samples are also reduced at 300 °C where all samples are rhombohedral or rhombohedral/cubic. The Mn⁴⁺/Mn³⁺ ratio in the surface of the sample, accompanied by the high-symmetry cation-vacant bulk structure, seems to be more important than any other characteristics of the studied catalysts.