Catalytic Performance of La 0 . 6 Ca 0 . 4 Fe 1-xNixO 3 Perovskites in DME Oxidation

Substituted La0.6Ca0.4Fe1-xNixO3 (0.0 ≤ x ≤ 1.0) perovskite-type oxides prepared by the amorphous citrate method have been investigated as catalysts in the dimethyl ether (DME) oxidation reaction. The samples with an x ≤ 0.3 composition show an orthorhombic structure, while the intermediate compositions with 0.5 ≤ x ≤ 0.9 show a less crystalline orthorhombic perovskite structure with segregated phases that increase upon increasing the Ni content. The progressive substitution of Fe3+ by Ni2+ in La0.6Ca0.4FeO3 entails the progressive formation of oxygen vacancies and Fe4+ species for Ni substitution degrees of x ≤ 0.5. For larger degrees of substitution with x > 0.7, segregated phases and a progressive loss of the perovskite structure are detected. The highest catalytic performance for the total combustion of DME is obtained from equimolar substitution in the B-site position (Fe0.5Ni0.5), in which a large extent of Fe4+ species and oxygen vacancies are present.


Introduction
Simple perovskites with a basic formula of ABO 3 have been consistently proposed since the 1970s [1] as alternative catalysts for the deep oxidation of hydrocarbons due to their lower cost in comparison with their noble metal counterparts [2].The perovskite lattice tolerates wide variations in the nature of the A and B cations and allows a large number of substitutions involving one or both of the cations [3] [4].Pure and substituted lanthanide perovskites with A = La and B = a first row transition element have received a great amount of attention as oxidation catalysts.It is well established that the nature and extent of the metal substituted in position A may also stabilize an unusual oxidation state of the cation in position B and/or produce oxygen vacancies [5]- [7].In LaFeO 3 , the substitution of the trivalent A-site metal ion with a bivalent metal cation (A') modifies the oxidation state of the B-site metal cation by the formation of structural defects, leading to non-stoichiometric ratios and thus modifying the catalytic activity.Pecchi et al. [8] [9] have reported an increase in the oxygen mobility, reducibility and catalytic reaction rate in partially A-cation substituted LaFeO 3 .Conversely, the partial substitution of cation B by B' also improves the perovskite stability or enhances its redox efficiency [10] [11].A significant enhancement in the catalytic activity in oxidation reactions of LaFeO 3 with a B cation partially substituted with Ni [12] and Zn [13] has also been reported.Noticeable increases in the oxidation reaction rate have also been reported for a double substitution in the A-and B-site positions of pure lanthanum manganite with 10% of Co and Ag, respectively [10].Among the oxidation reactions, dimethyl ether combustion (DME) has been recently studied on Ce-doped manganese oxide octahedral molecular sieves (OMS-2) [14], cryptomelane oxides [15] and layered pillared Al 2 O 3 [16] as an alternative low-cost replacement for liquefied petroleum gas (LPG) and diesel fuel due to its easy transportation, lower emission of polluting particles and higher thermal efficiency.The widely accepted mechanism for catalytic combustion on perovskite-type oxides at lower temperature (<400˚C) is closely related to the interaction of surface oxygen with the reactants.It was proposed that to improve the activity and enhance the redox efficiency of La 0.6 Ca 0.4 FeO 3 , a family of substituted La 0.6 Ca 0.4 Fe 1-x Ni x O 3 , 0 ≤ x ≤ 1.0, perovskites were synthesized and their surface, structural and catalytic properties were characterized.Based on this evidence, the synthesis, characterization and effect of the substitution in La 0.6 Ca 0.4 Fe 1-x Ni x O 3 (x = 0.0, 0.1, 0.3, 0.5, 0.7, 0.9, 1.0) perovskites calcined at 700˚C on the catalytic activity of DME combustion are addressed.

Preparation
Substituted La 0.6 Ca 0.4 Fe 1-x Ni x O 3 (0.0 ≤ x ≤ 1.0) perovskites were prepared by adding stoichiometric amounts of an aqueous solution of the metal nitrates to an aqueous solution of citric acid with a 10% of excess above the number of cations.The resulting solution was stirred for 15 min at room temperature and slowly evaporated at 70˚C under vacuum in a rotary evaporator until gel formation was reached.Then, the gel was dried in an oven, with a slow increase in the temperature up to 250˚C where it was maintained overnight in order to obtain a solid amorphous citrate precursor.The resulting powder was crushed and sieved to obtain the required particle size (<200 μm), and then, it was calcined at 700˚C in air for 10 h.

Characterization
Chemical analysis of the samples was performed by atomic absorption spectrometry (AAS) using a Perkin Elmer model 3100 instrument.The specific areas were calculated using the BET method from the nitrogen adsorption isotherms, which were recorded at the liquid nitrogen temperature on a Model ASAP 2010 Micromeritics apparatus.The amount of tetravalent iron contained in the calcined samples was determined by a redox titration according to the equation Fe 4+ + Fe 2+ = 2Fe 3+ [17] [18] by dissolving, in an acid medium, the samples in a known excess of a Mohr salt standard solution and titrating the excess of Fe(II) with K 2 Cr 2 O 7 .The titration was performed twice for each sample with the reproducibility of the results always within 2%.X-ray powder diffraction (XRD) patterns were obtained with nickel-filtered CuKα 1 radiation (λ = 1.5418Å) using a Rigaku diffractometer and were collected in a 2θ range of 20˚ to 70˚.TPR and oxygen TPD experiments were performed in a TPR/TPD 2900 Micromeritics system with a thermal conductivity detector.Samples of approximately 20 mg were placed in a U-shaped quartz tube, purged in a synthetic air stream for 1 h at 500˚C and then cooled to ambient temperature.The reduction profiles were recorded by passing 5% H 2 /Ar from ambient temperature to 700˚C.For the TPD experiments, the samples were preheated in an O 2 flow for 1 h at 700˚C and then cooled to room temperature in the same atmosphere.After switching the atmosphere to a helium flow, the oxygen desorption was monitored using a TCD.Surface analysis was performed on a VG Escalab 200R electron spectrometer equipped with an Mg Kα X-ray source and a hemispherical electron analyzer.Prior to analysis, all of the samples were degassed at 300˚C for 1 h inside the pre-treatment chamber of the spectrometer.The charging effects on the samples were corrected by taking the C1s peak of adventitious carbon at 284.9 eV.

Catalytic Activity
The evaluation of the catalytic activity in the total DME oxidation was performed in a conventional flow reactor at atmospheric pressure using 0.3 g of the catalyst powder (40 -60 mesh) diluted with 1.5 g of CSi.Prior to the catalytic tests, the samples were pretreated under a purge of the mixed gas at 120˚C for 30 min to stabilize the system.Then, the temperature was linearly increased up to the required temperature, maintained at that temperature for 20 min and raised to a new temperature using the same heating rate (1˚C•min −1 ).Several isothermal steps were performed until complete conversion.The reactant mixture was fed into the reactor to maintain a gas hourly space velocity (GHSV) of 10 dm 3 •g −1 •h −1 , [15] and a reaction mixture of DME:O 2 :He (molar ratio) of 1:10:89 was used.Effluents of the reactor were analyzed using a Hewlett Packard HP 4890D on-line gas chromatograph with a thermal conductivity detector.Helium was used as the carrier gas, and the column used was a 30-m Supelco 25,462 capillary.A Shimadzu GCMS-QP5050 quadrupole mass spectrometer was used to detect small traces of products.

Chemical Composition and Texture
The elemental compositions of the La 0.6 Ca 0.4 Fe 1-x Ni x O 3 perovskites determined by AAS show Fe and Ni contents similar to the nominal ones, as it can be seen in Table 1.The La and Ca contents (not shown) are much closer to the nominal contents of 40.5% and 7.4%, respectively.The specific BET areas, also summarized in Table 1, do not show large differences that are dependent upon the Ni or Fe content, although they are lower for the doubly substituted perovskites [19] [20].

Structural Analysis
X-ray diffraction patterns of the La 0.6 Ca 0.4 Fe 1-x Ni x O 3 oxides after calcination at 700˚C are displayed in Figure 1(a).For the non B-site substituted solids (La 0.6 Ca 0.4 FeO 3 and La 0.6 Ca 0.4 NiO 3 ), the sharp reflections located at the expected 2θ values confirm that Ca 2+ has been built-into the perovskite lattice for only La 0.6 Ca 0.4 FeO 3 , [8] due to that for La 0.6 Ca 0.4 NiO 3 , no perovskite structure is detected.For the A-and B-site substituted solids, two different structural behaviors are observed for lower (x ≤ 0.3) and higher (x ≥ 0.5) degrees of Ni-substitution.For x = 0.0, 0.1, and 0.3, isostructural perovskites corresponding to orthorhombic LaFeO 3 (37 -1493) are formed with no segregated phases of O 3 , thus supporting the hypothesis that small amounts of iron can be substituted by nickel cations [9]. Figure 1(b) shows a close up of the diffraction peaks at 2θ = 57˚.The shift towards higher 2θ values is indicative of a smaller interplanar distance or a decrease in the cell volume, which cannot be attributed to lanthanum or calcium due to their constant compositions.Therefore, the decreases in the cell volume can be achieved by the stabilization of metal ions in a high oxidation state in the perovskite structure.Doping with Ni 2+ into the lattice of the perovskite structure induces the partial oxidation of the B-site cation (Fe 3+ → Fe 4+ ) [21].The differences in the ionic radii of Fe 4+ (0.0585 nm) and Fe 3+ (0.0645 nm) explain the corresponding shift in Figure 1(b), which is in agreement with previous results reported for other iron-based perovskites [12] [18].For a higher Ni content, 0.5 ≤ x ≤ 0.9, the most intense diffraction peak of the perovskite structure appears at a higher 2θ value (58˚), indicative of the formation of a less crystalline orthorhombic La-NiO 3 (34 -1028) perovskite structure with a number of segregated phases, which were identified as CaO (28 -0775), La 2 NiO 4 (11 -557) and NiO (22 -1189).These results indicate that iron ions remain incorporated within an orthorhombic perovskite structure only for x ≤ 0.3.For larger Ni contents, a large amount of segregated phases as well as the formation of a less crystalline perovskite structure are obtained.

Surface Structure
The XPS experiments of O1s, Ca2p ) such as hydroxyl and/or carbonate groups at 532.3 eV [22]- [24].Table 2 summarizes the BEs of La3d 5/2 , Fe2p 3/2 and Ni2p 1/2 core-levels as well as the extent (%) of the contribution of each peak for the Ca2p 3/2 and O1s.As expected for La-containing perovskites, the BE of C1s at 289.8 eV, which is not included in the Table, does not change and corresponds to the presence of surface carbonates [25].The BE of the most intense peaks of doublets for La3d 5/2 at 834.8 eV and Fe2p 3/2 at 710.3 eV as well as the least intense peak for Ni2p 1/2 at 873.4 were almost constant even though the satellite line placed at ca. 881.2 eV is indicative of the presence of Ni 2+ ions on the surface.It must be emphasized that the higher BE contribution of the La3d 3/2 peak overlaps with the Ni2p 3/2 peak.This overlapping may mask not only the accurate measurement of the BE of nickel but also its intensity.To overcome this complication, the Ni2p 1/2 was measured.The surface Ca2p 3/2 at 346.8 eV, identified with a larger oxidized calcium content, increases upon increasing Ni contents, which is indicative of a larger extent of surface highly dispersed cal-    cium oxides [26].The changes in the extent of the stronger electrophilic weakly adsorbed oxygen species (O 2ads 2− and O ads − ) at BEs lower than 530.9 eV could be related to electronic changes [27] [28] in the perovskite structure, and the O1s species higher than 531 eV could be related to the thermal treatment before the XPS experiment.The surface atomic ratios for Ni/La + Ca, Ca/La + Ca and Fe/La + Ca calculated using atomic sensitivity factors for the Mg Kα source [22] are displayed in Figure 3.For comparison, the bulk ratios were also incorporated.It can be seen that the Ni/La + Ca surface ratio is always higher compared to the nominal ones, which is indicative of Ni-enrichment related to the increase of NiO and La 2 NiO 4 segregated phases.The Ca/La + Ca surface ratio is slightly higher than that from the nominal compositions and did not show significant differences that were dependent upon the Ni content.For the surface Fe/La + Ca ratio, two different trends can be seen.The ratio is lower than the nominal one for lower degrees of Ni substitution (x ≤ 0.3) and the same value as the nominal one for higher Ni contents (x ≥ 0.5).This result indicates a surface Ni-enrichment in the total studied extent of Ni, which is indicative that separate Ni-La(Ca)-O species are developed throughout the entire compositional range with no segregation of iron species.

Redox Properties
To gain insight into the changes in the oxidation state of the B-cations, redox titrations and O 2 -TPD profiles were carried out.In Table 1, the extent of Fe 4+ , calculated from redox titrations, shows larger values for x = 0.3 and 0.5.Due to the increases in the Ni content, which decrease the Fe extent, the normalized atomic (Fe 4+ /total Fe) ratio based upon the Ni content is shown in Figure 4. Again, the progressive increase of the Fe 4+ /Fe total ratio is a measurement of the progressive insertion of Ni 2+ into the perovskite structure up to x = 0.5.The larger extent of segregated phases for higher Ni contents (x > 0.7) and the smaller total Fe content increases both the Fe 4+ /total Fe ratio and the uncertainly of the value, and these measurements do not help to obtain clear conclusions.In the O 2 -TPD profiles shown in Figure 5, the expected physisorption occurred at approximately 90˚C, and two chemisorbed oxygen species between 200˚C and 600˚C [29] are observed.The lower temperatures (T < 300˚C) correspond to weakly chemisorbed oxygen on oxygen vacancies [30], and the highest temperatures (300˚C < T < 600˚C) correspond to oxygen species from boundaries or defects [31] called surficial oxygens [32].For lower Ni contents (x ≤ 0.3), only a broad desorption peak at 300˚C with a shoulder at 240˚C is observed, whereas for higher Ni contents (x ≥ 0.5), a new and well-defined oxygen desorption peak at 340˚C, with no shift   dependent upon the Ni content, indicates the appearance of a large extent of new chemisorbed oxygen species.These new chemisorbed oxygens for higher degrees of Ni substitution (x ≥ 0.5) can be associated with the large extent of segregated phases [31] in line with the XPS and XRD results.Because TPD-MS experiments confirm that the evolved gas and He flow only contain oxygen, the deconvolution of the oxygen desorption curves using a Lorentzian peak shape allow the calculation of the amount of desorbed oxygen.The total chemisorbed oxygen species between 200˚C and 600˚C based upon the Ni content shown in Table 1 indicate formation of oxygen vacancies achieved, at lower substitution levels of x ≤ 0.5 fundamentally by the oxidation of a portion of the iron ions from Fe 3+ to Fe 4+ .At higher Ni doping levels the oxygen vacancies can be related to the large extent of surface segregated phases.The maximum is reached for the x = 0.5 perovskite.The TPR profiles under a H 2 /Ar flow up to 700˚C are shown in Figure 6.The 0.1 ≤ x ≤ 0.9 solids show an intermediate behavior between the lowest La 0.6 Ca 0.4 FeO 3 and the largest La 0.6 Ce 0.4 NiO 3 reducible solids.The reduction profile for La 0.6 Ca 0.4 FeO 3 shows a single reduction peak at approximately 260˚C, which can be attributed to the reduction of Fe 4+ to Fe 3+ [9], which increases in the solids up to x ≤ 0.7.The shift towards lower temperatures indicates an increase in and the easier reducibility of Fe 4+ as a consequence of the presence of Ni 2+ in the perovskite lattice [18] [20] [33].For the higher Ni substituted perovskites (x ≥ 0.5) the TPR profiles do not show such a clear trend.

Catalytic Activity
The catalytic activity for DME combustion under an excess of oxygen was measured as a function of the temperature up to complete combustion.Water and CO 2 were the only observed products.The typical experimental sigmoidal curves displayed in Figure 7 indicate that the reaction starts at 170˚C until complete conversion at 300˚C.To quantify the catalytic behavior, in Table 3 the ignition temperatures (T i 50 ), defined as the temperatures to achieve 50% of reactant conversion, and the reaction rate at a low conversion level (<10%) evaluated at 195˚C are compiled.The larger catalytic activity, i.e., the lowest ignition temperature and larger reaction rate, is shown by the equimolar La 0.6 Ca 0.4 Fe 0.5 Ni 0.5 O 3 perovskite.A decrease in the reaction rate and a parallel increase in the ignition temperature for both the lower x ≤ 0.3 and higher x ≥ 0.7 values are also detected.It has been noted that structural defects and oxygen mobility are the most important factors controlling the catalytic activity of perovskites in oxidation reactions [34].Moreover, the catalytic activity for the combustion reaction has been found to be closely related to the activation of gaseous oxygen species and chemisorbed oxygen species [30] [35] as well as the redox properties of Fe 4+ species [8].Therefore, the fairly good correlation between the reaction rate and chemisorbed oxygens shown in Figure 8 supports the importance of desorbed oxygen in the studied systems.Due to that the Fe content is not constant in the studied samples; it has not been possible to search for a correlation between the reaction rate and the extent of Fe 4+ .

Conclusion
The improvement in the catalytic activity by the progressive substitution in the B-site position of Fe 3+ by Ni 2+ in La 0.6 Ca 0.4 Fe 1-x Ni x O 3 is driven by the formation of oxygen vacancies as well as iron ions in unusual oxidation states (+4).This improvement at lower substitution levels of x ≤ 0.5 is achieved fundamentally by the oxidation of a portion of the iron ions from Fe 3+ to Fe 4+ .At higher Ni doping levels, a large number of oxygen vacancies, segregated phases and a progressive decrease in the perovskite structure are found.The linear dependency of the reaction rate on the oxygen vacancies indicate that the active sites for the DME combustion are directly related to reversible adsorbed oxygen.

3 / 2 , 2 .
Fe2p 3/2 and Ni2p 1/2 spectra for a representative sample are shown in Figure The broad peak for Ca2p 3/2 can be deconvoluted into two peaks and the O1s into three peaks.The three components obtained from the fitting of the O1s peak can be assigned to the following: 1) surface lattice oxygen species (O latt 2− ) at 529.2 eV; 2) weakly adsorbed oxygen species (O 2ads 2− or O ads − ) at 531.0 eV arising from lattice [La-O-M] bonds; and 3) loosely adsorbed super peroxide species (O 2ads −

Table 1 .
Fe and Ni bulk composition (nominal values in parentheses) in wt.%, specific area, extent of Fe 4+ (%) and total desorbed oxygen.