Over the past several decades,catalytic oxidation of CO has attracted considerable attention due to its potential applications in many fields such as automobile exhaust,fuel cell technology,carbon dioxide laser and air purification in an enclosed atmosphere [1-3]. Among the catalysts suitable for this purpose,supported noble metals initially seemed to be good candidates. However,owing to high prices and low reserves of noble metals,nowadays,perovskite oxides such as the lanthanum manganites are potential substitutes for noble metal catalysts because of their unique properties such as their catalytic activity for CO oxidation [4].
In the LaMnO3 perovskite oxide,Mn3+ ions are coordinated with six oxide ions and the MnO6 units are connected in a corner-sharing manner making networks in which La3+ ions are located in the dodecahedral sites of the framework [5]. Due to its structural flexibility,the La3+ and/or Mn3+ ions in the lattice can be partially replaced by foreign cations (transition or rare metal ions) with no large change in the crystalline structure to produce substituted perovskites [6]. This property has frequently been exploited because the substitution usually promotes catalytic activity [7-10]. Substitution of La3+ ions with a divalent metal cation is accompanied by a change of the oxidation state of the manganese ions,thus leading to a modification in the catalytic activity for CH4 combustion [7]. Ponce et al. [8] studied the partial substitution Sr2+ for La3+ ions in La1-xSrxMnO3 (x = 0-0.5) oxides and found that with the substitution amount increasing,to balance the total charge,the charge compensation was achieved by the oxidation of Mn3+ to Mn4+. The presence of Mn4+ cations in La1-xSrxMnO3 oxide enhanced the catalytic activity by influencing the active oxygen near the Mn4+ that takes part in methane oxidation [9]. Nevertheless,the substitution of Mn3+ ions in LaMnO3 oxide proved to be a far more effective method to optimize the catalytic activity than choosing different lanthanides to substitute for La3+ ions in A sites. Geng [10] revealed that compared with partial substitution of A sites in La1-xSrxMnO3 oxide with Sr2+ ions,the ignition temperature of LaMn1-yCuyO3 oxides with an equal substitution amount of B sites with Cu2+ ions was far lower in the catalytic oxidation of CO.
On account of the high catalytic activity of the substituted perovskite,many studies on the improvement of catalytic activity have been carried out. Among these,increasing the specific surface area was considered to be an effective strategy. Hence,different preparation routes and technical conditions such as a solid state reaction,sol-gel,hydrothermal treatment,spray-drying,freeze-drying,co-precipitation,chemical vapor deposition,microemulsion [11] and templates [12] have been used to prepare substituted perovskite catalysts with a high specific surface area to boost the catalytic performance for CO oxidiation.
However,there have been few reports of perovskite-type oxides with specific morphologies prepared by the metal salt method. Liu et al. [13] reported that perovskite-type ACeO3 (A = Sr,Ba) was synthesized by the metal salt method at low temperatures in eutectic NaCl-KCl. Chen et al. [14] revealed that the perovskite single crystalline LaMnO3 nanocubes have been fabricated by a facile salt-assisted solution combustion process. Wang et al. [15] studied a morphology-controlled molten salt route developed to synthesize porous spherical LaMnO3 and cubic LaMnO3 nanoparticles using the as-prepared porous Mn2O3 spheres as a template.
In this work,first,hierarchical porous δ-MnO2 microspheres were synthesized by a hydrothermal reaction at 80 °C. Then,in order to decide the optimal calcination temperature,the LaMn0.8Fe0.2O3 catalysts were synthesized at different temperatures (450,550 and 650 °C) using the synthesized δ-MnO2 microspheres as a template in eutectic NaNO3-KNO3. Finally,La1-xSrxMn0.8Fe0.2O3 (0 ≤ x ≤ 0.6) catalysts obtained at the optimal calcination temperature were used to investigate the effect of Sr2+ doping on the catalytic activity for CO oxidiation.
All the chemical reagents were analytical purity and used without any further purification. Distilled water was used in the synthesis and treatment processes.
Hierarchical porous δ-MnO2 microspheres were synthesized by a hydrothermal reaction. In a typical procedure [16],first,0.9 g of KMnO4 and 2 ml of 37 wt% HCl were dissolved into 80 mL of distilled water at room temperature. After continuous magnetic stirring for 30 min,the homogeneous purple solution was sealed in a 100 mL Teflon-lined autoclave and maintained at 80 °C for 10 h in an oven. After cooling down to room temperature,the products were separated by filtration,washed with distilled water and absolute ethanol several times,and then dried in an oven at 60 °C for 12 h. The product was denoted as S80.
The spherical porous La1-xSrxMn0.8Fe0.2O3 (0 ≤ x ≤ 0.6) catalysts,referred to as LSMF (X) (0 ≤ X ≤ 0.6),were prepared by the molten salt method,using the synthesized hierarchical porous δ-MnO2 microspheres as a template. A mixture of NaNO3 and KNO3 with a molar ratio of 1.5:1 was adopted as the molten salt. Stoichiometric amounts of La(NO3)3•6H2O,Sr(NO3)2,as-synthesized MnO2 and Fe(NO3)3•9H2O were well mixed and ground in an agate mortar for 20 min. The nitrate and as-synthesized MnO2 to molten salt molar ratio was of 1:15. Then,in order to decide the optimal calcination temperature,the LaMn0.8Fe0.2O3 samples were heated at different temperatures (450,550 and 650 °C) with a heating rate of 5 °C/min for 4 h,respectively denoted as S450,S550 and S650. After cooling to room temperature in air,the products were immersed into deionized water in a beaker to transfer the product into an aqueous solution for the subsequent separation and purification. The remnant molten salts were removed from the products by washing with hot deionized water several times until there was no white precipitate on the surface. The synthesized products were finally dried at 80 °C for 12 h. The La1-xSrxMn0.8Fe0.2O3 (0 ≤ x ≤ 0.6) catalysts were synthesized at the optimal calcination temperature to investigate the effect of Sr2+ doping on the catalytic activity for CO oxidiation.
Powder X-ray diffraction (XRD) patterns were obtained using a D/max-RB type X-ray diffractometer with a step size of 0.02° and Cu Kα radiation (λ = 0.15418 nm). The 2θ scan range was 10°-80°. Fourier transform infrared spectroscopy (FT-IR) spectra were recorded on a Nicolet 6700 spectrometer (Thermo,USA) using KBr pellets in the wavenumber range of 4000-400 cm-1. Scanning electron microscope (SEM) images were obtained on a JSM-IT300 electron microscope (JEOL,Japan). Field emission SEM (FESEM) and transmission electron microscope (TEM) analysis as well as the corresponding selected area electron diffraction (SAED) patterns were conducted on a Zeiss Ultra Plus-43-13 (German) and JEM-2100F electron microscope (JEOL,Japan),respectively. The specific surface area (ABET) and porosity were determined by nitrogen adsorption isotherms at -196 °C using a Micromeritics ASAP 2020 M sorption analyzer (USA). The Brumauer-Emmett-Teller (BET) model was applied for specific surface area determination and the Barrett-Joyner-Halanda (BJH) model was used for porosity characterization using the adsorption branches of the isotherms. Sample analyses were performed after outgassing in vacuo (residual pressure ~10-2 mbar) for 5 h at 120 °C. X-ray photoelectron spectroscopy (XPS) was obtained on an ESCALAB 250 Xi system (Thermo,China) using the monochromatized Al Kα radiation (hυ = 1486.6 eV) at room temperature and a vacuum of 10-7 Pa (10-9 Torr). The error of the binding energy was ±0.2 eV and the C 1s (Eb = 284.6 eV) of the contaminated carbon was used as a standard for binding energy calibration.
Catalytic performance for CO oxidation was examined in a fixed-bed quartz microcatalytic flow reactor (i. d. = 8 mm) at atmospheric pressure. The reactor was positioned in an electrical furnace provided with a thermocouple on its wall. The temperature of the reactor was maintained by a temperature controller. Approximately 100 mg catalyst powders mixed evenly with glass-asbestos (catalyst:glass-asbestos = 1:5) were loaded into the reactor in each trail. Carbon monoxide,oxygen and argon were regulated using mass flow controllers. The molar composition of the feed was CO/O2/Ar (balance) = 1/1.25/97.75 and the total flow rate of the gas mixture was 438 mL/min giving a space velocity (GHSV) of 12000 h-1. Catalytic runs were performed at temperatures ranging from room temperature to 250 °C. The system outlet was connected to a gas chromatograph (GC-7890 Ⅱ,China) equipped with FID and TCD double detectors to separate reactants and products. The detected compounds were CO,O2 and CO2.
Shown in Fig. 1 are the XRD results of the synthesized precursor and LaMn0.8Fe0.2O3 samples by heating the mixture of nitrates and the precursor in the molten salt of NaNO3 and KNO3 at different temperatures for 4 h. It was clearly seen that there were only four very weak peaks indexed to pure layered birnessite-type MnO2 (JCPDS 80-1098,monoclinic,C2/m,a = 0.515 nm,b = 0.284 nm,c = 0.717 nm) in Fig. 1(1). MnO2 can form many polymorphs by interlinking the basic MnO6 octahedra in different configurations. Birnessite-type MnO2 with a two-dimensional (2D) lamellar structure is usually denoted as δ-MnO2. Generally,layered birnessite-type MnO2 needs more cations to stabilize the structure owing to an interlayer distance of 0.73 nm in the birnessite. The insertion/extraction of cations results in the coexistence of Mn4+ and Mn3+ ions in the δ-MnO2. In addition,it was clearly observed that there was only the LaMn0.8Fe0.2O3 phase similar to the cubic LaMnO3 (JCPDS 75-440,Pm-3m,a = 0.388 nm) and no other impurity was detected (Fig. 1(2),(3) and (4)). With a rise in calcination temperature from 450 to 650 °C,the intensity of the diffraction peaks decreased significantly,indicating that in eutectic NaNO3-KNO3,single phase LaMn0.8Fe0.2O3 can be generated from the mixture of nitrate and δ-MnO2 at a low temperature and a further increase in sintering temperature was beneficial for the improvement of crystallinity.
The morphology of S80 from the hydrothermal reaction was obtained by FESEM and TEM. As shown in Fig. 2 (a-1) and (a-2),S80 has a hierarchically porous structure consisting of 4.0 μm sized microspheres assembled from interleaving thin nanosheets in agreement with the study by Xiao et al. [17]. Seen from the TEM images (Fig. 2(a-3)),the darker/lighter contrast indicated that the microspheres clearly exhibited a core-corona structure and further demonstrated that pores exist not only on the surface but also in the interior. As can be seen from the HRTEM image in Fig. 2(a-4),an average lattice spacing was calculated to be 0.275 nm in good agreement with that (0.284 nm) of the (010) plane of δ-MnO2. From the corresponding SAED pattern (inset in Fig. 2(a-4)),one can see that there were only three weak diffraction spots,indicating that S80 with bad crystallinity is a single crystal in good agreement with the XRD results.
Fig. 2(b) shows the FESEM and TEM images of S450. Obviously,S450 still maintained the porous spherical structure composed of aggregated nanocrystallites and the size of the pores become smaller (Fig. 2(b-1)). The TEM image (Fig. 2(b-3)) further showed that these pores existed not only on the surface but also within the microspheres. In addition,the HRTEM image in Fig. 2(b-4) indicated that the lattice spacing of 0.27 nm was consistent with that (0.274 nm) of the S450 (110) plane and the single crystalline character is also demonstrated by the clear diffraction spots in the Fourier transform of a small well-ordered region (inset in Fig. 2(b-4)). However,as shown in Fig. 2(b-2),the hierarchically porous structure similar to δ-MnO2 was also found in S450 but the nanosheets were thicker and decorated with nanoparticles on the surface. Therefore,it can be deduced that the growth processes of the porous LaMn0.8Fe0.2O3 microspheres included several stages. In the initial stage,La(NO3)3,Sr(NO3)2 and Fe(NO3)3 dissolved into the salt and then diffused onto the surface of the nanosheets due to the much better solubility than the δ-MnO2 microspheres. Then the LaMn0.8Fe0.2O3 phase was formed in situ and because of the preservation of the hierarchically porous spherical structure,the morphology of S450 resembled that of δ-MnO2. Finally,the nanosheets were gradually transformed into nanoparticles.
In order to further understand the effect of calcination temperature on the morphology of LaMn0.8Fe0.2O3 oxides,the morphology of the products was observed by SEM. Fig. 3 gives the SEM images of S550 and S650. As observed from Fig. 3(a) and (b),in S550,part of LaMn0.8Fe0.2O3 microspheres were smashed into semi-spherical or axiohitic particles,surrounded by agglomerated nanocrystallites. As the calcination temperature continued increasing to 650 °C,fewer spherical LaMn0.8Fe0.2O3 oxides were observed and part of the remaining spherical LaMn0.8Fe0.2O3 oxide was covered with agglomerated nanocrystallites (Fig. 3(c) and (d)),indicating that the nanocrystallites were generated by peeling from the surface to the core of the LaMn0.8Fe0.2O3 microspheres at the high temperature.
Fig. 4 shows the nitrogen adsorption-desorption isotherms and corresponding pore size distribution curves of the δ-MnO2 and LaMn0.8Fe0.2O3 oxide obtained at different sintering temperatures. According to IUPAC classification,S80 and S450 samples showed a Type IV isotherm with H3 hysteresis loop in the relative pressure (p/p0) range of 0.4-1.0 (Fig. 4(a)),indicating the existence of a mesopore structure [18]. The N2 adsorption quantity of S80 was higher than S450,suggesting a more porous structure and thus larger surface area and pore volume in the S80 [19],which was also be confirmed by the results shown in Table 1. S550 and S650 gave a Type V isotherm with a very minor hysteresis loop at high relative pressures associated with a low pore volume (Fig. 4(c)). On the other hand,the isotherm shapes of the four samples differ from each other,also indicating that there exists differences in pore distribution. This deduction was substantiated by the results of the pore size distributions (Fig. 4(b) and (d)). For S80 and S450,the pore size distributions cover a wide range from 2 to 100 nm. In fact,the pores were mostly composed of mesopores (2-50 nm) and they also contain a small number of macropores (> 50 nm). Compared with S80,the average pore diameter of S450 was smaller. However,S550 and S650 both possessed two peaks including a sharp peak at 2.55 nm formed from aggregated nanocrystallites,and a weak peak covering a wider range from 12.5 to 125 nm,perhaps induced by the fracture of microspheres and collapse of the porous structure in fair agreement with the SEM results.
The textural parameters including the specific surface area,mean pore diameter and average pore volume are summarized in Table 1. Clearly,S450 has a larger specific surface area (55.73 m2/g),average pore diameter (9.38 nm) and pore volume (0.37 cm3/g). The specific surface area,average pore diameter and average pore volume decreased with the calcination temperature increase. Fig. 5 shows the X(CO) curves of S450,S550 and S650. It can be clearly seen that for S450,S550 and S650,X(CO) increased with a rise in reaction temperature,and as the calcination temperature increased,the catalytic performance of LaMn0.8Fe0.2O3 oxide decreased. Especially,for S650,X(CO) at 250 °C was less than 10% on account of a smaller surface area,which was detrimental to the adsorption of CO and O2 molecules and the diffusion of adsorbed CO species and adsorbed oxygen species [20]. Nevertheless,compared with S450,the temperature (190 °C) of 50% X(CO) (T50) for S550 was 3 °C higher. Hence,a conclusion can be obtained that spherical porous LaMn0.8Fe0.2O3 oxide with a better CO oxidation activity was synthesized at 450 °C for 4 h in eutectic NaNO3-KNO3.
To identify the effects of Sr2+ substitution amount on the catalyst structure,the FT-IR spectra of LSMF (X) (0 ≤ X ≤ 0.6) are shown in Fig. 6. The broad vibration band at 3404 cm-1 was attributed to the stretching vibration of the -OH group from the water molecules [21]. The weak band at 1635 cm-1 was normally associated with the O-H bending mode and the peak at 1470 cm-1 corresponded to the bending vibration of O-H bonds connected to Mn atoms [22]. The sharp band at 1348 cm-1 as well as the peak at 850 cm-1 was assigned to the vibration of residual NO3- on the surface of the catalysts [23] and the peak at 1050 cm-1 was ascribed to trans- and cis-hyponitrite (N2O22-) species [24]. Besides,the FT-IR peaks in the range of 400-1000 cm-1 revealed the existence of octahedral MnO6 [25]. However,compared with the stretching vibration peak of Mn-O bond in the MnO6 octahedron at 600 cm-1 in a previous study [26],the peak at 530 cm-1 associated with the vibrations of Mn-O bonds was slightly shifted towards lower wavenumbers,which was assigned to Fe3+ ion doping [27]. Furthermore,with the Sr2+ substitution amount increasing,the peak intensity at 530 cm-1 trailed off gradually,indicating that Sr2+ ions had indeed entered A sites,inducing the binding force between Mn and O to decrease. In other words,as Sr2+ ions substituted for La3+ ions,to ensure electrical neutrality,the oxidation state of manganese ion would increase from Mn3+ to Mn4+ and there existed not only DE interactions of Mn3+-O-Mn3+ or Fe3+-O-Mn3+ but also Mn3+-O-Mn4+. As a result,the variation of the Mn-O bond length and angle led to the structural distortion of FeO6/MnO6 (BO6) octahedrons [28],weakening the stability of octahedron structure.
Fig. 7 gives the results of the spherical porous LSMF (X) (0 ≤ X ≤ 0.6) oxide for CO catalytic oxidation. It is evident that a little substitution of Sr2+ for La3+ ions can slightly improve the catalytic activity for CO oxidation,while the catalytic activity was greatly improved when the substitution amount was more than 30%. As shown in Table 2,for LSMF (X) (X = 0,0.15,0.3),50% X(CO) was achieved at 186 °C and 90% X(CO) was achieved at 198 °C. However,for LSMF (X) (X = 0.45,0.6),T50 values were both far below 125 °C. The temperature of 90% X(CO) (T90) of LSMF (0.6) was still less than 125 °C,while the temperature of LSMF (0.45) was 181 °C. T90 of LSMF (X) (X = 0.45,0.6) was less than 275 °C in our previous study [10] and T90 of LSMF (0.6) was lower than that (175 °C) in Ref. [20]. This indicated that when the substitution amount of Fe3+ for Mn3+ ions in the B sites was 20%,it greatly enhanced X(CO) through increasing the substitution amount of Sr2+ for La3+ions in the A sites (more than 30%). This is because of the formation of more Mn4+ ions and oxygen vacancies at a higher substitution content to retain electric neutrality,which was confirmed by XPS analysis.
Fig. 8 illustrates the Mn 2p3/2 and O 1s XPS spectra of LSMF (X) (X = 0.45,0.6). It is observed from Fig. 8(a) and (b) that there was an asymmetrical Mn 2p3/2 peak,which could be decomposed into two components at Eb ≈ 641.6 and 642.4 eV. The former was assigned to surface Mn3+ species whereas the latter to surface Mn4+ species [29]. By curve fitting,one can decompose the asymmetrical O 1s XPS peak into three components at Eb ≈ 528.6,531.2 and 532 eV,which were due to surface lattice oxygen (Olat),adsorbed oxygen (Oads) and hydrated oxide species (Fig. 8(c) and (d)) [30]. It can be observed from Table 3 that the surface Mn4+/Mn3+ molar ratio (1.14) of LSMF (0.6) was much higher than that (0.92) of LSMF (0.45) and the Oads/Olatt molar ratio (1.09) of LSMF (0.6) was also higher than that (0.93) of LSMF (0.45). This result indicated that the surface of LSMF (0.6) exposed more abundant surface Mn4+ species and adsorbed oxygen species. The existence of more Mn4+ ions increased the oxygen vacancies of an oxide material and surface oxygen vacancies is beneficial to the adsorption,activation and migration of oxygen [31],which would accelerate the formation of more adsorbed oxygen species. In addition,adsorbed oxygen species played an important role in oxidation reactions [32, 33]. So,it can be concluded that the existence of more Mn4+ ions,surface oxygen vacancies and adsorbed oxygen species contributed to a higher catalytic performance for CO oxidation on LSMF (0.6).
To evaluate the catalytic stability,on-stream reaction experiments were carried out over LSMF (0.45) and LSMF (0.6) at 250 °C (where 100% X(CO) was achieved) and 500 °C (a temperature higher than the sintering temperature in the preparation of the samples) for 10 h,respectively. As shown in Fig. 9(a) and (b),for LSMF (X) (X = 0.45,0.6),only about 0.27% and 0.18% change in X(CO) at 250 °C for 10 h were detected. In other words,there was no significant decrease with reaction time and 100% X(CO) was indeed achieved at 250 °C on LSMF (X) (X = 0.45,0.6). In addition,the variation of X(CO) for both samples within 10 h at 500 °C was less than 0.025%. After reaction at 500 °C for 10 h,X(CO) with the increase of reaction temperature was tested again,the results shown in Fig. 9(c). Obviously,for LSMF (0.6),X(CO) under different reaction temperatures changed a little and even between 150 to 225 °C,X(CO) was higher than that before reaction at 500 °C for 10 h. For LSMF (0.45),after 125 °C,it also exhibited a higher catalyst activity. By comparison,it can be concluded that the spherical porous LSMF (X) (X = 0.45,0.6) samples were catalytically stable under the reaction conditions.
By the molten salt method,porous LaMn0.8Fe0.2O3 microspheres with good crystallinity,high specific surface area and better catalytic performance for CO oxidation were obtained in two stages at 450 °C for 4 h in eutectic NaNO3-KNO3 using hierarchical porous δ-MnO2 microspheres assembled from interleaving thin nanosheets as the precursor template. In addition,under the conditions of CO/O2/Ar (balance) molar ratio = 1/1.25/97.75 and GHSV = 12000 h-1,compared with the other samples prepared by the same method,the La0.4Sr0.6Mn0.8- Fe0.2O3 sample exhibited the best catalytic performance and stability for CO oxidation due to the existence of more Mn4+ ions,surface oxygen vacancies and adsorbed oxygen species.
2016,
Vol. 37 Issue (8): 1431-1439 
PDF (1476 KB)
2016,
Vol. 37 Issue (8): 1431-1439 PDF (1476 KB)