The physiochemical properties of the perovskite-type catalysts (taken from ref.17, reproduced by permission of The Royal Society of Chemistry).
Three series of perovskite-type catalysts, i.e., La–M–Mn–Cu–O (M = Mg, Y, Zn, Ce), La–M–Cu–Zn–O (M = Ce, Mg, Zr, Y), and La–Mn–Zn–Cu–O, were designed and applied in CO2 hydrogenation to methanol. The materials were characterized by XRD, N2-adsorption, N2O-adsorption, ICP-OES, XPS, and TPD techniques. Perovskite structures were observed and the ‘‘metal on oxide’’ could be realized via reduction. Upon the introduction of the fourth elements, more structure defects, smaller particles, higher Cu dispersion, larger amount of hydrogen desorption at low temperature, and more amount of basic sites were obtained. The selectivity for methanol and the TOF values were higher for the catalysts derived from perovskite-type precursors. The catalytic performance was related to Cuα+ and/or Cu0 species, low-temperature H2 adsorption on the unit, and the weak basic sites.
Perovskite-type oxides have received significant attention because of their important electric, magnetic, ferromagnetic, pyroelectric, and piezoelectric properties [1,2]. Recently, much attention has been paid to perovskite-type oxides as catalysts due to their high activity and thermal stability. For a typical ABO3 perovskite, A-site is a larger rare earth and/or alkaline earth cation and B-site is a smaller transition metal cation. In such structure, the A-site keeps the structure and the B-site provides the catalytic activity site. B-site cations could be reduced to well-dispersed metallic species supported on the A-site cations oxide, which leads to ideal catalyst precursors for many reactions that involve metal as active sites [3,4]. Besides, perovskite-type A2BO4 mixed oxides with the K2NiF4 structure, consisting of alternating layers of ABO3 perovskite and AO rock salt, have also been studied , which exhibit variable oxygen stoichiometry. The replacement of A-site and/or B-site cations by other metal cations often results in the formation of crystal microstrain and adjustable activity .
CO2 is the main greenhouse gas, and various strategies have been implemented to reduce its concentration [7-10]. An important CO2 utilization is the hydrogenation to methanol, which is considered as the most valuable product since it can be used as solvent, alternative fuel, and raw material, and it can be converted to olefins, aromatics, or gasoline derived from traditional petrochemical processes [11,12].
The synthesis of methanol over Cu/ZnO-type catalysts has been studied for many years. However, several important problems still remain open, such as the working oxidation of copper and the reaction mechanism [13-15]. In addition, the low activity and stability of catalysts, which are partly attributed to Cu sintering accelerated by the presence of the by-product, water vapor, create major barriers for practical application . It is found that the catalysts with higher Cu dispersion, easier reduction property, and better adsorption properties for relative gases could achieve better catalytic performance for methanol synthesis .
Few work on the application of Cu-based perovskite-type oxides for CO2 hydrogenation has been investigated. In the present work, three series of perovskite-type-based catalysts were prepared and tested for CO2 hydrogenation to methanol, and the relationship between physical–chemical property and catalytic performance was discussed.
2. Catalyst preparation
The perovskite-type oxides were prepared by sol-gel method using citric acid as complexing agent. The precursor salts were La(NO3)3. nH2O; Mn(NO3)2, 50% solution; Cu(NO3)2. 3H2O; Mg(NO3)2. 6H2O; Y(NO3)3. 6H2O; Zn(NO3)2. 6H2O; Ce(NO3)3. 6H2O; ZrO(NO3)2.2H2O. Adequate amounts of the precursor salts along with citric acid were dissolved in deionized water at a molar ratio of 2:1 (metal cations: citric acid). The solution was heated to 353 K to remove the water, and then the temperature was increased to 423 K until ignition. The resulting powder was finally calcined under air at 673 K for 2 h and then at 1073 K for 4 h. The three series of catalysts were: (1) For doped La–M-Mn–Cu–O based (M= Mg, Y, Zn, Ce) perovskite materials, the ratio for La, M, Mn, Cu is 0.8: 0.2: 0.5: 0.5. The La–Mn–Cu–O catalyst and Mg, Y, Zn, Ce doping catalysts were then denoted as P, Mg–P, Y–P, Zn–P, and Ce–P, respectively. (2) A series of La–M-Cu–Zn–O (M= Ce, Mg, Zr, Y) based perovskite-type catalysts, i.e., LaCu0.7Zn0.3Ox, La0.8Ce0.2Cu0.7Zn0.3Ox, La0.8Mg0.2Cu0.7Zn0.3Ox, La0.8Zr0.2Cu0.7Zn0.3Ox and La0.8Y0.2Cu0.7Zn0.3Ox samples were prepared, of which the subscripts were the nominal composition. The catalysts were then denoted as LCZ-173, LCCZ-8273, LMCZ-8273, LZCZ-8273 and LYCZ-8273, respectively. (3) The LaZn0.4Cu0.6Oy, LaMn0.1Zn0.3Cu0.6Oy, LaMn0.2Zn0.2Cu0.6Oy, LaMn0.3Zn0.1Cu0.6Oy, and LaMn0.4 Cu0.6Oy samples were prepared, of which the subscripts were the nominal composition. The catalysts were denoted as LZC-046, LMZC-136, LMZC-226, LMZC-316, and LMC-406, respectively.
3. Results and discussion
3.1. Performance of the La–M-Mn–Cu–O (M = Mg, Y, Zn, Ce) based perovskite precursors
3.1.1. Textural and structural properties
As shown in Figure 1a , for the La–M-Mn–Cu–O (M = Mg, Y, Zn, Ce) based perovskite precursors, the LaMnO3 (JCPDS # 75-0440) are the main phase. The diffraction peak at about 2θ=32.5° shift towards higher values when the fourth elements were doped. Small peaks at 2θ=35.6°and 38.9°assigned to CuO (JCPDS # 89-5899) appear in the doped samples but not in the P sample. In all diffraction patterns, no phase that ascribes to Mg, Y, or Zn is observed, while a new phase ascribed to CeO2 is found for the sample of Ce–P, which demonstrates that it is difficult for all the Ce to enter the perovskite lattice, which agrees with the conclusion by Weng et al. .
For all reduced samples (Figure 1b), LaMnO3 phase is still the main phase, which reveals that the reduction process does not destroy the perovskite structure. Meanwhile, the CuO phase disappears and the Cu phase emerges.
The physicochemical properties of the calcined perovskite-type catalysts are summarized in Table 1 . Low specific surface area for perovskite-type oxides is common. For this series samples, the largest one is only 11.3 m2 g-1 for Y–P and the lowest one is only 4.1 m2 g-1 for the Zn–P. The exposed Cu surface area and the Cu dispersion measured by N2O adsorption technique are also low, which even cannot be measured for both P and Ce–P. Y–P possesses the largest Cu surface area and the Cu dispersion by comparison. The lower copper surface area may not be favorable for the conversion of CO2 to methanol . The ICP results show that the experimental lanthanum amount is lower than the theoretical value, and other element amounts are similar to the nominal values.
3.1.2. The XPS investigations
The XPS results presented in Table 2  show that lanthanum ions of the reduced perovskite-type catalysts are present in the trivalent form because the La3d5/2 peak is close to the value of pure lanthana at 834.4 eV . However, the BE of La3d5/2 of P sample is higher than the other samples, which implies the increasing of the electron cloud density around La ions for the doped samples. It may due to the fourth elements affect the transfer of the electrons of La to O, since O has the highest electronegativity value among all elements . For O1s, the binding energy at around 528.9–529.1 eV is assigned to the lattice oxygen (O2-) [21,22] and the binding energy at around 530.8–533.0 eV is ascribed to the adsorbed oxygen species (Oad) in the surface, which contains hydroxyl (OH-), carbonate species (CO32-), and molecular water. The binding energy decreases after the fourth element except Ce adding, which indicates that there are more electrons around oxygen. It is likely that the fourth components transfer the electronic to the oxygen. The presence of surface adsorbed oxygen species suggests the formation of oxygen vacancies in the defected oxides , which is favorable for the activation of the catalyst. The Oad/O2- ratio is increased for the doped samples, which implies the improvement of catalysis activity. The binding energy values of Mn2p3/2 for the perovskite-type catalysts are located at 641.3 eV–642.2 eV. The peak positons of level of MnO, Mn2O3, and MnO2 are 640.6, 641.9, and 642.2 eV, respectively. The values are very similar, and the mean oxidation state of Mn ions at the surface layers is extremely difficult to detect by XPS, as reported in other studies [24,25]. However, the previous reports suggested that the BE difference between Mn2p3/2 and O1s emissions increases with about 0.6–0.7 eV for the change of the oxidation state between Mn3+ and Mn4+. As shown in Table 2, the BE difference is in the range of 112.3–113.0 eV, i.e., increasing with 0.7 eV, which means a change of the Mn4+/Mn3+ ratio for the perovskite-type catalysts [26,27].
Since the binding energy of the Cu2p3/2 band in the metal (932.6 eV) and in Cu+ (932.4 eV) is almost same, they can be distinguished by different kinetic energy of the Auger Cu LMM line position in Cu0 (918.6 eV), Cu+ (916.7 eV), or in Cu2+ (917.9 eV) [19,28]. The Auger electron spectroscopies of Cu LMM of reduced samples are shown in Figure 2 . The profiles are convoluted into two peaks. It can be seen that the majority of the copper species exist as Cu+ for all samples, which is in accordance with the report of Jia et al. . The weak Cu0 peak could be the explanation for the immeasurable of exposed Cu0 for P and Ce–P (Table 1).
3.1.3. H2-TPD and CO2-TPD analysis
Two CO2 desorption peaks are observed for all samples (Figure 3), which are denoted as peak α and peak β . The peak α at around 400 K could be assigned to weak basic sites and the peak β at around 600 K could be assigned to strong basic sites. With the introduction of the fourth components, the peak α shifts to higher temperature, while the peak β shifts to lower temperature, which indicate the increase of the weak basic sites’ strength but the decrease of the strong basic sites’ strength. The strength for the weak basic sites of the catalysts increases in the order of: P < Ce–P < Y–P < Mg–P < Zn–P. The amount of the basic sites is also changed with the fourth element doping. The quantitative analysis for the CO2-TPD based on the relative area of the profiles is listed in Table 3, in which the P sample is assigned as 1.00. Both the weak basic sites and the strong basic sites increase due to the alkalinity of Mg for Mg–P. For Y–P, the amount of total basic sites and strong basic sites improved remarkably with the amount of weak basic sites’ decreasing. Moreover, the amount of the weak basic sites increases, but the amount of the strong basic sites and total basic sites decrease for Zn–P and Ce–P samples.
|1.00 (383 )|
3.1.4. Catalytic performance
The catalytic performances of the catalysts for CO2 hydrogenation to methanol are summarized in Table 4 . Both the CO2 conversion and methanol selectivity are improved when the fourth element is added. With the introduction of Zn, the activity increased greatly, which might be due to the fact that the active site is Cuδ+–O–Zn2+ [31,32]. However, the activity improvement is slight for Ce-P. The relationship between the CO2 conversion and the amount of H2 desorption on unit surface area below 523 K (Table 3) is shown in Figure 4 . It can be seen that the more the H2 that is desorbed on the unit area, the more the CO2 that is converted. Lower CO2 conversion may result from lower copper content as well as lower surface area of copper in the system. The results of Figure 5  show that the trend of the weak basic sites’ strength and methanol selectivity is similar, which indicates their dependency.
3.2. Performance of the La-M-Cu-Zn-O (M = Ce, Mg, Zr, Y) based perovskite precursors
3.2.1. Textural and structural properties
Figure 6  shows that the orthorhombic perovskite structure (A2BO4) with high degree of crystallinity of the La2CuO4 (JCPDS # 82-2142) is the main phase and two small peaks at 2θ=35.6˚ and 38.9˚ ascribed to the CuO phase (JCPDS # 89-5899) present in all samples. The weak peak at 2θ=36.3˚ attributed to the ZnO phase (JCPDS # 80-0075) appears in all samples except LMCZ-8273, which implies that Mg seems to have a special effect on the structure. The phase of Ce7O12 and La2Zr2O7 appear in the sample of LCCZ-8273 and LZCZ-8273, respectively, while there are no new phases containing Mg and Y appearing in the sample of LMCZ-8273 and LYCZ-8273.
The crystallographic parameters of the prepared materials were calculated by employing least-squares refinement, assuming an orthorhombic crystal system for the samples, and the results are listed in Table 5 . A certain degree of changes of the lattice parameters occured after the fourth component was introduced. The lattice parameters a, b, and c were lower than those of LCZ-173, which can be attributed to the shrinkage of the La2CuO4 due to the introduction of the fourth elements. The mean grain size of La2CuO4 calculated by the Scherrer equation shows that the particles size of the La2CuO4 decreased remarkably for LCCZ-8273, LZCZ-8273, and LMCZ-8273, but slightly for LYCZ-8273. The physicochemical properties of the calcined catalysts are summarized in Table 6 . The BET surface area for all calcined samples are rather low (SBET < 3 m2 g-1), which is common for perovskite-type of materials . It can be seen that the highest specific surface area is just only 2.3 m2 g-1 for LZCZ-8273. Moreover, the tendency of the exposed Cu surface area and the Cu dispersion measured by N2O adsorption is the same for the materials. The LZCZ-8273 shows the highest Cu surface area (SCu) and the best dispersion of copper (DCu). The physicochemical properties of LYCZ-8273 are similar to that of the LCZ-173, which indicates that the influence of the Y doping is negligible.
|LYCZ-8273||0.7||4.5||3.2|3.2.2. XPS investigations
The reduced perovskite-type catalysts are analyzed by XPS, and the binding energies (BE) of La3d5/2 and Zn2p3/2 are presented in Table 7 . According to the literature, La3d5/2 features in perovskite structure are located at 837.5 and 834.3 eV [20,34] which are close to the values of pure lanthana at 837.8 and 834.4 eV, indicating that lanthanum ions are present in the trivalent form. A slight shift in the La3d5/2 binding energy is observed upon introduction of the fourth elements and the values are in the range of 837.86–838.01 eV and 834.06–834.36 eV, respectively. Small differences may relate to the changes in crystal structure and/or electronic structure. In addition, small changes are also observed for the binding energy of Zn at around 1021.7 eV for different samples. The Auger electron spectroscopies of Cu LMM of reduced samples are shown in Figure 7 . A broad peak appears in the range of 915.0 eV–920.0 eV for all samples, and it is hard to distinguish the Cu+, Cu2+, and Cu0 apparently. However, the peaks at around 918.6 eV attributed to Cu0 are distinct for all samples. A new peak appears at around 911.2–914.3 eV, lower than that of Cu+, which is defined as peak α, implying that a special Cu species Cuα+ exists in the perovskite system. The presence of Cu+ may accelerate the reduction of CO2 to CO (RWGS) ,while the Cuα+ (not Cu2+, Cu+, Cu0) plays an important role for the methanol synthesis from CO2/H2 [31,35].
The X-ray photoelectron spectroscopies of the fourth elements in the reduced samples suggest that both Ce3+ and Ce4+ exist in the LCCZ-8273 and the +4 oxidation state is predominant. The result agrees with the XRD analysis. The Zr in the LZCZ-8273 sample exists in the phase of La2Zr2O7.
3.2.3. Catalytic performance
The catalytic performance for La-M-Cu-Zn-O (M = Ce, Mg, Zr, Y) catalysts are listed in Table 8 . The LMCZ-8273 shows the highest CO2 conversion and the maximum yield of methanol despite the lowest selectivity among all the samples. The LMCZ 8273 shows the highest methanol selectivity and the LCCZ-8273 shows moderate CO2 conversion and methanol selectivity. The lowest CO2 conversion and less improvement for methanol selectivity are observed for LYCZ-8273. The varying of the CO2 conversion had the same tendency as the surface area of copper (Figure 8 ), indicating more surface copper existing in the catalysts may lead to higher activity, i.e., Cu0 is the active site for CO2 hydrogenation to methanol [10,16,19,22,35]. It is noteworthy that all catalysts show promising CH3OH selectivity, especially for LMCZ-8273. The order of the selectivity to CH3OH is as follows: LMCZ-8273 > LCCZ-8273 > LYCZ-8273 > LCZ173 > LZCZ-8273. The relationship between CH3OH selectivity and the Cuα+ Auger peaks is shown in Figure 9 . It can be seen that Cuα+ had a strong effect on the selectivity for methanol: the lower the binding energy of the peak α, the higher is the CH3OH selectivity. Cu+ is favorable for the reduction of CO2 to CO (RWGS), so it can be derived that the farther away from 916.6 eV (the binding energy of Cu+ in Cu LMM) for the peak α, the higher the CH3OH selectivity that can be obtained. As discussed above, doping of Mg leads to the proper oxide state of copper, which results in the best selectivity for methanol. For LCCZ-8273 and LYCZ-8273, Ce and Y substitute La in the A-site with the same charge (+3) and similar ionic radius, which produces more defects in the perovskite structure that causes the special oxide state for copper species. With the special structure of La2CuO4 perovskite, the high dispersed copper species can be realized and stronger physical and electric interaction between the copper and other metal oxides can be obtained, which may lead to the formation and stabilization of the copper species with special valence . However, for the LZCZ-8273, the formation of lanthanum zirconium pyrochlore has little influence on the perovskite structure but a great influence on the content of La2CuO4 perovskite, which may lead to the lowest selectivity of methanol. The turnover frequency (TOF), which represents the number of CO2 molecules hydrogenated in a unit site per second (s-1), is calculated from the exposed copper surface area for the perovskite-type catalysts (Table 8 ). The TOF values of the perovskite-type catalysts were very high compared with other catalytic systems [22,35], indicating the better efficiency for copper atoms on perovskite-type catalysts.
3.3. Performance of the La–Mn–Zn–Cu–O based perovskite precursors
3.3.1. Textural and structural properties
The XRD patterns of the fresh and reduced perovskites are presented in Figure 10a and b, respectively . La2CuO4 perovskite-like structure can be observed for all fresh samples. LaMnO3 phase emerges and the La2CuO4 phase transfers from tetragonal (JPCDS 81-2450) to orthorhombic (JPCDS 81-0872) as the manganese is introduced, which indicates that the manganese introduction distorts the structure of the La2CuO4. With the increasing of the manganese amount, the intensity of the La2CuO4 phase decreases while that of the LaMnO3 phase increases, which implies that the formation of LaMnO3 is easier than that of La2CuO4. This phenomenon reveals that the structure of LaMnO3 is more stable than La2CuO4. Small peaks for both CuO and ZnO can also be observed except LMC-046, which indicates the perovskite structure has certain tolerance for the involved elements for this perovskite-type system. Moreover, the peak intensity of the separated CuO decreases when the value of Mn/Zn decreases, which means the formation of LaMnO3 can lead to the separation of copper from the La2CuO4 perovskite structure. For the reduced sample (Figure 10b), the La2CuO4 perovskite structure disappears and the metallic copper and La2O3 is observed, which indicates that the ‘‘metal-on-oxide’’ can be attained. The appearance of La0.974Mn0.974O3 phase reveals that the reduction progress can result in defects rather than destruction for the Mn-based perovskite.
The crystallographic parameters of the prepared materials were calculated by employing least-squares refinement and the results are listed in Table 9 . It can be found that the axes are elongated for the four element samples, which means that doping may strut the perovskite structure. The size of La2CuO4 crystallites becomes smaller with the introduction of manganese. The LaMnO3 phase changes from cubic to orthorhombic structure as zinc is introduced. Moreover, the LMZC-136 and LZC-046 possess the smallest LaMnO3 and the largest La2CuO4 crystallites among all the samples, which implies that the abundant zinc species can improve the formation of small LaMnO3 and large La2CuO4 crystallites. The change of lattice parameters of the samples implies that interaction between the involved elements might be different.
The BET specific surface area along with the exposed surface copper area and the copper dispersion measured by N2O adsorption are summarized in Table 10 . The specific surface area for all samples is low, which is common for perovskite-type materials . LMC-406 possesses the largest specific surface area (SBET), the exposed surface copper area (SCu) as well as the copper dispersion (DCu), while the LZC-046 sample shows the lowest SCu and DCu, which indicate that the existence of LaMnO3 perovskite structure is favorable for increasing the surface copper area due to the extension of the space structure for the samples with manganese (Table 9).
3.3.2. XPS investigations
The XPS results of the reduced perovskite-type catalysts are listed in Table 11 . The values of La3d5/2 binding energy (BE) are located in the range of 834.2–834.6 eV, demonstrating that La ions are in the trivalent form for all samples. Small changes of Zn (around 1021.7 eV) and Mn (around 642.0 eV) BE may relate to the small distortions in electronic structure and/or crystal structure. For the O1s patterns, the peak at around 528.2–529.3 eV can be attributed to the oxygen ions in the crystal lattice (O2-) and the peak at around 531.1–531.5 eV can be assigned to the adsorbed oxygen species (Oad) derived from the defects or oxygen vacancies in the structure . The O1s BE shifts to lower value with the decreasing of the Mn/Zn ratio, which suggests the increasing of electron cloud density around O element. The value of Oad/O2- is max for the LZC-046, which decreases for the Mn containing samples, indicating that the LaMnO3 could reduce the structural defects.
For this series catalyst, the binding energies of Cu2p3/2 are lower than that for the copper oxide (933.0 eV) apparently, indicating that the Cu atoms are not in the simple copper oxides form. Figure 11  shows the Auger electron spectroscopies of Cu LMM for the reduced samples. A broad peak can be observed, which can then be deconvoluted into three peaks. The peak at around 916.5 and 919.0 eV matched with kinetic binding energy of Cu+ and Cu0 within the error limit, respectively. However, a new peak at around 911.2–913.2 eV is observed which may be ascribed to the Cuα+. According to literatures and our works, Cuα+ can be appeared in perovskite-type system.
3.3.3. Catalytic performance
The performance of the La–Mn–Zn–Cu–O based perovskite catalysts for methanol synthesis from CO2 hydrogenation is shown in Table 12 . The LMC-406 shows the worst performance despite the largest surface area and exposed copper surface area. The LZC- 046, which also contains three metal elements, but Zn instead of Mn, shows a moderate catalytic performance. It is well-known that the site of Cu+-O-Zn2+ favors the adsorption of hydrogen that can transport to the bulk copper species via spillover [20,39]. So the lack of the site of Cu+-O-Zn2+ may be the reason for the poor catalytic performance of LMC-406. Moreover, the TOFCu value increases sharply upon Zn introduction, which verifies that the copper sites are not the only active sites for the reaction . With more reducible copper species and high TOFCu values, the four-component samples show a good catalytic performance. The ratio of both Cuα+ and Cu0 species to the total copper species (calculated from the Auger spectroscopy (Figure 11)) shows the same change tendency with the CO2 conversion (Figure 12) , indicating that both Cuα+ and Cu0 species could be the active sites for the conversion of CO2. In addition, with the change of the ratio for Mn/Zn, the synergy between copper and the other components might vary, and then the reduction state of copper species (Cuα+ and Cu0) changes. The four-component samples with two kinds of perovskites show better methanol selectivity, which implies that the strong synergy for different elements and different perovskite phases are significant for the improvement of the catalytic performance. In addition, it is also found that the lower the BE of the Cuα+, the higher is the CH3OH selectivity (Figure 13) .
Three series of catalysts derived from perovskite-type precursors were prepared by sol-gel method, which were applied in the CO2 hydrogenation to methanol. The conclusions are as follows:
The perovskite-type ABO3, A2BO4, and ABO3+A2BO4 can be purposive obtained in our work. Cu-based perovskite-type catalyst shows good methanol selectivity in CO2 hydrogenation to methanol.
For La–M–Mn–Cu–O (M = Mg, Y, Zn, Ce) catalysts, the introduction of the fourth elements leads to the separation of copper from the LaMnO3 perovskite lattice and thus produces more oxygen vacancies. Because of the increasing of defects, temperature adsorption properties are improved for the doped samples. The CO2 conversion is related to the amount of absorbed H2 on the unit area under 523 K, and the methanol selectivity corresponds to the strength of the weak basic sites. The catalytic performance enhanced considerably for Zn–P based on this, which also implies that Zn is important to the CO2 hydrogenation to methanol.
La2CuO4 (A2BO4) perovskite structure is obtained for the La–M–Cu–Zn–O (M = Ce, Mg, Zr, Y) samples. With the addition of Ce, Mg, and Zr, good properties can be obtained: smaller particle size, higher Cu dispersion, larger amount of hydrogen desorption at low temperature, higher concentration of basic sites, and so on. The excellent methanol selectivity originates from the special copper valence that presents in the perovskite structure after reduction, and the CO2 conversion is in correlation with the surface area of metallic copper.
Both La2CuO4 and LaMnO3 perovskite structure can be observed in the LaMn0.4-xZnxCu0.6Oy (x = 0, 0.1, 0.2, 0.3, 0.4) catalysts. The ‘‘metal-on-oxide’’ can be realized after reduction. With decreasing of Mn/Zn, more oxygen defects were formed. The perovskites exhibited better methanol selectivity due to the appearance of Cuα+ derived from the abundant defects of perovskite structure and the strong interaction between different elements. Moderate LaMnO3 can balance the defects in the structure, and then lead to the perfect Cuα+, which is important for the methanol selectivity.
This work was financially supported by the Key Science and Technology Program of Shanxi Province, China (MD2014-10), the National Key Technology Research and Development Program of the Ministry of Science and Technology (2013BAC11B00), and the Natural Science Foundation of China (21343012).