Recently, the harmful nature of nitrogen oxides (NO_x) has become a serious environmental issue that has attracted public attention. It has been recognized that NO_x are major atmospheric pollutants, mainly emitted from stationary sources, such as coal-fired power plants, and mobile sources, such as motor vehicles. These emissions not only cause acid rain and photochemical smog, but also act as a precursor for airborne particulate matter (PM_2.5), which can damage human health and plant growth ^[1-3]. Elimination of NO_x is an urgent environmental concern. Among currently available technology, selective catalytic reduction of NO_x by NH₃ (NH₃-SCR) has been proved to be an efficient, reliable, and economical way to control NO_x emissions ^[4].

A V₂O₅/TiO₂ sample promoted by WO₃ or MoO₃ is the commercial catalyst for NH₃-SCR process, which exhibits excellent catalytic performance between 300 and 400 ℃ ^[5-7]. However, to reach these operating temperatures, the NH₃-SCR device has to be located upstream of electric precipitator and desulfurizer units in a power station setting. Under these conditions, the catalyst may become blocked and deactivated by dust and sulfur-containing compounds in flue gas ^{[3, 5, 8]}. Furthermore, there may not be enough space to place an NH₃-SCR device before the electric precipitator and desulfurizer units in old coal-fired power plants. Therefore, there is demand for low-temperature (below 200 ℃) NH₃-SCR catalysts that would allow the NH₃-SCR device to be placed downstream of the electrostatic precipitator and desulfurizer units, after dust and SO₂ have been eliminated ^{[3, 9, 10]}.

In recent years, considerable efforts have been devoted to developing novel NH₃-SCR catalysts that show good low-temperature catalytic performance. Some examples include supported noble-metal catalysts ^[11], metal-ion exchanged molecular sieves ^[12-14], and manganese-based catalysts ^{[5, 8, 9, 15-18]}. Among these, manganese-based catalysts have attracted considerable attention owing to their good redox properties and excellent oxygen migration ability, which are beneficial for oxidation of NO to NO₂, and lead to relatively high conversion of NO_x at low temperatures through a "fast NH₃-SCR" route ^{[3, 19]}. However, the low N₂ selectivity, poor H₂O resistance, and narrow operating temperature window of manganese-based catalysts require further improvements ^[20-22]. Efficient low-temperature NH₃-SCR catalysts are required, which not only exhibit high catalytic activity at low temperatures, but also have a wide operation temperature window, good H₂O resistance and high N₂ selectivity at low temperatures.

Ceria (CeO₂) has been widely used to eliminate NO_x owing to its good redox properties and high oxygen storage/release capacity associated with oxygen vacancies in the material and the Ce⁴⁺/Ce³⁺ redox couple ^{[2, 23-29]}. It has also been reported that CeO₂ can promote adsorption of NO_x, provide stronger Brnsted acidic sites, enhance oxidation of NO to NO₂, and improve the water and sulfur resistance of catalysts, which would be particularly beneficial for the NH₃-SCR model reaction ^{[10, 19, 26, 30]}. Therefore, a combination of manganese oxide (MnO_x) and CeO₂ to form MnO_x-CeO₂ catalysts might improve the materials catalytic performance for the NH₃-SCR of NO_x. This is because of the synergistic interaction between MnO_x and CeO₂, and other advantages. Shen et al. ^[31] synthesized a supported MnO_x/CeO₂ catalyst for the low-temperature NH₃-SCR model reaction, which showed more than 90% NO conversion between 120 and 220 ℃ owing to a high dispersion of MnO_x, favorable redox properties, and good adsorption of oxygen species. However, the operating temperature window, H₂O resistance, and N₂ selectivity of this catalyst need further improvement.

It has been widely reported that the physicochemical properties and catalytic performance of redox catalysts are highly dependent on their preparation methods ^{[22, 32]}. In the present work, we synthesized a series of MnO_x-CeO₂ (MnCe) catalysts for a low-temperature NH₃-SCR model reaction, using several different preparation methods. The obtained samples were characterized by X-ray diffraction (XRD), Raman spectroscopy, N₂ physisorption, H₂ temperature-programmed reduction (TPR), NH₃ temperature-programmed desorption (TPD), X-ray photoelectron spectroscopy (XPS), and in situ diffuse reflectance Fourier-transformed infrared spectroscopy (DRIFTS). Moreover, a NO+NH₃+O₂ model reaction was chosen to evaluate the catalytic performance of these catalysts. The purpose of this paper is to first examine the influence of different preparation methods on the physicochemical properties and catalytic performance of MnO_x-CeO₂ catalysts and to screen for an optimal preparation method. Second, we investigate the interaction of NH₃ and NO+O₂ with these catalysts by in situ DRIFTS in the temperature range of 25-350 ℃ to further understand the reaction mechanism of NH₃-SCR over MnO_x-CeO₂ catalysts.

The MnCe-MMM catalyst was prepared by a mechanical mixing method. MnO_x and CeO₂ were obtained by separate thermal decompositions of Mn(NO₃)₂ and Ce(NO₃)₃·6H₂O at 500 ℃ for 5 h in air, after grinding the precursors in an agate mortar. The desired amounts of MnO_x and CeO₂ were then mixed together in an agate mortar for 1 h to prepare the sample.

The MnCe-IM catalyst was prepared by an impregnationmethod. CeO₂ (obtained by thermal decomposition of Ce(NO₃)₃·6H₂O at 500 ℃ for 5 h in air) was impregnated with an aqueous solution containing the desired amount of Mn(NO₃)₂ for 2 h. The resulting solid was then heated at 100 ℃ using an oil bath to evaporate residual water. The sample was dried at 110 ℃ in an oven for 12 h and finally calcined at 500 ℃ for 5 h in air.

The MnCe-HTM catalyst was prepared by a hydrothermal treatment method. The required amounts of Mn(NO₃)₂ and Ce(NO₃)₃·6H₂O were dissolved in deionized water and stirred together for 2 h, followed by introduction of NaOH (6 mol/L) to achieve a solution of pH 10. The resulting solution was stirred for another 3 h to obtain a suspension and then transferred to a Teflon autoclave. The mixture was heated to 120 ℃ at a rate of 2 ℃/min and held at that temperature for 24 h. The obtained precipitate was centrifuged, washed several times with deionized water and one time with anhydrous ethanol until the solution pH did not change. After drying at 60 ℃ for 12 h in an oven, the sample was finally calcined at 500 ℃ for 5 h in air.

The MnCe-CPM catalyst was prepared by a co-precipitation method. The required amounts of Mn(NO₃)₂ and Ce(NO₃)₃·6H₂O were dissolved in deionized water and stirred together for 2 h, and then added dropwise to an excess of ammonia (25%) with magnetic stirring until the solution reached pH 10. The resulting solution was stirred for another 3 h, aged for 24 h, then filtered, and washed with deionized water until the solution pH did not change. The obtained precipitate was dried at 110 ℃ for 12 h and then calcined at 500 ℃ for 5 h in air.

The MnCe-SGM catalyst was prepared by a sol-gel method. Briefly, the desired amounts of Mn(NO₃)₂ and Ce(NO₃)₃·6H₂O were dissolved in deionized water and stirred together for 2 h. Citric acid (C₆H₈O₇/(Mn+Ce) = 0.3 molar ratio) was then introduced into the mixed solution, and stirred for another 2 h. Subsequently, the obtained solution was heated to 50 ℃ to react for 3 h, and then heated at 100 ℃ with stirring to evaporate residual water. Finally, the sample was dried at 110 ℃ for 12 h and calcined at 500 ℃ for 5 h in air.

For all catalysts, the molar ratio of Ce to Mn was fixed at 7:3. The chemical reagents used in the present work were of analytical grade and used without further purification.

X-ray diffraction (XRD) patterns of the synthesized samples were recorded on a Philips X’Pert3 Powder diffractometer using Ni-filtered Cu K_α radiation (λ = 0.15418 nm). The X-ray tube was operated at 40 kV and 40 mA. The intensity data were collected over a 2θ range of 10°-80°. The scan speed was set at 10°/min with a step size of 0.02°. The mean crystallite size (D_β) and lattice parameters (a) were determined from the strongest peak of these samples by D_β = Kλ/βcosθ (Debye-Scherrer equation) andd = 3^1/2nλ/2sinθ (Bragg equation), respectively. In these equations K is the particle shape factor, taken as 0.89, λ is the X-ray wavelength, β is the full-width at half maximum height (FWHM) in radians, θ is the diffraction angle, d is the interplanar spacing, and n is the diffraction order number.

Raman spectra of the synthesized samples were collected on a Renishaw inVia Reflex Laser Raman spectrometer using an Ar⁺ laser beam. The Raman spectra were recorded with an excitation wavelength of 532 nm and a laser power of 5 mW. The intensity data were collected from 200 to 1000 cm^-1.

Surface characteristics of the synthesized samples were obtained by N₂ physisorption at -196 ℃ on a Belsorp-max analyzer. We used the Brunauer-Emmet-Teller (BET) method to determine the specific surface area and the Barrett-Joyner-Halenda (BJH) method to determine the pore distribution. Prior to each analysis, the catalyst was degassed under vacuum at 300 ℃ for 4 h.

The H₂-TPR experiments were performed in a quartz reactor connected to a thermal conductivity detector with an H₂-Ar mixture (7.0% H₂ by volume, 30 mL/min) as the reductant. Prior to reduction, the sample (50 mg) was pretreated in a highly purified N₂ stream at 300 ℃ for 1 h and then cooled to room temperature. The H₂-TPR was started from 100 ℃ and increased to the target temperature at a rate of 10 ℃/min.

The NH₃-TPD experiments were performed on a multifunction chemisorption analyzer with a quartz reactor, and measured by a thermal conductivity detector. About 200 mg of catalyst was pretreated by a flow of high purity N₂ (30 mL/min) at 300 ℃ for 1 h. After the pretreatment, the catalyst was saturated with NH₃-N₂ mixture (1.0% NH₃ by volume, 30 mL/min) at 50 ℃ for 1 h and then flushed with a flow of high purity N₂ (30 mL/min) at the same temperature for 1 h to remove gaseous NH₃, before allowing the catalyst to cool to room temperature. The catalyst was then heated from room temperature to 600 ℃ at a rate of 10 ℃/min in a flow of high purity N₂ (30 mL/min).

X-ray photoelectron spectra (XPS) of the synthesized samples were performed on a PHI 5000 VersaProbe system, with monochromatic Al K_α radiation (1486.6 eV) operating at an accelerating voltage of 15 kW. Before the measurement, the catalyst was outgassed at room temperature in a UHV chamber (< 5×10^-7 Pa). Sample charging effects were compensated by calibrating all binding energies (BE) to the adventitious C 1s peak at 284.6 eV, which gave BE values with an accuracy of ±0.1 eV.

The in situ DRIFTS of the representative samples were collected on a Nicolet 5700 FT-IR spectrometer equipped with a high-sensitivity mercury cadmium telluride detector cooled by liquid N₂. The DRIFTS cell (Harrick) was fitted with a ZnSe window and a heating cartridge that allowed the sample to be heated to 350 ℃. A fine catalyst powder placed on the sample holder was carefully flattened to enhance IR reflection. Before the DRIFTS test, the sample was pretreated by a flow of high purity N₂ at 300 ℃ for 1 h to remove physisorbed water. A sample background at each target temperature was collected during the cooling process. At room temperature, the sample was exposed to a controlled stream of NH₃-N₂ (1% NH₃ by volume) or/and NO-N₂ + O₂-N₂ (1% NO and 5% O₂ by volume) at a rate of 50 mL/min for 1 h to achieve saturation. For in situ DRIFTS of NH₃ adsorption, residual gaseous NH₃ was purged by a high purity N₂ stream (50 mL/min) for 1 h at room temperature. Desorption/reaction studies were performed by heating the samples and the spectra of adsorbed species were recorded at a rate of 10 ℃/min from room temperature to 350 ℃, by subtraction of the corresponding background reference.

The catalytic performance of the synthesized samples for the NH₃-SCR model reaction in the presence of excess oxygen was determined in a steady state, involving a feed stream with a fixed composition: 500 ppm NO, 500 ppm NH₃, 5% O₂, 5% H₂O (when used), and N₂ in balance. The catalyst (100 mg) was placed a quartz tube, pretreated in a high purity N₂ stream at 300 ℃ for 1 h and then cooled to room temperature, before switching on the mixed reaction gases. The reactions were carried out at different temperatures with a space velocity of 60000 mL g^-1 h^-1. The concentrations of NO, NH₃, NO₂, and N₂O were detected at 150 ℃ by a Thermofisher IS10 FTIR spectrometer equipped with a 2 m path-length gas cell (250 ml volume). The NO conversion (%) and N₂ selectivity (%) were calculated from the following equations:

(1)

(2)

The catalytic activity and N₂ selectivity of these synthesized catalysts for the selective catalytic reduction of NO by NH₃ (NH₃-SCR model reaction) in the presence of excess oxygen are shown in Fig. 1. The catalytic activity of CeO₂ was negligible (< 30%) over the whole reaction temperature range (50-350 ℃), as shown in Fig. 1(a), and increased only slightly at higher temperatures. MnO_x exhibited a considerable catalytic activity in the range 50-350 ℃, but this did not exceed 60%. Compared with CeO₂ and MnO_x, the prepared MnO_x-CeO₂ catalysts featured remarkably high catalytic performance (except for MnCe-MMM) owing to the synergistic interaction between MnO_x and CeO₂. Moreover, the catalytic activity of the MnO_x-CeO₂ catalysts obtained by different preparation methods exhibited similar temperature dependences, first increasing to a maximum value and then declining as temperature was increased. The temperature increases provided energy which enhanced the interactions between MnO_x and CeO₂ and in turn promoted transformation of NO, accelerating the catalytic reduction of NO by NH₃. The increased movement and collision of reactant molecules at higher temperatures caused a clear enhancement of the catalytic activity in the low temperature range (below 200 ℃). However, an NH₃ oxidation side reaction, which consumed the NH₃ reductant to generate N₂O, NO, and NO₂, dominated the reaction at higher temperature, leading to decreased catalytic activity and selectivity for N₂. In addition, we found that the catalytic activity of these MnO_x-CeO₂ catalysts depended on the method used to prepare the catalyst, in the order MnCe-HTM > MnCe-SGM > MnCe-CPMM > MnCe-IM > MnCe-MMM, over the whole reaction temperature range (50-350 ℃). The MnCe-HTM catalyst not only showed the highest catalytic activity but also possessed the widest operating temperature window, exhibiting a catalytic activity of nearly 100% in the range 100-325 ℃. Fig. 1(b) shows the N₂ selectivity of the synthesized samples and exhibited a similar trend, which declined with increasing temperature due to generation of N₂O and NO₂ from NH₃ oxidation. The MnO_x catalyst showed the worst N₂ selectivity, but was clearly improved when combined with CeO₂ to form MnO_x-CeO₂ catalysts. Notably, the N₂ selectivity of the MnCe-HTM catalyst (having the highest catalytic activity) was greater than 70% when the temperature was no higher than 200 ℃.

Water is an inevitable contaminant in the practical denitrification (deNO_x) process, so it is important to investigate the influence of water on the catalytic performance of our deNO_x catalysts. Therefore, the H₂O resistance of the MnCe-HTM catalyst during the NH₃-SCR model reaction at 200 ℃ was evaluated through a long-time test, and the corresponding results are displayed in Fig. 2. For the first 3 h, without H₂O, we observed that the MnCe-HTM catalyst showed full NO conversion and greater than 70% N₂ selectivity in the NH₃-SCR of NO. When H₂O was introduced into the catalytic reaction system, the NO conversion declined slightly from 100% to around 80%-90%, while the N₂ selectivity increased from 70% to 90% in the presence of 5% H₂O. This was attributed to the effects of competitive adsorption among NO, NH₃, and H₂O ^[33]. Furthermore, both the NO conversion and N₂ selectivity were stable during the whole test range of H₂O resistance (3-45 h), which indicated good H₂O resistance of the MnCe-HTM catalyst. When H₂O was removed (45-48 h), the NO conversion and N₂ selectivity of MnCe-HTM catalyst were recovered to some extent. Thus, the MnCe-HTM catalyst exhibited nearly 100% NO conversion in the range 100-325 ℃, above 70% N₂ selectivity in the range 50-200 ℃, and good H₂O resistance at 200 ℃ for the selective catalytic reduction of NO by NH₃. These results indicate that the hydrothermal treatment is an effective method for preparing MnO_x-CeO₂ catalysts for low-temperature deNO_x processes. To gain further insight into the catalytic performance of the MnCe-HTM catalyst and to understand the differences in the performance of catalysts produced by different methods, we characterized their structural and morphological properties.

Fig. 3(a) shows the XRD patterns of the MnO_x-CeO₂ catalysts obtained by different preparation methods. The results of the MnO_x and CeO₂ catalysts are also shown in this figure for comparison. Diffraction peaks were detected in MnO_x, which were attributed to a large amount of Mn₂O₃ (PDF-ICDD 41-1442) and a small amount of MnO₂ (PDF-ICDD 44-0141). The CeO₂ samples showed a typical cubic fluorite structure (PDF-ICDD 34-0394) with several characteristic diffraction peaks. The results of the MnO_x-CeO₂ catalysts showed some interesting phenomena. For the MnCe-MMM catalyst, strong diffraction peaks assigned to CeO₂ and several weaker characteristic peaks attributed to Mn₂O₃ were detected. These resulted from the mechanical mixing of CeO₂ and MnO_x. However, a weak diffraction peak from MnO₂ appeared at 37.3° instead of that of Mn₂O₃, which was detected in the MnCe-IM catalyst. We attribute this to Mn species on the surface of CeO₂ that were fully oxidized during the calcination process. Interestingly, only weakened and broadened diffraction peaks from CeO₂ were observed from MnCe-HTM, MnCe-CPM, and MnCe-SGM catalysts. The absence of MnO_x diffraction peaks indicated that Mn species were highly dispersed on the CeO₂ surface and/or incorporated into the CeO₂ lattice. Moreover, we found that the diffraction peaks from the MnCe-HTM, MnCe-CPM, and MnCe-SGM catalysts shifted to slightly higher diffraction angles, compared with those in CeO₂. This result further confirmed that Mn species were partially incorporated into the CeO₂ lattice forming a uniform ceria-based solid solution (containing a Mn-O-Ce structure) ^[34-36].

The structural data of these synthesized catalysts are summarized in Table 1. The lattice parameters of CeO₂, MnCe-MMM, and MnCe-IM catalysts were very close, indicating that the mechanical mixing and impregnation methods had a similar influence on the bulk structure of the CeO₂. Interestingly, the lattice parameters of MnCe-HTM, MnCe-CPM, and MnCe-SGM catalysts were smaller than those of bulk CeO₂. This was likely because the ionic radii of Mn⁴⁺ (0.54 Å), Mn³⁺ (0.66 Å), and Mn²⁺ (0.80 Å) are smaller than that of Ce⁴⁺ (0.92 Å). Incorporation of these Mn ions into the CeO₂ lattice led to lattice contraction and distortion ^[37-39]. The crystallite size of these synthesized catalysts was also studied. Both CeO₂ and MnCe-MMM showed similar crystallite sizes, while the MnCe-IM catalyst was larger than that of CeO₂ due to the second calcination. Notably, the MnCe-HTM, MnCe-CPM, and MnCe-SGM catalysts exhibited crystallite sizes that were smaller than those of bulk CeO₂, which suggests that the incorporation of Mnⁿ⁺ inhibited the grain growth of CeO₂ ^{[37, 40]}.

The structures of the different MnO_x-CeO₂, MnO_x, and CeO₂ catalysts were further characterized by Raman spectroscopy, and the corresponding results are shown in Fig. 3(b). Two weak Raman bands were observed at 310 and 702 cm^-1 for MnO_x, which were attributed to out-of-plane bending and symmetric stretching modes of Mn₂O₃, respectively ^{[41, 42]}. CeO₂ exhibited a strong Raman band at 463 cm^-1 and a shoulder at ~600 cm^-1, which were assigned to the F_2g vibration (labeled as I) and defect-induced modes (D band, labeled as II) of the cubic fluorite structure, respectively ^[43-45]. The D band is related to oxygen vacancies in ceria-based materials, the area ratio of F_2g and D (i.e., S_II/S_I) reflects oxygen vacancy concentration ^{[19, 23, 45]}. The F_2g vibration mode and defect-induced mode (D band) of CeO₂ were also detected in the MnO_x-CeO₂ catalysts, as shown in Fig. 3(b). The F_2g band of the MnCe-MMM catalysts was not shifted with respect to that of CeO₂, while that of the MnCe-IM catalyst shifted slightly from 463 to 460 cm^-1, owing to electronic interactions between MnO_x and CeO₂ (Table 1). Table 1 and Fig. 3(b) show that the F_2g band of MnCe-HTM, MnCe-CPM, and MnCe-SGM catalysts also clearly shifted to 448, 459, and 448 cm^-1, respectively. This shift was accompanied by an enhancement of the D band, which further indicated that Mn species were incorporated into the CeO₂ lattice to form a uniform ceria-based solid solution (containing Mn-O-Ce structures). The substitution of Ce⁴⁺ by Mnⁿ⁺ (such as Mn⁴⁺, Mn³⁺, and Mn²⁺) into the CeO₂ lattice induced additional lattice perturbations and structural strain, which generated surface defects (e.g., oxygen vacancy). The FWHM values of F_2g for the synthesized catalysts are listed in Table 1. The FWHM of F_2g for the MnCe-HTM, MnCe-CPM, and MnCe-SGM catalysts were larger than those of CeO₂, MnCe-MMM and MnCe-IM catalysts. This may be partially related to the crystallite size and oxygen vacancies in these samples. Importantly, oxygen vacancies (V_O) can lead to transformation of V_O + O₂ + e^- → adsorbed oxygen species (including O₂⁻, O₂²⁻, and O⁻). This improved oxygen migration ability was beneficial for the oxidation of NO to NO₂, and enhanced the NH₃-SCR catalytic performance by introducing a "fast NH₃-SCR" route ^[19]. The oxygen vacancy concentration for these MnO_x-CeO₂ catalysts is also summarized in Table 1, and followed the trend: MnCe-HTM > MnCe-SGM > MnCe-CPM > MnCe-IM > MnCe-MMM. This pattern corresponds with the catalytic performance results for the NH₃-SCR model reaction. The MnCe-HTM catalyst exhibited the largest oxygen vacancy concentration among the synthesized samples because of incorporation of Mnⁿ⁺ into the CeO₂ lattice. This sample formed a uniform ceria-based solid solution (containing the Mn-O-Ce structure), which strengthened electronic interactions between MnO_x and CeO₂, driven by the high-temperature and high-pressure conditions during the hydrothermal treatment process.

The BET specific surface area and pore volume of the synthesized catalysts measured from N₂ physisorption are listed in Table 1. The MnCe-MMM and MnCe-IM catalysts exhibited a smaller BET specific surface area and pore volume than those of CeO₂, likely because MnO_x particles blocked pores in the CeO₂. However, the BET specific surface area and pore volume of MnCe-HTM, MnCe-CPM, and MnCe-SGM catalysts were obviously larger than those of CeO₂, which is partially related to the crystallite size of these samples. The CeO₂ crystallite size was decreased by incorporation of Mnⁿ⁺ into the CeO₂ lattice, which resulted in an increased BET specific surface area and pore volume. These features also likely contributed to the enhanced catalytic performance by increasing the number of reaction sites on the catalysts.

The reduction properties of the catalysts were investigated by H₂-TPR, as shown in Fig. 4(a). The H₂-TPR profile of CeO₂ as a function of reduction temperature is also included in this figure, and presents two reduction peaks at 510 and 816 ℃ corresponding to reduction of CeO₂ surfaces and bulk CeO₂, respectively ^{[24, 40, 46]}. All of the MnO_x-CeO₂ catalysts exhibited two broad reduction peaks (a low-temperature peak labeled I, and a high-temperature peak labeled II) in the range 100-550 ℃, which are attributed to stepwise reduction of MnO_x, i.e., MnO₂/Mn₂O₃ → Mn₃O₄ → MnO, overlapped by reduction of surface Ce^{4+ [19, 35, 36]}. Compared with the MnCe-MMM catalyst, the reduction peaks of the MnCe-IMM, MnCe-HTM, MnCe-CPM, and MnCe-SGM catalysts shifted to lower temperatures, which indicated that some electronic interactions existed between MnO_x and CeO₂ for these samples. This phenomenon is very similar to the findings of He et al. ^{[47, 48]} in cerium-gold oxide and cerium-vanadium oxide systems. Quantitative analysis of the H₂-TPR data of the synthesized catalysts are summarized in Table 2. The total H₂ consumption (peaks I + II) of the MnCe-IM, MnCe-HTM, MnCe-CPM, and MnCe-SGM catalysts were larger than that of the MnCe-MMM catalyst, which suggests that the electronic interactions between MnO_x and CeO₂ promoted the reduction of surface Ce⁴⁺. The total H₂ consumption of these MnO_x-CeO₂ catalysts followed the order: MnCe-HTM > MnCe-SGM > MnCe-CPM > MnCe-IM > MnCe-MMM. On the basis of these results and the XRD and Raman findings, it is likely that incorporation of Mnⁿ⁺ into the CeO₂ lattice generated a uniform ceria-based solid solution (containing Mn-O-Ce structures). This promoted the reduction of surface Ce⁴⁺ more effectively than simple surface loading of MnO_x to form a supported catalyst, as is the case for the MnCe-IM catalyst. Furthermore, it has been widely reported that Mn⁴⁺ is the most effective redox state of manganese-based catalysts and is beneficial for NO conversion ^{[19, 49]}. Therefore, we focused on the low-temperature reduction peak I of these MnO_x-CeO₂ catalysts and found that the corresponding H₂ consumption could be ranked in the order: MnCe-HTM > MnCe-SGM > MnCe-CPM > MnCe-IM > MnCe-MMM, while the peak temperature exhibited the reverse order. This relationship was also consistent with our NH₃-SCR results.

The adsorption and activation of NH₃ at active sites of the catalysts play an important role in the NH₃-SCR model reaction. Therefore, NH₃-TPD was used to probe the amount of acid sites and their acidic strength in the MnO_x-CeO₂ catalysts. The corresponding results are shown in Fig. 4(b). The NH₃-TPD profile of CeO₂ as a function of desorption temperature is also shown in this figure for comparison. The area and position of these desorption peaks directly relate to the amount of acidic sites and their acidic strength, respectively ^{[26, 50]}. Fig. 4(b) presents the NH₃ desorption signals of the catalysts over a wide temperature range. Differences in the peak positions and heights are caused by variations in the thermal stability of adsorbed NH₃ species. All the samples exhibited a narrow peak around 100 ℃ (labeled as α), and a broad peak between 150 and 550 ℃ (labeled as β). The former (α) at low temperature was attributed to desorption of physisorbed NH₃ and NH₃ at weak acidic sites, while the latter (β) at high temperature resulted from NH₃ adsorbed to strong acid sites of the catalysts ^{[2, 19, 26, 50]}. The NH₄⁺ ions anchored at Brnsted acid sites were less thermally stable than NH₃ species bonded at Lewis acid sites. This suggests that the desorption peak (α) at low temperature was related to NH₄⁺ ions anchored at Brnsted acid sites, while the desorption peak (β) at high temperature was related to NH₃ species bonded at Lewis acid sites ^{[26, 50, 51]}. The temperatures of the β desorption peak of the MnCe-HTM, MnCe-CPM, and MnCe-SGM catalysts were obviously higher than those of the MnCe-MMM and MnCe-IM catalysts, as shown in Fig. 4(b). This indicates that the first three catalysts were more strongly acidic than the latter two catalysts. This phenomenon suggests tha t incorporation of Mnⁿ⁺ into the CeO₂ lattice generated a uniform ceria-based solid solution (containing Mn-O-Ce structure), which enhanced the acidic strength of the MnO_x-CeO₂ catalysts. Quantitative analysis of the NH₃-TPD data from the synthesized catalysts are listed in Table 2. These data show that the total amount of acidic sites in the MnCe-IM, MnCe-HTM, MnCe-CPM, and MnCe-SGM catalysts was considerably larger than those in the CeO₂ and MnCe-MMM samples owing to electronic interactions between MnO_x and CeO₂. The total amount of acid sites followed the order: MnCe-HTM > MnCe-SGM > MnCe-CPM > MnCe-IM > MnCe-MMM, which corresponded with the order of catalytic performances in the NH₃-SCR model reaction. The MnCe-HTM catalyst exhibited the largest amount of acidic sites and the strongest acidic strength among the synthesized samples. The acidity was increased by strengthening of the electron interactions between MnO_x and CeO₂ driven by the high-temperature and high-pressure conditions during the hydrothermal treatment process.

The surface chemical composition and oxidation states of these synthesized catalysts were measured by XPS. High resolution spectra from Ce 3d, Mn 2p, and O 1s are displayed in Fig. 5. For the Ce 3d spectrum (Fig. 5(a)), all samples exhibited eight characteristic peaks in the range 880-920 eV, which are labeled as uʹʺ, uʺ, uʹ, u⁰ (u), vʹʺ, vʺ, vʹ, and v⁰ (v). The uʹ and vʹ peaks can be attributed to Ce³⁺ species, while the uʹʺ, uʺ, u⁰ (u), vʹʺ, vʺ, and v⁰ (v) peaks can be assigned to Ce⁴⁺ species. The spectra indicate that Ce was mainly in the +4 oxidation state, with some Ce³⁺ present in the synthesized catalysts ^{[40, 46, 52, 53]}. The relative content of surface Ce³⁺ (i.e., Ce³⁺/(Ce³⁺+Ce⁴⁺) was calculated as shown in Table 3. This was determined by the following formula based on the area of the uʹ and vʹ peaks ^{[24, 54]}:

The data in Table 3 show that the relative content of surface Ce³⁺ in the MnCe-MMM catalyst was very close to that of CeO₂, while the MnCe-IM catalyst exhibited a higher relative surface Ce³⁺ content than the CeO₂ and MnCe-MMM samples. This was attributed to electronic interactions between MnO_x and CeO₂. Interestingly, the relative contents of surface Ce³⁺ in the MnCe-HTM, MnCe-CPM, and MnCe-SGM catalysts were clearly larger than those of CeO₂, MnCe-MMM, and MnCe-IM samples. This is also related to the incorporation of Mnⁿ⁺ into the CeO₂ lattice to form a uniform ceria-based solid solution (containing Mn-O-Ce structures) and electron interactions between MnO_x and CeO₂. The relative surface Ce³⁺ content is another widely used indicator for defect structures (e.g., oxygen vacancy) in ceria-based materials ^[44]. The surface Ce³⁺ content for the MnO_x-CeO₂ catalysts varied in the order MnCe-HTM > MnCe-SGM > MnCe-CPM > MnCe-IM > MnCe-MMM. This also followed the order of oxygen vacancy concentrations determined by Raman analysis. Oxygen vacancies driven by the transformation of Ce⁴⁺ to Ce³⁺ can promote activation and migration of oxygen molecules, which is beneficial to the oxidation of NO to NO₂, and further enhances the catalytic performance through a "fast NH₃-SCR" route ^[19].

The Mn 2p spectrum (Fig. 5(b)) suggested that Mn species existed in a mixture of valence states on the surface of the MnO_x-CeO₂ catalysts. By performing peak-fitting deconvolution, we divided the Mn 2p_3/2 signal into three characteristic peaks, which were assigned to Mn⁴⁺ (643.3 eV), Mn³⁺ (641.3 eV), and Mn²⁺ (639.6 eV), respectively ^{[40, 46]}. The Mn 2p_1/2 signal exhibited corresponding peaks from three Mn species in the binding energy range 650-660 eV. Previous studies have shown that Mn⁴⁺ species have favorable redox properties for NO conversion ^{[19, 49]}. Therefore, the relative content of surface Mn⁴⁺ (i.e., Mn⁴⁺/(Mn²⁺+Mn³⁺+Mn⁴⁺)) of these MnO_x-CeO₂ catalysts was calculated, and listed in Table 3. We found that the surface Mn⁴⁺ content of the MnCe-IM, MnCe-HTM, MnCe-CPM, and MnCe-SGM catalysts was larger than that of the MnCe-MMM catalyst. This was attributed to electronic interactions between MnO_x and CeO₂, which resulted in the redox equilibrium of Mn³⁺ + Ce⁴⁺ ↔ Mn⁴⁺ + Ce³⁺ shifting to right ^[19]. The surface composition of these synthesized catalysts is summarized in Table 3. The surface Mn/Ce molar ratios of MnCe-MMM and MnCe-IM catalysts were clear smaller than their respective theoretical values (given in the parentheses) because of the occurrence of crystalline Mn₂O₃ and MnO₂ (as determined by XRD). However, the surface Mn/Ce molar ratios of the MnCe-HTM, MnCe-CPM and MnCe-SGM catalysts were larger than the theoretical values, indicating that the Mn species were highly dispersed on the surface of the catalysts. The surface Mn/Ce molar ratio of these MnO_x-CeO₂ catalysts was ranked as MnCe-HTM > MnCe-SGM > MnCe- CPM > MnCe-IM > MnCe-MMM, which was in agreement with the NH₃-SCR results.

Fig. 5(c) shows the O 1s spectrum of these synthesized catalysts, which exhibited a strong peak at 529.1 eV (labeled as Oʹ), and a shoulder at a higher binding energy of 531.2 eV (labeled as Oʺ), which can be attributed to lattice oxygen and adsorbed oxygen on the surface of these samples, respectively ^{[1, 55]}. It has been widely reported that oxygen in the gas phase can be activated by oxygen vacancies on the surface of deNO_x catalysts to form surface adsorbed oxygen. This effect may also enhance catalytic performance in the NH₃-SCR model reaction ^{[19, 56]}. Furthermore, a process of oxygen storage/release between Ce³⁺ and Ce⁴⁺ has been proposed in the following equations: (a) 2CeO₂ → Ce₂O₃ + O* (adsorbed oxygen) and (b) Ce₂O₃ + 1/2O₂ → 2CeO₂ ^[19]. The increase of Oʺ content can promote the oxidation of NO to NO₂ and enhance the catalytic performance of the low-temperature deNO_x catalysts through a "fast NH₃-SCR" route ^{[19, 57]}. The relative content of Oʺ (i.e., Oʺ/(Oʹ+Oʺ)) of these synthesized catalysts was calculated, and is given in Table 3. The Oʺ content of the MnO_x-CeO₂ catalysts was higher than that in CeO₂ owing to the higher of oxygen vacancy concentration in the MnO_x-CeO₂ catalysts. The Oʺ content followed the order MnCe-HTM > MnCe-SGM > MnCe-CPM > MnCe-IM > MnCe-MMM, which was consistent with the catalytic performance of the catalysts in the NH₃-SCR model reaction. To summarize, the MnCe-HTM catalyst exhibited the highest surface Ce³⁺, Mn⁴⁺ and adsorbed Oʺ contents among the synthesized samples. This was attributed to the incorporation of Mnⁿ⁺ into the CeO₂ lattice to form a uniform ceria-based solid solution (containing Mn-O-Ce structures). Furthermore this catalyst featured a strong electron interaction between MnO_x and CeO₂ (i.e., Mn³⁺ + Ce⁴⁺ ↔ Mn⁴⁺ + Ce³⁺) driven by the high-temperature and high-pressure conditions during the hydrothermal treatment process.

The MnCe-HTM catalyst exhibited excellent physicochemical properties and optimal catalytic performance for the low-temperature NH₃-SCR model reaction. Thus, the MnCe-HTM catalyst and CeO₂ (for comparison) were chosen as representative samples to investigate interactions between low-temperature deNO_x catalysts and reactant molecules. The in situ DRIFTS technique is a useful tool for this purpose. NH₃ adsorption measurements by in situ DRIFTS were performed on representative samples, and the results as a function of temperature are shown in Fig. 6. For CeO₂ (Fig. 6(a)), several characteristic bands were observed at 1560, 1299, 1140, and 1051 cm^-1 at room temperature (25 ℃), which were attributed to the vibration modes of coordinated NH₃ adsorbed to Lewis acid sites (labeled as L acid) ^{[58, 59]}. The bands at 1641 and 1430 cm^-1 were assigned to symmetric and asymmetric bending vibration modes of NH₄⁺ species adsorbed to Brnsted acid sites (labeled as B acid) ^{[19, 60]}. Some interesting phenomena were observed as temperature was increased. At 100 ℃, the bands from the B acid disappeared completely due to the poor thermal stability of these species. Furthermore, the intensity of the L acid band weakened at higher temperature, but the band remained even at 350 ℃, which was consistent with the results of NH₃-TPD. A new band was detected at 1320 cm^-1 at 200 ℃, which was attributed to adsorbed -NH₂ species, resulting from dissociation of NH₃ coordinated to L acid sites ^{[59, 61]}. For the MnCe-HTM catalyst (Fig. 6(b)), the L acid bands were observed at 1604, 1167, and 1051 cm^-1 at room temperature (25 ℃), and the B acid exhibited two bands at 1668 and 1404 cm^-1. Compared with CeO₂, we found that the B acid bands of the MnCe-HTM catalyst disappeared at 200 ℃. The inte nsity decrease of the L acid bands in the MnCe-HTM catalyst was considerably slower than that in CeO₂. Furthermore, the L acid band had a considerable intensity up to 350 ℃. The adsorbed -NH₂ species appeared as two bands at 1428 and 1357 cm^-1 at a lower temperature of 150 ℃. The corresponding intensity of these bands was clearly much stronger than those in CeO₂. Fig. 6(a) and (b) show that the intensity of the L acid and B acid bands of the MnCe-HTM catalyst was stronger than that of CeO₂ at all temperatures. In summary, these observations indicated that the amount of acidic sites and the acidic strength of the MnCe-HTM catalyst were larger and stronger than those of CeO₂. This was attributed to the incorporation of Mnⁿ⁺ into the CeO₂ lattice to form a uniform ceria-based solid solution (containing Mn-O-Ce structures). Strengthening of the electron interaction between MnO_x and CeO₂ (i.e., Mn³⁺ + Ce⁴⁺ ↔ Mn⁴⁺ + Ce³⁺) driven by the high-temperature and high-pressure conditions during the hydrothermal treatment process, was also beneficial for the adsorption and activation of NH₃ molecules.

NO+O₂ co-adsorption in situ DRIFTS studies were performed on these representative samples to investigate their oxidation ability for NO to NO₂ and the adsorption behavior of NO species, as shown in Fig. 7. For CeO₂ (Fig. 7(a)), following introduction of NO+O₂ mixed gases into the in situ DRIFTS cell at room temperature (25 ℃), several vibration bands appeared in the range of 900-1700 cm^-1. Bridging bidentate nitrate exhibited an N=O stretching vibration band at 1611 cm^-1, and a NO₂ symmetric vibration mode at 1006 cm^-1; the linear nitrite displayed a vibrational band at 1278 cm^-1; moreover, the chelating bidentate nitrate showed a vibration mode at 1229 cm^{-1 [2, 62, 63]}. As temperature was increased, the band of the linear nitrite weakened and disappeared at 350 ℃ due to desorption/decomposition/transformation during the heating process. We found that the bands of the bridging bidentate nitrate and chelating bidentate nitrate increased in intensity as the temperature was increased owing to oxidation of NO to NO₂. The NO₂ was adsorbed to the CeO₂ surface and formed different kinds of nitrates. With regard to the MnCe-HTM catalyst (Fig. 7(b)), all the vibrational bands of bridging bidentate nitrate, linear nitrite, and chelating bidentate nitrate were also detected at the same corresponding positions at 25 ℃. The changes of these bands with temperature for the MnCe-HTM catalyst were similar to those which occurred for CeO₂. Interestingly, compared with CeO₂, the linear nitrite band for the MnCe-HTM catalyst disappeared at a lower temperature of 200 ℃. This indicates that the desorption/decomposition/transformation of linear nitrite was promoted by the electronic interaction between MnO_x and CeO₂. Fig. 7(a) and (b) show that bands of bridging bidentate nitrate and chelatin g bidentate nitrate on the surface of the MnCe-HTM catalyst were stronger than those on the CeO₂ surface at all temperatures. Moreover, the intensity of these bands increased more rapidly for the MnCe-HTM catalyst than CeO₂ as temperature was increased. These phenomena may be attributed to the incorporation of Mnⁿ⁺ into the CeO₂ lattice to form a uniform ceria-based solid solution (containing Mn-O-Ce structures). The strengthening of the electronic interaction between MnO_x and CeO₂ (i.e., Mn³⁺ + Ce⁴⁺ ↔ Mn⁴⁺ + Ce³⁺), driven by the high-temperature and high-pressure conditions during the hydrothermal treatment process, also effectively improved adsorption of NO species and the ability of the catalyst to oxidize NO to NO₂. This was attributed to good migration ability of oxygen, which resulted from the high oxygen vacancy concentration in the MnCe-HTM catalyst. This effect enhanced the catalytic performance for selective catalytic reduction of NO by NH₃ through a "fast NH₃-SCR" route.

To explore the reaction mechanism of the selective catalytic reduction of NO by NH₃ over MnO_x-CeO₂ catalysts, NO+NH₃+O₂ co-adsorption was measured by in situ DRIFTS under the reaction conditions for these representative samples, and the corresponding results are presented in Fig. 8. When CeO₂ was exposed to NO+NH₃+O₂ mixed gases at 25 ℃, bridging bidentate nitrate and linear nitrite were observed at 1605 and 1284 cm^-1, respectively ^{[62, 63]}. Furthermore, NH₄⁺ species adsorbed to B acid sites and NH₃ species coordinated to L acid sites were also detected at 1435 and 1200 cm^-1, respectively ^{[59, 60]}. Compared with the in situ DRIFTS results for NH₃ adsorption (Fig. 6(a)), the B and L acid bands were more intense in the reaction mixture because the presence of NO+O₂ increased the amount of acidic sites in CeO₂. However, as the temperature was increased, the B and L acid bands disappeared at 100 and 300 ℃, respectively. Bands of chelating bidentate nitrate appeared at 1530 and 1220 cm^-1, and the bridging bidentate nitrate exhibited a NO₂ symmetric vibrational band at 1017 cm^{-1 [62]}. For the MnCe-HTM catalyst (Fig. 8(b)), in the presence of the mixed gases in the in situ DRIFTS cell at room temperature (25 ℃), bridging bidentate nitrate (1605 and 1023 cm^-1), chelating bidentate nitrate (1542 and 1266 cm^-1), monodentate nitrate (1495 cm^-1), linear nitrite (1301 cm^-1), B acid (1435 cm^-1), and L acid (1188 cm^-1) bands were detected ^{[23, 59, 60, 62]}. These results suggested that both of NO and NH₃ were simultaneously adsorbed on the surface of MnCe-HTM catalyst. This indicated that selective catalytic reduction of NO by NH₃ over the MnCe-HTM catalyst followed a Langmuir-Hinshelwood (L-H) mechanism. Furthermore, we found that the intensity of bands from adsorbed NO species on the MnCe-HTM catalyst were stronger than that on CeO₂. This can again be attributed to the introduction of MnO_x. This observation is supported by reports from Li et al. ^[64], who reported that Mn species can improve the adsorption of NO species to CeO₂ catalysts. At elevated temperatures, the B and L acid band disappeared at 200 and 300 ℃, respectively. Changes in the bands of the adsorbed NO species were not obviously owing to the balance between NO conversion and NO adsorption. However, the bands for adsorbed NO species became more intense above 300 ℃ owing to declining NO conversion. These findings are also supported by the results of the NH₃-SCR model reaction.

On the basis of our characterizations, we propose a possible reaction mechanism (schematic diagram) for the selective catalytic reduction of NO by NH₃ under the current reaction conditions, as shown in Fig. 9. The results obtained indicated that the MnCe-HTM catalyst exhibited excellent physicochemical properties, optimal catalytic performance, and good H₂O resistance during a low-temperature NH₃-SCR model reaction. The reaction mechanism over the MnCe-HTM catalyst was examined. When exposed to a mixture of NO+NH₃+O₂ gases, the reactant molecules can simultaneously adsorb to the MnCe-HTM catalyst surface, and the selective catalytic reduction of NO by NH₃ over MnCe-HTM catalyst follows a L-H mechanism. This is supported by the results of our in situ DRIFTS studies. First, oxygen vacancies associated with Ce³⁺ enhanced the migration abilities of oxygen and activated O₂ molecules to generate surface adsorbed oxygen species (such as O^-_(ad)), as confirmed from our Raman and XPS results. Second, it has been reported that Mn species promote adsorption of NO species to these catalysts ^[64]. Any NO molecules adsorbed to Mn species can be transformed into NO⁺_(ad) by donation of electrons to Mn⁴⁺. The oxidized species can then further react with surface adsorbed oxygen species to form NO₂ or adsorbed nitrate (ad-NO₃^-). These findings are consistent with the results of our in situ DRIFTS studies. The intensities of the vibration bands for NH₄⁺ and NH₃ at B and L acid sites, respectively, were clearly stronger for the MnCe-HTM than those for CeO₂, indicating that the adsorption of NH₃ molecules occurred mainly on Mn species. Adsorbed NH₃ species (NH₄⁺ and NH_{3 (ad)}) react with surface adsorbed oxygen spec ies to generate adsorbed -NH₂ (ad-NH₂). Finally, NO₂ or ad-NO₃^- reacts with ad-NH₂ to form the final products, N₂ and H₂O. To summarize, the redox interaction between Mn³⁺ + Ce⁴⁺ ↔ Mn⁴⁺ + Ce³⁺ can remarkably increase Ce³⁺, Mn⁴⁺, oxygen vacancies, the amount of acidic sites and their acidic strength. These features enhance the catalytic performance of MnO_x-CeO₂ catalysts in the NH₃-SCR model reaction. The MnCe-HTM catalyst exhibited the best catalytic performance owing to the strongest electron interaction between MnO_x and CeO₂.

In the present work, a series of MnO_x-CeO₂ catalysts were synthesized by different preparation methods to explore the effects on physicochemical properties and catalytic performance for the selective catalytic reduction of NO by NH₃ at low temperature. Several conclusions can be drawn as follows:

(1) Mnⁿ⁺ species can be incorporated into the lattice of CeO₂. The presence of these cations inhibits the grain growth, resulting in smaller crystallite sizes and larger BET specific surface areas for the MnO_x-CeO₂ catalysts prepared by the hydrothermal treatment method, co-precipitation method, and sol-gel method.

(2) The MnCe-HTM catalyst exhibited the best catalytic performance and good H₂O resistance for the low-temperature NH₃-SCR model reaction. This was attributed to the high content of surface Ce³⁺, Mn⁴⁺, and adsorbed oxygen species, as well as a high concentration of oxygen vacancies and a large amount of acidic sites with high acidic strength. These effects resulted from incorporation of Mnⁿ⁺ (n = 2-4) into the CeO₂ lattice to form a uniform ceria-based solid solution (containing Mn-O-Ce structures). Furthermore, these features were linked to strengthening of the electronic interaction between MnO_x and CeO₂ (i.e., Mn³⁺ + Ce⁴⁺ ↔ Mn⁴⁺ + Ce³⁺) driven by the high-temperature and high-pressure conditions during the hydrothermal treatment process.

(3) NO, NH₃, and O₂ can interact with MnCe-HTM catalyst simultaneously, which indicates that the selective catalytic reduction of NO by NH₃ over MnCe-HTM catalyst follows a L-H mechanism.