Recently, low-temperature denitration techniques [1-3] have attracted increasing attention for the following reasons. First, the denitration unit can be fixed after the precipitator and desulfurizer without reheating the flue gas, which is beneficial for energy-saving and inhibition of catalyst poisoning caused by dust and SO₂. Second, denitration reform of coal-fired power plants can be conveniently achieved without the need to adjust the precipitator and desulfurizer. Finally, low-temperature denitration can be used for some non-electrical industry denitration processes with low-temperature flue gas (below 300 ℃) from waste incinerators, glass furnaces, coking furnaces, etc. Low-temperature denitration catalysts are the most important factor in this technique; thus, the investigation of low-temperature denitration catalysts has become highly topical.

Commercial V₂O₅-WO₃/TiO₂ catalysts exhibit excellent catalytic performance for NH₃-selective catalytic reduction (i.e., NH₃-SCR) of nitrogen oxides (i.e., NO_x) from the flue gas of stationary sources at 300-400 ℃, whereas their performance in low-temperature denitration is not satisfactory [4, 5]. Moreover, V₂O₅ exhibits biological toxicity, which can result in secondary pollution [6, 7]. Therefore, investigations of low-temperature denitration catalysts have mainly focused on environment-friendly non-vanadium-based transition metal oxides. Specifically, manganese oxides (i.e., MnO_x) have been widely studied for low-temperature NH₃-SCR due to their good denitration activity at low temperature [8-14]. In order to further improve the denitration activity and N₂ selectivity, MnO_x species have been loaded onto various supports, such as CeO₂, TiO₂, carbon nano-tubes (CNTs), SiO₂, and γ-Al₂O₃ [15-19]. MnO_x/CeO₂ catalysts have attracted much more attention for low-temperature denitration due to their good redox ability and the high oxygen storage/release capacity of CeO₂ [20-25]. However, it is well known that the specific surface area and thermal stability of pure CeO₂ are not satisfactory. Fortunately, the incorporation of other cations (such as Zr⁴⁺, Sn⁴⁺, Ti⁴⁺, Si⁴⁺, Al³⁺, and Mg²⁺) into the lattice of CeO₂ can effectively be used to overcome these drawbacks [26-28].

With the development of techniques for the preparation of nanomaterials, CeO₂ with different morphologies (such as nano-cubes, nano-polyhedrons, nano-rods, and nano-spheres) has been successfully synthesized [21, 29-31]. Some research results indicate that CeO₂ nano-rods are more suitable for use as a support than CeO₂ with other morphologies due to the predominantly exposed {110} facet of the CeO₂ nano-rods, which facilitates the formation of oxygen vacancies and interactions with the surface dispersed components [29, 30, 32]. Therefore, in the present work, we incorporate the thermally stable Zr⁴⁺, Al³⁺, and Si⁴⁺ cations into the lattice of CeO₂ nano-rods to improve their specific surface area and thermal stability. MnO_x species are then loaded onto these doped supports with the purpose of exploring the doping effect of cations (Zr⁴⁺, Al³⁺, and Si⁴⁺) on the MnO_x/CeO₂ nano-rod catalyst for low-temperature NH₃-SCR, and finally, the optimal dopant is selected.

These doped ceria-based nano-rod supports were synthesized by a hydrothermal method. The detailed experimental steps are as follows. First, the required amounts of Ce(NO₃)₃·6H₂O and Zr(NO₃)₄·5H₂O or Al(NO₃)₃·9H₂O or C₈H₂₀O₄Si (i.e., TEOS) were dissolved in deionized water and homogenized by magnetic stirring for 30 min. The above solution was added dropwise to 6 mol L^-1 NaOH solution with magnetic stirring for another 30 min. The obtained suspension was placed in a Teflon-lined stainless steel autoclave and kept at 100 ℃ for 24 h. The cooled precipitate was centrifuged and washed several times with deionized water until no pH change could be observed. The filter cake was then dried at 70 ℃ for 12 h in a vacuum oven. Finally, these samples were calcined at 400 ℃ for 3 h in a muffle furnace. The Zr⁴⁺, Al³⁺, and Si⁴⁺ doped supports are respectively denoted as CZ-NR, CA-NR, and CS-NR, where the molar ratio of Ce/Zr (Al, Si) was 9/1. Furthermore, the CeO₂-NR support was also prepared by the same procedure for comparison.

MnO_x was loaded onto the surface of the above supports by a wet impregnation method. In brief, the desired amount of Mn(NO₃)₂ was dissolved in deionized water with magnetic stirring. The required amount of CeO₂-NR or CZ-NR or CA-NR or CS-NR support was added to the Mn(NO₃)₂ solution and magnetically stirred for 1 h. Subsequently, the suspension was heated at 100 ℃ in an oil bath to vaporize the water, and oven dried at 100 ℃ for 12 h. Finally, all the samples were calcined at 400 ℃ for 3 h in the muffle furnace. The prepared samples are denoted as MnO_x/CeO₂-NR, MnO_x/CZ-NR, MnO_x/CA-NR, and MnO_x/CS-NR, respectively; the MnO_x loading was fixed at 1.0 mmol Mnⁿ⁺/100 m² support.

Transmission electron microscopy (TEM) and high-resolution transmission electron microscopy (HRTEM) characterizations were carried out on a JEM-2100 instrument operating at 200 kV. Each sample was dispersed in ethanol and kept in an ultrasonic bath for 30 min, then deposited on a carbon-covered copper grid for each test.

X-ray diffraction (XRD) patterns were recorded on a Philips X'pert Pro diffractometer with a Ni-filtered Cu-K_α target (λ = 0.15418 nm) and the X-ray tube was operated at 40 kV and 40 mA.

Raman spectra were collected on a Renishaw Invia Laser Raman spectrometer with an Ar⁺ laser beam. The excitation wavelength and laser power were 532 nm and 5 mW, respectively.

Brunauer-Emmett-Teller (BET) specific surface area was obtained from the N₂-physisorption data acquired at -196 ℃ on a Belsorp-max analyzer. Before the analysis, the sample was degassed under vacuum at 300 ℃ for 4 h.

H₂ temperature-programmed reduction (H₂-TPR) was performed on a chemisorption analyzer (TP-5076) with a 7% H₂-93% Ar mixture as the reductant. Before the reduction, 50 mg of catalyst was pretreated with N₂ at 300 ℃ for 1 h. After cooling to ambient temperature, the TPR process was started at a rate of 10 ℃ min^–1.

NH₃ temperature-programmed desorption (NH₃-TPD) was also performed on the chemisorption analyzer (TP-5076). The catalyst (200 mg) was pretreated with N₂ at 300 ℃ for 1 h. Thereafter, the catalyst was saturated with a 1% NH₃-99% N₂ mixture at 100 ℃, and then flushed with N₂ for 1 h to remove the gas-phase NH₃. Finally, the TPD process was carried out at a rate of 10 ℃ min^–1.

In situ diffuse reflectance infrared Fourier transform spectra (in situ DRIFTS) for NH₃ adsorption were recorded on a Nicolet 5700 FT-IR spectrometer equipped with an MCT detector. Prior to the test, the sample was pretreated with N₂ at 350 ℃ for 1 h. The sample background was collected in the cooling process. At room temperature, the sample was exposed to a 1% NH₃-99% N₂ mixture until it became saturated. Finally, the spectrum was recorded at each target temperature by subtracting the corresponding sample background.

X-ray photoelectron spectra (XPS) were acquired on a PHI 5000 VersaProbe system with a monochromatic Al-K_α target (1486.6 eV). Before the test, the sample was outgassed in a UHV chamber (< 5 × 10^–7 Pa). Moreover, the charging effect was compensated by calibration relative to the binding energy of adventitious C 1s at 284.6 eV.

Measurements of the denitration performance and H₂O+SO₂ tolerance were carried out at steady-state. The reaction gas comprised NO (500 ppm), NH₃ (500 ppm), O₂ (5%), SO₂ (0 or 100 ppm), and H₂O (0 or 5%), with N₂ balance. First, 200 mg of the sample was placed in the reaction tube and pretreated with N₂ at 300 ℃ for 1 h. The reactions were then allowed to take place at each temperature at a space velocity of 60000 ml g^–1 h^–1. NO and NO₂ were detected by a flue gas analyzer, while N₂O was measured using a N₂O analyzer. The NO_x conversion was determined as follows:

The morphologies and exposed crystal planes of these doped CeO₂-based supports and the corresponding supported MnO_x-based catalysts were determined by TEM and HRTEM characterization, as shown in Fig. 1 (and Fig. S1 in the Supporting Information). The CeO₂-NR support exhibited a regular rod morphology with lengths from 30 to 150 nm and diameters in the range of 5-10 nm (Fig. 1(a)). The lattice fringes with an inter-planar spacing of 0.190 nm were detected in Fig. 1(b), which are attributed to the (220) crystal plane of the {110} facet [29, 30, 32, 33]. Moreover, the rod morphology was also observed for the Zr⁴⁺, Al³⁺, and Si⁴⁺ doped samples (i.e., CZ-NR, CA-NR, and CS-NR supports) in Fig. 1(c, e, g), where the rod length of CZ-NR was obviously shorter than that of the other supports, which suggests that the incorporation of Al³⁺ and Si⁴⁺ can maintain the rod morphology of CeO₂ more efficiently. Furthermore, Fig. 1(d, f, h) shows that the {110} facet was the mainly exposed face for the CZ-NR, CA-NR, and CS-NR supports. These results indicate that the incorporation of Al³⁺ and Si⁴⁺ had no obvious influence on the morphology and exposed crystal plane of CeO₂, while Zr⁴⁺ doping leads to a shorter rod length.

Fig. S1 (in the Supporting Information) demonstrates that all the MnO_x/CeO₂-NR, MnO_x/CZ-NR, MnO_x/CA-NR, and MnO_x/CS-NR catalysts display the original morphologies and exposed crystal planes of the CeO₂-NR, CZ-NR, CA-NR, and CS-NR supports, respectively. In addition, the lattice fringes of MnO_x were absent. These phenomena suggest that the MnO_x species were well dispersed on the surface of the CeO₂-NR, CZ-NR, CA-NR, and CS-NR supports, and did not change the original morphologies and exposed crystal planes of these supports. Furthermore, it is important to gain information about the morphologies of the used catalysts, which is beneficial for understanding the structural stability of these catalysts. Therefore, TEM experiments were carried out on these catalysts after use from room temperature to 250 ℃ in the NH₃-SCR reaction (Fig. S2; Supporting Information). All of the used catalysts exhibited a rod morphology (the rod length of the used MnO_x/CZ-NR catalyst was obviously shorter than that of the other three samples), which is similar to the TEM images of the corresponding fresh catalysts. The reason may be that these fresh catalysts were obtained by calcination at 400 ℃, while the highest temperature used in the reaction was only 250 ℃. Thus, the original morphologies could be maintained after use from room temperature to 250 ℃ in the NH₃-SCR reaction, which suggests that the rod structure of these catalysts exhibits good stability.

The XRD patterns of these supports and the supported MnO_x-based catalysts are displayed in Fig. 2. It can be seen from Fig. 2(a) that the CeO₂-NR support exhibits the typical peaks of the cubic fluorite structure between 20° and 80° (PDF-ICDD 34-0394) [34-36]. For the CZ-NR, CA-NR, and CS-NR supports, only the peaks attributed to cubic fluorite-type CeO₂ were detected, while the diffraction signals of ZrO₂, Al₂O₃, and SiO₂ were absent. In addition, the peaks of the CZ-NR, CA-NR, and CS-NR supports shifted slightly to higher angles compared to those of the CeO₂-NR support, which is because the ionic radii of Zr⁴⁺ (0.80 Å), Al³⁺ (0.51 Å), and Si⁴⁺ (0.42 Å) are smaller than that of Ce⁴⁺ (0.92 Å). The incorporation of Zr⁴⁺, Al³⁺, and Si⁴⁺ results in shrinkage of the CeO₂ lattice, and further led to a right-shift of the peaks according to the Bragg equation [27, 28]. The lattice parameters of these supports were calculated and are listed in Table 1. The lattice parameters of the CZ-NR, CA-NR, and CS-NR supports were smaller than that of the CeO₂-NR support, which indicates that Zr⁴⁺, Al³⁺, and Si⁴⁺ (with smaller ionic radii) were indeed incorporated into the lattice of the CeO₂-NR support. Moreover, the intensity of the XRD peaks of the CZ-NR, CA-NR, and CS-NR supports was weaker than that of the CeO₂-NR support, whereas the full-width at half maximum height (FWHM) of these doped supports was wider than that of the CeO₂-NR support, which suggests that the incorporation of Zr⁴⁺, Al³⁺, and Si⁴⁺ inhibits growth of the CeO₂ grains, and results in a smaller grain size according to the Debye-Scherrer equation [28, 36]. All of these results indicate that Zr⁴⁺, Al³⁺, and Si⁴⁺ were successfully doped into the CeO₂ lattice to generate the Ce-O-M (M = Zr, Al, and Si) solid solution. Fig. 2(b) shows that the MnO_x/CeO₂-NR, MnO_x/CZ-NR, MnO_x/CA-NR, and MnO_x/CS-NR catalysts also exhibit the cubic fluorite structure of CeO₂, while no diffraction signals of MnO_x were observed. These results suggest that the MnO_x species were well dispersed on the surface of the CeO₂-NR, CZ-NR, CA-NR, and CS-NR supports, or/and exist in the clustered state, and are not within the detection-limit of the XRD technique.

Raman spectroscopy was employed to further investigate the structures of these supports and supported MnO_x-based catalysts, as presented in Fig. 3. As shown in Fig. 3(a), the CeO₂-NR support exhibits a strong Raman scattering peak at 464 cm^–1 and a broad weak band around 610 cm^–1. The first strong peak is attributed to the F_2g vibration mode of cubic fluorite-type CeO₂, while the latter is related to defect-induced modes (denoted as D, e.g., oxygen vacancies) on the surface of CeO₂ [32, 37, 38]. With regard to the CZ-NR, CA-NR, and CS-NR supports, only the F_2g and D bands could be observed, while the signals of ZrO₂, Al₂O₃, and SiO₂ were not detected. Moreover, compared to the profile of the CeO₂-NR support, the F_2g peak of these doped supports shifted slightly to the low-wavenumber direction and weakened to some extent. These findings suggest that Zr⁴⁺, Al³⁺, and Si⁴⁺ are incorporated into the lattice of CeO₂, which is consistent with the XRD results. Interestingly, it is widely recognized that the intensity ratio of the D and F_2g (i.e., I_D/I_F2g) bands may indicate the relative concentration of oxygen vacancies on the surface of ceria-based materials [22, 32, 37]. The I_D/I_F2g ratio of these supports is also presented in Fig. 3(a), where the values for the CZ-NR, CA-NR, and CS-NR supports are slightly larger than that of the CeO₂-NR support, which indicates that the incorporation of Zr⁴⁺, Al³⁺, and Si⁴⁺ can increase the concentration of oxygen vacancies on the surface of CeO₂. From Fig. 3(b), it was found that the MnO_x/CeO₂-NR, MnO_x/CZ-NR, MnO_x/CA-NR, and MnO_x/CS-NR catalysts also presented a strong peak (F_2g) and a broad weak band (D) without the signals of MnO_x, which suggests that the MnO_x species were well dispersed on the surface of the CeO₂-NR, CZ-NR, CA-NR, and CS-NR supports (supported by the XRD results). Furthermore, the I_D/I_F2g values of these MnO_x-based catalysts (inset in Fig. 3(b)) are remarkably larger than those of the corresponding supports, which indicates that the loaded MnO_x species can interact with these supports to generate more oxygen vacancies. Notably, the MnO_x/CS-NR catalyst had the most oxygen vacancies among these MnO_x-based catalysts.

Textural data for these supports were obtained from N₂-physisorption experiments, as listed in Table 1. The BET specific surface area and total pore volume of the CZ-NR, CA-NR, and CS-NR supports were obviously larger than those of the CeO₂-NR support, which is related to the fact that the incorporation of Zr⁴⁺, Al³⁺, and Si⁴⁺ inhibits growth of the CeO₂ grains, and thus results in a smaller grain size (supported by the XRD results). Moreover, the increase in the BET specific surface area and total pore volume is beneficial for dispersion of the MnO_x species on the surface of these doped supports.

H₂-TPR was used to explore the redox properties of these MnO_x-based catalysts, and the results are displayed in Fig. 4. The data show that the MnO_x/CeO₂-NR, MnO_x/CZ-NR, MnO_x/CA-NR, and MnO_x/CS-NR catalysts all exhibit three reduction peaks in the temperature range of 200-500 ℃, attributed to gradual reduction of the MnO_x species (i.e., MnO₂ → Mn₂O₃ → Mn₃O₄ → MnO, denoted as α, β, and γ, respectively) [2, 39-41]. The quantitative H₂-TPR data are summarized in Table 2. The peak temperature of the MnO_x/CZ-NR, MnO_x/CA-NR, and MnO_x/CS-NR catalysts was higher than that of the MnO_x/CeO₂-NR catalyst, which indicates that the incorporation of Zr⁴⁺, Al³⁺, and Si⁴⁺ leads to greater difficulty in reducing the MnO_x species. The lessened reduction ability is beneficial for inhibiting the non-selective catalytic oxidation of NH₃ during the NH₃-SCR reaction, which results in decreased generation of N₂O as a by-product. Furthermore, Table 2 shows that the practical H₂ consumption of these MnO_x-based catalysts is clearly larger than the corresponding theoretical H₂ consumption (presented in brackets), which indicates that the reduction peaks of α, β, and γ not only correspond to the reduction of MnO_x species, but also have contributions from the reduction of these ceria-based supports. Interestingly, the actual H₂ consumption of the MnO_x/CZ-NR, MnO_x/CA-NR, and MnO_x/CS-NR catalysts was significantly larger than that of the MnO_x/CeO₂-NR catalyst, which suggests that the incorporation of Zr⁴⁺, Al³⁺, and Si⁴⁺ increases the amount of reducible species. It is generally recognized that the NH₃-SCR reaction involves a redox cycle and an acid cycle [42]. Thus, the increase in the reducible species over the MnO_x/CZ-NR, MnO_x/CA-NR, and MnO_x/CS-NR catalysts may contribute to the redox cycle in the NH₃-SCR reaction.

The NH₃-TPD profiles of these MnO_x-based catalysts are presented in Fig. 5. This Fig. shows that the MnO_x/CeO₂-NR and MnO_x/CZ-NR catalysts exhibit four desorption peaks between 100 and 600 ℃ (labeled as I, II, III, and IV, respectively), which are attributed to the desorption of physisorbed NH₃ species, as well as the desorption of chemisorbed NH₃ species, from the weak acid sites, medium-strong acid sites, and strong acid sites, respectively [18, 43-46]. For the MnO_x/CA-NR and MnO_x/CS-NR catalysts, apart from the four aforementioned desorption peaks, an extra shoulder peak was observed at high temperature (denoted as V), which may be related to the desorption of chemisorbed NH₃ species from the hydroxyl groups bonding to Al³⁺ and Si⁴⁺ ions. Some interesting deductions can be made from the quantitative NH₃-TPD analysis (Table 3). The total number of acidic sites of the MnO_x/CZ-NR, MnO_x/CA-NR, and MnO_x/CS-NR catalysts is obviously higher than that of the MnO_x/CeO₂-NR catalyst, which indicates that the incorporation of Zr⁴⁺, Al³⁺, and Si⁴⁺ can efficiently improve the surface acidity of the MnO_x/CeO₂-NR catalyst. Notably, the MnO_x/CS-NR catalyst has the largest number of acidic sites among these MnO_x-based catalysts, which is conducive to the adsorption and activation of NH₃ molecules, and further enhances the denitration performance for NH₃-SCR reaction.

Therefore, the MnO_x/CS-NR catalyst was selected as the representative sample to further distinguish the Brønsted (B) acid sites and Lewis (L) acid sites on the catalyst surface through in situ diffuse reflectance infrared Fourier-transform spectroscopy (DRIFTS) analysis of the NH₃ adsorption, as shown in Fig. 6. For comparison, the in situ DRIFTS data for NH₃ adsorption over the MnO_x/CeO₂-NR catalyst are also presented in this Fig.. Fig. 6(a) shows that when the MnO_x/CeO₂-NR catalyst was saturated with a 1% NH₃-99% N₂ mixture at room temperature, signals of NH₃ coordinated to the L acid sites appeared at 1056, 1117, 1291, 1549, and 1589 cm^–1, while the vibration band of NH₄⁺ bonded to the B acid sites appeared at 1425 cm^–1 [36, 47-51]. Interestingly, the intensity of the vibration bands related to the B acid sites and L acid sites weakened with an increase in temperature, and some signals of the L acid sites disappeared at 300 and 350 ℃, which suggests that the adsorbed NH₃ species on the surface of the MnO_x/CeO₂-NR catalyst can be desorbed and transformed with an increase in temperature. For the MnO_x/CS-NR catalyst (Fig. 6(b)), upon saturation with a 1% NH₃-99% N₂ mixture at room temperature, the vibration bands of the B acid sites and L acid sites could be detected. Compared with Fig. 6(a), the relative intensity of the band of the B acid sites of the MnO_x/CS-NR catalyst was remarkably stronger than that of the MnO_x/CeO₂-NR catalyst, which indicates that the incorporation of Si⁴⁺ effectively increases the number of B acid sites. It is generally accepted that B acid sites play a more important role in the low-temperature NH₃-SCR reaction than L acid sites [19, 52]. Furthermore, the intensity change of all the vibration bands of the MnO_x/CS-NR catalyst with an increase in temperature was more pronounced than for the MnO_x/CeO₂-NR catalyst (especially from 100 to 150 ℃), which suggests that the MnO_x/CS-NR catalyst is more effective for activation of the adsorbed NH₃ species. These results suggest that the MnO_x/CS-NR catalyst may exhibit better denitration performance than the MnO_x/CeO₂-NR catalyst for the low-temperature NH₃-SCR reaction.

The surface properties of the denitration catalysts play an important role in the NH₃-SCR reaction. Hence, the surface chemical states of these MnO_x-based catalysts were analyzed by XPS, as presented in Fig. 7. The Mn 2p spectra in Fig. 7(a) show six bands for the MnO_x/CeO₂-NR, MnO_x/CZ-NR, MnO_x/CA-NR, and MnO_x/CS-NR catalysts, ascribed to the Mn 2p_1/2 and Mn 2p_3/2 signals of Mn⁴⁺, Mn³⁺, and Mn²⁺, respectively [53, 54]. It is commonly recognized that Mn⁴⁺ is more effective for the elimination of NO_x than the Mn³⁺ and Mn²⁺ ions [55]. Therefore, the relative content of Mn⁴⁺ on the surface of these MnO_x-based catalysts was calculated, as presented in Table 4. The Mn⁴⁺ content of the MnO_x/CZ-NR, MnO_x/CA-NR, and MnO_x/CS-NR catalysts was higher than that of the MnO_x/CeO₂-NR catalyst. The reasons may be as follows: first, the larger BET specific surface area of the CZ-NR, CA-NR, and CS-NR supports is beneficial for dispersion of the Mn-precursor, which can promote the oxidation of Mn²⁺ to high-valence Mn⁴⁺ during the calcination process; second, the incorporation of Zr⁴⁺, Al³⁺, and Si⁴⁺ strengthens the electronic interaction between the doped supports and the surface-dispersed MnO_x species, which can lead to the generation of more Mn⁴⁺. Notably, the MnO_x/CS-NR catalyst had the highest Mn⁴⁺ content among these MnO_x-based catalysts, which may contribute to the excellent denitration performance for the low-temperature NH₃-SCR reaction.

Fig. 7(b) shows the O 1s spectra of these MnO_x-based catalysts, demonstrating the presence of two types of oxygen species on the surface of the MnO_x/CeO₂-NR, MnO_x/CZ-NR, MnO_x/CA-NR, and MnO_x/CS-NR catalysts, i.e., lattice oxygen species with a signal centered at 529.2 eV (denoted as O_β) and surface adsorbed oxygen species with a signal at 531.7 eV (denoted as O_α) [50, 56]. It is generally recognized that the ratio of the areas of the O_α and O_α+O_β peaks (i.e., O_α/(O_α+O_β)) is an indicator of the relative content of surface adsorbed oxygen species, and an increase in the O_α content can promote the oxidation of NO to NO₂, and further enhances the denitration performance for the low-temperature NH₃-SCR reaction through a fast selective catalytic reduction (SCR) pathway [50]. Therefore, the O_α content of the surface of these MnO_x-based catalysts was calculated, and is also listed in Table 4. The O_α content of the MnO_x/CZ-NR, MnO_x/CA-NR, and MnO_x/CS-NR catalysts was smaller than that of the MnO_x/CeO₂-NR catalyst, which suggests that the incorporation of Zr⁴⁺, Al³⁺, and Si⁴⁺ may inhibit the fast SCR pathway, and further negatively affects the low-temperature NH₃-SCR reaction.

The denitration performance of these CeO₂-based supports and corresponding MnO_x-based catalysts is displayed in Fig. 8. The NO_x conversion data in Fig. 8(a) show that the CZ-NR, CA-NR, and CS-NR supports exhibit slightly higher catalytic activity than the CeO₂-NR support, and loading with MnO_x could obviously enhance the catalytic activity of the supports. Notably, the MnO_x/CS-NR catalyst gave the best catalytic activity among these catalysts above 100 ℃. The reasons may be as follows: first, the MnO_x/CS-NR catalyst possesses the largest amount of oxygen vacancies, which may enhance dissociation of the NO_x molecules; second, it has the largest number of acid sites, which can promote the adsorption and activation of NH₃ molecules; finally, it has the highest Mn⁴⁺ content, which is conducive to the elimination of NO_x. However, the MnO_x/CZ-NR catalyst exhibits the lowest NO_x conversion among these catalysts in the 50-200 ℃ range, which may be because the incorporation of Zr⁴⁺ into the lattice of CeO₂ shortens the length of the nano-rods and further decreases the extent of exposure of the active {110} facet. Moreover, the NO_x conversion curves of these catalysts exhibit a similar trend as a function of temperature, i.e., an initial increase with an increase in temperature and a subsequent decline with a further increase in the temperature, because the non-selective catalytic oxidation of NH₃ results in a decrease in the amount of reductant, which is maximal at 200 ℃.

It is well known that MnO_x-based catalysts generally exhibit poor N₂ selectivity as their very strong redox properties lead to the non-selective catalytic oxidation of NH₃ to N₂O (by-product). Therefore, the concentration of N₂O generated in the reaction process was monitored, as shown in Fig. 8(b). Almost no N₂O was formed over these CeO₂-based supports over the entire temperature range, whereas the N₂O concentration increased significantly for the reaction over these MnO_x-based catalysts in the 100-250 ℃ range. In addition, the MnO_x/CS-NR catalyst presented the lowest N₂O concentration among these catalysts over the entire temperature range, which may be related to its lower redox ability and effective activation of NH₃. These two factors partially inhibit the non-selective catalytic oxidation of NH₃ to generate N₂O. Combining the results of Fig. 8(a) and (b), it is found that the MnO_x/CS-NR catalyst exhibits the best denitration performance in the present study.

Fig. 9 shows the H₂O+SO₂ tolerance of the MnO_x/CeO₂-NR, MnO_x/CZ-NR, MnO_x/CA-NR, and MnO_x/CS-NR catalysts at 200℃. Steady-state NO_x conversion over these MnO_x-based catalysts was easily achieved in the absence of H₂O and SO₂ (within the first 120 min). However, all the catalysts exhibited an obvious decrease in the NO_x conversion (about 20%) once H₂O and SO₂ were introduced (from 120 to 600 min). According to the literature [2, 44, 57], this change may be derived from competitive adsorption between the reactants and toxic molecules, the deposition of sulfates, and sulfation of the active species. Finally, the NO_x conversion with these MnO_x-based catalysts could be partially restored after the removal of H₂O and SO₂, with a decline in the activity of only about 3% relative to the original values (the last 120 min). These experimental results suggest that the decrease in the NO_x conversion over these catalysts caused by H₂O and SO₂ occurs mainly by reversible reactions, with some irreversible deactivation. Furthermore, the similar trend in the NO_x conversion based on measurement of the H₂O+SO₂ tolerance indicates that these catalysts might follow the same deactivation mechanism. Overall, the MnO_x/CS-NR catalyst generally presented the highest NO_x conversion among these catalysts despite the absence or presence of H₂O and SO₂, which suggests that MnO_x/CS-NR is the most promising catalyst in the present work.

The thermally stable Zr⁴⁺, Al³⁺, and Si⁴⁺ cations were incorporated into the lattice of CeO₂-NR to prepare MnO_x/CeO₂-NR, MnO_x/CZ-NR, MnO_x/CA-NR, and MnO_x/CS-NR catalysts for low-temperature NH₃-SCR with the purpose of exploring the doping effect and screening the optimal dopant. These samples were characterized in detail, and the key conclusions are as follows: (1) the incorporation of Zr⁴⁺, Al³⁺, and Si⁴⁺ does not change the rod-like structure and exposed active {110} facet of CeO₂-NR, but Zr⁴⁺ doping shortens the rod length; (2) the BET specific surface area of CeO₂-NR can be effectively enlarged by Zr⁴⁺, Al³⁺, and Si⁴⁺ doping, which is beneficial for dispersion of the MnO_x species; (3) the MnO_x/CS-NR catalyst has the largest number of oxygen vacancies, the largest number of acid sites, and the highest Mn⁴⁺ content, which result in the best denitration performance; (4) the Si⁴⁺-doped catalyst not only exhibits the best denitration performance, but also displays good H₂O+SO₂ tolerance, which indicates that Si⁴⁺ is the optimal dopant for the MnO_x/CeO₂-NR catalyst in the present work.