In recent years, NH₃-selective catalytic reduction (NH₃-SCR) of nitrogen oxides (NO_x) has been recognized as the most efficient way to eliminate NO_x from coal-fired flue gas and diesel exhaust emissions [1-3]. V₂O₅-WO₃/TiO₂ and V₂O₅-MoO₃/TiO₂ catalysts were conventionally used for NH₃-SCR because of their excellent catalytic performance and H₂O + SO₂ resistance during elevated reaction temperatures of 300-400 ℃ [4-6]. However, vanadium-based catalysts offer drawbacks such as high activity for SO₂ oxidation to SO₃ and the non-selective catalytic oxidation of NH₃ to N₂O at high temperatures, as well as the biological toxicity of vanadium, etc. [7, 8]. Therefore, non-vanadium-based NH₃-SCR catalysts currently attract significantly more attention in this field.

CeO₂ has been widely investigated in NH₃-SCR as it offers good redox properties and high oxygen storage/release capacity associated with the facile change in the surface atomic ratio of Ce³⁺/Ce⁴⁺, which is a key function to eliminate NO_x [9-12]. However, the catalytic performance of pure CeO₂ for NH₃-SCR is very poor. Therefore, CeO₂ is commonly prepared into ceria-based mixed metal-oxide catalysts and supported ceria-based catalysts, which exhibit excellent catalytic performance. For example, Liu et al. [13] synthesized a novel FeO_x-CeO₂-TiO₂ catalyst by a hydrothermal method, which exhibits good NH₃-SCR activity, high N₂ selectivity, and strong resistance to H₂O and SO₂. They attributed the desirable catalytic performance to a synergetic effect between FeO_x and CeO₂ that induced a Ce⁴⁺ + Fe²⁺ ↔ Ce³⁺ + Fe³⁺ redox cycle, which improves NO and NH₃ activation. Li et al. [4] prepared a series of supported ceria-based catalysts, and observed that the CeO₂/TiO₂-SiO₂ catalyst with a Ti/Si mass ratio of 3/1 showed optimum catalytic activity and N₂ selectivity for NH₃-SCR. They highlighted that the interaction between TiO₂ and SiO₂ promoted the transformation of Ce⁴⁺ to Ce³⁺, while the introduction of SiO₂ increased the number of acid sites in the CeO₂/TiO₂-SiO₂ catalyst, which resulted in enhanced catalytic performance. Recently, supported ceria-based catalysts have attracted a wealth of interdependent research because of their excellent catalytic performance, large Brunauer-Emmett-Teller (BET) specific surface areas, high thermal stability, and low utilization of active components.

It is widely reported that the support material can indirectly influence the physicochemical properties, and concomitantly, the catalytic performance of the active component in the supported metal-oxide catalyst [14-16]. Therefore, the study of the interaction between surface-dispersed active components and the support is widely reported. For example, Hong et al. [17] investigated the influence of Ce⁴⁺ doping on the interaction between surface-dispersed MnO_x and a CeO₂-TiO₂ support, and observed that the introduction of Ce⁴⁺ promoted MnO_x dispersion and increased Mn⁴⁺ content. Furthermore, the presence of Ce⁴⁺ enhanced surface acidity, which improved NH₃-SCR activity of the catalyst. Zhang et al. [18] explored the relationship between the catalytic activity of a NiO/CeO₂ nanorod catalyst and the interaction of the NiO with CeO₂ nanorods support. They reported that compared with the pure NiO and CeO₂ nanorods, the NiO/CeO₂ nanorod catalyst exhibited a higher Ce³⁺ content, larger active oxygen concentration, lower temperature reducibility, and a stronger interaction with reactant molecules, all of which contribute to the enhancement in low-temperature NH₃-SCR activity.

CeO₂ can be used as an active component, additive, or support for NH₃-SCR. In the present work, we mainly employ CeO₂ as an active component. To investigate the influence of the support—having different crystal structures—on the physicochemical properties and catalytic performance of supported ceria-based catalysts for NH₃-SCR, which is conducive to screen a suitable support, we prepared a series of CeO₂/SiO₂, CeO₂/γ-Al₂O₃, CeO₂/ZrO₂, and CeO₂/TiO₂ catalysts. The obtained catalysts were characterized by X-ray diffraction (XRD), Raman spectroscopy, BET surface area measurements, X-ray photoelectron spectroscopy (XPS), H₂-temperature programmed reduction (H₂-TPR), NH₃-temperature programmed desorption (NH₃-TPD), and were catalytically evaluated by the NH₃-SCR model reaction to explore the influence of the support on the supported ceria-based catalysts.

SiO₂ was synthesized by a sol-gel method. The desired amounts of tetraethyl orthosilicate, ethanol, and distilled water having a mole ratio of 1:5:10, respectively, were mixed together at room temperature and magnetically stirred at 50 ℃ for 10 h to generate a transparent gel. The obtained gel was dried at 110 ℃ overnight, and subsequently calcined at 550 ℃ for 5 h.

γ-Al₂O₃, ZrO₂, and TiO₂ samples were obtained by a precipitation method. Separately, the required amounts of Al(NO₃)₃∙9H₂O, Zr(NO₃)₄∙5H₂O, and TiCl₄ were dissolved in distilled water prior to adding into excess concentrated ammonia under magnetic stirring. The pH was controlled at 10 by HNO₃ solution. The resulting suspensions were further stirred for 3 h, aged for 12 h, and then filtered, washed with distilled water until the filtrate was pH = 7. Finally, the as-prepared samples were dried at 110 ℃ overnight, and subsequently calcined at 550 ℃ for 5 h.

The supported ceria-based catalysts were prepared by impregnating SiO₂, γ-Al₂O₃, ZrO₂, and TiO₂ supports with Ce(NO₃)₃∙6H₂O solution. CeO₂ loading was fixed at 0.4 mmol/g_support. The suspension was stirred for 3 h, and evaporated at 100 ℃ using an oil bath. Thereafter, the obtained cake-like products were dried at 110 ℃ overnight, and finally calcined at 500 ℃ for 5 h.

XRD patterns of the catalysts were collected on a Philips X'Pert3 Powder diffractometer with Ni-filtered Cu K_α radiation (λ = 0.15418 nm). The operating voltage and current of the X-ray tube was 40 kV and 40 mA, respectively.

Raman spectra of the catalysts were recorded on a Renishaw inVia Reflex Laser Raman spectrometer with an Ar⁺ laser beam having an excitation wavelength of 532 nm. The laser power was set at 5 mW.

Textural properties of the catalysts were obtained by the BET method using a Belsorp-max analyzer. Prior to each measurement, samples were degassed under vacuum at 300 ℃ for 4 h.

XPS spectra of the catalysts were recorded using a PHI 5000 VersaProbe system with monochromatic Al K_α radiation (1486.6 eV) having an accelerating power of 15 kW. Prior to the test, samples were outgassed in an ultra-high vacuum chamber ( < 5 × 10^-7 Pa). The sample charging effect was compensated by calibrating the binding energy with adventitious C 1s peak at 284.6 eV.

H₂-TPR experiments were performed using a chemisorption analyzer (TP-5076) in a H₂-Ar mixture (7% H₂ by volume, 30 mL/min) as the reductant. Before the reduction, 50 mg catalyst was pretreated with N₂ at 300 ℃ for 1 h, and then cooled to ambient temperature. Thereafter, the H₂-Ar mixture was introduced, and the H₂-TPR profile collected from 100-900 ℃ at a ramp rate of 10 ℃/min.

NH₃-TPD experiments were also performed on the chemisorption analyzer (TP-5076). 200 mg catalyst was pretreated with N₂ at 300 ℃ for 1 h prior to saturating the catalyst with a NH₃-N₂ mixed gas (1% NH₃ by volume, 30 mL/min) at 100 ℃ for 1 h. The catalyst was subsequently flushed with N₂ at 100 ℃ for an additional 1 h to eliminate any NH₃ in the gas phase. Finally, the NH₃-TPD profile was recorded from room temperature to 800 ℃ under a N₂ atmosphere (30 mL/min) at a ramp rate of 10 ℃/min.

Catalytic performance and H₂O + SO₂ tolerance of the catalysts were evaluated by the NH₃-SCR model reaction under steady state, which involved a feed stream having a fixed composition of 500 ppm NO, 500 ppm NH₃, 5% O₂, 5% H₂O (when used), 100 ppm SO₂ (when used), and N₂ in balance. 200 mg catalyst was fitted in a quartz reaction tube and pretreated with N₂ at 300 ℃ for 1 h, before cooling to ambient temperature. Thereafter, the reaction gases were introduced, and the space velocity was fixed at 60000 mL g^-1 h^-1. NO concentration of the inlet and outlet was detected by a flue gas analyzer, and NO conversion of the catalysts was determined from the following equation: NO conversion = ([NO]_in -[NO]_out)/[NO]_in × 100%.

The catalytic performance of the supported ceria-based catalysts was evaluated by the NH₃-SCR model reaction, and the corresponding NO conversion is given in Fig. 1(a). It can be seen that each catalyst follows a similar NO conversion trend: first, NO conversion increases as a function of reaction temperature related to the input of thermal energy; thereafter, NO conversion decreases with further increases of reaction temperature because of the consumption of the NH₃ reducing agent by non-selective oxidation at high temperature [19, 20]. Furthermore, we observe that the catalytic activity of the catalysts is highly dependent on the support type. The CeO₂/γ-Al₂O₃ catalyst exhibits the best catalytic performance among the catalysts, providing high NO conversion ( > 90%) at 300-500 ℃. However, the innate character of how the support influences the catalytic performance of the supported ceria-based catalysts still needs further investigation.

Additionally, H₂O and SO₂ are known to suppress NO conversion in practical applications. Therefore, we explored the H₂O + SO₂ tolerance of the supported ceria-based catalysts at 400 ℃, and the corresponding results are presented in Fig. 1(b). For the duration of the initial 120 min, prior to the introduction of H₂O and SO₂, NO conversion of the catalysts is very stable, which indicates that the NH₃-SCR reaction reaches a steady state. With the introduction of H₂O and SO₂ (from 120 to 600 min), NO conversion decreases slightly because of the competitive adsorption between reactants and H₂O + SO₂, the generation of surface hydroxyl groups, and the deposition of sulfates [5, 21]. Finally, NO conversion can be recovered partially after removing H₂O and SO₂ from the feed stream (in the last 120 min), which suggests that deactivation of the supported ceria-based catalysts resulted from H₂O + SO₂ poisoning and follows a two-part reversible and irreversible deactivation mechanism. Additionally, we observe that the loss of catalytic activity for the catalysts is always less than 25% in the presence of H₂O and SO₂. The CeO₂/γ-Al₂O₃ catalyst exhibits above 80% NO conversion across the entire measurement of H₂O+SO₂ tolerance, which is higher than the other three catalysts, indicating it as a potential de-NO_x candidate for coal-fired flue gas applications.

To further understand how the support influences the catalytic performance of the supported ceria-based catalysts in the NH₃-SCR reaction, their physicochemical properties were carefully characterized. XRD and Raman were chosen to investigate the structural properties of the catalysts and supports, the corresponding results are displayed in Fig. 2. It can be seen from Fig. 2(a) that the SiO₂ support presents a broad peak around 22.26° (PDF-ICDD 29-0085), while several new diffraction peaks appear at 28.44°, 32.97°, 47.37°, and 56.26° after CeO₂ loading on the surface of SiO₂, indicating the existence of crystalline cubic fluorite-type CeO₂ (PDF-ICDD 43-1002). The diffraction peaks of the defective spinel structure for γ-Al₂O₃ are clearly observed for the γ-Al₂O₃ support (PDF-ICDD 10-0425), and the CeO₂/γ-Al₂O₃ catalyst also exhibits diffraction peaks attributed to CeO₂ at the corresponding position. The ZrO₂ support displays a series of complex diffraction signals, which contains both monoclinic (labeled as "m") ZrO₂ (PDF-ICDD 37-1484) and tetragonal (labeled as "t") ZrO₂ (PDF-ICDD 50-1089) phases, while the diffraction peaks assigned to CeO₂ are absent for the CeO₂/ZrO₂ catalyst. All diffraction peaks relating to the TiO₂ support can be assigned to anatase TiO₂ (PDF-ICDD 21-1272), and crystalline CeO₂ is also observed for the CeO₂/TiO₂ catalyst. XRD results suggest that crystalline CeO₂ is on the surface of SiO₂, γ-Al₂O₃, and TiO₂ supports, while highly dispersed on the surface of the ZrO₂ support.

Fig. 2(b) shows no observable Raman signals for the SiO₂ and γ-Al₂O₃ supports, while the F_2g vibration mode of CeO₂ can be detected at 463 cm^-1 for the CeO₂/SiO₂ and CeO₂/γ-Al₂O₃ catalysts [7, 22, 23]. The complex Raman vibration bands of the ZrO₂ support can be attributed to monoclinic (labeled as "m") ZrO₂ and tetragonal (labeled as "t") ZrO₂ phases [24, 25], however, the F_2g vibration mode of CeO₂ is absent over the CeO₂/ZrO₂ catalyst. The TiO₂ support exhibits Raman vibration bands associated with anatase TiO₂ at 146, 197, 397, 516, and 639 cm^-1 [9, 13], and the Raman signal of CeO₂ can be also observed for the CeO₂/TiO₂ catalyst. The obtained Raman results further confirm that CeO₂ is highly dispersed on the surface of the ZrO₂ support, while aggregates into a crystalline phase over the SiO₂, γ-Al₂O₃, and TiO₂ supports, which is in accordance with XRD results. However, the NH₃-SCR catalytic performance of the CeO₂/ZrO₂ catalyst is less superior to that of the CeO₂/γ-Al₂O₃ catalyst above 250 ℃, which indicates that besides the dispersion of CeO₂, there are other key factors to influence the catalytic performance.

Textural properties of the supported ceria-based catalysts and supports were characterized by BET measurements, and the corresponding results are listed in Table 1. We observe that the BET specific surface area and total pore volume of the catalysts are obviously smaller than the corresponding supports, which suggests that CeO₂ mainly exists in the pore structure and on the surface of these supports. Furthermore, the CeO₂/SiO₂ catalyst exhibits a larger BET specific surface area and total pore volume than the CeO₂/γ-Al₂O₃, CeO₂/ZrO₂, and CeO₂/TiO₂ catalysts, while it presents the worst catalytic performance for the NH₃-SCR reaction from 250-400 ℃, which suggests that the textural properties of the catalyst are not a key factor.

It is well known that catalyst surface properties can remarkably influence the adsorption and activation of reactant molecules, and concomitantly, the catalytic performance for NH₃-SCR reactions. Therefore, XPS was chosen to investigate the surface properties of the supported ceria-based catalysts, and the corresponding results are presented in Fig. 3. We can observe from Fig. 3(a) that the Ce 3d spectra of these catalysts are fitted with eight binding energy peaks (labeled as u‴, u″, u′, u⁰(u), and v‴, v″, v′, v⁰(v), respectively), in which u′ and v′ represent Ce³⁺ species, while the others are related to Ce⁴⁺ ions [22, 26-28]. The obtained results indicate that Ce³⁺ and Ce⁴⁺ co-exist on the surface of these catalysts. Additionally, Ce³⁺ content is calculated and listed in Table 2. It can be seen that the CeO₂/γ-Al₂O₃ catalyst exhibits the highest Ce³⁺ ratio among these catalysts. We know that oxygen vacancies are usually generated as a function of surface Ce³⁺ in ceria-based catalysts. Furthermore, oxygen vacancies can weaken N-O bonds to promote the dissociation of NO molecules [29, 30], which is beneficial to the enhancement of NH₃-SCR catalytic performance.

Fig. 3(b) shows that the O 1s spectra of the catalysts (except the CeO₂/γ-Al₂O₃ catalyst) can be fitted with two components, labeled as O_Ⅰ and O_Ⅱ, respectively. According to the literature [10, 13, 18], O_Ⅰ is attributed to the lattice oxygen of metal-oxides, while O_Ⅱ is related to the adsorbed oxygen species on the catalyst surface. The binding energy of the O 1s spectra is different in each catalyst because of the different support structures of the supported ceria-based catalysts.

Reduction behavior of catalysts is an important factor for NH₃-SCR as this directly influences the oxidation of NO to NO₂, and further promotes the reaction through a "fast SCR" route [20]. H₂-TPR is a useful tool to characterize the reduction behavior of the supported ceria-based catalysts, and the results are displayed in Fig. 4. We observe that the H₂-TPR profiles of the catalysts are very similar, all of which exhibit three reduction peaks, labeled as α, β, and γ, respectively. The obtained results—complementary to related literature—indicate that the α reduction peak is attributed to the reduction of surface CeO₂ species; the β and γ reduction peaks are related to the reduction of sub-surface CeO₂ and bulk CeO₂, respectively [11, 31]. Interestingly, quantitative analysis of H₂-TPR (Table 3) shows that H₂ consumption of the CeO₂/γ-Al₂O₃ catalyst is significantly larger than that of the CeO₂/SiO₂, CeO₂/ZrO₂, and CeO₂/TiO₂ catalysts, which suggests that the CeO₂/γ-Al₂O₃ catalyst possesses the desired reduction behavior among the catalysts. Furthermore, the improved reduction behavior of the CeO₂/γ-Al₂O₃ catalyst indicates that the migration ability of oxygen species over the CeO₂/γ-Al₂O₃ catalyst is the strongest among the catalysts during the heating process, which enhances the oxidation of NO to NO₂ (i.e., oxygen species of the CeO₂/γ-Al₂O₃ catalyst can migrate to oxidize NO to NO₂ during the heating process), and further promotes NH₃-SCR catalytic performance through a "fast NH₃-SCR" route (i.e., NO + NO₂ + 2NH₃ → 2N₂ + 3H₂O).

To further confirm that the reduction behavior of the catalysts can influence the oxidation of NO to NO₂, the catalytic activity of the supported ceria-based catalysts for the NO + O₂ model reaction was measured, and the corresponding results are presented in Fig. 5. First, the catalytic activity observed from this figure shows that NO oxidation to NO₂ increases as a function of reaction temperature for the CeO₂/γ-Al₂O₃ and CeO₂/ZrO₂ catalysts, and thereafter declines with further increases of reaction temperature, while the activity of the CeO₂/SiO₂ and CeO₂/TiO₂ catalysts increases as the reaction temperature is elevated across the entire temperature range of 50-500 ℃. Especially, we can find that the catalytic activity of the CeO₂/γ-Al₂O₃ catalyst is obviously higher than that of the CeO₂/SiO₂, CeO₂/ZrO₂, and CeO₂/TiO₂ catalysts, which indicates that the CeO₂/γ-Al₂O₃ catalyst favors the oxidation of NO to NO₂ among the supported ceria-based catalysts, which is attributed to the enhanced reduction behavior. The obtained results are in good agreement with that of the H₂-TPR.

Surface acidity of the supported ceria-based catalysts was investigated by NH₃-TPD, as shown in Fig. 6. We observe that the CeO₂/γ-Al₂O₃, CeO₂/ZrO₂, and CeO₂/TiO₂ catalysts exhibit four desorption peaks (labeled as Ⅰ, Ⅱ, Ⅲ, and Ⅳ) during the heating process, which are related to physisorbed NH₃, weak acids, medium-strong acids, and strong acids, respectively [20, 32-34]. However, only two desorption peaks (i.e., Ⅰ and Ⅳ), resulting from the desorption of physisorbed NH₃ and the desorption of chemisorbed NH₃ on strong acid sites, can be observed for the CeO₂/SiO₂ catalyst. The desorption peak areas in the NH₃-TPD profiles are proportional to the acid concentration in de-NO_x catalysts. Therefore, quantitative analysis of the NH₃-TPD profiles for the supported ceria-based catalysts was calculated, and the results are given in Table 4. Interestingly, it can be seen from Table 4 that the total acid amount in the CeO₂/γ-Al₂O₃ catalyst is remarkably larger than that of the CeO₂/SiO₂, CeO₂/ZrO₂, and CeO₂/TiO₂ catalysts, which promotes the adsorption and activation of reactant NH₃ molecules, and further enhances the NH₃-SCR catalytic performance.

In the present work, a series of supported ceria-based catalysts were prepared with SiO₂, γ-Al₂O₃, ZrO₂, and TiO₂ chosen as the supports. The purpose was to investigate how the physicochemical properties and NH₃-SCR catalytic performance were influenced as a function of the support. Some interesting conclusions can be drawn from the characterization results as follows: the results of XPS, H₂-TPR, and NH₃-TPD show that the surface properties, reduction behavior, and surface acidity of the supported ceria-based catalysts are highly dependent on the used supports. First, the ratio of Ce³⁺/Ce⁴⁺ on the surface of the CeO₂/γ-Al₂O₃ catalyst is larger than that of the CeO₂/SiO₂, CeO₂/ZrO₂, and CeO₂/TiO₂ catalysts, which promotes the generation of oxygen vacancies, and concomitantly, the dissociation of NO molecules, and further leads to enhanced NH₃-SCR catalytic performance. Second, the CeO₂/γ-Al₂O₃ catalyst possesses optimum reduction behavior among the supported ceria-based catalysts, which improves oxidation of NO to NO₂, and by extension, the NH₃-SCR catalytic performance through a "fast NH₃-SCR" route. Third, the total acid concentration of the CeO₂/γ-Al₂O₃ catalyst is the highest among the catalysts, which promotes the adsorption and activation of reactant NH₃ molecules, and further results in enhanced NH₃-SCR catalytic performance. Finally, catalytic performance and H₂O + SO₂ tolerance results show that the CeO₂/γ-Al₂O₃ catalyst exhibits the best catalytic performance for NH₃-SCR in the absence/presence of H₂O and SO₂, which indicates its de-NO_x potential as a candidate for coal-fired flue gas applications.