Significant Effects of Adding Mode on Low-Temperature De-NO_x Performance and SO₂ Resistance of a MnCeTiO_x Catalyst Prepared by the Co-Precipitation Method

Abstract

Three MnCeTiO_x catalysts with the same composition were prepared by conventional co-precipitation (MCT-C), reverse co-precipitation (MCT-R), and parallel co-precipitation (MCT-P), respectively, and their low-temperature SCR performance for de-NO_x was evaluated. The textural and structural properties, surface acidity, redox capacity, and reaction mechanism of the catalysts were investigated by a series of characterizations including N₂ adsorption and desorption, XRD, SEM, XPS, H₂-TPR, NH₃-TPD, NO-TPD, and in situ DRIFTs. The results revealed that the most excellent catalytic performance was achieved on MCT-R, and more than 90% NO_x conversion can be obtained at 100–300 °C under a high GHSV of 80,000 mL/(g_cat·H). Furthermore, MCT-R possessed optimal tolerance to H₂O and SO₂ poisoning. The excellent catalytic performance of MCT-R can be attributed to its larger BET specific surface area; higher contents of Mn⁴⁺, Ce³⁺, and adsorbed oxygen species; and more adsorption capacity for NH₃ and NO. Moreover, in situ DRIFTs results indicated that the NH₃-SCR reaction follows simultaneously the Langmuir–Hinshelwood and Eley–Rideal mechanisms at 100 °C. By adjusting the adding mode during the co-precipitation process, excellent low-temperature de-NO_x activity of MCT-R can be obtained simply and conveniently, which is of great practical value for the preparation of a MnCeTiO_x catalyst for denitrification.

1. Introduction

The selective catalytic reduction with NH₃ (NH₃-SCR) has been proven to be one of the most effective technologies for controlling NO_x emissions [1,2,3]. The research of low-temperature SCR catalysts has attracted considerable concerns owing to the lower temperatures of flue gas from non-electric industries [4]. Among the numerous low-temperature NH₃-SCR catalysts, MnO_x-based catalysts received tremendous attention due to their excellent de-NO_x activity at low temperatures [5,6,7]. TiO₂, as a typical NH₃-SCR catalyst carrier, not only offers excellent dispersion of active components but also promotes the decomposition of sulfate species formed in the process of SO₂ poisoning, thus alleviating the SO₂ poisoning effect of the catalysts [8]. The rare earth element Ce can induce surface electron imbalance of the catalyst, forming unsaturated chemical bonds and oxygen vacancies and increasing the concentration of adsorbed oxygen on the catalyst surface [9]. Therefore, the catalyst of MnCeTiO_x is considered a promising low-temperature de-NO_x catalyst. Nevertheless, MnCeTiO_x still has some obvious disadvantages, such as its relatively poor H₂O/SO₂ tolerance, narrow operation temperature window, and low N₂ selectivity [9,10]. Hence, it is important to further develop MnCeTiO_x catalysts with excellent low-temperature SCR activity, high N₂ selectivity, and strong tolerance to H₂O/SO₂.

As is well-known, in addition to the component effect, the catalyst preparation method also has a significant effect on the catalytic performance [11]. Compared with other preparation methods, the co-precipitation method is widely used due to its simple preparation process, low cost, and the ability to refine the catalyst grain and increase the BET specific surface area [12]. Chao et al. [13] compared the effects of sol–gel, citric acid complexing, and co-precipitation methods on the catalytic performance of Mn-Ce-O_x/TiO₂ for NH₃-SCR. The results showed that the catalyst prepared by co-precipitation methods exhibited the best low-temperature de-NO_x activity and resistance to SO₂ poisoning. Wu et al. [14] also investigated the influence of preparation methods on the SCR activity of Mn-Ce/TiO₂ and found that the SCR activity of catalysts decreased in the following sequence of co-precipitation method > citric acid method > impregnation method. Due to the large BET specific surface area and highly dispersed active component particles, the catalyst prepared by the co-precipitation method showed superior SCR performance.

The preparation parameters employed during the catalyst synthesis via the co-precipitation method significantly influence the NH₃-SCR activity. For instance, Chen et al. [11] found that MnFeO_x prepared with (NH₄)₂CO₃ as the precipitating agent had the best low-temperature de-NO_x activity compared to those prepared with NH₄OH and NaOH. Sheng et al. [15] investigated the effect of different Ti sources on Mn-Ce-O_x/TiO₂ prepared by co-precipitation, and their findings revealed that Mn-Ce-O_x/TiO₂ prepared using Ti(NO₃)₄ as the Ti source exhibited superior SCR activity compared to those synthesized using Ti(SO₄)₂. On the other hand, the adding mode can also have a significant impact on the performance of catalysts prepared by the co-precipitation method. For example, Rahmani et al. [16] synthesized pure Yttrium Aluminum Garnet (YAG) nanoparticles by co-precipitation, and the effects of adding mode (normal and reverse) on the phase evolution, thermal behavior, the composition and morphology of precipitation, and chemical bonding of the powders were investigated. The results showed that the reverse adding mode obtained a more homogeneous fluffy precipitate with higher carbonate content. Thus, it is expected that various adding modes during the preparation of MnCeTiO_x using the co-precipitation method may lead to significant differences in the structure and performance of the catalysts. However, to the best of our knowledge, the preparation of MnCeTiO_x by the co-precipitation method with different adding modes has not been explored to date.

In this work, the MnCeTiO_x catalysts were prepared by a co-precipitation method with three different adding modes, and their NH₃-SCR activities and the tolerance to H₂O/SO₂ were investigated. The reasons for the significantly different low-temperature activity and resistance to H₂O/SO₂ over MnCeTiO_x prepared by co-precipitation with different adding modes were explored by extensive characterizations. Furthermore, the reaction mechanisms of the related catalysts for NH₃-SCR were investigated using in situ DRIFTs.

2. Results and Discussion

2.1. Textural and Structural Properties

2.1.1. XRD

Figure 1a shows the XRD patterns of MCT-C, MCT-R, and MCT-P. As can be seen, MCT-P exhibits diffraction peaks at 25.16°, 37.74°, 47.99°, 55.01°, and 62.62°, which can be indexed to anatase TiO₂ (PDF#21-1272) [17,18]. Furthermore, the weak peaks of MnO₂ (PDF#72-1982) and CeO₂ (PDF#65-5923) are, respectively, observed at 27.95° and 32.9° for MCT-P. Similarly, both TiO₂ and CeO₂ diffraction peaks are clearly observed in MCT-C. Dissimilarly, only two weak diffraction peaks of TiO₂ (25.16° and 55.01°) and one weak diffraction peak of MnO₂ (27.95°) are observed on MCT-R, while no distinct diffraction peak of CeO₂ can be observed. These results indicate that CeO₂ is highly dispersed on the catalyst surface or in an amorphous state in MCT-R [19]. These phenomena reveal that reverse co-precipitation enhances the dispersion of active species, thereby increasing the BET specific surface area of MCT-R, as shown in

Table 1. BET specific surface area and pore characteristics of fresh and SO₂ poisoned catalysts.

Catalyst	BET Specific Surface Area (m2/g)	Pore Volume (cm3/g)	Average Pore Diameter (nm)
MCT-C	97.9	0.22	8.8
MCT-R	141.6	0.46	12.9
MCT-P	54.0	0.23	17.2
MCT-C-S	79.3	0.18	9.2
MCT-R-S	112.7	0.36	12.6
MCT-P-S	39.4	0.19	18.9

.

2.1.2. N₂ Adsorption–Desorption

Figure 1b shows the N₂ adsorption–desorption isotherm of MCT-C, MCT-R, and MCT-P. The BET specific surface area, pore volume, and pore diameter of the catalysts are listed in Table 1. According to the IUPAC classification, all three catalysts present typical type IV with an H3 hysteresis loop, which indicates that they are mesoporous materials. Compared to other catalysts, the larger BET specific surface area (141.6 m²/g) and pore volume (0.46 cm³/g) can be observed on MCT-R.

2.1.3. SEM/EDX

Figure S1 shows the SEM images of all investigated catalysts, all of which can be observed as rough, irregular particles, stacked and agglomerated. Figure S2 shows the elemental mapping of Mn, Ce, Ti, and O. It can be found that the elements of Mn, Ce, Ti, and O are uniformly distributed on the surface of MCT-R, while these elements are somewhat aggregated on that of MCT-C and MCT-P. These observations illustrate that the reverse co-precipitation method alleviates the disadvantage of uneven distribution of component aggregates to a certain extent.

2.2. Catalytic Activity

2.2.1. NH₃-SCR Activity

Figure 2 shows the low-temperature SCR performance of MCT-C, MCT-R, and MCT-P at 50–300 °C. As shown in Figure 2a, the NO_x conversion of MCT-C and MCT-P first increases and then decreases with the increase in the reaction temperature. However, the NO_x conversion of MCT-R increases at 50–100 °C, remains the same at 100% at 100–200 °C, and then decreases slightly at 200–300 °C. Notably, MCT-R shows the widest operation temperature window (above 90% NO_x conversion), of 100–300 °C, whereas the corresponding values of MCT-C and MCT-P are 150–300 °C and 200–250 °C, respectively. These results illustrate that the reverse adding mode remarkably improves the low-temperature activity and broadens the operation temperature window of the MCT catalyst.

The N₂ selectivity of all the catalysts is provided in Figure 2b. The N₂ selectivity of MCT-C remained at 100% in the range of 50–200 °C and decreased at 200 °C. However, the N₂ selectivity of MCT-R and MCT-P decreases from 150 °C and 100 °C, respectively. In the whole temperature range, the N₂ selectivity of MCT-C and MCT-R are higher than 75%, showing that MCT-C and MCT-R exhibit better N₂ selectivity than MCT-P.

A comparison of the de-NO_x performance of Mn-Ce-Ti catalysts prepared by the most common impregnation and co-precipitation methods is summarized in Table S1. It was found that MCT-R prepared in this work shows a wider operation temperature window (>90% NO_x conversion at 100–300 °C) and excellent N₂ selectivity (~100% at 50–150 °C). The results illustrate a promising prospect in the practical application of low-temperature flue gas stemmed from non-electricity industries.

2.2.2. H₂O and SO₂ Resistance

Figure 3 shows the resistance to H₂O and SO₂ of the catalysts for NH₃-SCR. The effect of H₂O on SCR performance at 200 °C is shown in Figure 3a. The NO_x conversion of MCT-R remains close to 100% after 10% H₂O is introduced to the NH₃-SCR system for 6 h, while that of MCT-C and MCT-P gradually decreases to 96% and 88% for 6 h, respectively. Furthermore, the SCR activities of three catalysts could be fully recovered to the initial level after stopping the introduction of H₂O. These results suggest that the deactivation of MCT by H₂O is reversible, which may be attributed to the competitive adsorption between H₂O and reactants rather than the destruction of the active sites [20].

Figure 3b shows the effect of SO₂ on the SCR performance of three catalysts at 200 °C. NO_x conversion of MCT-R is below 90% after SO₂ introduction for 4 h, whereas that of MCT-C and MCT-P is below 90% after SO₂ introduction for 2 h. Notably, the NO_x conversion of MCT-R reduces to 70% after 8 h, whereas that of MCT-C and MCT-P decreases to 56% and 51%, respectively. Additionally, the NO_x conversion of all the catalysts remains unchanged after stopping the introduction of SO₂. These results indicate that the catalyst’s deactivation resulting from SO₂ poisoning is irreversible, consistent with the results reported by Li et al. [21]. It is widely recognized that there are two main reasons for the irreversible deactivation of the catalyst poisoned by SO₂. One is that the formation of ammonium sulfate ((NH₄)₂SO₃) and/or ammonium bisulfate (NH₄HSO₄) covers the active sites and blocks the pores of the catalyst. The other one is that the active component of the catalysts undergoes sulfation, forming inactive sulfates. For the Mn-based catalysts, it is generally accepted that the latter poisoning effect is more serious than the former [22,23]. Notably, Shi et al. [24] found that the metal sulfates are more easily formed on MnO_x-CeO₂-TiO₂, whereas (NH₄)₂SO₃ and NH₄HSO₄ are rarely formed. Furthermore, Zhang et al. [25] reported that the formation of sulfate species is temperature-dependent. At a low temperature (100 °C), the deposition of (NH₄)₂SO₃ and NH₄HSO₄ on MnO_x-CeO₂-TiO₂ by SO₂ is dominant, whereas the deactivation predominantly caused by the sulfation of the active components occurs at a high temperature (200 °C). Therefore, the irreversible deactivation of Mn-Ce/TiO₂ observed in the work is mainly due to the sulfation of active components, leading to the formation of inactive sulfates.

In summary, MCT-R prepared in this work exhibits excellent low-temperature de-NO_x activity, a broader operational temperature window, and superior resistance to H₂O and SO₂. It provides a good reference for efficiently solving the existing shortcomings of an MCT catalyst mentioned in Section 1 (Introduction).

2.3. Separate Oxidation of NO and NH₃

As shown in Figure 4a, negligible conversion of NH₃ can be observed over MCT-C, MCT-R, and MCT-P at temperatures below 150 °C. The NH₃ conversion of MCT-C and MCT-P increases at 200–300 °C and reaches almost 90% at 300 °C. Conversely, MCT-R has much higher activity in NH₃ oxidation, which rises sharply in the range of 150–300 °C. Overall, the NH₃ oxidation conversion of the catalysts abided by the order of MCT-R > MCT-C ≥ MCT-P, indicating that MCT-R possessed the strongest NH₃ oxidation ability.

It has been widely recognized that the oxidation of NO to N₂O can improve the de-NO_x activity through the “Fast SCR” reaction (2NH₃ + NO + NO₂ → 2N₂ + 3H₂O) [8]. It is noteworthy that the NO oxidation conversion of all the catalysts is similar at 50–150 °C (Figure 4b), which demonstrates that the “Fast SCR” reaction is not predominant in this temperature region. A similar phenomenon has also been reported on Sm-MnO_x [26] and MnCoVO_x [27]. Nevertheless, the NO oxidation conversion significantly differs at higher temperatures of 150–300 °C. Specifically, the NO oxidation conversion on MCT-C and MCT-P are similar, ranging from 10% to 40%, whereas MCT-R spans a broader range, from 10% to 70%. The NO oxidation conversion followed the order of MCT-R > MCT-C ≥ MCT-P, indicating that MCT-R exhibits the strongest activity of NO oxidation at higher temperatures.

2.4. XPS Analysis

The XPS analysis was employed to investigate the surface composition and chemical nature of MCT-C, MCT-R, and MCT-P. The Mn 2p, Ce 3d, O 1s, and Ti 2p spectra of the catalysts are exhibited in Figure 5. The corresponding peaks of each catalyst were fitted to the split peaks and the peak areas were used to calculate the relative concentration ratios of the different valence states of the elements, and the results are summarized in Table S2.

Figure 5a shows the XPS patterns of Mn 2p. The two distinct peaks centered at about 642.0 and 653.5 eV could be ascribed to Mn 2p_3/2 and Mn 2p_1/2, respectively [21]. The Mn 2p_3/2 spectra could be fitted into three peaks of Mn⁴⁺ (644.86 ± 0.1 eV), Mn³⁺ (642.66 ± 0.1 eV), and Mn²⁺ (641.34 ± 0.1 eV) [21,28]. As can be seen from Table S2, the Mn⁴⁺/Mnⁿ⁺ of MCT-R (29.8%) is higher than that of MCT-C (27.2%) and MCT-P (27.4%), suggesting that the reverse co-precipitation method improves the dispersion of high valence Mn species on the catalyst surface. According to the literature [29], Mn⁴⁺ species play a crucial role in the SCR reaction and are conducive to the de-NO_x performance in low-temperature regions.

The Ce 3d XPS spectra of the catalysts are shown in Figure 5b. The peaks labeled as u′ (903.2 ± 0.1 eV) and v′ (885 ± 0.2 eV) correspond to Ce³⁺, while the peaks denoted as u (900.7 ± 0.1 eV), u″ (906.9 ± 0.2 eV), u″′ (916.4 eV), v (882.1 ± 0.1 eV), v″ (888.7 eV), and v″′ (897.9 ± 0.1 eV) correspond to Ce⁴⁺ [30,31]. It is widely recognized that Ce³⁺ facilitates the formation of oxygen vacancies and charge imbalance on the catalyst surface, which in turn promotes oxygen migration from the bulk to the surface, thereby improving SCR activity [31]. From Table S2, the ratio of Ce³⁺/Ceⁿ⁺ (23.5%) over MCT-R is markedly higher than that of MCT-C (19.7%) and MCT-P (18.9%), demonstrating that the formation of Ce³⁺ species is promoted by the reverse co-precipitation method and consequently improves the NH₃-SCR activity.

For the Ti 2p spectrum, it can be fitted into two kinds of peaks, as shown in Figure 5c. The binding energies located at 464.5 eV and 458.6 eV correspond to Ti 2p_1/2 and Ti 2p_3/2, respectively, which are characteristic of the Ti⁴⁺ species [21].

Figure 5d shows the XPS spectra of O 1s. It can be deconvoluted into two characteristic peaks at 530–530.5 eV and 531.5–531.8 eV, which are attributed to the lattice oxygen (denoted as O_β) and surface chemisorbed oxygen (denoted as O_α), respectively [10,28]. It is well known that compared to O_β species, the higher oxygen mobility and reactivity are obtained on O_α species [19]. As shown in Table S2, the O_α/O (36.1%) ratio of the MCT-R catalyst is significantly higher than that of MCT-C (25.7%) and MCT-P (28.1%), indicating that the catalyst prepared by the reverse co-precipitation method has much more surface chemisorbed oxygen species (O_α).

2.5. H₂-TPR

H₂-TPR profiles of MCT-C, MCT-R, and MCT-P catalysts are illustrated in Figure 6. For the MCT-C and MCT-P, the peaks at 335 °C (375 °C) and 365 °C (410 °C) may be attributed to the reduction of MnO₂ → Mn₂O₃ and Mn₂O₃ → Mn₃O₄, respectively, and the reduction peak at 430 °C (465 °C) may be attributed to the overlapping reduction of Mn₃O₄ → MnO and surface CeO₂. Notably, MCT-R shows a broader overlapping peak at 150–450 °C, which can be attributed to the successive reduction of MnO_x species (MnO₂ → Mn₂O₃ → Mn₃O₄) and the reduction of surface CeO₂ species [32]. The appearance of seriously overlapping peaks over MCT-R can be attributed to the strong interaction between MnO_x, CeO_x, and TiO₂ as evidenced by the results of XRD.

For the temperatures of the reduction peaks, the highest reduction temperature is observed on MCT-P, indicating the worst redox ability of MCT-P. Evidently, the reduction temperature of MCT-R is the lowest. The shift of the reduction peak towards lower temperatures implies better redox properties of MnO_x and CeO₂ of MCT-R [31]. Clearly, the optimal redox ability is obtained on MCT-R due to the synergistic effect of MnO_x, CeO_x, and TiO₂ [4].

2.6. TPD Analyses for NH₃ and NO

The adsorption abilities of NH₃ on MCT-C, MCT-R, and MCT-P catalysts were investigated by the NH₃-TPD experiment. As can be observed from Figure 7a, there are two distinct NH₃ desorption peaks for each catalyst, which correspond to weak desorption peaks (100–300 °C) and middle-strong desorption peaks (300–500 °C), respectively [33]. Additionally, the normalized desorption peak areas S (a.u.) of all the investigated catalysts follows the order of MCT-R (1.00) > MCT-C (0.97) > MCT-P (0.85), indicating the more acidic peak sits on MCT-R.

Figure 7b shows the NO-TPD profiles of MCT-C, MCT-R, and MCT-P catalysts. Apparently, NO desorption on all catalysts can be divided into two zones between 50 and 650 °C. The peaks below 200 °C could be attributed to the decomposition of monodentate nitrite/nitrate species, whereas the broad peaks above 200 °C are due to the decomposition of bridging nitrate species and bidentate nitrate species with higher thermal stability [34]. Additionally, the desorption peak areas of NO at 100–200 °C were calculated, and the normalized peak areas (a.u.) follow the sequence of MCT-R (1.00) > MCT-C (0.87) > MCT-P (0.60).

2.7. In Situ DRIFTs Characterization

2.7.1. Adsorption of NH₃ Species

As shown in Figure 8, in situ DRIFTs spectra of NH₃ adsorption on MCT-R and MCT-P catalysts at different temperatures were recorded. It can be observed that there are several bands in the range of 1000–2000 cm⁻¹. The bands at 1180 cm⁻¹ (1185 cm⁻¹) and 1610 cm⁻¹ (1605 cm⁻¹) are assigned to NH₃ species on Lewis acid sites, and those at 1425 cm⁻¹ and 1480 cm⁻¹ belong to NH₄⁺ species on Brønsted acid sites [13,35]. The bands of NH₃ species on the Lewis acid site remained present at the high temperature (300 °C), whereas those on the Brønsted acid sites disappeared gradually with increasing temperature and completely after 200 °C. Furthermore, the peak intensity of NH₃ adsorbed on Lewis acid sites is significantly stronger than on Brønsted acid sites. These results indicate that Lewis acid sites have stronger thermal stability and dominate the acid sites on both catalysts. Clearly, under the same scale, the band intensities of adsorbed NH₃ species on MCT-R are all significantly stronger than those on MCT-P, proving that MCT-R exhibits more active sites for NH₃ adsorption [36].

2.7.2. Adsorption of NO_x Species

As shown in Figure 9, in situ DRIFT spectra of NO adsorption on MCT-R and MCT-P at different temperatures were recorded. The bands at 1278 cm⁻¹ (1290) cm⁻¹, 1355 cm⁻¹, 1510 cm⁻¹ (1140 cm⁻¹ and 1580 cm⁻¹), 1620 cm⁻¹ (1615 cm⁻¹), and 1700 cm⁻¹ can be ascribed to monodentate nitrate, monodentate nitrite, bidentate nitrate, bridged nitrate, and adsorbed NO species, respectively [37,38]. The bands of monodentate nitrate (1278(1290) cm⁻¹) weakened with increasing temperature and finally disappeared at 250 °C owing to the poor thermal stability. In contrast, the bands of bidentate nitrate (1510 cm⁻¹) and bridging nitrate (1620(1615) cm⁻¹) species are still present at 300 °C. Notably, the bands of bidentate nitrate species at 1140 cm⁻¹ and 1580 cm⁻¹ appeared on MCT-R at higher temperatures, which may be due to the conversion of monodentate nitrite and monodentate nitrate to thermally more stable bidentate nitrate species at higher temperatures [39]. Additionally, the number of adsorption peaks on MCT-R is significantly greater than that on MCT-P, indicating the presence of more different types of NO_x intermediates on MCT-R.

2.7.3. Reaction between NO_x and Pre-Adsorbed NH₃ Species

Figure 10 shows in situ DRIFT spectra of the reaction between NO_x and pre-adsorbed NH₃ species on MCT-R and MCT-P catalysts at 100 °C. As presented in Figure 10a, the NH₃ species absorbed on Lewis acid sites (1180 and 1610 cm⁻¹) and Brønsted acid sites (1425 cm⁻¹) can be detected after NH₃ was introduced for 60 min. The peaks at 1180, 1425, and 1610 cm⁻¹ disappeared at 30 min, 10 min, and 10 min after the introduction of NO and O₂ mixtures, respectively. With the continuous introduction of NO and O₂ mixtures, the characteristic bands of monodentate nitrate (1270 cm⁻¹) and bidentate nitrate (1525 and 1535 cm⁻¹) species appear on the catalyst surface. These results suggest that all NH₃ species can react with NO_x via the Eley–Rideal (E-R) mechanism [26,40]. A similar situation for the reaction between NO_x and pre-adsorbed NH₃ species can be observed on MCT-P. Differently, the band intensity at 1180 cm⁻¹ on MCT-R is stronger than that on MCT-P, suggesting more adsorption of NH₃ species on MCT-R and thus promoting the elimination of NO_x by the E-R route for NH₃-SCR.

2.7.4. Reaction between NH₃ and Pre-Adsorbed NO_x Species

Figure 11 shows in situ DRIFT spectra of the reaction between NH₃ and pre-adsorbed NO_x species on MCT-R and MCT-P at 100 °C. After the catalysts were pretreated in a mixture of NO and O₂, the bands of monodentate nitrate (1290 cm⁻¹), bidentate nitrate (1490 cm⁻¹), and bridged nitrate (1610(1615) cm⁻¹) species can be detected. With the introduction of NH₃, the bridged nitrate species at 1615(1610) cm⁻¹ is gradually consumed, whereas the monodentate nitrate at 1290 cm⁻¹ and the bidentate nitrate species at 1490 cm⁻¹ remain almost unchanged. These observations demonstrate that bridged nitrate with higher reactivity can react with NH₃ by the Langmuir–Hinshelwood (L-H) route for NO_x elimination [41]. Significantly, the number of bridged nitrates species on MCT-R is significantly larger than that on MCT-P, indicating that MCT-R could provide more adsorption sites, and thus improve the activation capability.

2.8. Characterization of SO₂-Poisoned Samples

As described in Section 2.2.2, MCT-R exhibits better resistance to SO₂ poisoning than MCT-C and MCT-P. To explore the reason for the superior SO₂ resistance of MCT-R, the SO₂-poisoned MCT catalysts (MCT-C-S, MCT-R-S, and MCT-P-S) were characterized and discussed in the following sections.

2.8.1. N₂ Adsorption–Desorption

As shown in Table 1, BET specific surface areas of MCT-C-S, MCT-R-S, and MCT-P-S decreased after SO₂ poisoning, which may be attributed to the formation of metal sulfates that block the catalyst pore structure [23,42]. However, BET specific surface area and the pore volume of MCT-R-S (112.7 m²/g, 0.36 cm³/g) are still significantly higher than those of MCT-C-S (79.3 m²/g, 0.18 cm³/g) and MCT-P-S (39.4 m²/g, 0.19 cm³/g). This phenomenon might be explained by the fact that the reverse co-precipitation method could enhance the dispersion of sulfate species due to a larger BET specific surface area. Further evidence will be introduced in the XPS analysis described in Section 2.8.2.

2.8.2. XPS Analysis

To investigate the impact of SO₂ on the chemical states of Mn, Ce, Ti, and O atoms on the surfaces of MCT-R, MCT-C, and MCT-P, XPS techniques were utilized and the results are depicted in Figure S3. After peak fitting deconvolution, the XPS spectra for various elements were resolved into multiple peaks. The relative ratios of different oxidation states were calculated based on peak areas and summarized in Table S2.

The Mn 2p spectra of SO₂-poisoned catalysts are shown in Figure S3a. Compared to fresh catalysts, the higher ratios of Mn²⁺/Mnⁿ⁺ are observed on the corresponding poisoned catalysts due to the generation of MnSO₄ [23], which induces a decrease in the relative ratios of Mn⁴⁺/Mnⁿ⁺ and Mn³⁺/Mnⁿ⁺ (Table S2). Therefore, a decreasing trend of NH₃-SCR activity occurred after SO₂ was introduced into the SCR system. Mn⁴⁺ species as an active species; the relative ratios of Mn⁴⁺/Mnⁿ⁺ of MCT-R-S (26.9%) were also higher than that of MCT-C-S (26.2%) and MCT-P-S (25.7%).

Figure S3b shows the XPS patterns of Ce 3d on SO₂-poisoned catalysts. Ce⁴⁺ species can react with SO₂ to generate Ce₂(SO₄)₃ species by 2CeO₂ + 3SO₂ + O₂ → Ce₂(SO₄)₃ [43]. Ce₂(SO₄)₃ species can deposit on the catalyst’s surface or active sites, obstructing de-NO_x performance [21]. As shown in Table S2, the relative ratio of Ce³⁺/Ceⁿ⁺ increased for all the SO₂-poisoned catalysts. Especially, compared to MCT-C-S and MCT-P-S, the smallest increase in the relative concentration of Ce³⁺ species is observed on MCT-R-S, revealing the least formation of Ce₂(SO₄)₃ species.

Figure S3c shows Ti 2p spectra of SO₂-poisoned catalysts, and the form of Ti⁴⁺ and the corresponding binding energy were not affected by introducing SO₂.

In Figure S3d, the O 1s XPS spectra of the SO₂-poisoned catalysts shift to higher binding energies compared to the fresh catalysts, and more surface chemisorbed oxygen (O_α) species are detected on the SO₂-poisoned catalysts due to the formation of sulfate species [44,45]. Notably, the increased degree of O_α in MCT-R-S (3.3%) is remarkably lower than that of MCT-P-S (11.3%) and MCT-C-S (11.1%). This result reasonably accounts for the superior sulfur resistance of MCT-R-S since the formed sulfate species would cover the active sites, consequently resulting in a decrease in catalytic activity compared to the fresh catalysts [45,46].

Figure S3e shows the XPS profile of S 2p over SO₂-poisoned catalysts. Two peaks assigned to SO₄²⁻ (~169.3 eV) and SO₃²⁻ (~168.2 eV) are observed on the S 2p spectra [23,47]. The S content of MCT-C-S, MCT-R-S, and MCT-P-S is 2.8%, 3.2%, and 3.1%, respectively, as shown in Table S2. This observation indicates that metal sulfate substances do exist in the catalysts after SO₂ poisoning, and the amount of metal sulfates is similar. However, the BET specific surface area of MCT-R-S (112.7 m²/g) is much larger than MCT-C-S (79.3 m²/g) and MCT-P-S (39.4 m²/g) in Table 1, contributing to the dispersion of sulfate species and alleviating the poisoning effect of SO₂ on the catalyst [23]. Additionally, the results of S atom/m² were calculated as follows: MCT-P-S (0.079%) > MCT-C-S (0.035%) > MCT-R-S (0.028%). The results indicated that the lowest S atomic ratio per unit surface area is found on MCT-R-S, suggesting that S coverage is lower on MCT-R-S, thereby exhibiting the best SO₂ resistance.

2.9. Promotional Effect of Reverse Co-Precipitation on the Catalytic Performance of MCT

Compared to MCT-C and MCT-P, excellent catalytic performance in NH₃-SCR is exhibited on MCT-R, with above 90% NO_x conversion achieved at 100–300 °C. Additionally, optimal resistance to H₂O/SO₂ is also demonstrated by MCT-R. The excellent catalytic performance of MCT-R can be attributed to highly dispersed active sites and the higher BET specific surface area; higher contents of Mn⁴⁺, Ce³⁺, and O_α; stronger reducibility; and more amounts of active NH₃ and active bidentate nitrate species, which will be discussed in detail as follows.

Ⅰ. Activity

(1) Textural and structural properties: It is well-known that the dispersion of active sites is crucial for catalyst performance [48,49]. No diffraction peaks corresponding to Ce species and only weak diffraction peaks for Mn and Ti species are detected on the MCT-R in Figure 1a, suggesting that the Mn, Ce, and Ti oxides are highly dispersed on MCT-R [19,50]. On the other hand, the higher dispersion of the components favors the increase in the BET specific surface area [18,48,49]. The large BET specific surface area contributes to the contact between the active sites and reactive gases, enhancing the adsorption and activation of the reactants [41,50]. Evidently, the BET specific surface area of MCT-R (141.6 m²/g) is significantly higher than that of MCT-C (97.9 m²/g) and MCT-P (54.0 m²/g), as shown in Table 1. Thus, an efficient NO_x conversion was obtained over MCT-R.

(2) Adsorption and activation of reactants: The superior redox capability of Mn⁴⁺ enhances NO_x removal at low temperatures, and Ce³⁺ species quickly engage in the Ce-Mn redox cycle (Mn³⁺ + Ce⁴⁺ ↔ Mn⁴⁺ + Ce³⁺) with O_α species to accelerate NO_x conversion [10,31,51]. Obviously, the XPS results reveal that the relative ratios of Mn⁴⁺, Ce³⁺, and O_α species on MCT-R are significantly higher than those in MCT-C and MCT-P (Table S2), suggesting that more active sites are exhibited on MCT-R. Meanwhile, the abundant active sites are in favor of NO_x and NH₃ adsorption [31,52]. Both NH₃ and NO desorption amounts are higher over MCT-R compared to MCT-C and MCT-P, which can be reflected in the NH₃/NO-TPD results (Figure 7). Moreover, the lower reduction temperatures are also observed on MCT-R (Figure 6), which implies the stronger reducibility on MCT-R. Notably, the excellent redox properties can activate adsorbed species and the activated active species to generate active intermediates [32,53]. Compared with MCT-P, the amount of active NH₃ species and active bidentate nitrate species is much higher over MCT-R, as shown in Figure 8 and Figure 9, which is also one of the reasons for achieving excellent de-NO_x activity.

Ⅱ. SO₂ resistance

The formation and coverage of metal sulfates on the active sites have been regarded as the main reasons for the deactivation of Mn-based catalysts by SO₂ poisoning [41,54,55]. Additionally, the larger BET specific surface can reduce the coverage of sulfate species on active sites owing to their higher dispersion [56,57]. In this study, the larger BET specific surface of MCT-R-S reduces the coverage of sulfate species, as shown in Table 1. As a result, the relative proportions of Ce³⁺/Ceⁿ⁺ and Mn⁴⁺/Mnⁿ⁺ are higher on MCT-R-S compared to MCT-C-S and MCT-P-S (Table S2), resulting in a higher exposure of the active sites and maintaining relatively high activity.

3. Materials and Methods

3.1. Catalyst Preparation

MnCeTiO_x catalysts were prepared by the co-precipitation method with different adding modes. The conventional co-precipitation method is as follows: Firstly, TiCl₄ (≥99%, Adamas) was slowly added to the C₂H₅OH (≥99.7%, Adamas) in an ice water bath with vigorous stirring until well mixed. Subsequently, a certain amount of Mn(NO₃)₂ (50% aqueous solution, Adamas) and Ce(NO₃)₃·6H₂O (≥99.99%, Adamas) were dissolved in the above solution. The aqueous solution formed by the dissolution of (NH₄)₂CO₃ (≥40%, Adamas) was used as the precipitant and slowly added to the above mixture and stirred at the temperature of 30 °C until the pH value reached 8. After aging at 30 °C for 24 h, the resulting precipitate was filtered and washed with deionized water several times. Finally, the sample was dried in an oven at 80 °C for 12 h and then calcined at 500 °C for 4 h in a muffle furnace, and the obtained catalyst was denoted as MCT-C. Similar to the above steps, the reverse co-precipitation method involves the slow dropwise addition of the above metal salt solution to the (NH₄)₂CO₃ solution, and the resulting catalyst was denoted as MCT-R. The parallel co-precipitation method involves simultaneously adding the above metal salt solution and the (NH₄)₂CO₃ solution dropwise to a beaker containing a small amount of deionized water, and the resulting catalyst was denoted as MCT-P. According to previous works [13,31], the optimal ratio of Mn:Ce:Ti is 0.4:0.1:1 in the MnCeTiO_x catalyst. Therefore, the molar ratio of Mn:Ce:Ti is a constant (0.4:0.1:1) in the work. Considering the toxicity of TiCl₄, the experiments were conducted in a fume hood. Meanwhile, personal protective equipment including gloves, masks, and goggles was worn to reduce the risk.

3.2. Characterization

The N₂ adsorption and desorption isotherm curves of the catalysts were measured using an ASAP 2020 HD88 adsorption device manufactured by Micromeritics, Inc. (Norcross, GA, USA). About 0.1 g of sample was degassed at 200 °C for 10 h under vacuum, followed by N₂ adsorption and desorption at –196 °C in liquid nitrogen. The specific surface area was determined using the BET model, while the pore volume and pore diameter were determined using the BJH model.

The X-ray diffraction patterns of the catalysts were measured using an X-ray diffractometer model XRD-6100 from Shimadzu, Japan. The diffractometer was scanned using Cu-Kα rays at an operating current and voltage of 40 mA and 40 kV with a rate of 7°/min and an angle 2θ of 10–70°. A Zeiss Gemini 300 scanning electron microscope was used for SEM analysis of the samples with a maximum operating voltage of 200 kV.

The X-ray photoelectron spectra of the catalysts were measured by an ESCALAB 250Xi photoelectron spectrometer from Thermo Scientific (Waltham, MA, USA) for the analysis of the surface elemental composition and valence states of each catalyst, using an Al Kα (1486.6 eV) target as the excitation source. The binding energy of each element measured was corrected using the C 1s peak (284.6 eV).

The temperature-programmed reduction experiments of the catalysts with H₂ (H₂-TPR) were carried out on a homemade device. First, 0.1 g of sample was loaded into a quartz reaction tube and pretreated with high purity nitrogen (99.999%) for 30 min at 300 °C to remove impurities and moisture from the catalyst surface, followed by cooling to 50 °C, switching the nitrogen to reducing gas (10% H₂/N₂) at a flow rate of 50 mL/min. The temperature was increased from 50 °C to 650 °C at a rate of 5 °C/min. The reduction curve was obtained from the change in the signal detected by the thermal conductivity cell detector (TCD) of the gas chromatograph (GC 9750, Fuli, China).

The temperature-programmed desorptions of NH₃ or NO_x pre-adsorbed on the catalysts were carried out on a homemade apparatus. Firstly, each catalyst was pretreated according to the same procedure as in the H₂-TPR experiment. For NH₃-TPD experiments, after the catalyst cooled to room temperature, the nitrogen gas was switched to an ammonia–nitrogen mixture (10% NH₃/N₂) at a flow rate of 30 mL/min. After 30 min, the physically adsorbed NH₃ on the catalyst surface was blown off with nitrogen gas, and then the gas was passed into the gas chromatograph. The temperature was increased from 50 °C to 650 °C with a ramp rate of 10 °C/min. For the NO-TPD experiments, He was used for both pretreatment and purging, and after the catalyst was cooled to room temperature, the He was replaced by a mixture of NO/N₂ (500 ppm NO) at a flow rate of 30 mL/min. The temperature was increased from 50 °C to 650 °C at a rate of 10 °C/min. The change of gas was detected on a mass spectrometer (Pfeiffer Vacuum Quadstar, 32-bit, Aßlar, Germany) during the temperature increase to obtain the desorption curve.

The in situ diffuse reflectance infrared Fourier transform spectra (DRIFTs) of the catalysts were measured using a Nicolet 6700 IR spectrometer (Waltham, MA, USA) with a cadmium mercury telluride detector, 64 spectral scans, and a resolution of 4 cm⁻¹. The catalysts were loaded into a reaction cell with a window sheet of calcium fluoride (CaF₂), and pretreatment in a nitrogen atmosphere occurred at 300 °C for 30 min. The spectra of samples were recorded under flowing N₂ at the required temperatures and set as backgrounds. Afterward, the gas mixture (500 ppm NH₃ or 500 ppm NO + 5 vol% O₂, and N₂ as the balanced gas) was introduced to the cell with a flow rate of 100 mL/min for 1 h, followed by N₂ purge. For the reaction between pre-adsorbed NH₃ and the mixture of NO and O₂, the catalysts were first exposed to NH₃ for 60 min, then N₂ purging occurred for 30 min before introducing the mixture of NO and O₂ into the reaction system. Similarly, the spectra of the transient reaction between pre-adsorbed NO_x and NH₃ on the catalysts at 100 °C were obtained by changing the introduction sequence of NH₃ and the mixture of NO and O₂.

3.3. Activity Tests

The NH₃-SCR reaction was carried out in a fixed-bed quartz reaction tube (6 mm inner diameter, and 400 mm length). First, the catalyst was pressed and sieved to obtain particles with a grain size of 40–60 meshes to eliminate the impacts of mass transfer limitation. Then, 0.15 g of catalyst was placed into the reactor for each experiment. The simulated flue gas consisted of 500 ppm NO, 500 ppm NH₃, 5 vol% O₂, 10% H₂O (when used), 50 ppm SO₂ (when used), and N₂ as balance gas. The total flow rate of the simulated flue gas was 200 mL/min, corresponding to a space velocity of 80,000 mL/(g_cat·h). The concentrations of NO and NO₂ in the outlet stream of the reactor were analyzed via a flue gas analyzer (Testo 340, Lenzkirch, Germany), and N₂O concentration was detected by online gas chromatography (Panna, A91, China; Polymer Quaternary Ammonium (PQ) column). Catalytic activity testing was performed from 50 °C to 300 °C with a step of 50 °C, and reactant and product concentration in the reactor were determined after 30 min of stabilization at each temperature point. For the resistance experiments to H₂O or SO₂, the SCR reaction was required to stabilize for 2 h at 200 °C, and then H₂O or SO₂ was introduced into the simulated flue gas. The samples after the SO₂ resistance testing were recovered and donated as MCT-C-S, MCT-R-S, and MCT-P-S, respectively. The NO and NH₃ oxidation reaction performance test was also performed on the SCR reaction setup, but the reaction gas is slightly different, i.e., 500 ppm NO (when used), 500 ppm NH₃ (when used), 5 vol% O₂, and N₂ as carrier gas.

NO_x conversion (NO_x = NO + NO₂) and N₂ selectivity were calculated by the following equations:

$${\mathrm{NO}}_x \mathrm{conversion} \left(\%\right)={\displaystyle \frac{{\left[{\mathrm{NO}}_x\right]}_{\mathrm{in}}-{\left[{\mathrm{NO}}_x\right]}_{\mathrm{out}}}{{\left[{\mathrm{NO}}_x\right]}_{\mathrm{in}}}}\times 100\%$$

(1)

$${\mathrm{N}}_2 \mathrm{selectivity} \left(\%\right)=\left(1-{\displaystyle \frac{2\times {\left[{\mathrm{N}}_2\mathrm{O}\right]}_{\mathrm{out}}}{{\left[{\mathrm{NO}}_x\right]}_{\mathrm{in}}+{\left[{\mathrm{NH}}_3\right]}_{\mathrm{in}}-{\left[{\mathrm{NO}}_x\right]}_{\mathrm{out}}-{\left[{\mathrm{NH}}_3\right]}_{\mathrm{out}}}}\right)\times 100\%$$

(2)

The conversion of NO oxidation, NH₃ oxidation, and N₂O selectivity were calculated by the following equations:

$$\mathrm{NO} \mathrm{conversion} \left(\%\right)={\displaystyle \frac{{\left[\mathrm{NO}\right]}_{\mathrm{in}}-{\left[\mathrm{NO}\right]}_{\mathrm{out}}}{{\left[\mathrm{NO}\right]}_{\mathrm{in}}}}\times 100\%$$

(3)

$${\mathrm{NH}}_3 \mathrm{conversion} \left(\%\right)={\displaystyle \frac{{\left[{\mathrm{NH}}_3\right]}_{\mathrm{in}}-{\left[{\mathrm{NH}}_3\right]}_{\mathrm{out}}}{{\left[{\mathrm{NH}}_3\right]}_{\mathrm{in}}}}\times 100\%$$

(4)

$${\mathrm{N}}_2\mathrm{O} \mathrm{selectivity}\left(\%\right)={\displaystyle \frac{2\times { \left[{\mathrm{N}}_2\mathrm{O}\right]}_{\mathrm{out}}}{{\left[{\mathrm{NH}}_3\right]}_{\mathrm{in}}-{\left[{\mathrm{NH}}_3\right]}_{\mathrm{out}}}}\times 100\%$$

(5)

The subscripts “in” and “out” indicate the inlet and outlet of the gas, respectively.

4. Conclusions

In summary, different adding modes have a great influence on the low-temperature catalytic activity and H₂O/SO₂ resistance of the MnCeTiO_x catalyst prepared by the co-precipitation method. Among them, MCT-R possesses excellent catalytic performance at 100–300 °C and optimal resistance to H₂O or SO₂ at 200 °C. This remarkable SCR performance is attributed to the fact that the reverse co-precipitation can enlarge the specific surface area, promote the dispersion of MnO_x and CeO_x, and improve the redox capacity, as well as the activation of NO_x and NH₃. Excellent low-temperature de-NO_x activity of the MnCeTiO_x catalyst can be obtained simply and conveniently by changing the adding mode during the co-precipitation process, which is of great practical value for the preparation of the MnCeTiO_x catalyst for denitrification.