3.1. NH3-SCR activity
Figure 1 shows catalytic activities of the montmorillonite clay, Fe-PILC, 4Mn/Fe-PILC, and yCe-4Mn/Fe-PILC (y = 2.0, 4.0, 6.0 wt%) samples for the NH3-SCR reaction. NO conversion over each sample first increased significantly with the rise in temperature from 150 to 350 ℃, and then decreased gradually from 350 to 400 ℃. The raw clay sample exhibited the lowest NO conversion with a maximum of only about 10%. The Fe-PILC and 4Mn/Fe-PILC samples showed higher NO conversions than the raw clay, which reached about 70 and 85% at 350 ℃, respectively. Over the yCe-4Mn/Fe-PILC samples, NO conversion increased with the rise in Ce doping less than 4.0 wt%, but an excessive doping of Ce (> 4.0 wt%) to 4Mn/Fe-PILC led to a decrease in catalytic activity. Among all of the samples, 4Ce-4Mn/Fe-PILC performed the best, with NO conversions of over 80% being achieved in the temperature range of 250−400 °C and a maximum conversion of 9 % at 350 °, ,which was much higher than those over the commercial catalysts [20−22].
According to our previous study [17], the increase in catalytic activity of the PILC-based samples (compared with than that of the raw clay) was attributed to introduction of the active sites and increase in surface area of the sample after pillaring. When ceria was doped to 4Mn/Fe-PILC, catalytic activity over yCe-4Mn/Fe-PILC was significantly higher than that over 4Mn/Fe-PILC. This result was associated with the strong interaction between MnOx and CeO2. Moreover, the gradual decrease in NO conversion from 350 to 400 ℃ might result from the phenomenon reported previously [23, 24], i.e., hindering the oxidation of NH3 by O2 to generate NO at high temperatures.
Owing to the yCe-4Mn/Fe-PILC sample with a Ce loading of 4.0 wt% showed the best NH3-SCR performance and the widest catalytic activity window, we choose this catalyst to study its resistance to SO2 poisoning. Effects of SO2 on SCR activity of 4Mn/Fe-PILC and 4Ce-4Mn/Fe-PILC at 350 ℃ are shown in Fig. 2. Before introducing SO2, NO conversions over the 4Mn/Fe-PILC and 4Ce-4Mn/Fe-PILC samples were 85 and 95%, respectively. When 200 ppm SO2 was introduced into the reaction system, NO conversions over both samples started to decrease, but their drop trends were significantly different. Over the 4Mn/Fe-PILC sample, NO conversion dramatically declined and was finally stabilized at 54%, while NO conversion over the 4Ce-4Mn/Fe-PILC sample decreased only slightly and was maintained at ca. 83%. After cutting off SO2, NO conversion over 4Mn/Fe-PILC was recovered to only 68%, while that over 4Ce-4Mn/Fe-PILC was rapidly restored to 89%, indicating that the 4Ce-4Mn/Fe-PILC sample exhibited better SO2 resistance. The results demonstrate that the introduction of SO2 can inhibit NH3-SCR catalytic activity of the 4Mn/Fe-PILC and 4Ce-4Mn/Fe-PILC samples, and the doping of ceria with an appropriate amount can improve its SO2 resistance.
3.2. Crystal phase composition
In order to further study the doping of ceria to improve SO2-resistant performance of the catalyst, XRD techniques were used to identify crystal phases of the samples before and after SO2 poisoning. XRD patterns of the fresh and SO2-poisoned 4Mn/Fe-PILC-S and 4Ce-4Mn/Fe-PILC-S samples are illustrated in Fig. 3. The XRD peaks of all of the samples were consistent with the characteristic peaks at 2θ = 19.8°, 26.6°, 28°, and 34.9° of montmorillonite clay, and the peak at 2θ < 10° was characteristic of pillar [25]. No characteristic peaks due to the other impurity phases were detected, indicating that the catalysts before and after sulfur dioxide poisoning maintained the clay structure. The characteristic peaks assignable to the manganese oxide and ceria phases were not obviously seen, indicating that the two metal oxides were uniformly dispersed on the surface of Fe-PILC. No diffraction peaks due to the sulfate species were observed in the 4Mn/Fe-PILC-S and 4Ce-4Mn/Fe-PILC-S samples, demonstrating that the sulfate species exists mainly in an amorphous state on the surface of the samples. For the 4Mn/Fe-PILC-S sample, intensity of XRD peaks was significantly decreased although their positions did not alter compared with the fresh sample. However, XRD peak intensity of the 4Ce-4Mn/Fe-PILC-S sample did not change significantly compared with that of the fresh counterpart. Therefore, the results of XRD characterization reveal that the introduction of ceria protects the 4Ce-4Mn/Fe-PILC catalyst from being poisoned by SO2 during the SO2 treatment.
3.3. Morphology
According to the literature [24], we can realize that after sulfur dioxide poisoning, the metal elements in the catalyst combine with SO2 to form the sulfate species on the surface. The decomposition temperature of manganese sulfate is ca. 850°C, whereas that of cerium sulfate is ca. 920°C. Therefore, via calcination at different temperatures, the existence states of metal sulfates in the samples after SO2 poisoning can be distinguished. The 4Mn/Fe-PILC-S and 4Ce-4Mn/Fe-PILC-S samples were calcined in air at 850 and 950°C for 1 h, respectively, which were characterized by the SEM and EDS techniques, and their results are shown in Fig. 4. It is seen from the SEM images that the samples before and after calcination display disordered flaky clay structures without no obvious difference between them, indicating that there was no significant alteration in morphology of the samples after SO2 poisoning. Furthermore, element contents of these samples were measured by the EDS technique, and the S contents in 4Mn/Fe-PILC-S and 4Ce-4Mn/Fe-PILC-S changed differently before and after calcination. Before calcination, the S content in 4Mn/Fe-PILC-S was 24 wt%, while that in 4Ce-4Mn/Fe-PILC-S was 29 wt%. After calcination at 850°C, however, the S content in 4Mn/Fe-PILC-S decreased to 0.5 wt%, which might be due to the decomposition of manganese sulfate via calcination at 850°C. After the 4Ce-4Mn/Fe-PILC-S sample was calcined at 850°C, its S content decreased to 14 wt%, which was significantly higher than that in the 4Mn/Fe-PILC-S sample, owing to the fact that the manganese sulfate species was decomposed while the cerium sulfate species was not decomposed under the adopted calcination condition (850°C). After further calcination at 950°C, only a trace amount (around 0.5 wt%) of sulfur was present in both samples, indicating that the species of manganese and cerium sulfates were all decomposed. Therefore, it can be deduced that the S content in cerium sulfate of the 4Ce-4Mn/Fe-PILC-S sample was about 14 wt%, while that in manganese sulfate of this sample was 15 wt%, which was significantly lower than that in manganese sulfate (24 wt%) of the 4Mn/Fe-PILC-S sample. The above results show that after sulfur dioxide poisoning of the 4Ce-4Mn/Fe-PILC sample, most of S existed as the cerium sulfate species. That is to say, SO2 preferentially binds with Ce, thus protecting the active Mn sites from being poisoned by SO2.
3.4. Pore structure and surface area
N2 adsorption−desorption was used to characterize pore structures and surface areas of the samples. N2 adsorption−desorption isotherms of the untreated and SO2-treated samples are shown in Fig. 5. According to the International Union of Pure and Applied Chemistry (IUPAC) classification, the isotherm of each sample was characteristic of microporous materials [26], which was in good consistency with montmorillonite clay reported in the literature [25]. Surface areas and pore volumes of the samples are summarized in Table 1. Surface area and pore volume of the clay were significantly increased after the pillaring treatment, but slightly decreased after loading of manganese oxide and ceria, which was attributed to partial blocking of the pores in Fe-PILC by the oxides of manganese and cerium. Furthermore, surface areas and pore volumes of the 4Mn/Fe-PILC-S and 4Ce-4Mn/Fe-PILC-S samples after SO2 treatment were declined compared with those of their fresh counterparts, which might be due to formation of Mn and Ce sulfates. Moreover, surface area of the 4Mn/Fe-PILC sample was decreased by 64 % (from 204 m2/g of 4Mn/Fe-PILC to 191 m2/g of 4Mn/Fe-PILC-S), while it was only dropped by 26 % (from 192 m2/g of 4Ce-4Mn/Fe-PILC to 187 m2/g of 4Ce-4Mn/Fe-PILC-S). This result indicates that the doping of ceria can favor the protection of the sample from being poisoned by SO2, which is in good agreement with the XRD results.
Table 1
BET surface areas and pore volumes of the samples.
Sample | BET surface area (m2 g−1) | Pore volume (cm3 g−1) |
Clay | 48 | 0.019 |
Fe-PILC | 245 | 0.117 |
4Mn/Fe-PILC | 204 | 0.095 |
4Ce-4Mn/Fe-PILC | 192 | 0.079 |
4Mn/Fe-PILC-S | 191 | 0.088 |
4Ce-4Mn/Fe-PILC-S | 187 | 0.075 |
3.5. Surface property
XPS characterization was used to analyze concentrations and valence states of elements on the surface of the samples before and after SO2 poisoning. Figure 6A shows Mn 2p XPS spectra of the 4Mn/Fe-PILC and 4Ce-4Mn/Fe-PILC samples before and after SO2 poisoning. Mn 2p XPS spectrum of each sample contained two main peaks with the spin-orbit doublets of Mn 2p3/2 and Mn 2p1/2, respectively. Furthermore, Mn 2p3/2 signal of each sample could be deconvoluted into three components at BE = 643.5−644.7, 641.7−642.9, and 640.5−641.7 eV, which were attributed to the surface Mn4+, Mn3+, and Mn2+ species [27−29], respectively. As reported previously, the Mn4+ species was more active for the low-temperature SCR reaction [27]. The Mn4+/(Mn2+ + Mn3+ + Mn4+) molar ratios on the sample surface were calculated, as listed in Table 2. The Mn4+/(Mn2+ + Mn3+ + Mn4+) molar ratio decreased according to the sequence of 4Ce-4Mn/Fe-PILC (8.7 %) > 4Ce-4Mn/Fe-PILC-S (6.2 %) > 4Mn/Fe-PILC (3.9 %) > 4Mn/Fe-PILC-S 9.3 %). Apparently, SO2 poisoning gave rise to a decrease in surface Mn4+ species concentration. At the same time, the concentration of Mn4+ species on the surface of 4Ce-4Mn/Fe-PILC-S was higher than that on the surface of 4Mn/Fe-PILC-S. The results reveal that during the SO2 poisoning process, the doped ceria protects the surface active Mn4+ species from being poisoned by SO2 in the samples.
Figure 6B shows O 1s XPS spectra of the 4Mn/Fe-PILC and 4Ce-4Mn/Fe-PILC samples before and after SO2 poisoning. O 1s XPS spectrum of each sample could be divided into three components: The first component in the BE range of 528.9−529.7 eV corresponded to the surface lattice oxygen (Oα) species, the second one in the BE range of 530.8−531.8 eV was ascribed to the surface chemisorbed oxygen (Oβ) species, and the third one in the BE range of 529.7−532.7 eV was attributed to the surface lattice oxygen (Oγ) species of the Si−O bonds in SiO2 of montmorillonite clay [19, 30]. It is well known that the surface chemisorbed oxygen (Oβ) species was beneficial for NO removal in the NH3-SCR reaction [31]. The Oβ species concentrations were calculated, as listed in Table 2. The (Oβ)/(Oα + Oβ + Oγ) molar ratio decreased in the order of 4Ce-4Mn/Fe-PILC (3.2 %) > 4Ce-4Mn/Fe-PILC-S (2.1 %) > 4Mn/Fe-PILC (7.4 %) > Mn/Fe-PILC-S (5.9 %). Apparently, the concentration of the Oβ species dropped significantly after sulfur dioxide poisoning, which was consistent with the surface Mn4+ concentrations. Moreover, the surface Oβ concentration on the 4Ce-4Mn/Fe-PILC-S sample was still higher than that on the 4Mn/Fe-PILC-S sample, proving that the doping of ceria suppresses the decrease in surface Oβ concentration during the SO2 poisoning process.
Shown in Fig. 6C are S 2p XPS spectra of the 4Mn/Fe-PILC and 4Ce-4Mn/Fe-PILC samples after SO2 poisoning. By using the Gaussian fitting method, S 2p XPS spectrum of each sample was divided into two components at BE = 168.4 and 166.3 eV, which corresponded to the different sulfate species [32]. Intensity of the S 2p XPS peak on the 4Ce-4Mn/Fe-PILC-S sample was significantly weaker than that on the 4Mn/Fe-PILC-S sample, indicating that there was a les amount of sulfate species on the surface of the former sample.
Fe 2p and Ce 3d spectra of the samples are shown in Figs. S1 and S2, respectively. Fe 2p was fitted into five peaks (around 727.2 eV, 724.3 eV, 713 eV, 711.2 eV and 718 eV), which can be assigned to Fe 2p1/2 and Fe 2p3/2 of Fe2+, Fe3+ and satellite [31]. Ce 3d was fitted into eight peaks (from 930 − 870 eV). The peaks labeled by v′, u′ can be assigned to Ce3+, and those marked by v, v″, v‴, u, u″, u‴ can be assigned to Ce4+ [27].
3.6. Redox ability
The H2-TPR technique was used to evaluate redox properties of the samples. H2-TPR profiles of the fresh and sulfur dioxide-poisoned samples are shown in Fig. 7, and their H2 consumption are listed in Table 3. For the 4Mn/Fe-PILC sample, three reduction peaks centered at 388, 453, and 638 ℃ were distinguished in the H2-TPR curve: The first reduction peak at 388 ℃corresponded to the reduction of Mn4+ to Mn3+, the second one at 453 ℃ was due to the reduction of Mn3+ to Mn2+, and the third one at 638 ℃ could be identified as the reduction of the Fe species [28, 33]. The 4Ce-4Mn/Fe-PILC sample displayed four reduction peaks at 382, 435, 503, and 621 ℃, which were attributed to the reduction of Mn4+ to Mn3+, Mn3+ to Mn2+, Ce4+ to Ce3+, and the Fe species [33, 34], respectively. Compared with the 4Mn/Fe-PILC sample, 4Ce-4Mn/Fe-PILC displayed a significantly lower reduction temperature, which indicates that redox ability of the sample was enhanced after ceria doping.
Reducibility of the samples (4Mn/Fe-PILC-S and 4Ce-4Mn/Fe-PILC-S) after sulfur dioxide poisoning was also measured by the H2-TPR technique, and their results are also illustrated in Fig. 7. Compared with the fresh samples, the two sulfur dioxide-poisoned samples showed bigger reduction peaks in the range of 400−550 ℃, which were considered as the reduction of sulfate species after sulfur dioxide poisoning [32, 35], indicating that the sulfate species were present in the 4Mn/Fe-PILC-S and 4Ce-4Mn/Fe-PILC-S samples. It is well known that the sulfate species was not conducive to the progress of the catalytic reaction. Furthermore, the peaks in the 4Ce-4Mn/Fe-PILC-S sample could be divided into two peaks at 441 and 470°C, corresponding to the sulfate species of manganese and cerium [35], respectively. The peak area of manganese sulfate in the 4Ce-4Mn/Fe-PILC-S sample was significantly smaller than that in the 4Mn/Fe-PILC-S sample, which indicates that the doping of ceria reduces the content of manganese sulfate species in the 4Ce-4Mn/Fe-PILC sample. Hydrogen consumption of the two sulfur dioxide-poisoned samples is summarized in Table 3. The H2 consumption below 400°C of the two SO2-poisoned samples was decreased, indicating that redox ability of the samples was weakened. A further comparison shows that the change in H2 consumption from 0.207 mmol g−1 of 4Ce-4Mn/Fe-PILC-S to 0.193 mmol g−1 of 4Ce-4Mn/Fe-PILC was significantly smaller than that from 0.255 mmol g−1 of 4Mn/Fe-PILC-S to 0.172 mmol g−1 of 4Mn/Fe-PILC. This result further confirms that the doping of ceria enhances sulfur dioxide-resistant ability of the sample.
Table 2
Surface element compositions and H2 consumption of the samples.
Sample | Oxygen speciesa | | Mn speciesa | | H2 consumption (< 400 ℃)b (mmol g−1) |
| Oα (mol%) | Oβ (mol%) | Oγ (mol%) | | Mn2+ (mol%) | Mn3+ (mol%) | Mn4+ (mol%) | |
4Mn/Fe-PILC | 14.5 | 37.4 | 48.1 | | 45.9 | 40.2 | 13.9 | | 0.255 |
4Ce-4Mn/Fe-PILC | 11.2 | 43.2 | 45.6 | | 39.7 | 41.6 | 18.7 | | 0.207 |
4Mn/Fe-PILC-S | 19.3 | 35.9 | 44.8 | | 49.2 | 41.5 | 9.3 | | 0.172 |
4Ce-4Mn/Fe-PILC-S | 10.1 | 42.1 | 47.8 | | 41.7 | 42.1 | 16.2 | | 0.193 |
a Data were obtained by quantitatively analyzing the peaks in XPS spectra of the samples; |
b Data were obtained by quantitatively analyzing the reduction peaks in H2-TPR profiles of the samples.
3.7. In situ DRIFTS of NH3 adsorption
In situ DRIFTS characterization is usually used to study the changes in the reactants, intermediates, and products on the sample surface during the catalytic reaction processes. Figure 8A shows in situ DRIFTS spectra of NH3 adsorption at 150°C on the 4Mn/Fe-PILC sample. After NH3 adsorption, several different NH3 species appeared on the surface of the sample: the bands at 1229 and 1680 cm−1 was attributed to the coordinating NH3 species combined to the Lewis acid sites, while the one at 1450 cm−1 was identified as the linking NH4+ species at the Brønsted acid sites. The band at 3000−3400 cm−1 could be attributed to the NH3 species adsorbed on the Lewis acid sites [36−38]. Intensity of the above bands enhanced with the extension of adsorption time, indicating that the amount of NH3 species adsorbed on the sample surface increases gradually.
In situ DRIFTS spectra of NH3 adsorption on the 4Mn/Fe-PILC-S sample are shown in Fig. 8B. The positions of the observed bands of the SO2-poisoned sample were basically similar to those of the fresh sample. For a further comparison with the 4Mn/Fe-PILC sample, the characteristic bands at 1229 and 1680 cm−1 of NH3 adsorption on the SO2-poisoned 4Mn/Fe-PILC-S sample corresponding to the Lewis acid sites were significantly weakened in intensity, and the characteristic band at 1450 cm−1 corresponding to the binding of NH4+ species to the Brønsted acid site was significantly enhanced in intensity. The decrease in amount of the Lewis acid sites was attributed to the fact that some Lewis acid sites provided by the manganese oxide species are sulfated, while the increase in amount of the Brønsted acid sites was due to formation of the NH4+ sulfates after NH3 adsorption in sulfur-containing samples. It is generally believed that the increased amount of NH4+ sulfate strongly inhibits the reaction of ammonium species with the adsorbed NO2 during the NH3-SCR reaction [36]. Therefore, the new Brønsted acid sites generated by the samples after SO2 poisoning were one of the reasons for inhibiting NH3-SCR catalytic activity of the samples at low temperatures.
In situ DRIFTS spectra of NH3 adsorption on 4Ce-4Mn/Fe-PILC and 4Ce-4Mn/Fe-PILC-S are shown in Fig. 8C and D, respectively. Consistent with the 4Mn/Fe-PILC sample, the characteristic bands at 3400−3000, 1680, 1450, and 1229 cm−1 were also observed in the spectra of NH3 adsorption on the 4Ce-4Mn/Fe-PILC and 4Ce-4Mn/Fe-PILC-S samples. For the SO2-treated 4Ce-4Mn/Fe-PILC-S sample, the characteristic band located at 1450 cm−1 due to the Brønsted acid sites was also significantly increased compared with that on the fresh sample 4Ce-4Mn/Fe-PILC sample, indicating that the NH4+ sulfates were also formed at the Brønsted acid sites. There were still obvious characteristic bands at 1680 and 1229 cm−1 belonging to the Lewis acid sites on the 4Ce-4Mn/Fe-PILC-S sample, and the band intensity did not change significantly. After further comparing the in situ DRIFTS spectra of NH3 adsorption on the 4Mn/Fe-PILC-S and 4Ce-4Mn/Fe-PILC-S samples, we can realize that there are differences in the characteristic band changes of the two samples. Intensity of the characteristic band at 1450 cm−1 of the NH4+ sulfate on the 4Ce-4Mn/Fe-PILC-S sample was significantly lower than that on the 4Mn/Fe-PILC-S sample. The main reason for such differences was that the doping of ceria inhibited sulfation of the NH3 species.
3.8. In situ DRIFTS of (NO + O2) and (SO2 + NO + O2) adsorption
Figure 9A and B shows the in situ DRIFTS spectra of (NO + O2) and (SO2 + NO + O2) co-adsorption at 150 ℃ on the 4Mn/Fe-PILC sample, respectively. After (NO + O2) was adsorbed on the 4Mn/Fe-PILC sample, three bands were detected at 1603, 1548, and 1295 cm−1: The band at 1603 cm−1 was characteristic of the adsorbed NO2 molecules, the one at 1548 cm−1 was due to the bidentate nitrate species, and the one at 1295 cm−1 was owing to the monodentate nitrate species [36−38]. When SO2 was introduced into the sample system, three characteristic bands located at 1603, 1548, and 1295 cm−1 were gradually weakened in intensity and finally disappeared. New bands appeared at 1630 and 1307 cm−1, those intensity was gradually enhanced with adsorption time. These two bands were characteristic of the adsorbed SO2 molecule and bidentate sulfate species [39], respectively. The changes in characteristic bands in the spectrum indicates that there is a competitive adsorption between SO2 and NO, and SO2 exerts a great influence on the adsorption of NO on the 4Mn/Fe-PILC sample.
Figure 9C and D shows the in situ DRIFTS spectra of (NO + O2) and (SO2 + NO + O2) co-adsorption at 150 ℃ on the 4Ce-4Mn/Fe-PILC sample, respectively. When (NO + O2) was introduced for 30 min, the same characteristic bands at 1603, 1548, and 1295 cm−1 appeared, and intensity of the bands on 4Ce-4Mn/Fe-PILC was basically consistent with that on 4Mn/Fe-PILC. When SO2 was introduced into the sample system, intensity of the bands at 1603 and 1548 cm−1 decreased with the adsorption time, and two weak characteristic bands also appeared at 1630 and 1307 cm−1, which were also characteristic of the adsorbed SO2 molecule and bidentate sulfate species, respectively. Although the characteristic bands of NO adsorbed species on the 4Ce-4Mn/Fe-PILC sample were weakened in intensity, these bands still appeared clearly in the spectrum, while the bands due to the NO adsorbed species on the 4Mn/Fe-PILC sample almost disappeared. On the other hand, the characteristic bands of sulfur species on the 4Ce-4Mn/Fe-PILC sample appeared later, and their intensity was also significantly lower than that on the 4Mn/Fe-PILC sample. This result indicates that ceria doping can effectively suppress SO2 adsorption on the surface of the sample.
3.9. TG/DSC curves of NH4HSO4- and (NH4)2SO4-deposited catalysts
The difference in SO2-resistant performance of the 4Mn/Fe-PILC and 4Ce-4Mn/Fe-PILC samples was further verified using the TGA/DSC technique. Two fresh samples were treated in a (1100 ppm NH3 + 200 ppm SO2 + 4 vol% O2 + N2 (balance)) mixture of 500 mL min−1 at 150°C for 10 h before testing. Since the pretreatment atmosphere contained NH3 and SO2, it can be considered that there are (NH4)2SO4 and NH4HSO4 present in the 4Mn/Fe-PILC and 4Ce-4Mn/Fe-PILC samples. Obviously, the TG/DSC curves of the two samples were different. As shown in Fig. 10A, there were three weight losses in the TGA curve of 4Mn/Fe-PILC, accompanying with three corresponding peaks in the DSC curve. The first peak around 120°C was attributed to the loss of water in the sample, the second one at 232°C was due to the decomposition of (NH4)2SO4, and the third one at 354°C was ascribed to decomposition of NH4HSO4 [40]. As shown in Fig. 10B, three peaks also appeared in the TG/DSC curve of the 4Ce-4Mn/Fe-PILC sample. However, the peak due to decomposition of (NH4)2SO4 in the 4Ce-4Mn/Fe-PILC sample appeared at 221°C, whereas that owing to decomposition of NH4HSO4 appeared at 312°C, both of which were lower than the corresponding decomposition temperatures on the 4Mn/Fe-PILC sample. This result shows that ammonium sulfate is more easily decomposed on 4Ce-4Mn/Fe-PILC than on 4Mn/Fe-PILC, further confirming that the 4Ce-4Mn/Fe-PILC sample possesses a stronger SO2-resistant ability.
3.10. Mechanism of ceria doping to enhance SO2 resistance
SO2 is an acid gas, which is easily adsorbed and combined with the active metal elements on the surface of the catalyst sample to be further converted into the sulfate or sulfite species in the presence of gas-phase oxygen [41]. For the 4Mn/Fe-PILC sample with Mn as the main active element, the MnOx was loaded in the interlayer or on the surface of Fe-PILC. These MnOx can interact with the adsorbed SO2 and O2 to form the manganese sulfate species on the catalyst surface. It was reported that the formation rate of sulfate or sulfite on the surface of the Ce-containing catalyst was faster during the NH3-SCR reaction in the presence of SO2, i.e., cerium could bind with SO2 more easily than manganese [39]. That is to say, cerium can act as a trap for SO2. The activity evaluation and TGA/DSC results showed that the doping of ceria significantly improved NH3-SCR performance and SO2 resistance of the 4Mn/Fe-PILC sample. In situ DRIFTS characterization results demonstrated that for the 4Mn/Fe-PILC sample, part of the Lewis acid sites was formed by manganese, and the sulfate species were formed with manganese after the introduction of SO2, which reduced amount of the Lewis acid sites, thereby resulting in a decrease in catalytic activity. In the 4Ce-4Mn/Fe-PILC sample, however, the sulfate species were preferentially formed at the cerium site rather than at the manganese site, thus protecting the active manganese from being poisoned by SO2. Therefore, the sample containing cerium showed better resistance to SO2 poisoning than the sample without cerium. From the results of XRD, SEM, XPS, and H2-TPR characterizations, we conclude that cerium in the 4Ce-4Mn/Fe-PILC sample acts as a SO2 trap, which can be preferentially sulfated. In other words, SO2 adsorbed on the surface of the sample preferentially combined with cerium to form the cerium sulfate species. This might be the main reason why catalytic activity of the 4Mn/Fe-PILC sample after ceria doping was less affected by SO2 for the NH3-SCR reaction.