Improved Sulfur Resistance of COMMERCIAl V₂O₅-WO₃/TiO₂ SCR Catalyst Modified by Ce and Cu

Abstract

The accumulation of NH₄HSO₄ leads to the deactivation of commercial V₂O₅-WO₃/TiO₂ catalyst (VWTi) in practical application. The commercial catalyst is modified with 0.3 wt. % Ce and 0.05 wt. % Cu (donated as VWCeCuTi), and its sulfur resistance is noticeably improved. After loading 20 wt. % NH₄HSO₄, the NO_x conversion of VWCeCuTi-S remains 40% at 250 °C, higher than that of VWTi-S (25%). Through a series of characterization analyses, it was found that the damaged surface areas and acid sites are the key factors for the deactivation of S-poisoned samples. However, surface-active oxygen and NO adsorption are increased by NH₄HSO₄ deposition, and the L–H mechanism is promoted over S-poisoned samples. Due to the interaction between V, Ce and Cu, the surface-active oxygen over VWCeCuTi-S is increased, and then NO adsorption is promoted. In addition, VWCeCuTi-S obtains a higher V⁵⁺ ratio and a better redox property than VWTi-S, which in turn accelerates the NH₃-SCR reaction. More NO adsorption and encouraged reaction contribute to the better sulfur resistance of VWCeCuTi.

1. Introduction

Nitrogen oxides (NO_x) emitted from stationary and mobile sources have attracted much attention since they can cause a series of environmental issues, such as greenhouse effects, acid rain, ozone depletion and photochemical smog [1]. Selective catalytic reduction of NO_x with ammonia (NH₃-SCR) is considered to be the most effective and common technology in controlling NO_x emissions [2]. As a typical commercial NH₃-SCR deNO_x catalyst, the V₂O₅-WO₃/TiO₂ catalyst shows excellent medium temperature activity, high N₂ selectivity and good thermal stability [3,4]. Despite the low sensitivity to SO₂, deactivation resulting from SO₂ poisoning does occur over V₂O₅-WO₃/TiO₂ catalyst with service time [5,6,7]. The accumulation of NH₄HSO₄ (ABS) forming at low load has been reported as the primary cause for the catalyst deactivation in practical application [8,9,10].

To prolong the service life of the catalyst, great efforts have been invested to improve the sulfur resistance of V₂O₅-WO₃/TiO₂ catalyst [11,12,13,14,15]. Xu et al. [13] reported that Sb-modified monolithic catalyst (V₂O₅-Sb₂O₃/TiO₂) exhibited a better sulfur resistance than the commercial V₂O₅-WO₃/TiO₂ catalyst due to the lower deposition of ABS. Li et al. [14] doped Ce to VWTi catalyst, and the results showed that the addition of Ce could weaken the stability of ABS and promote the reaction between NH₄⁺ species from ABS and NO, leading to a better ABS resistance. Ye et al. [15] found that the introduction of Sb₂O₅ and Nb₂O₅ to VWTi SCR catalyst could enhance the decomposition of ABS, and promote the reaction between ABS and NO, contributing to a better sulfur resistance. Kang et al. [16] physically mixed V₂O₅-WO₃/TiO₂ catalyst with Fe₂O₃ samples, and the physically mixed catalysts exhibited higher catalytic stability and SO₂ resistance than the V₂O₅-WO₃/TiO₂ catalyst. However, the sulfur content in these studies is far less than that in the practical industry after running for a long time.

Recently, because of their unique features, Ce and Cu have been used to improve the SCR performance of commercial VWTi catalyst. Liang et al. [17] added 3% of the CeO₂ to V₂O₅-WO₃/TiO₂ catalyst and found that its low-temperature NO conversion and sulfur and water resistance were improved. Peng et al. [18] and Liu et al. [19] modified V₂O₅-WO₃/TiO₂ catalyst with Ce and reported that its NO conversion and alkali resistance were enhanced. Wang et al. [20] introduced Cu and Fe to V₂O₅-WO₃/TiO₂ catalyst and found that Cu or Fe addition, especially Cu, improved the catalytic performance of V₂O₅-WO₃/TiO₂. Additionally, Ali et al. [21] suggested that the addition of Cu, even in fairly small amounts, could improve the NH₃-SCR performance of CeO₂/TiO₂ catalyst due to the interaction between Ce and Cu. All of the above studies demonstrate that some interactions exist between V, Ce and Cu, contributing to better SCR performances. However, to our knowledge, there is little research on investigating the cooperative effect of Ce and Cu over the V₂O₅-WO₃/TiO₂ catalyst for the selective catalytic reduction of NO by NH₃.

Taking the advantage of interactions between Ce, Cu and V, in this work, commercial V₂O₅-WO₃/TiO₂ catalyst was first modified with Ce(SO₄)₂ and Cu(NO₃)₂, and then the sulfur resistance of modified catalyst was investigated by loading a certain amount of NH₄HSO₄ on it through a wet impregnation method. Finally, the structural properties as well as chemical properties were systematically studied.

2. Results and Discussion

2.1. NO_x Conversion

Considering the decomposition temperature of ABS, the NH₃-SCR performance of fresh and ABS-deposited catalysts were evaluated at 250 °C and the results are presented in Figure 1. The NO_x conversion of VWTi catalyst stabilizes at about 73% in the whole test process, and that of VWCeCuTi is about 84%. The activities of VWTi and VWCeCuTi in a larger temperature range of 200–400 °C were also tested. As shown in Figure S1, VWCeCuTi shows a general trend of better performance than VWTi. These indicate that the addition of Ce and Cu improves the activity of VWTi catalyst. After being deposited by 20 wt. % ABS, the NO_x conversion of VWTi-S declines sharply to 25%, and that of VWCeCuTi-S declines to 40%, implying that ABS deposition severely damages the activity of catalysts. Nevertheless, VWCeCuTi-S possesses a higher activity than VWTi-S. N₂O is a harmful byproduct of the NH₃-SCR reaction, and its concentration in the de-NO_x process is presented in Figure S2. During the whole process, the concentration of N₂O over all catalysts is no more than 2 ppm, indicating that ABS deposition has no distinct influence on N₂O formation.

2.2. Structure Investigation

XRD patterns of fresh and S-poisoned catalysts are displayed in Figure 2. Only the peaks belonging to anatase TiO₂ could be detected on all samples, indicating that neither modification nor ABS deposition makes a difference to the crystalline structure of TiO₂ support. Active species and ABS are highly dispersed or amorphous on the surface. N₂ physisorption isotherms of fresh and S-poisoned catalysts are presented in Figure S3. Similar isotherms attributed to type IV with H1 hysteresis loops are observed on all samples, suggesting that modification and ABS deposition do not change the pore type of VWTi catalyst. BET surface area, total pore volume and average pore diameter based on N₂ physisorption isotherm are summarized in

Table 1. Textural properties of fresh and S-poisoned catalysts.

Sample	BET Surface Area (m2·g−1)	Total Pore Volume (cm3·g−1)	Average Pore Diameter (nm)
VWTi	44.1	0.23	20.7
VWCeCuTi	41.3	0.22	21.4
VWTi-S	23.7	0.14	24.3
VWCeCuTi-S	23.5	0.15	26.2

. No big differences in surface areas exist between VWTi and VWCeCuTi. However, their specific surface areas are seriously damaged by ABS deposition, dropping from 44.1 and 41.3 m²·g⁻¹ to 23.7 and 23.5 m²·g⁻¹, respectively. Meanwhile, the total pore volumes of S-poisoned samples decrease significantly, while their average pore diameters increase to some extent. This may be because some small pores are blocked by ABS. Though ABS deposition blocks pores, which is a main factor for sulfur poisoning, there is no big differences between the two S-poisoned samples.

Sulfur species over these catalysts were analyzed by TGA, and the relative profiles are shown in Figure 3. The profiles of all samples are divided into three stages: Step I (below 200 °C) could be ascribed to the desorption of adsorbed H₂O; Step II (200–535 °C) is assigned to the decomposition of NH₄HSO₄ or (NH₄)₂SO₄; Step III (535–900 °C) is attributed to the decomposition of bulk sulfate species such as vanadium sulfate [22,23,24]. The weight losses at each stage are summarized in

Table 2. Weight losses of catalysts during the TGA processes.

Samples	Step I (wt. %)	Step II (wt. %)	Step III (wt. %)
VWTi	0.7	0.2	0.8
VWCeCuTi	0.5	0	1.6
VWTi-S	1.7	13.6	2.3
VWCeCuTi-S	1.7	13.9	2.6

. It was found that there is a small amount of sulfur species on fresh samples, coming from the preparation processes of catalysts. After loading ABS, the weight losses over both samples increase significantly at Stage II, and show a slight increase at Stage III, indicating sulfur species over the two S-poisoned samples are mainly in the form of NH₄HSO₄ or (NH₄)₂SO₄, and the contents of sulfur species on the two samples are almost the same.

2.3. Surface Analysis by XPS

XPS analysis was carried out to gain an insight into the near surface atomic compositions and chemical states of elements. The obtained XPS spectra of Ti 2p, S 2p, V 2p, Ce 3d and O 1s are displayed in Figure 4, and their relative atomic concentrations are listed in

Table 3. Surface atomic ratio and concentration from XPS.

Samples	Atomic Ratio (%)		Atomic Concentration (at%)
	V5+/(V5+ + V4+)	Oβ/(Oα + Oβ + Oγ)	V	Ce
VWTi	42.2	12.1	0.5	/
VWCeCuTi	54.1	13.1	0.6	0.2
VWTi-S	58.5	22.9	0.2	/
VWCeCuTi-S	63.9	27.4	0.2	0.1

and Table S1. As shown in Figure 4a, the binding energies of Ti 2p_3/2 (458.8 eV) and Ti 2p_1/2 (464.5 eV), which are characteristic peaks of Ti⁴⁺ in TiO₂, are detected over a VWTi sample, illustrating that Ti atoms are in the +4 oxidation state on the catalyst surface [25,26]. Compared with VWTi, no obvious variations in the binding energies of Ti 2p peaks are observed over VWCeCuTi, indicating that the addition of Ce and Cu nearly has no effect on the TiO₂ support, which may be due to the low contents of these additives. Given the deposition of ABS, the peaks shift towards higher binding energies, indicating that the electron cloud density around Ti atoms is decreased. It should be caused by the interaction between TiO₂ and ABS, which may decrease the coordination number of Ti and shorten the Ti-O bond through the Ti–O–S bond, leading to an electron-deficit state of Ti atoms [27,28,29].

In Figure 4b, after deconvolution, two peaks centered at ca. 169.7 and ca. 168.6 eV are obtained, which are the characteristic peaks of S⁶⁺ in bidentate sulfate, and could be assigned to S 2p_1/2 and S 2p_3/2, respectively [6,15,23,30]. The S 2p spectra observed on fresh samples are due to the small amount of S contained in commercial catalyst. After loading ABS, the intensities of S 2p spectra increase significantly, suggesting the deposition of sulfate species on S-poisoned samples, which is in good agreement with the data collected from TGA. No characteristic peaks belonging to S⁴⁺ are detected over these catalysts, indicating that sulfite species likely do not exist on these catalysts’ surface, which is probably because all samples are calcined in the air during their preparation processes.

V 2p spectra are shown in Figure 4c. After deconvolution, two peaks located at ca. 516.7 and ca. 515.7 eV are observed, which could be attributed to V⁵⁺ and V⁴⁺, respectively [31,32]. It has been reported that a superior content of V⁵⁺ exerts a promotion effect on the dehydrogenation process of ammonia species, which is a key step in the NH₃-SCR reaction [33,34]. V⁵⁺ ratio is calculated by V⁵⁺/(V⁴⁺ + V⁵⁺), and the results are shown in Table 2. After introducing Ce and Cu, the V⁵⁺ ratio of VWTi increases from 42.2 to 54.1%, which is due to the interaction between V, Ce and Cu through the redox cycles of V⁵⁺/V⁴⁺, Ce⁴⁺/Ce³⁺ and Cu⁺/Cu²⁺ [35,36]. Moreover, the V⁵⁺ ratio is increased by loading ABS, which are 58.5 and 63.9% over VWTi-S and VWCeCuTi-S respectively. It indicates that ABS deposition increases the valence state of surface V atoms, in accordance with other studies [15,34,37]. These studies demonstrate that electrons from V atoms would be attracted by sulfate anions with stronger electronegativity, which is responsible for the increased valence state of V atoms on the ABS-deposited samples. Nevertheless, the surface concentration of V decreases by more than half after ABS deposition, which could be attributed to the coverage of surface V sites by ABS [23,29]. As a result, the contents of V⁵⁺ on S-poisoned samples are still lower than those of fresh ones. In addition, VWCeCuTi-S obtains a higher V⁵⁺ content in favor of the catalytic activity when compared with VWTi-S in isolation.

Ce 3d spectra of VWCeCuTi and VWCeCuTi-S are presented in Figure 4d. The bands labeled as “v” and “u” correspond to Ce 3d_5/2 and Ce 3d_3/2 spin–orbit components, respectively. Furthermore, the bands v^’ and u^’ are related to the 3d¹⁰4f¹ initial electronic state of Ce³⁺ species, while the rest of the bands are ascribed to the 3d¹⁰4f⁰ initial electronic state of Ce⁴⁺ species, revealing the co-existence of Ce³⁺ and Ce⁴⁺ on both samples [38,39,40]. Furthermore, Ce³⁺ is demonstrated to produce oxygen vacancies by generating a charge imbalance, and then promote NO oxidation [14,41]. For the concentration of surface Ce atoms, it also decreases after loading ABS due to the coverage of catalyst surface. No obvious Cu 2p peaks are observed because of the low content of Cu species, as shown in Figure S4.

Figure 4e displays O 1s spectra of the series samples. After deconvolution, O 1s spectra are fitted into three characterization peaks, attributed to the lattice oxygen (O_γ) at ca. 530.1–530.3 eV, surface adsorbed oxygen (O_β) like hydroxyl species at ca. 531.6–531.8 eV and weakly bonded oxygen species (O_α) at ca. 532.8–533.0 eV, respectively [5,42]. It has been reported that O_β is more active in oxidation reaction and beneficial to NO adsorption [43,44,45]. The ratio of O_β is calculated by O_β/(O_α + O_β + O_γ) and is summarized in Table 2. After introducing Ce and Cu, the O_β ratio increases from 12.1 to 13.1%. Compared with fresh samples, the ratios of O_β over both S-poisoned samples increase significantly, the main reason for which is the formation of abundant hydroxyl species through hydration of SO₄²⁻ [23,46]. In addition, the O 1s bands shift to higher binding energies after ABS deposition, indicating that the electron cloud density around O atoms is decreased by ABS, owing to the stronger electronegative of SO₄²⁻. However, comparing the two poisoned samples, O_β ratio over VWCeCuTi-S is 27.4%, higher than that over VWTi-S (22.9%). Then, the more surface-active oxygen over VWCeCuTi-S may contribute to more NO adsorption, which will be further clarified below.

2.4. Redox Property

It is generally considered that redox property plays an important role in NH₃-SCR reaction, and is usually investigated by H₂-TPR. As Figure 5 shows, two reduction peaks gradually appear on the profile of VWTi catalyst with temperature increasing. The first one centered at 603 °C could be assigned to the reduction of V⁵⁺ to V³⁺, and the second one centered at 690 °C could be ascribed to the reduction of W⁶⁺ to W⁴⁺ [31,47]. After loading Ce and Cu, a new peak at 523 °C appears, which could be ascribed to the reduction of Ce⁴⁺ to Ce³⁺ [47,48,49]. Because of its low content, no peaks belonging to the reduction of Cu²⁺ are detected. It is generally recognized that a lower reduction peak implies a stronger reducibility [37,50], so VWCeCuTi exhibits a stronger redox property than VWTi.

A strong H₂ consumption peak at 597 °C is detected over VWTi-S, which could be attributed to the couple reduction of V⁵⁺ and SO₄²⁻ species [6]. Compared with VWTi, the peak position of V⁵⁺ to V³⁺ remains nearly unchanged over VWTi-S, but the intensity increases significantly due to the reduction of surface SO₄²⁻ species. Compared with VWCeCuTi, both reduction peaks of VWCeCuTi-S shift to higher temperatures (523–548 °C and 603–624 °C), indicating that its reducibility is slightly inhibited by ABS deposition. The intensity of the first consumption peak also increases sharply due to the overlapped reduction of Ce⁴⁺ and SO₄²⁻ species. With a lower reduction peak at 548 °C, VWCeCuTi-S still exhibits a better redox property than VWTi-S.

2.5. Surface Acidity and NO Adsorption

In situ DRIFT spectra of NH₃ adsorption were collected to investigate the acid property of fresh and S-poisoned catalysts. In each test, the sample pre-adsorbs 2000 ppm NH₃ for 30 min, followed by purging with N₂ for 10 min, and the spectra of four catalysts are summarized in Figure 6. On VWTi catalyst, five NH₃ adsorption bands are detected between 1800 and 1000 cm⁻¹. The bands centered at 1604 and 1259 cm⁻¹ could be assigned to asymmetric and symmetric bending vibrations of the N-H bonds in NH₃ coordinately linked to Lewis acid sites, the bands at 1668 and 1450 cm⁻¹ could be attributed to asymmetric and symmetric bending vibrations of NH₄⁺ species on the Brønsted acid sites, and the band at 1365 cm⁻¹ could be ascribed to the amide (NH₂) species [51,52]. The same bands are observed over the VWCeCuTi sample, but the intensities of bands at 1668 and 1450 cm⁻¹ are much higher than those over VWTi, suggesting that Brønsted acid sites are increased by the modification with Ce and Cu. However, only one weak band that belongs to amide (NH₂) species could be detected over S-poisoned samples. This result is different from the findings of some SO₂ in situ poisoning studies, in which Brønsted acid sites increase after SO₂ and H₂O treatment due to the formation of surface OH species through the hydration of SO₄²⁻ [31,53]. In this work, the seriously damaged surface acid sites may be a result of ABS cover [54,55], which is another important factor for the deactivation of S-poisoned samples.

NO adsorption over fresh and S-poisoned catalysts were investigated by in situ DRIFTS. Each sample was first exposed to 2000 ppm NO + 5% O₂ for 30 min, purged by N₂ for 10 min, and then spectra were collected. As displayed in Figure 7, only a weak band at 1615 cm⁻¹ attributed to gaseous NO₂ is recorded over VWTi and VWCeCuTi, indicating that NO adsorption over fresh samples is weak [56]. However, three obvious bands at 1676, 1484 and 1348 cm⁻¹ are observed over S-poisoned samples, which could be ascribed to N₂O₄, monodentate nitrate and cis-N₂O₂²⁻, respectively [57,58,59]. The result suggests that the adsorption ability of NO is efficiently improved after ABS deposition. Furthermore, the intensity of the band at 1484 cm⁻¹ over VWCeCuTi-S is much stronger than VWTi-S, indicating that the formation of monodentate nitrate is largely increased by modification with Ce and Cu. Combined with the result of O 1s spectra, the stronger NO adsorption could be attributed to the more surface-active oxygen over VWCeCuTi-S.

2.6. Reaction between Ammonia and Adsorbed Nitrogen Oxides Species

As discussed above, NO adsorption over S-poisoned catalysts is improved, so further experiments were conducted to investigate if these adsorbed NO_x species could participate in the NH₃-SCR reaction. Figure 8 presents the DRIFT spectra of VWTi-S and VWCeCuTi-S in a flow of 2000 ppm NH₃ after catalysts pre-adsorbed NO + O₂ for 30 min and purged with N₂ for 10 min. As shown in Figure 8a, after introducing NH₃, the intensities of NO adsorption peaks over VWTi-S decrease gradually, and disappear completely within 5 min. Subsequently, a new band (1327 cm⁻¹) ascribed to the NH₂ species appears. It indicates that all the adsorbed NO_x species could take part in the SCR reaction. Similarly, as shown in Figure 8b, NO adsorption peaks over VWCeCuTi-S decrease and disappear within 3 min after the introduction of NH₃, and a new peak (1327 cm⁻¹) assigned to the NH₂ species appears. These results demonstrate that the SCR reaction over S-poisoned samples involves adsorbed ammonia species as well as adsorbed NO_x species, following the L–H mechanism. Given the stronger intensity and faster disappearance of NO adsorption peaks over VWCeCuTi-S, it is inferred that the NH₃-SCR reaction over VWCeCuTi-S is faster than that over VWTi-S.

According to our previous work [35], the NH₃-SCR reaction over VWTi and VWCeCuTi mainly follows the Eley–Rideal mechanism, in which adsorbed ammonia species react with gas-phase NO. However, after loading ABS, the Langmuir–Hinshelwood mechanism (adsorbed ammonia species react with adsorbed NO_x species) is promoted by enhanced NO_x adsorption. Furthermore, the reaction is enhanced by Ce and Cu modification.

2.7. Possible Mechanism of Ce- and Cu-Promoting Effects on S-Poisoned Commercial Catalyst

According to the results of in situ DRIFTS experiments, Brønsted acid sites over VWTi are significantly increased by the modification with Ce and Cu, promoting NH₃ adsorption. Meanwhile, NO adsorption over VWTi and VWCeCuTi is weak, indicating that Ce and Cu modification does not change the mechanism of VWTi (following the E–R mechanism), but enhances the activity by providing more Brønsted acid sites. After introducing ABS, on one hand, acid sites over S-poisoned samples decrease sharply; on the other hand, NO adsorption is significantly enhanced, especially that over VWCeCuTi-S, indicating that the L–H mechanism is promoted over S-poisoned samples. Combined with the results of BET, TGA and XPS, the decrease in acid sites is mainly due to the coverage of catalyst surface by ABS, leading to the deactivation of S-poisoned samples. However, based on the XPS result of O 1s, the surface-active oxygen is increased by ABS deposition, which is suggested to be beneficial to NO adsorption [45,60,61]. Additionally, the surface-active oxygen over VWCeCuTi-S is more than that over VWTi-S. In our previous study, it is inferred that electron transfer exists between V, Ce and Cu through the following redox processes [35]:

V⁵⁺ + Ce³⁺→Ce⁴⁺ + V⁴⁺

(1)

V⁴⁺ + Cu²⁺→Cu⁺ + V⁵⁺

(2)

Cu⁺ + Ce⁴⁺→Ce³⁺ + Cu²⁺

(3)

These redox cycles contribute to the co-existence of Ce⁴⁺/Ce³⁺ and Cu⁺/Cu²⁺ on VWCeCuTi as well as VWCeCuTi-S. Meanwhile, it has been reported that Ce³⁺ and Cu⁺ could create oxygen vacancy, and produce active oxygen [21,23], as shown below:

$${\mathrm{O}}_2+2{\mathrm{Ce}}^{3+}\to 2{\mathrm{O}}_{\mathrm{ad}}^{-}+2{\mathrm{Ce}}^{4+}$$

(4)

$${\mathrm{O}}_2+2{\mathrm{Cu}}^{+}\to 2{\mathrm{O}}_{\mathrm{ad}}^{-}+2{\mathrm{Cu}}^{2+}$$

(5)

Therefore, because of the interaction between V, Ce and Cu, VWCeCuTi-S obtains more surface-active oxygen than VWTi-S, and then performs stronger NO adsorption. Additionally, the interaction between V, Ce and Cu makes VWCeCuTi-S obtain a higher V⁵⁺ ratio and better redox property than VWTi-S. Ammonia activation is a key step in the NH₃-SCR reaction, which mainly happens on V⁵⁺ = O for commercial VWTi catalyst, and a superior redox property would be better for the activation [3,33,34]. So, the higher V⁵⁺ ratio and better redox property would promote the NH₃-SCR reaction over VWCeCuTi-S. Overall, Ce and Cu modification promotes NO adsorption and NH₃-SCR reaction on VWCeCuTi-S, leading to a better performance than VWTi-S. The proposed mechanism of sulfur resistance over VWCeCuTi is generalized in Scheme 1.

3. Experimental

3.1. Catalyst Preparation

A commercial V₂O₅-WO₃/TiO₂ SCR catalyst (donated as VWTi below) from a coal-fired power plant in China was used in this investigation, of which the contents of V₂O₅ and WO₃ were 2% and 3%, respectively. First, the modified catalyst containing 0.3 wt. % Ce and 0.05 wt. % Cu (donated as VWCeCuTi below) was prepared by the wet impregnation method according to our previous work [35]. Second, for the S-poisoned samples, 20 wt. % ABS, which was determined according to the results in Figure S5, were separately deposited on VWTi and VWCeCuTi by the impregnation method. The detailed steps were as follows: impregnating both samples with a certain amount of NH₄HSO₄ solution for 12 h at room temperature, drying the mixtures at 105 °C for 12 h, and then calcining them at 300 °C for 5 h in air. The obtained catalysts are crushed and sieved to 40–60 mesh and donated as VWTi-S and VWCeCuTi-S, respectively.

3.2. NH₃-SCR Activity Test

NO_x conversion was measured in a fixed-bed quartz reactor (ø10 mm × 600 mm). In each test, 0.6 mL of sample was placed in the middle of the reactor, proceeding for 12 h at 250 °C. The reaction gas mixture consisted of 500 ppm NO, 500 ppm NH₃, 5 vol. % O₂, and N₂ as the balance gas. The total flow rate was 600 mL/min, corresponding to the GHSV of 60,000 h⁻¹. The inlet and outlet concentrations of NO, NO₂ and N₂O were monitored by an Antaris IGS flue gas analyzer (Thermo Scientific, America), and NO_x conversion is defined as follows:

$${\mathrm{NO}}_x\mathrm{conversion}=\frac{N{O}_{x,in}-N{O}_{x,out}}{N{O}_{x,in}}$$

where $N{O}_{x,in}$ and $N{O}_{x,out}$ represent the inlet and outlet concentrations of NO_x (NO + NO₂), respectively.

3.3. Catalyst Characterization

The powder X-ray diffraction (XRD) patterns were determined by a X’Pert Pro XRD diffractometer (PANalytical B.V., Holland) using Cu Kα radiation. N₂ adsorption-desorption experiments were conducted on a NOVA 2000e surface area and pore size analyzer (Quantachrome, USA) at 77 K. Prior to each test, the sample was degassed under vacuum at 250 °C for 6 h. Thermal Gravimetric Analysis (TGA) was carried out on a Netzch thermoanlyzer (TG-209-F3) with a heating rate of 10 °C/min from 40 to 950 °C in a N₂ flow (60 mL/min). X-ray photoelectron spectroscopy (XPS) was performed on an AXIS Supra by Kratos Analytical Inc., using Al Kα radiation (hv = 1486.6 eV, 150 W). Narrow region scans were acquired using a pass energy of 40 and a 0.1 eV step size, and all spectra were calibrated by C 1s (284.8 eV).

A temperature-programmed reduction of H₂ (H₂-TPR) was recorded on a ChemBET-3000 TPR-TPD chemisorption analyzer (Quantachrome, USA). Prior to each test, about 0.05 g of sample was treated in Ar for 1 h at 300 °C, and then cooled to room temperature. Following this, 5% H₂/Ar was switched on, and the temperature was heated to 800 °C linearly with a rate of 10 °C/min. The outlet gas was detected by an online mass spectrometer (MS, DYCOR LC-D100, Ametek Company, USA).

In situ diffuse reflectance infrared Fourier transform spectroscopy (in situ DRIFTS) experiments were carried out by a Bruker Vertex 70 infrared spectrometer, equipped with KBr optics and a liquid-nitrogen-cooled MCT detector. Prior to each experiment, the sample was preheated in a N₂ flow (80 mL/min) at 300 °C for 1 h and then cooled to 250 °C. Background spectrum was recorded before introducing reactant gases into the cell. DRIFT spectra were recorded through accumulating 64 scans at a resolution of 4 cm⁻¹.

4. Conclusions

The sulfur tolerance of commercial V₂O₅-WO₃/TiO₂ catalyst is visibly enhanced by modification with Ce(SO₄)₂ and Cu(NO₃)₂. After depositing 20 wt. % NH₄HSO₄, the NO_x conversion of VWCeCuTi-S remains 40% at 250 °C, higher than that of VWTi-S (25%). To explore the contributing factors, various characterization experiments were carried out. It was found that the specific surface areas and surface acid sites are severely damaged by ABS deposition, which are the key factors for the deactivation of S-poisoned samples. However, surface-active oxygen and NO adsorption are increased, and the L–H mechanism is promoted over S-poisoned samples. Based on the results of XPS and H₂-TPR, the V⁵⁺ ratio, surface-active oxygen and redox property over VWCeCuTi-S are enhanced by the interaction between V, Ce and Cu when compared with VWTi-S. More surface-active oxygen would promote greater NO adsorption. In addition, the higher V⁵⁺ ratio and better redox property would accelerate the NH₃-SCR reaction over VWCeCuTi-S, in accordance with the results of in situ DRIFTS. The increased NO adsorption and promoted L–H reaction contribute to the better sulfur resistance of VWCeCuTi.