Protection Effect of Ammonia on CeNbTi NH₃-SCR Catalyst from SO₂ Poisoning

Abstract

CeNbTi catalyst was poisoned in different sulfur poisoning atmospheres at 300 °C for 6 h and then was evaluated for selective catalytic reduction (SCR) of NO_x with NH₃. The catalyst deactivation upon SO₂ exposure was effectively inhibited in the presence of NH₃. Temperature-programmed decomposition (TPD) analyses were applied to identify deposit species on the poisoned catalysts by comparison with several groups of reference samples. Diffuses reflectance infrared Fourier transform spectroscopy (DRIFTS) over CeNbTi catalysts with different poisoning pretreatments and gas purging sequences were designed to investigate the roles of NH₃ in the removal of surface sulfites and sulfates. More ammonium sulfates including ammonium bisulfate and ammonium cerium sulfate were generated instead of inert cerium sulfate in these conditions. The mechanisms about the formation and transformation of surface deposits upon sulfur poisoning w/wo NH₃ were explored, which provided a basis for developing Ce-based mixed oxides as SCR catalysts for stationary sources.

1. Introduction

Over the past decades nitrogen oxides (NO_x) that are produced from mobile and stationary sources have led to many environment problems [1]. The selective catalytic reduction (SCR) of NO_x using ammonia as the reductant is one of the most promising deNO_x technologies [2]. Ceria-based catalysts have attracted plenty of interest due to some advantageous features including high NO_x conversion, nontoxicity, and low cost [3,4,5,6]. Recently, the activity and hydrothermal stability of ceria-based catalysts have been improved by inducing WO₃ and SnO₂ as a promoter, realizing over 90% NO_x conversion at 300−550 °C after hydrothermal-aging at 1000 °C, as reported by He et al. [7,8]. This means the application scenarios of ceria-based catalysts are broadened, such as used downstream of the diesel particulate filter (DPF). However, in typical application scenarios, especially coal-fired power plants or the aftertreatment system of marine diesel engines, the effluent gases always contain a certain concentration of SO₂. Thus, SO₂ resistance is one of the most important characteristics of SCR catalysts. The deposition of cerium sulfates and ammonium bisulfate has been recognized as the main factor of deactivation of ceria-based catalysts. The deactivation of these catalysts in the presence of SO₂ is sensitive to reaction temperatures. Zhang et al. [9] reported that with the increase of reaction temperature from 180 to 300 °C, the content of cerium sulfates increased significantly, accompanied with the decreased deposition of ammonium sulfates/bisulfates. Zhang et al. [10] found that the sulfation process was gradually worsened with raising the treating temperature, and the sulfate species over CeO₂ changed from surface sulfates to bulk-like ones, and then to bulk ones. These studies indicate that different sulfur species are formed on ceria-based catalysts at different temperature ranges, which affect NH₃-SCR activity significantly.

In addition to reaction temperatures, other gases, especially ammonia, influence the type of sulfur species that is formed during SO₂ poisoning. Research on the effects of ammonia on SO₂ poisoning have been focused on Cu-CHA catalysts [11,12,13]. For example, as investigated by Wijayanti et al. [11,12], ammonium-sulfur species formed over Cu-SSZ-13 during sulfur poisoning under SCR conditions at 300 °C, reducing the availability of copper sites for the redox SCR cycle due to the site-blocking effect of the deposits [12]. Wang et al. [13] found that copper sulfate was present in all the Cu/SAPO-34 samples that were sulfated at various temperatures, whereas ammonia sulfate was only found over the sample at 250 °C. They suggested that both copper sulfate and ammonia sulfate species decrease the SCR reaction rate of the Cu/SAPO-34 catalyst by reducing the number of Cu²⁺ sites. Moreover, Jangjou et al. [14] found that ammonia can drive sulfur from a more thermodynamically copper sulfate to less stable ammonium sulfate over Cu/SAPO-34 catalysts. The results showed that ammonium affected the types of sulfur-containing species. This appears favorable, as NH₃ can react with pre-adsorbed sulfur on the catalyst to form ammonium-sulfur species, which decomposes at lower temperatures in comparison to the other sulfate forms.

Besides, similar phenomena have also been observed over Fe-W and Nb-V/Ce catalysts in recent studies [15,16], that the introduction of NH₃ in the sulfation process at 300 °C dramatically reduced the formation of surface metal sulfate species. Fe₂(SO₄)₃ can decompose in NH₃ atmosphere at approximately 220 °C, helping reduce sulfate deposit on the catalyst [15]. The NH₃-SCR performances of metal oxide catalysts (such as γ-Fe₂O₃, CeO₂, and γ-MnO₂) in a sulfur-containing atmosphere were closely related to the difference in the deposition/decomposition ability of the formed sulfate species as reported by An et al. [17]. Adsorbed NH₃ and H₂O can promote the decomposition of metal sulfate Me(SO₄)_y, proven by temperature-programmed surface reaction (TPSR) of NH₃.

Despite the above studies, the SO₂ poisoning over ceria-based catalysts in an SCR environment, to our knowledge, has not been comprehensively studied. Different from Cu-CHA, ceria-based catalysts tend to deactivate more quickly in sulfur poisoning due to the formation of cerium sulfate. This may have a positive effect on the activity if the formation of cerium sulfate can be inhibited in the NH₃-containing atmospheres. Therefore, sulfur poisoning in different atmospheres is worth studying for ceria-based catalysts. In our previous work [18,19], the addition of niobium on CeWTi catalyst resulted in high NO_x conversion in a wide temperature window and significantly improved sulfur resistance at low and medium temperatures. In this work, CeNbTi was chosen as a typical ceria-based oxides catalyst, and 300 °C was chosen as a typical temperature for sulfur poisoning of SCR catalysts. The sulfur poisoning mechanism of CeNbTi catalyst in an SCR environment and specifically the role of NH₃ were explored.

2. Experimental

2.1. Catalyst Preparation

CeNbTi catalyst was prepared by a co-precipitation method. C₁₂H₇NbO₂₄ (Aladdin, China) and Ce(NO₃)₃·6H₂O (Aladdin) were dissolved in deionized water with a Nb₂O₅:CeO₂ mass ratio of 2:1. The precursors were mixed with commercial TiO₂ powders (DT51, Cristal, Saudi Arabia) in deionized water according to a (Nb₂O₅ + CeO₂):TiO₂ molar ratio of 3:7. Diluted ammonium hydroxide (Beijing Chem., Beijing, China) was then added in the mixed solution as a precipitating agent under vigorous stirring. The obtained precipitate was dried at 110 °C overnight and calcined in a muffle furnace at 600 °C for 5 h. The loading of CeO₂ and Nb₂O₅ are 12 wt.% and 22 wt.%, respectively.

The poisoned catalysts were obtained by treating the as-received CeNbTi in a gas flow consisting of 500 ppm NO (when used), 500 ppm NH₃ (when used), 200 ppm SO₂ (when used), 5% H₂O (when used), 5% O₂, and balanced N₂ at 300 °C for 6 h.

For reference, the deposited catalysts were prepared by an incipient wetness impregnation method. Ammonium sulfite (Aladdin), ammonium hydrogen sulfate (Aladdin), cerium sulfate (Aladdin), and ammonium cerium sulfate (Aladdin) were impregnated on the CeNbTi catalyst with a nominal loading of 2 wt.%. The aqueous solution was added dropwise to the catalyst powders. The mixture was dried at 80 °C for 12 h, and the obtained samples were denoted as (NH₄)₂SO₃/CeNbTi, NH₄HSO₄/CeNbTi, Ce(SO₄)₂/CeNbTi, and (NH₄)₄Ce(SO₄)₄/CeNbTi, respectively.

2.2. Activity Measurement

The NH₃-SCR activity measurement was carried out in a fixed bed reactor with powder catalyst (200 mg, 40–60 mesh). The reaction gas mixture consisted of 500 ppm NO, 500 ppm NH₃, 5% O₂, 5% H₂O, and N₂ in balance. The gas hourly space velocity (GHSV) was 80,000 h⁻¹ and the reaction temperature was set at 300 °C. The NO_x (NO, NO₂, and N₂O) concentrations in the outlet gases were measured by a Nicolet 380 infrared (IR) spectrometer (Thermo Fisher, Waltham, MA, USA). The NO_x conversion was calculated as follows:

$${\mathrm{NO}}_x\mathrm{conversion}(\%)=\frac{\left[\mathrm{NO}{]}_{\mathrm{in}}+{\left[{\mathrm{NO}}_2\right]}_{\mathrm{in}}-\right[\mathrm{NO}{]}_{\mathrm{out}}-{\left[{\mathrm{NO}}_2\right]}_{\mathrm{out}}}{{[\mathrm{NO}]}_{\mathrm{in}}+{\left[{\mathrm{NO}}_2\right]}_{\mathrm{in}}}\times 100$$

(1)

2.3. Catalyst Characterization

Nitrogen adsorption isotherms were measured on a JW-BK200(JWGB, China) instrument. All the samples were degassed at 220 °C for 1 h before the nitrogen adsorption measurements. The Brunner–Emmett–Teller (BET) surface area was calculated by a multipoint BET method.

Temperature-programmed decomposition (TPD) analyses were performed on an MKS 2030 Fourier transform infrared (FT-IR) gas analyzer (MKS Instruments, USA). A total of 100 mg sample was pretreated in N₂ at 100 °C for 30 min. Then, the sample was ramped to 1000 °C at a rate of 10 °C/min in N₂.

In situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) was performed on a Nicolet 6700 IR spectrometer (Thermo Fisher). The spectra were calculated by the Kubelka–Munk function. The reaction atmosphere consisted of 200/500 ppm NH₃ (when used), 100/200 ppm SO₂ (when used), 5% O₂, and N₂ in balance. The catalyst in a diffuse reflectance IR cell was pretreated by N₂ (100 mL/min) at 500 °C for 30 min to remove traces of organic residues. Then, the sample was cooled down to 100 °C. Afterward, the samples were purged with the reaction atmosphere and the spectra were collected as a function of time.

3. Results

3.1. NH₃-SCR Activity

The NO_x conversions of the used and poisoned catalysts were obtained from the isothermal activity test at 300 °C and the results are shown in Figure 1. After running in a standard NH₃-SCR condition for 6 h, the used CeNbTi catalyst still exhibited high SCR activity, with a temperature window of 195–420 °C in which the NO_x conversion exceeded 80% (Figure S1). When only SO₂ and O₂ are applied for a sulfur poisoning test, the catalyst is deactivated severely with the NO_x conversion declining sharply to 62%. Such a poisoning effect is maintained in the presence of water. Nevertheless, the runnings in NH₃-containing atmospheres prevent the catalyst deactivation to a great extent, and these poisoned catalysts show ca. 90% NO_x conversion. Again, the influence of water can be omitted. To judge the role of NO, it was removed from the treating atmosphere and the NO_x conversion of the obtained catalyst increased slightly from 88% to 91%. A similar negligible role of H₂O can also be observed. All these demonstrate that ammonia is responsible for the greatly improved sulfur resistance of the ceria-based catalyst.

3.2. TPD Analyses

3.2.1. Identification of Surface Deposits

To identify the deposits on the poisoned catalysts, the TPD experiments were performed in N₂. Almost no NO_x (including NO, NO₂, and N₂O) were observed within the whole temperature range (not shown), indicating their poor adsorption when competing with sulfur oxides. Figure 2 shows the evolution of SO₂ and NH₃ signals during the TPD processes. The NH₃ and SO₂ desorption were estimated by peak fitting and the results are summarized in

Table 1. Amounts of NH₃ and SO₂ that were released from the treated catalysts during the TPD tests.

Sample	SBET (m2/g)	NH3 (μmol/g)	SO2 (μmol/g)
			Peak Temperature (°C)				Deposit *
			565	685	730	900
NH3 + NO + O2	113	16	0	0	98	34	-
SO2 + O2	85	0	35	185	98	34	220
SO2 + O2 + H2O	87	0	35	212	98	38	247
SO2 + O2 + NH3 + NO	93	35	17	131	98	37	148
SO2 + O2 +NH3 + NO + H2O	86	33	15	129	98	32	144
SO2 + O2 + NH3	89	34	17	125	98	32	142

* Estimated by the sum of peaks at 565 and 685 °C.

. Since the sulfur poisoning tests were performed at 300 °C, the NH₃ signal also starts at this temperature in Figure 2a. The NH₃ release (16 μmol/g) for the catalyst that was treated in a simple SCR condition (without SO₂) is derived from desorption of ammonia that is adsorbed at surface acid sites. The ammonia desorption is doubled in the SO₂-containing treating atmospheres. According to the reports of Xu et al. [9] and Song et al. [20], the increased ammonia can be related to the formation of ammonium sulfates including ammonium bisulfate, ammonium sulfate, and ammonium cerium sulfate. The presence of H₂O or NO does not affect the ammonium adsorption on the catalyst upon the sulfur poisoning treatments. Since the specific surface areas of catalysts almost have no changes with or without NH₃-containing SO₂ poisoning atmosphere and the maximal amount of NH₄HSO₄ (about 17 μmol/g) is much smaller than that of metal sulfates and ammonium metal sulfates, the physical poisoning (covering effect) of the deposited NH₄HSO₄ should not be the main factor to deactivate the catalyst.

Compared with NH₃, the SO₂ signal appears at much higher temperatures (450–1000 °C) in Figure 2b. Interestingly, there are two SO₂ peaks over the catalyst that was treated in the simple SCR condition without SO₂. The distinct SO₂ peak centered at 730 °C with a shoulder at 900 °C is associated with the decomposition of titanium sulfates in the original material [18,21,22]. After sulfur poisoning, the major peak shifts towards a lower temperature (685 °C) and another small peak appears at 565 °C. According to the previous studies [23,24], the low-temperature peak is ascribed to the decomposition of metal sulfites. The overlapped peak at 685 °C is attributed to the decomposition of ammonium sulfates and cerium sulfates [25,26]. When ammonia is added into the poisoning atmosphere, both the SO₂ peaks at 565 and 685 °C decrease in intensity significantly. As listed in Table 1, about two thirds of sulfites and about one third of sulfate deposits are reduced on those catalysts.

To further verify the deposits on the treated catalysts, one group of reference samples was prepared and their TPD curves are shown in Figure 2c,d. Among these samples, (NH₄)₂SO₃, NH₄HSO₄, Ce(SO₄)₂, and (NH₄)₄Ce(SO₄)₄ were chosen as the representatives of ammonium sulfites, ammonium bisulfates, cerium sulfates, and ammonium ceria sulfates, respectively. The NH₃ desorption occurs at a much lower temperature (around 250 °C) over (NH₄)₂SO₃/CeNbTi. Comparatively, the NH₃ desorption peak appears at 340 and 380 °C for the NH₄HSO₄ and (NH₄)₄Ce(SO₄)₄ impregnated catalysts, respectively, which are close to those (400 °C) over the catalysts that were poisoned in the NH₃-containing atmospheres Figure 2a.

As shown in Figure 2d, a small SO₂ desorption peak at 525 °C is only found over (NH₄)₂SO₃/CeNbTi, indicating the origin of the 565 °C peak in Figure 2b. It is interesting that a distinct peak at 700 °C is also observed for this sample. One possible explanation is the re-adsorption of SO₂ from sulfite decomposition with ceria acting as strong basic sites, which releases at higher temperatures. Another possibility that cannot be excluded is that sulfites may be readily oxidized to sulfates by active oxygen in ceria during the impregnation and TPD processes. Similar phenomena about delayed release of SO₂ have been observed over many catalysts including VTi, VWTi, and Cu-SSZ-13 [26,27,28]. All the impregnated sulfates show a similar SO₂ desorption peak at 600–750 °C. Therefore, it is difficult to identify specific sulfates from the 685 °C desorption peaks in Figure 2b in this way. Nevertheless, some important information can still be deduced. Considering the fact that almost only cerium sulfates (Ce(SO₄)₂ and Ce₂(SO₄)₃) are generated in the poisoning atmospheres without NH₃ and the total SO₂ desorption declines over the SO₂ + NH₃ co-treated catalysts (Figure 2b), the additional NH₃ release over the latter catalysts (Figure 2a) should be ascribed to the decomposition of ammonium sulfates such as NH₄HSO₄ and (NH₄)₄Ce(SO₄)₄. All these indicate that the formation of cerium sulfates is inhibited to a large extent in the presence of NH₃.

3.2.2. Effect of Ammonia on Surface Deposits

To investigate the mechanism of NH₃ addition on the sulfur resistance of CeNbTi catalyst, several experiments were performed to characterize structural changes after the durability tests. No obvious differences were observed in the XRD patterns (Figure S2) of the fresh and poisoned catalysts. Only the characteristic peaks of anatase TiO₂ are observed, and those of ceria, niobia, or sulfates do not appear due to their low crystallinity. In order to further identify the deposits on the poisoned catalyst, another group of reference samples was prepared by treating CeNbTi catalyst with SO₂ + O₂ and NH₃ in different sequences before TPD experiments. As shown in Figure 3, a short-time (30 min) pretreatment in SO₂ + O₂ resulted in similar TPD curves of NH₃ and SO₂ as those for a much longer period (6 h). The sulfite-related SO₂ desorption peak of the former catalyst is even stronger than that in Figure 3b, indicating the transformation of sulfites to sulfates proceeded with time-on-stream. After that, NH₃ alone was introduced to react with the sulfates/sulfites on the SO₂ + O₂ pretreated catalyst and then the TPD measurement was performed. Here, the sulfite-related peak almost diminishes due to the formation and following decomposition of ammonium sulfite. As shown by the TG/DSC curves in Figure S3, ammonium sulfite decomposes easily even at temperatures below 120 °C. It implies that once it was formed by a reaction of metal sulfites with the introduced ammonia, ammonia sulfite would decompose subsequently during the pretreatment. As a result, few metal sulfites exist on the catalyst after NH₃ purging (Figure 3b). Meanwhile, a small NH₃ desorption peak occurs at 400 °C (Figure 3a), assigned to the decomposition of ammonium sulfates and chemisorbed ammonia on the surface acid sites.

Then, a reverse sequence was adopted for purging gases. The catalyst was first pretreated in NH₃ for 30 min, aiming to saturate NH₃ adsorption, and then the feeding gas was switched to SO₂ + O₂. It is interesting to find that the sulfite-related SO₂ peak significantly decreases in intensity and almost no NH₃ desorbs. In this case, the adsorbed NH₃ is firstly consumed by gaseous SO₂ via forming ammonium sulfite and subsequent decomposition, as well as driven by competitive adsorption of SO_x. Then, the continuous passing with SO₂ + O₂ produces ceria-based sulfite and sulfate species. It is noted that the sulfate-related SO₂ desorption peak decreases slightly in intensity after purging with NH₃ (Figure 3b). It demonstrates that NH₃ can remove surface sulfites but release the deposited sulfates to a small extent within the limited time. Thus, the reduced sulfate deposits in the presence of NH₃ (Figure 3b) should be ascribed to a great extent to other factors such as competitive adsorption between NH₃ and SO₂.

3.3. Infrared Studies

3.3.1. Surface Groups Identification

Ex situ IR spectra of the treated CeNbTi catalysts were compared with reference samples to determine surface deposit species. As shown in Figure 4a, four main bands are found in the spectrum of ammonium sulfite. The bands at 1183, 1027, and 930 cm⁻¹ are assigned to sulfite species [29], and the one at 1502 cm⁻¹ is assigned to NH₄⁺ on Brønsted acid sites [30]. In the case of ammonium hydrogen sulfate (Figure 4b), the band at 1478 cm⁻¹ is attributed to asymmetric bending vibrations of NH₄⁺ and the bands at 1326 and 1260 cm⁻¹ are assigned to the S=O vibration absorption singles of HSO₄⁻ [31]. For cerium sulfate (Figure 4c), the bands at 1367 and 1300 cm⁻¹ are related to the asymmetric stretching frequencies of O=S=O species [32]. The band at 1660 cm⁻¹ is assigned to O-H stretching modes from the reaction between SO₂ and surface hydroxyl groups [33,34]. As for ammonium cerium sulfate, five characteristic bands are observed in Figure 4d. The distinct band at 1275 cm⁻¹ with a shoulder at 1200 cm⁻¹ is ascribed to bidentate sulfate species [35]. The band at 1067 cm⁻¹ is attributed to the S-O v₃ vibrations of Ce-bonded bidentate SO₄²⁻ in C_2v symmetry [36]. The bands at 1606 and 1468 cm⁻¹ are assigned to the NH₄⁺ on Brønsted acid sites [37].

The IR spectrum of the CeNbTi that was treated in a simple SCR condition is shown in Figure 4e as the background. The band at 1617 cm⁻¹ is assigned to adsorb H₂O from the reaction between SO₂ and surface hydroxyl groups [33]. The band at 965 cm⁻¹ is attributed to characteristic vibrations of Nb=O bond in surface NbO_x species [23] and the one at 1128 cm⁻¹ is due to modes of Nb-O-Nb stretching vibration [38,39]. The stretching modes of Ce-O-Ce bands generally results in a band at 729 cm⁻¹ [34], which shifts to 782 cm⁻¹ due to interacting with Nb₂O₅ [38]. As for the spectrum that was collected over the catalyst upon exposure to SO₂ + O₂ (Figure 4f), the sulfate-related bands can be observed. There are two broad bands ranging at 1370–1300 and 1275–1180 cm⁻¹ that are ascribed to the asymmetric stretching frequencies of the O=S=O bond and S-O bond of bidentate sulfate species, respectively. SO₂ + O₂ + NH₃ +NO was chosen as a representative of three NH₃-containing poisoning atmospheres. As shown in Figure 4g, the spectrum of the obtained catalyst exhibits a characteristic band at 1446 cm⁻¹. It is assigned to the asymmetric bending vibrations of ionic NH₄⁺ that is bound to Brønsted acid sites that are provided by surface sulfates (i.e., NH₄HSO₄ and (NH₄)₄Ce(SO₄)₂). Considering the similar intensities of the 1478 and 1326 cm⁻¹ bands for NH₄HSO₄ (Figure 4b), the extra intensified band at 1446 cm⁻¹ may mean the formation of ammonium cerium sulfate to a great extent. Another evidence is the band at 1067 cm⁻¹ that is attributed to Ce-bonded bidentate sulfates. The broad band at 1275–1180 cm⁻¹ may come from HSO₄⁻¹, CeSO₄, and (NH₄)₄Ce(SO₄)₄. Other adsorption bands of sulfates at 1381 and 1354 cm⁻¹ are attributed to asymmetric stretching frequencies of O=S=O species that are related to the formation of Ce(SO₄)₂.

3.3.2. In Situ SO₂ + O₂ Adsorption

Figure 5a shows the DRIFT spectra of adsorbed species over CeNbTi catalyst in a SO₂ + O₂ flow at 100 °C as a function of time. As the signals that were obtained by performing the DRIFTS experiments at 300 °C are much weaker, the adsorption of reactants and products becomes more significant by reducing the operating temperature to 100 °C. Hereby, the sulfur poisoning and ammonium protecting effects are amplified, although the specific mechanisms may be somewhat different at two temperatures. The band at 1268 cm⁻¹ is assigned to bidentate sulfates [35], while several small features at 1180, 1024, and 925 cm⁻¹ are attributed to sulfite species that are associated with different adsorption sites [33]. The band at 1086 cm⁻¹ could be assigned to v₃ vibrations of bidentate SO₄²⁻ in C_2v symmetry [36], while those at 1383 and 1300 cm⁻¹ could be attributed to the asymmetric stretching frequencies of O=S=O species for sulfates that were adsorbed on metal oxides [32]. The band at 1430 cm⁻¹ is attributed to the formed S−OH from the reaction between SO₂ and metal cations [18]. The band at 1630 cm⁻¹ is assigned to O-H stretching modes from the reaction between SO₂ and surface hydroxyl groups [33,34]. All these bands increase sharply in intensity with time, indicating the deposition of a large amount of sulfite/sulfate species on CeNbTi catalyst upon sulfur poisoning.

3.3.3. In Situ Surface Reaction between Ad-Species and NH₃/SO₂

As mentioned above, fewer sulfates and sulfites formed when SO₂ poisoning tests were performed in the NH₃-containing atmospheres. In order to understand how NH₃ affects the adsorbed sulfate/sulfite species, DRIFTS was performed to characterize the evolution of surface species on the SO₂ + O₂ pretreated catalyst upon exposure to NH₃ and the results are shown in Figure 5b. The sulfate bands (1428, 1383, 1300, 1268, and 1086 cm⁻¹) and sulfite bands (1024 and 925 cm⁻¹) were observed for the pretreated catalyst. After purging with NH₃ for 30 min, a new band appears at 1676 cm⁻¹, assigned to NH₄⁺ species that are bound to Bronsted acid sites [38,40]. Meanwhile, one band that was assigned to sulfate species shifts from 1430 to 1452 cm⁻¹. The regular IR spectrum of ammonium bisulfate and ammonium cerium sulfate shows a NH₄⁺ characteristic band at 1478 (Figure 4b) and 1468 cm⁻¹ (Figure 4c), respectively. Both the asymmetric bending vibrations of NH₄⁺ and the formed S−OH from metal sulfates contribute to this band. Thus, it implies the generation of ammonium cerium sulfate and ammonium bisulfate after the introduction of NH₃. Meanwhile, the metal sulfate bands (1383 and 1300 cm⁻¹) disappear finally, suggesting the transformation of Ce(SO₄)₂ to (NH₄)₄Ce(SO₄)₄ since NH₄⁺ bonding with SO₄²⁻ consumes O=S=O in Ce(SO₄)₂ [41], and the latter species decomposes subsequently (Figure S4).

A reverse DRIFT spectroscopy was also performed to explore the evolution of surface species on the NH₃ pretreated catalyst upon exposure to SO₂ + O₂. As shown in Figure 4c, the Brønsted acid site-related bands (1660 and 1451 cm⁻¹) and Lewis acid site-related bands (1600, 1210, and 1024 cm⁻¹) are observed initially for the pretreated catalyst [10,34,42,43]. After the introduction of SO₂, the sulfate-related bands (1434, 1266, and 1068 cm⁻¹) form with time-on-stream. Importantly, the bands at 1383 and 1300 cm⁻¹ that were attributed to CeSO₄ can hardly be observed, illustrating that introduced SO_x prefers to react with adsorbed NH₃ to form NH₄HSO₄ and (NH₄)₄Ce(SO₄)₄. Meanwhile, the sulfite-related bands at 1022 and 925 cm⁻¹ appear slowly. By comparison with the case of the as-received catalyst (Figure 5c), it indicates that the adsorbed ammonia reduces the formation of metal sulfates effectively. That is, the competitive adsorption between NH₃ and SO₂ would improve the sulfur resistance of the catalyst greatly.

Furthermore, in order to judge the discrepancy of NH₃ and SO₂ adsorption capacity over the CeNbTi catalyst, a competitive adsorption test was performed in a gas stream containing 200 ppm NH₃, 200 ppm SO₂, and 5% O₂. At the initial stage (1 min) in Figure 5d, NH₃ is preferentially adsorbed on Brønsted acid sites with the symmetric bending vibrations of NH₄⁺ at 1618 cm⁻¹, and then (2 min) the signal of S-O bond in SO₄²⁻ appears at 1300 cm⁻¹. After that (3 min), overlapped signals of the asymmetric bending vibrations of adsorbed NH₄⁺ and the S−OH vibrations of SO₄²⁻ appear at 1430 cm⁻¹. The vibration signals of NH₄⁺ and SO₄²⁻ show a blue shift to a higher wavenumber and a red shift to lower wavenumber, respectively. This was derived from more NH₃ adsorption forming NH₄⁺[NH₃]_n structure [44], O-H from the reaction between SO₂ and surface hydroxyl groups, and SO₄²⁻ influenced by ammonia. These results demonstrate that the ammonia co-adsorbs on the CeNbTi catalyst with SO₂ and could promote metal sulfates species transforming to ammonium sulfate-type and ammonium metal sulfate-type species.

4. Discussion

Intrinsically, NH₃ as an alkaline gas can facilitate SO₂ adsorption on ceria-based catalysts by the formation of ammonium sulfate-type species. Figure 6 shows the evolution of the typical sulfate bands (1300–1256 cm⁻¹) and sulfite band (925 cm⁻¹) over the CeNbTi catalyst that was treated by different poisoning procedures that were derived from Figure 5a,c,d, respectively. The peak areas of these bands were integrated and drawn as a function of time. In a simplified atmosphere (SO₂ + O₂), the sulfate bands increase in intensity continuously (Figure 6a), with cerium sulfate as the main product (Equation (2)), proven by the infrared spectra in Figure 4f and Figure 5a. It also correlates with the TG results (Table 1) that after the treatment in SO₂+O₂ for 6 h, about 13% of CeO₂ transforms to cerium sulfate by assuming cerium sulfate as the dominating metal sulfate. Similarly, the sulfite band also obviously increases in intensity (Figure 6b), with the transformation of about 3% of CeO₂ to cerium sulfites.

Significantly, the formation of sulfates/sulfites is inhibited on the NH₃ pretreated catalyst and reaches a quasi-equilibrium quickly. One possible elimination route of sulfites by ammonia is the reaction between the pre-absorbed NH₃ and sulfur dioxide to form ammonia sulfite (Equation (3)) which can decompose feasibly at much lower temperatures (Figure S3). When NH₃, SO₂, and O₂ co-purge over the catalyst, almost no sulfite band is observed.

Meanwhile, the deposition amount of the sulfates decreases sharply on the NH₃ pretreated catalyst. Ammonium bisulfate is more likely to form on niobium-modified catalysts [18], which can protect ceria from sulfur poisoning. As shown by the DRIFTS result of SO₂ + O₂ adsorption (Figure 5c), typical signals of HSO₄²⁻ (1268 cm⁻¹) and NH₄⁺ that are adsorbed on Brønsted acid sites that are provided by surface sulfates (1434 cm⁻¹) are observed. Therefore, it is reasonable that NH₄HSO₄ deposits on the pretreated catalyst according to Equation (4).

Previous studies [21,45] discovered that the decomposition of NH₄HSO₄ occurs at the temperature range of 320–450 °C and NH₄HSO₄ can react with ceria to form ammonium cerium sulfate at lower temperatures according to Equation (5). Jangjou [14] found that the ammonia could cause the transformation from Cu_xSO_y to (NH₄)_xCu_ySO_z at 210 °C on the Cu-SSZ-13 catalyst. Similarly, it can be seen in Figure 5b that (NH₄)₄Ce(SO₄)₄ species forms on the sulfated CeNbTi catalyst via Equation (6) upon NH₃ purging. Compared with Ce(SO₄)₂, this material can serve as an active species to a certain extent via the redox cycle between (NH₄)₄Ce(SO₄)₄ and NH₄Ce(SO₄)₂ (Figure S5). The competitive adsorption of NH₃ and SO₂ on the catalyst further reduces the surface sulfate deposition. NH₃ adsorbs preferentially on the catalyst (Figure 5d), and this preferential occupation of surface active sites by NH₃ reduces SO_x adsorption. As listed in Table 1, it can be calculated that about 4–5% of CeO₂ is deactivated by assuming (NH₄)₄Ce(SO₄)₄ as the distinct deposit when treating the catalyst in the SO₂ + O₂ + NH₃ + NO atmosphere.

CeO₂ + 2SO₂ + O₂ → Ce(SO₄)₂

(2)

$$2{\mathrm{NH}}_3+{\mathrm{SO}}_2+{\mathrm{H}}_2\mathrm{O}\leftrightharpoons {({\mathrm{NH}}_4)}_2{\mathrm{SO}}_3$$

(3)

2NH₃ + 2SO₂ + O₂ + 2H₂O → 2NH₄HSO₄

(4)

4NH₄HSO₄ + CeO₂ → (NH₄)₄Ce(SO₄)₄ + 2H₂O

(5)

Ce(SO₄)₂ + 4NH₃ + 2SO₂ + 2H₂O + O₂ → (NH₄)₄Ce(SO₄)₄

(6)

The degradation degree of CeNbTi catalyst by SO₂ poisoning under SCR conditions (NH₃ + NO + O₂) is much slighter than SO₂ + O₂ exposure alone. Based on the above results and analysis, ammonia inhibits sulfur poisoning in NH₃-SCR to a great extent via two main ways. On the one hand, ammonia would react with sulfites and form ammonium-sulfite which can subsequently decompose. Consequently, there is much less sulfite residue on the catalyst in the presence of ammonia. On the other hand, ammonia can pre-occupy surface adsorptions sites separating SO₂ and active sites to reduce the sulfation of active metal. Besides, ammonia can transform cerium sulfate to ammonium cerium sulfate with relatively higher reactivity than cerium sulfate. The corresponding SO₂ deactivation mechanisms in the presence and absence of ammonia are presented primarily in Figure 7. The present work discloses the inhibition effect of NH₃ on sulfur poisoning of CeNbTi catalyst, which provides a comprehensive mechanistic insight and enables the development of ceria-based oxide catalysts for nitrogen oxide emission control in stationary sources. It should be noted that the above findings are based on the TPD and DRIFTS results, in which the suggestion of surface species still suffers from some uncertainty and remains to be further validated.

5. Conclusions

The deactivation of CeNbTi catalysts by exposure to SO₂ in different atmospheres was investigated. The TPD and DRIFTS techniques were employed to characterize different species that were deposited on the catalyst and investigate their formation/decomposition/transformation mechanisms. Several major conclusions were drawn as follows:

(1)

After exposure to SO₂ + O₂ for 6 h, the NO_x conversion at 300 °C of the catalyst decreases from almost 100% to ca. 60%, while the presence of NH₃ in the poisoning atmospheres can reserve the catalyst with about 90% NH₃-SCR activity. Additionally, humid conditions do not result in any obvious changes in the deactivation degree of sulfur poisoning with or without NH₃.

(2)

The types and amounts of sulfates/sulfites depend importantly on the sulfur poisoning atmospheres. When the poisoning is performed in the presence of NH₃, the total amount of sulfites and sulfates that are deposited on the catalyst is reduced by 44% compared with that in absence of ammonia. Competitive adsorption between NH₃ and SO₂ is suggested to be one of the dominating factors for the decreased surface deposits. The pre-occupying NH₃ is confirmed to protect ceria active sites from reacting with SO₂.

(3)

With the introduction of ammonia, not only the amounts of sulfates/sulfites decrease significantly, but also the types of sulfates change a lot from metal sulfates to ammonium sulfates. The ready decomposition behaviors of NH₄HSO₄ (and (NH₄)₂SO₄), as well as the transformation of cerium sulfates with more sulfate radicals that are bonded per cerium atom, can also transform to cerium ammonium sulfates which facilitate the decrease in sulfate deposits. Compared with inert metal sulfates, cerium ammonium sulfates even contribute to NH₃-SCR reaction with the redox cycle between Ce³⁺ and Ce⁴⁺ in the ammonium sulfates. Additionally, metal sulfites can be partially converted into ammonium sulfite in a NH₃-containing atmosphere which is easy to decompose even at low temperatures, resulting in fewer sulfite deposits.