Study on the Mechanism of SO₂ Poisoning of MnO_x/PG for Lower Temperature SCR by Simple Washing Regeneration

Abstract

Manganese oxide-supported palygorskite (MnO_x/PG) catalysts are considered highly efficient for low-temperature SCR of NO_x. However, the MnO_x/PG catalyst tends to be poisoned by SO₂. The effect of SO₂ on activity of the SO₂-pretreated poisoning catalysts under ammonia-free conditions was explored. It was determined that the MnO_x/PG catalyst tends to be considerably deactivated by SO₂ in the absence of ammonia and that water-washed regeneration can completely recover activity of the deactivated catalyst. Based on these results and characterizations of the catalysts, a reasonable mechanism for the deactivation of MnO_x/PG catalyst by SO₂ was proposed in this study. SO₂ easily oxidized to SO₃ on the surface of the catalyst, leading to the formation of polysulfuric acid, wrapping of the active component and blocking the micropores. The deactivation of the MnO_x/PG catalyst is initially caused by the formation of polysulfuric rather than the deposition of ammonia sulfate, which occurs later.

1. Introduction

NO_x is a series of reactive gases, including NO₂, NO, and N₂O. Considerable NO_x emissions due to human activities cause significant environmental problems, such as acid rain, the ozone hole, and photochemical smog [1,2,3,4]. Currently, selective catalytic reduction using NH₃ (NH₃-SCR) technology is widely used in thermal power plants, industrial boilers, and other fixed-source flue gas-control [5], owing to its high efficiency and low cost. A catalyst is essential for NH₃-SCR technology. Metal oxide catalysts are widely used in NH₃-SCR for their excellent denitrification performance and availability. Currently, V₂O₅/TiO₂ and V₂O₅-WO₃/TiO₂ catalysts are widely used. Because they need a relatively high reaction temperature of above 300 °C, so the process is usually performed in the gas upstream [6]. An SCR denitrification device is commonly located in front of the dust removal and desulfurization device, which results in the catalyst being easily inactivated by the poisoning of SO₂ and dust. Therefore, low-temperature SCR technology is attractive. Owing to its advantages of low energy consumption and low impact of SO₂ and dust, SCR technology has gradually attracted the attention of researchers [7,8,9,10].

To date, a series of low-temperature metal oxide catalysts (e.g., cerium (Ce), cobalt (Co), copper (Cu), iron (Fe), manganese (Mn), molybdenum (Mo), nickel (Ni), and vanadium (V)) have been studied. Of these, manganese oxide (MnO_x) has various oxidation states, a high valence state, and characteristic crystal form, and it shows excellent NO conversion and N₂ selectivity. Pena et al. [11] showed that MnO_x/TiO₂ had the highest activity compared to Co, Cr, Cu, Fe, Mn, Ni, and V oxides supported on TiO₂ at a low temperature. Among the characteristic crystal forms of MnO_x, the active components of the catalysts MnO₂ and Mn₂O₃ show the highest NO conversion and N₂ selectivity [12]. However, residual sulfur dioxide after desulfurization has a strong poisoning effect on metal oxidation catalysts at low temperatures. Thus, it is important to clarify the poisoning mechanism of the catalyst. There are different opinions about the poisoning mechanism of SO₂. Kijlstra et al. [13] speculated that the main reason for the decrease in the activity of the MnO_x/Al₂O₃ catalyst after passing SO₂ was the sulfation of the active center atom, Mn, to form manganese sulfate. Luo et al. [14] believed that an SO₂-induced decrease in the MnO_x/MWCNTs catalyst’s activity was attributed, in part, to the formation of ammonium sulfate on the catalyst’s surface that blocks the catalyst’s pores and is also attributed, in part, to the sulfation of the active center atom, Mn, to form MnSO₄. Qi and Yang [15] thought that the main reason for MnO_x/TiO₂ deactivation is the competitive adsorption of SO₂ and NO. This deactivation is reversible, and the activity of the catalyst can be completely restored after removing SO₂.

The structural formula of palygorskite (PG) is Mg₅Si₈O₂₀(OH)₂(OH₂)₄·4H₂O. In our previous study, MnO_x/PG showed a high NO_x conversion rate at a low temperature window of 100–300 °C [16]. Because of its easily available raw materials, abundant channel structure, and large specific surface area, MnO_x/PG has a good industrialization prospect. However, its NO_x conversion rate is easily reduced under low concentrations of SO₂. Currently, as there are no unanimous conclusions on the poisoning mechanism of Mn-based catalysts, our group explored the poisoning mechanism of MnO_x/PG. Zhang et al. [17] determined the poisoning mechanism of MnO_x/PG catalysts by simulating the conditions of poisoning and the recovery of thermal regeneration. They thought that the reason for the deactivation of the catalyst was the collaborative effect of ammonium sulfate and metal sulfate and that the deposition of ammonium sulfate was the main reason for the deactivation of the catalyst at low temperatures. The addition of Ce enables MnO_x/PG to resist sulfur [18]. The abovementioned researchers believed that the fundamental reason for the resistance to Ce is to inhibit MnSO₄ production. As a transition metal, Ce easily combines with SOx to form stable Ce₂(SO₄)₃, which protects the sulfation of Mn. The formation of MnSO₄ is an important reason for the deactivation of the catalyst above 150 °C. However, there is no clear conclusion about the deactivation mechanism of the Mn₈/PG catalysts. Therefore, it is necessary to explore the cause of low-temperature SCR catalyst inactivation caused by SO₂ so as to further find a solution to SCR-catalyst sulfide poisoning. In addition to ammonium sulfate salts and ammonium sulfite salts, the effect of metal sulfate salts on the deactivation of catalysts is controversial. In this study, the reasons why SO₂ poisoning leads to a decrease in the catalytic activity and water-washed regeneration leads to the recovery of activity were studied under ammonia-free conditions, combined with specific surface-area and pore-size determination (BET), elemental analysis (EA), X-ray diffraction (XRD), temperature-programmed reduction (TPR), temperature-programmed desorption (TPD), and temperature-programmed surface reaction (TPSR). After discussing the mechanisms of SO₂-caused poisoning and water-washed regeneration in depth, we proposed a possible explanation for the catalyst deactivation by SO₂ and regeneration of the deactivation catalyst by a water-washing process.

2. Results and Discussion

2.1. Elemental Analysis, Specific Surface-Area and Pore-Volume Determination, and X-ray Diffraction

To explore the composition of soluble sulfur compounds and their effect on the physical environment of the catalyst surface, elemental and BET analysis tests were performed.

Table 1. S content, specific surface area and pore volume of the three catalysts.

Catalysts	S Content (wt%)	SBET (m2/g)	Vp (cm3/g)
Mn8/PG	0.025	121.44	0.436
Mn8/PG-S	0.736	92.76	0.421
Mn8/PG-S-W	0.123	115.21	0.432

shows that the S content of the Mn₈/PG-S catalyst is significantly higher than that of fresh the Mn₈/PG catalyst. A negligible portion of sulfur was observed in the Mn₈/PG catalyst due to the measurement accuracy of the method or impurities in the instrument. The S content on the catalyst surface was significantly decreased after water-washed regeneration, which indicates that the accumulation and removal of S from the catalyst were the key factors for the deactivation and regeneration of the catalyst.

The BET results show that the specific surface area and pore volume of the catalyst decreased after the SO₂-pretreated poisoning at 250 °C; the specific surface area and pore volume decreased by 24% and 3%, respectively. Clearly, the formation of sulfur blocked pores and reduced the specific surface area, which was an important factor that inhibited the SCR activity of the Mn₈/PG-S catalyst. The specific surface area and pore volume of the Mn₈/PG-S-W catalyst considerably increased compared to that of the Mn₈/PG-S catalyst. Thus, the SCR activity of the Mn₈/PG-S-W catalyst recovered even more than that of fresh samples. This suggests that sulfur compounds on the catalyst were efficiently removed by the water-washing process, which resulted in a large extent of recovery of specific surface area. A small portion of sulfur was observed in the Mn₈/PG catalyst due to the measurement accuracy of the method or impurities in the instrument because no sulfur-containing precursors were used during the preparation of the catalyst.

To explore the changes in the crystal shape of the catalyst before and after the SO₂-pretreated poisoning and water-washed regeneration, XRD tests were performed on the three catalysts. The XRD patterns of the three catalysts are shown in Figure 1. All of the catalysts obtained by different treatments maintained the primary structure of palygorskite.

The main forms of manganese oxide on the surfaces of Mn₈/PG catalysts were MnO₂ (i.e., 37.6°, 42.8°, 57.0°) [PDF#30-0820] and Mn₂O₃ [PDF#10-0069]. It is easier to form amorphous Mn₂O₃ than MnO₂; thus, it is difficult to observe Mn₂O₃ by XRD. The intensity of the MnO₂ diffraction peak in the catalyst was weakened after Mn₈/PG catalysts were SO₂-pretreated, and the intensity of the diffraction peak of the MnO₂ catalyst was restored after the water-washed treatment. The abovementioned results were obtained because sulfur compounds, which were formed on the surface of the catalyst, covered MnO₂ after the treatment in the SO₂ atmosphere, and the characteristic diffraction peak intensity of MnO₂ was weakened. The water-washed regeneration can remove sulfur compounds and restore the intensity of the diffraction peaks of MnO₂.

2.2. Scanning Electron Microscopy

According to the abovementioned analysis, it is believed that the presence of sulfur compounds likely causes a significant change in the morphology of the catalyst; thus, SEM analysis was performed to investigate morphological changes of the three catalysts. Figure 2a shows that the Mn₈/PG catalyst has a one-dimensional rod-shaped structure with individual rods being uniformly dispersed. These properties are considered to be essential to obtain superior catalyst activity. For the Mn₈/PG-S catalyst (Figure 2b), a clear agglomeration of catalysts was observed. A paste-like compound wrapped the catalyst’s surface. For the Mn₈/PG-S-W catalyst (Figure 2c), the individual rods were highly dispersed, and the paste-like compound on the rods was clearly removed. The appearance of the Mn₈/PG-S-W catalyst was similar to that of the Mn₈/PG catalyst. Combined with the BET analysis, it is suggested that the paste-like compound, which is formed during the sulfur treatment, causes a decrease in specific surface area and catalytic deactivation.

2.3. In Situ DRIFT

To further study the internal functional groups of the catalyst after the SO₂-pretreated poisoning and water-washed regeneration, the three catalysts were evaluated by FT-IR spectroscopy, as shown in Figure 3. Two peaks can be seen in the wavelength range of 1300–800 cm⁻¹. All the samples show the same broad peak at 1020 cm⁻¹, which is attributed to the anti-symmetrical vibration of quartz Si–O–Si, which is contributed by the PG support. The peak vibration at 876 cm⁻¹ is attributed to the S–O stretching vibration that was observed on Mn₈/PG-S and Mn₈/PG-S-W samples [19]. There was no change detected in the 1020 cm⁻¹ peak intensity for the three catalysts, which indicates that Mn₈/PG-S poisoning and Mn₈/PG-S-W regeneration did not affect the carrier, with SiO₂ as the main component of the catalyst. However, the peak intensity of S–O at 876 cm⁻¹ considerably changed for the three samples. For the Mn8/PG-S catalyst, there was a clear peak at that frequency, which indicates that sulfur compounds formed on the surface of the catalysts. After water-washed regeneration, the peak at that frequency became weak, which indicates that some of the sulfur compounds were washed away, and some metal sulfate salts remained on the catalyst surface.

On the basis of the abovementioned tests and characterization analyses, it can be suggested that the deactivation of the catalyst after the SO₂ pretreatment is due to the formation of soluble sulfur compounds rather than MnSO₄. Those sulfur compounds can be easily removed by water washing to completely restore the SCR activity of the catalyst. According to reports by Seong et al. [20], at a higher SO₄²⁻ concentration, the isolated sulfate will be converted into a polynuclear sulfate type, which is larger and will wrap around the active component. It is reasonable to suggest that the soluble sulfur species is actually an SO₄²⁻ polymer.

2.4. Atomic Absorption Spectroscopy and X-ray Fluorescence Spectroscopy

Because of the compositional characteristics of the palygorskite carrier (magnesium aluminum silicate), the sample contains a large amount of O, Si, Mg, and Al elements. By comparing the Mn content of the three catalysts, we determined that the Mn content does not considerably change (as shown in

Table 2. Content of elements in different catalysts.

	Element Content (wt%)
Catalysts	Si	O	Mn	Mg	Al	Fe	C	K	Ca	Ti	S	R
Mn8/PG	36.2	30.8	8.32	8.51	5.44	4.70	1.79	1.03	1.42	0.82	0.08	0.91
Mn8/PG-W-1	35.5	33.4	8.65	7.43	5.25	4.21	2.63	0.85	0.82	0.54	0.42	0.28
Mn8/PG-W-2	36.4	33	9.18	7.46	4.86	3.63	3.13	0.86	0.45	0.48	0.34	0.24
Mn8/PG-W-3	36	33.7	8.65	7.44	4.97	3.56	2.85	0.87	0.27	0.48	0.33	0.84
Mn8/PG-W-4	36.8	33.8	9.29	7.24	4.92	3.84	3.10	0.86	0.17	0.49	0.30	0.22

) and that the overall growth trend is attributed to the water-washing process, which washes away a part of the carrier.

Combined with the atomic absorption spectroscopy analysis, shown in

Table 3. Elements’ content in different water-washed solutions of catalysts.

Catalysts	Mn	Mg	PH
Mn8/PG	2.1 mg/L	27.7 mg/L	7.56
Mn8/PG-W-1	4.6 mg/L	56.2 mg/L	5.30
Mn8/PG-W-3	10.8 mg/L	38 mg/L	7.23
MnO2-W-1	0.24 mg/L	-	6.5

, we determined that there are small amounts of Mn and Mg in the water-washed solution of the Mn₈/PG catalyst because of the incompletely decomposed precursor Mn(NO₃)₂ and soluble magnesium salt on the carrier. A small amount of Mn was observed in the water-washed solutions of the Mn₈/PG-S catalyst that was washed once and thrice; this is attributed to the fact that SO₃ reacts with active manganese and eventually produces MnSO₄, which is not strongly bonded to the carrier. The Mn element that is present during the SO₂-pretreatment process is converted into the content of lost manganese. It can be determined that the amount of manganese removed during the SO₂ pretreatment process is considerably small (approximately 1% of manganese), which suggests that Mn is not lost during the SO₂ pretreatment process. The SO₂ pretreatment process does not wash away manganese present on the catalyst, which is an important prerequisite for the recovery of the catalyst activity. To further determine the loss of Mn during the SO₂ pretreatment process, we excluded other factors that may affect the experimental results. Thus, a simulation experiment with pure MnO₂ was performed. MnO₂ was used as a SO₂ pretreatment with the same process as for Mn₈/PG. It was determined that MnO₂ had the same trend of activity as that of the Mn₈/PG-S catalyst. The analysis of the water-washed solution of SO₂-pretreated MnO₂ showed that the loss of manganese was in the amount of 0.002%, which can be neglected. The elimination of deactivation of the catalyst via sulfation is not the main cause during the SO₂ pretreatment process.

To investigate the acidity and alkalinity of the three catalysts, the pH of the water-washed solution of the three catalysts was determined. The pH of the water-washed solution of the Mn₈/PG catalyst was 7.56, which is neutral. The pH of the water-washed solution of the Mn₈/PG-S catalyst was 5.3, which was weakly acidic. The pH of the water-washed solution of the Mn₈/PG-S-W catalyst was 7.23. Clearly, compared to the fresh and water-washed samples, the Mn₈/PG-S catalyst was more covered by soluble sulfur compounds. It can be concluded that the removal of acidic sulfur compounds during the SO₂ pretreatment process is the main cause of the recovery of the deactivated catalyst.

2.5. Effect of SO₂ Pretreatment and Water Washing on the Activity of Mn₈/PG Catalysts

There are two reasons for the deactivation of metal oxides by SO₂: the deposition of ammonium sulfate salt on the catalyst and the sulfation of active sites. To eliminate the effect of ammonium sulfate on the activity of the catalyst, Mn₈/PG was SO₂-pretreated under ammonia-free conditions to investigate the effect of metal sulfates on catalytic activity.

Figure 4 shows the NO conversion of different Mn₈/PG catalysts in the studied temperature range. The NO conversion of Mn₈/PG-S was significantly inhibited compared with Mn₈/PG, indicating that the activity of catalysts was remarkably affected by SO₂ pretreatment. The activity of the catalyst was considerably restored after Mn₈/PG-S catalysts were water-washed, and the NO conversion was even slightly higher than that of the Mn₈/PG catalyst. The results show that SO₂ pretreatment causes irreversible deactivation in the low-temperature range of the catalyst, which is consistent with the results of the research on the reversible and irreversible deactivation of the Cu-CHA NH₃-SCR catalyst by Hammershøia et al. [21]. The NH₃-SCR activity of the Mn₈/PG-S catalyst considerably increased from 250 °C to 300 °C. This occurs because the oxidation efficiency of manganese increases with an increase in temperature, which directly leads to an increase in the SCR reaction rate per unit of time.

The NO conversion of the Mn₈/PG-S catalyst is considerably inhibited compared with that of the Mn₈/PG catalyst in the low-temperature range, indicating that the formation of ammonium sulfate on the catalyst surface is not the main cause of catalyst deactivation. The catalyst was possibly deactivated before the formation of ammonium sulfate. We tentatively concluded that the stratification of ammonium sulfate was not the main reason for the deactivation of the Mn₈/PG catalyst.

2.6. X-ray Photoelectron Spectroscopy (XPS)

The S atom concentration of the Mn₈/PG-S catalyst is 3.18%, whereas the S atom concentration of the Mn₈/PG-S-W catalyst is 0.67% (

Table 4. XPS results for surface Mn atomic percentage.

Catalysts	Surface Atomic Concentration/%
	Mn				S	O
		Mn4+	Mn3+	Mn2+
Mn8/PG	1.29	68.2%	38.8%	-	-	98.71
Mn8/PG-S	1.07	25.2%	38.6%	36.2%	3.18	95.75
Mn8/PG-S-W	1.44	32.7%	43.7%	23.6%	0.67	97.9

). Combined with the result shown in Figure 5, it was determined that the denitration activity of the Mn₈/PG-S catalyst was restored to that of the fresh catalyst by washing with water. This observation suggests that the removal of S atoms during the water-washing process is essential for recovering the catalyst’s denitration activity.

Figure 5a shows the XPS spectrum of the S 2p orbital of the Mn₈/PG-S catalyst. The S element mainly exists in the forms of S⁴⁺ and S⁶⁺ in the Mn₈/PG-S catalyst. The peak at 168.9 eV is attributed to SO₃²⁻, which contributes to the sulfurous-acid generation by the SO₂ and H₂O adsorption on the surface of the catalyst. The peak at 169.7 eV is attributed to SO₄²⁻, which contributes to the sulfuric-acid generation by SO₃ and H₂O on the surface of the catalysts and surface-generated MnSO₄ [22,23]. The S peak can be divided into S⁶⁺ and S⁴⁺. According to the calculated area of the two peaks, it was determined that the proportions of S⁶⁺ and S⁴⁺ are 49% and 51%, respectively. Compared to the catalytic reaction without a sulfur atmosphere, more S⁴⁺ was oxidized to S⁶⁺ due to the presence of MnO₂ (Mn⁴⁺) and SO₂ when the catalytic reaction proceeded in the sulfur-rich atmosphere [24]. Thus, sulfur was primarily present as a sulfuric acid and MnSO₄ on the surface of the catalyst. After water-washed regeneration, most S was washed away. The XPS spectrum of the S 2p orbital of the washed regenerated catalyst cannot be observed because of the extremely low content of S.

The XPS spectra of the Mn 2p orbitals for the three catalysts are shown in Figure 5b. The Mn 2p 3/2 peak of the Mn₈/PG catalyst can be divided into two peaks: MnO₂ (643.1 eV) and Mn₂O₃ (641.6 eV) [25]. Based on the calculation of the peak area, the proportions of MnO₂ and Mn₂O₃ are 68.2% and 38.8%, respectively. The peak of Mn 2p3/2 can be divided into three peaks, which correspond to MnO₂ (643 eV), Mn₂O₃ (641.8 eV), and manganese sulfate (644.9 eV), owing to the sulfur-containing atmosphere [26]. The proportions of the abovementioned compounds were 25.2%, 38.6%, and 36.2%, respectively. The presence of MnO₂ and manganese sulfate in the Mn₈/PG-S catalyst indicate that the surface high-valence Mn is reduced to Mn²⁺ by S, and then S and Mn combine to form stable MnSO₄ on the surface of Mn₈/PG-S catalysts. After splitting and fitting the orbital peak of the Mn₈/PG-S-W catalyst Mn 2p, the peak of Mn²⁺ (MnSO₄) still exists in the Mn₈/PG-S-W catalyst. However, the activity of the catalyst was significantly restored, which suggests that the main cause of catalyst deactivation may not be MnSO₄.

2.7. H₂-Temperature-Programmed Reduction (H₂-TPR)

The redox capacity of manganese-based catalysts is a key factor in the catalytic reduction of NH₃-SCR [27]. The reduction temperature can be used to evaluate the redox ability of catalysts. The low reduction temperature leads to the stronger redox ability of the catalysts [28]. The redox properties of the three catalysts were investigated by performing the H₂-TPR analysis in Figure 6 and measuring the produced H₂ concentration during the temperature increase from 50 °C to 800 °C. The profile of H₂ consumption on each catalyst, along with the temperature rising in the H₂-TPR experiment, is shown in Figure 6.

For fresh samples, there are two prominent reduction peaks (307 °C and 456 °C), as shown in the H₂-TPR analysis. The two peaks correspond to the reduction of Mn⁴⁺ and Mn³⁺, respectively [29,30]. In the case of the Mn₈/PG-S catalyst, the reduction peak at 412 °C (Mn⁴⁺) was clearly weakened, and a new reduction peak at 647 °C (Mn²⁺) appeared. During the SO₂ pretreatment process of the Mn₈/PG catalyst, Mn⁴⁺ and Mn³⁺ were reduced to Mn²⁺ by SO₂, S⁴⁺ was oxidized to S⁶⁺, and S⁶⁺ was present on the surface of catalysts in the form of SO₄²⁻. A shift in the reduction peak of the Mn₈/PG catalyst was caused by the sulfur compounds that are formed on the catalyst and that suppress the oxidizing activity of MnO_x, which is in agreement with the XPS result. In the case of the Mn₈/PG-S-W catalyst, the reduction peaks appeared at 390 °C and 501 °C. The intensity of the reduction peak of the Mn₈/PG-S-W catalyst Mn⁴⁺ could not be completely restored to that of Mn₈/PG catalyst because a part of Mn₈/PG catalyst Mn⁴⁺ was converted to Mn²⁺, which could not be reoxidized to Mn⁴⁺ by the water-washing process. The water-washing process removes the sulfur species from the sulfated catalyst, thereby restoring the redox capacity of the catalyst. This conclusion further proves that sulfur species are an important cause for catalyst deactivation.

2.8. Activity Test on Catalysts Treated by HCl

The results of atomic absorption spectroscopy showed that a small amount of Mg²⁺ exists in the water-washed solution of different catalysts. To eliminate the potential influence of sulfurization of elements from the carrier, we pretreated the PG carrier with 1.2 mol/L HCl to make sure the metal oxides in the carrier were completely removed. Then, manganese oxide was reloaded by the wetness co-impregnation method to prepare different Mn₈/PG-HCl catalysts, and the results are shown in Figure 7. Because PG contains large amounts of alkali earth metal oxides (e.g., MgO and Al₂O₃) that can contribute to a lower SCR activity, PG may be sulfurized by SO₂ to block the surface of the carrier. After pretreatment with an HCl solution, PG was considerably inactive throughout the low-temperature range, which indicates that alkali earth metal oxides, which provided activity in the carrier, were considerably removed. When MnO_x was reloaded, the activity of the Mn₈/PG-HCl catalyst was almost the same as that of Mn₈/PG, which indicates that Mn, as an active component, provided the main denitrification activity in the lower temperature range. Mn₈/PG-S-HCl and Mn₈/PG-S-W-HCl catalysts showed a trend that was similar to that which occurred after the Mn₈/PG treatment, which indicates deactivation due to the Mn₈/PG sulfurization of alkali earth metal from the carrier. However, it is debatable whether the deactivation of the catalyst was caused by the formation of sulfur compounds from the active component.

2.9. SCR Activity Test of MnSO₄-Doped PG

To evaluate the potential deactivation of the catalyst by the formation of MnSO₄, the PG carrier was doped with MnSO₄ and then tested for SCR. Figure 8 shows the NO conversion of PG; MnSO₄ (1%)/PG and MnSO₄ (8%)/PG were compared. The experimental results show that PG has a certain catalytic activity, which is attributed to the existence of oxygen-containing functional groups, which are an active component and can promote the redox process [17]. After the loading of 1% MnSO₄ (simulates SO₂-pretreated poisoning with S content during the elemental analysis), a considerable increase in the SCR activity was observed. However, for MnSO₄ (8%)/PG (assuming all active components of MnO_x are completely deactivated), a clear decrease in the SCR activity appeared. This result shows that Mn²⁺ has a certain activity, and a small amount of MnSO₄ is easily distributed on the catalyst, which is favorable for catalytic activity. When the content of MnSO₄ is 8%, the MnSO₄ bulk phase is large. Thus, an excessively high MnSO₄ content is not advantageous for the uniform dispersion on the catalyst, aggregation tends to easily occur, and the activity of PG decreases. This result further confirms that the small amount of MnSO₄ that is generated during the SO₂ pretreatment process is not the main reason for the deactivation of the catalyst.

2.10. SO₂ Adsorption and TPD

Considering that the oxidation of SO₂ is frequently induced by the adsorption of SO₂ on catalysts, the performance of SO₂ adsorption on Mn₈/PG, Mn₈/PG-S, and Mn₈/PG-S-W catalysts was measured. Each sample was exposed to an SO₂-containing flowing gas for 100 min until equilibrium. The TPD of SO₂ for each sample from 50 °C to 1000 °C was also performed. The profiles of SO₂ adsorption equilibria and SO₂ TPD for the three catalysts are compared in Figure 9.

As shown in Figure 9, SO₂ adsorption equilibrium profiles obtained at 50 °C for 100 min vary significantly. For the Mn₈/PG-S catalyst, SO₂ penetrates immediately and reaches the equilibrium. However, 6 and 12 min diffusion time was observed for the Mn₈/PG-S-W and Mn₈/PG catalysts, respectively. This implies that the Mn₈/PG-S catalyst has almost no SO₂ adsorption capacity, and the water-washing process can restore a part of its SO₂ adsorption capacity. By calculating the corresponding area of each curve, the SO₂ adsorption capacity of each catalyst can be compared more directly. The SO₂ adsorption capacities of the Mn₈/PG, Mn₈/PG-S and Mn₈/PG-S-W catalysts are 1.34 × 10⁻⁴, 4.14 × 10⁻⁵, and 6.7 × 10⁻⁵ mol·g⁻¹, respectively. Adsorbed SO₂ is oxidized to SO₃ by MnO_x during the adsorption process. Figure 6 shows that there is a difference between the SO₂ adsorption equilibrium concentration and the inlet concentration, and the difference between the SO₂ equilibrium concentration and the inlet concentration (400 ppm) is believed to be attributed to SO₂ oxidizing to SO₃. The SO₂ oxidation rates of the Mn₈/PG, Mn₈/PG-S, and Mn₈/PG-S-W catalysts are 6.25%, 1.25%, and 2%, respectively. SO₂ is continuously adsorbed and oxidized by the active component. A part of SO₂ forms MnSO₄ with the active component, and the other part forms a removable sulfur compound. These removable sulfur compounds are responsible for the decreased SO₂ adsorption and oxidation capacity. Removable sulfur compounds that cover the active component can be removed by the water-washing process, thereby recovering the adsorption capacity of the catalyst. Because the water-washing process does not wash away MnSO₄, the oxidizing capability is not significantly recovered. Most of the soluble sulfur compounds can be removed by the water-washing process; thus, the adsorption capacity of the catalyst is recovered. Because MnSO₄ is still present after the water-washing process (the presence of SO₄²⁻ may inhibit the adsorption of SO₂), the adsorption and oxidation capacities of the catalyst are not fully recovered.

For direct comparison with SO₂ adsorption, the SO₂-TPD profiles of each catalyst are shown in Figure 9. The figure shows that there is no SO₂ desorption peak for the Mn₈/PG catalyst. A considerable SO₂ desorption peak was obtained on the Mn₈/PG-S catalyst in the temperature range of 600 °C–800 °C. This may be attributed to the joint desorption action of a soluble sulfur compound and MnSO₄. The desorption peak of the water-washed regenerated catalyst was observed at 700 °C–800 °C, and the peak was weaker than that of the sulfur Mn₈/PG-S sample. This desorption peak may be attributed to the decomposition of MnSO₄, which remained after water washing. By comparing the SO₂ desorption peak areas of the Mn₈/PG-S and Mn₈/PG-S-W catalysts, it can be observed that the soluble sulfur compound is the main product of SO₂, which is also the main reason for the deactivation of the Mn₈/PG-S catalyst.

2.11. NH₃ Adsorption and TPD

It is widely believed that the adsorption and activation of NH₃ on the catalyst surface is the first important step in the SCR reaction, according to either E–R or L–H mechanism [31]. In general, an increase in the acidity of the catalyst surface is favorable for the adsorption of NH₃ and SCR activity. The NH₃ adsorption ability of each sample was measured, followed by the TPD experiments. During the adsorption, the catalyst was exposed to a gas mixture containing 600 ppm of NH₃ balanced by N₂ at 50 °C for more than 100 min until equilibrium. Then, after a sweeping with N₂, TPD profiles were also recorded, as shown in Figure 10.

From the comparison of the NH₃ adsorption equilibrium curves for the three catalysts, it was observed that the diffusion time of the Mn₈/PG-S sample was 4 min faster than that of the fresh sample. The area enclosed by the NH₃ curve of the Mn₈/PG-S catalyst was significantly smaller than that of the fresh sample during the process of adsorption equilibrium, which indicates that the NH₃ adsorption on the Mn₈/PG-S catalyst was considerably inhibited. However, the diffusion time and the area enclosed by the NH₃ curve of the Mn₈/PG-S-W catalyst were found to be slightly greater than those of fresh samples, which indicates that the NH₃ adsorption of the catalyst was efficiently recovered during the removal of sulfate species from the poisoned catalyst.

Figure 10 shows the NH₃ TPD curve after 125 min of adsorption. There are several clear desorption peaks of NH₃ in the range of 50–900 °C: a weak acidic site desorption peak at 132–380 °C, a moderate acidic site desorption peak at 380–640 °C, and a strong acidic site desorption peak at 640–775 °C. The desorption peak of the Mn₈/PG-S catalyst is lower than that of the Mn₈/PG catalyst due to the reduction of stored NH₃ during the adsorption. However, the temperature corresponding to the desorption peak did not change, which indicates that the binding position of NH₃ on the catalyst did not change. The amount of desorption of NH₃ on the water-washed regeneration catalyst was slightly higher than that of the Mn₈/PG catalyst, which indicates that the water-washing process can remove the SO₄²⁻ polymer, which inhibits the adsorption of NH₃.

The NH₃ adsorption-desorption curve confirms that the formation of ammonium sulfates is not the main cause of the decrease in catalytic activity. The deactivation of the catalyst occurred before the possible formation of ammonium sulfates. Based on the result that the NH₃ adsorption of the catalyst was considerably inhibited by the SO₂ pretreatment of the catalyst, it can be speculated that the formation of an SO₄²⁻ polymer caused the pore blockage of the catalyst and the occupation of active sites for NH₃ adsorption.

2.12. NO Adsorption and TPD

Currently, some researchers believe that the SCR reaction follows the L–H mechanism, and the NO adsorption behavior in the L–H mechanism considerably affects the SCR reaction [32,33]. To investigate the effect of soluble sulfur compounds (SO₄²⁻ polymer) on the adsorption and desorption of NO on the catalyst, we performed the adsorption and desorption analyses of NO on the three catalysts. During the absorption period, the catalyst was placed in an atmosphere containing 600 ppm of NO with N₂ as the carrier gas at 50 °C, and the outlet NO concentration is shown in Figure 11.

The adsorption equilibrium curves for the three catalysts at 50 °C before 80 min are shown in Figure 11. The diffusion time of the Mn₈/PG-S catalyst by NO was 50 min faster than that of the Mn₈/PG catalyst. The enclosed area of the Mn₈/PG-S catalyst was significantly smaller than that of the Mn₈/PG catalyst during the adsorption period, which indicates that the capacity of stored NO was considerably affected by the generated sulfur compounds. The adsorption equilibrium time of the Mn₈/PG-S-W catalyst is approximately the same as that of the Mn₈/PG catalyst. However, the amount of adsorbed NO is slightly lower than that of the Mn₈/PG catalyst, which may be due to the inhibition of NO adsorption by the residual SO₄²⁻.

There are two distinct desorption peaks for fresh Mn₈/PG. The first desorption peak appears at 250–320 °C. The desorption peak is attributed to the decomposition of the NO adsorption product, which is mainly involved in SCR. It may be formed by the decomposition of bridging nitrate salts [34]. The second peak belongs to the high-temperature peak at approximately 400 °C. It may be formed by the oxidation of NO to form nitrite salts and nitrate salts, which are decomposed and desorbed at high temperatures [35]. Because the decomposition temperature of this nitrate is considerably high, it cannot easily participate in the SCR reaction at low temperatures. When the catalyst is reacted in a sulfur-containing atmosphere, the low- and high-temperature peaks are both attenuated. The NO storage is reduced during the adsorption period due to the presence of sulfur compounds. The water-washed regeneration process can remove the SO₄²⁻ polymer from the catalyst surface, but MnSO₄ still exists. The presence of MnSO₄ inhibits the activation of NO; thus, the peak intensity does not recover at 400 °C.

2.13. Temperature-Programmed Surface Reaction (TPSR)

To explore the effect of sulfuric acid and sulfurous acid on the surface environment of the catalyst, a further investigation was conducted using the TPSR, as shown in Figure 12.

As a result, the concentration of NO was quickly increased to an inlet concentration of 570 ppm at 4 min for the Mn₈/PG catalyst without NH₃ adsorption. When the temperature of the system increased above 230 °C, the NO concentration gradually decreased to a low concentration. After 80 min, the NO concentration did not reach the inlet concentration due to the consumption of the NO heat reaction under high temperature conditions.

For the Mn₈/PG catalyst with NH₃ adsorption, there is a clear inverted peak in the TPSR curve. The larger NO depletion peak at 20–80 min is attributed to the consumption of the NO surface reaction with adsorbed NH₃. The TPSR curve of the Mn₈/PG-S catalyst also shows a distinct consumption peak (1.94 × 10⁻⁴ mol·g⁻¹), but the depletion peak attributed to the adsorbed NH₃ reaction at the same time period was lower than the Mn₈/PG consumption peak (2.26 × 10⁻⁴ mol·g⁻¹). The presence of soluble sulfur compounds inhibits the adsorption of NH₃ at certain specific acidic sites, which are the most critical active sites during low-temperature SCR reactions. The TPSR curve for the Mn₈/PG-S-W catalyst shows the same consumption peak (2.69 × 10⁻⁴ mol·g⁻¹) as that of the Mn₈/PG catalyst in the same time period because the water-washing process can remove the SO₄²⁻ polymer at the key active sites in the low-temperature SCR reaction. Meanwhile, it was observed that the amount of activated adsorption of NH₃ by the Mn₈/PG-S-W catalyst is slightly higher than that of the Mn₈/PG catalyst, which may be an important reason why the activity of the Mn₈/PG-S-W catalyst is slightly higher than that of the Mn₈/PG catalyst.

2.14. Proposed SO₂-Pretreated Poisoning and Water-Washed Regeneration Mechanism Model

To better illustrate and summarize the abovementioned characterization analysis, in Figure 13, we propose a mechanism model for SO₂-pretreated poisoning and water-washed regeneration.

The process of SO₂-pretreated poisoning of the catalyst is as follows. First, SO₂ adsorbs on the active component, MnO_x, of the Mn₈/PG catalyst. Then, SO₂ is oxidized by the active component to SO₃ and adsorbed on the surface of the active component. When the concentration of SO₄²⁻ on the active site reaches a certain level, sulfate is converted into an SO₄²⁻ polymer. Thus, the specific surface area of the catalyst is decreased, which corresponds to a reduction in the adsorption and activation of NH₃ and NO.

The mechanism of water-washed regeneration is as follows. The water-washing process can easily wash away the SO₄²⁻ polymer from the active component. Thus, the catalytic activity of the active components can be restored.

3. Experimental

3.1. Materials

The palygorskite, supplied by Xinzhou New Material Factory (Anhui, China), was ground and sieved through 0.42–0.84 mm mesh. The chemical composition of the palygorskite was 69.44 wt.% SiO₂, 11.84 wt.% Al₂O₃, 5.14 wt.% Fe₂O₃, 12.16 wt.% MgO, 0.43 wt.% TiO₂, 0.5 wt.% K₂O, 0.21 wt.% CaO, and 0.28 wt.% others. The surface area, pore volume, and pore size of palygorskite were 228.5 m²·g⁻¹, 0.41 cc·g⁻¹, 2.8 nm, respectively. All used chemicals, including manganese nitrate, manganese sulfate and hydrochloric acid, were analytical grade and purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, SCR, China) Simulated flue gas provided by Nanjing Shangyuan Industrial Gas Plant, (Nanjing, China). MnO₂ is produced by calcination of manganese nitrate.

3.2. Catalyst Preparation

Fresh Mn₈/PG catalyst: The catalysts were prepared by the wetness co-impregnation loading of active species. Palygorskite, as a support, was pore-volume impregnated by a mixed aqueous solution of manganese nitrate as a precursor. The mass ratios of manganese to palygorskite were 8 wt%. The catalysts were initially dried at 50 °C for 6 h, followed by drying at 110 °C for 12 h. The dried catalysts were ground and sieved through 20–40 mm meshes and subsequently calcined in air at 300 °C for 3 h. Thus, the obtained catalysts were designated as Mn₈/PG, where “8” refers to the mass percentage of each element.

SO₂-pretreated catalyst: A total of 2-g fresh Mn₈/PG catalyst was continuously ventilated under the following reaction conditions: 400 ppm SO₂, 3 vol% O₂, N₂ as an equilibrium gas, total flow rate of 400 mL·min⁻¹, and reaction temperature of 250 °C for 5 h until SO₂ adsorption was saturated. The sample was labeled as Mn₈/PG-S.

Water-washed catalyst: A total of 2-g Mn₈/PG-S catalyst was added to an Erlenmeyer flask containing 50-mL deionized water; the mixture was placed in an SHA-B-type water bath, stirred at a constant temperature (25 °C) for 10 min, and then dried at 105 °C to obtain a water-washed catalyst. The sample was labeled as Mn₈/PG-S-W.

Multi-cycles of SO₂ pretreatment followed by water washing of the catalyst were also carried out to evaluate the potential weight loss of manganese species. Concrete steps are as follows: first, in the absence of NH₃, SO₂ was piped in, the catalyst was presulfurized, again according to the process of the catalyst presulfurization water experiment, and dried after a bath. The catalyst was then washed in SO₂, secondary vulcanization processing took place, followed by water processing, the cycle. The samples were labeled as Mn₈/PG-W-x, where “x” means the number of cycles.

MnSO₄ (x)/PG catalyst: MnSO₄ was impregnated with the equal volume method. Firstly, 10 g PG and a certain amount of MnSO₄ were weighed, and a certain amount of deionized water was added to MnSO₄ to be stirred and dissolved. Then, manganese sulfate solution was poured into PG, stirred for 15 min, and allowed to stand for 24 h to obtain the MnSO₄ (X)/PG catalyst. The mass of manganese sulfate was calculated according to the load ratio, X.

3.3. Catalytic Activity Test

Catalytic performance of the catalysts for NH₃-SCR was evaluated in a fixed-bed quartz tube (with an inner diameter of 15 mm and a height of 80 mm) continuous-flow reactor containing 3 mL of the catalysts. The evaluation device consists of three parts: simulated flue-gas region, fixed-bed reactor, and analysis region. The device composition is shown in Figure 14. The typical feed-gas composition was as follows: 600 ppm of NO, 600 ppm of NH₃, 400 ppm of SO₂, 3% O₂, and the balance of N₂. The concentrations of NO and SO₂ in the inlet and outlet streams of the reactor were monitored online by an NO analyzer (Testo 350XL, Testo SE & Co. KGaA, Lenzkirch, Germany) and an SO₂ (MRU-OPTIMA7) analyzer (Testo SE & Co. KGaA, Germany), respectively. All data were collected after 40 min when the SCR reaction reached a steady state.

The denitrification performance of catalysts was based on the NO conversion (X_NO). The calculation formula is as follows:

$${\mathrm{X}}_{{\mathrm{NO}}_{\mathrm{x}}}=\frac{{[{\mathrm{NO}}_{\mathrm{x}}]}_{\mathrm{in}}-{[{\mathrm{NO}}_{\mathrm{x}}]}_{\mathrm{out}}}{{[{\mathrm{NO}}_{\mathrm{x}}]}_{\mathrm{out}}}$$

where [NO_x]_in is the inlet concentration of NO and [NO_x]in is the outlet concentration of NO.

3.4. Catalyst Characterization

The surface area of the catalyst was measured using a Quantachrome Nove (2200 ev) analyzer (Anton Paar Co. Ltd, Graz, Austria). The microscopic surface composition of the catalysts was analyzed using a new Hitachi SU8020 high-resolution field-emission scanning electron microscopy (SEM) instrument (Carl Zeiss AG, Oberkochen, Germany). The powder XRD measurement was performed on a Rigaku D/MAx2500V system (PANalytical B.V., Almelo, Netherlands) with a Cu Kα radiation. The XPS analysis was performed on an ESCALAB 250-type electron energy spectrometer (Thermo Fisher Scientific, Waltham, America) with Al Kα, h_ν = 1486.6 eV and a power of 150 W. The X-ray fluorescence (XRF) spectroscopic analysis was performed on an XRF-1800 X-ray fluorescence spectrometer (Shimadzu, Kyoto, Janpan), Rh target, a maximum power of 4 kW, with a minimum area of 0.5 mm.

The NH₃ isothermal adsorption and temperature-programmed desorption measurement of ammonia (NH₃-TPD) catalysts was performed in fixed quartz reaction tubes. A total of 0.5-g catalysts was loaded into a quartz tube, pretreated at 110 °C for 3 h in N₂ flow (100 mL·min⁻¹), and cooled to 50 °C before the measurement. The adsorption of NH₃ was performed at 50 °C. After adsorption saturation, the catalysts were heated at a rate of 5 °C·min⁻¹, from 50 °C to 900 °C under a constant N₂ flow of 200 mL·min⁻¹. The desorbed amount of NH₃ was monitored online by mass spectrometry (Hiden QIC-200, Beijing Enghead Analysis Technology Co., Ltd, Beijing, China). SO₂ and NO isothermal adsorption and TPD experiments were consistent with NH₃-TPD experiments.

Hydrogen temperature-programmed reduction (H₂-TPR) experiments were performed using a homemade H₂-TPR device (Tianjin Xianquan Instrument Co. Ltd, Tianjin, China). A total of 50.0-mg catalyst with a particle diameter of 0.38–0.83 mm was placed into the device, a 5% H₂–Ar gas mixture was passed, and the gas-flow rate was set at 20 mL·min⁻¹. The baseline was purged at 50 °C until the baseline stabilized and switched to the analytical state. Then, the temperature was increased from 50 °C to 850 °C at a ramping rate of 10 °C·min⁻¹. Hydrogen concentration was measured using a thermal conductivity cell (TCD detector, Tianjin Xianquan Instrument Co. Ltd, Tianjin, China) to produce a TPR spectrum. The detection temperature was 80 °C, and the bridge flow was 60 mA.

Temperature-programmed surface reactions (TPSR) were carried out to study the reactivity of catalyst-pre-adsorbed ammonia with gaseous NO. The catalysts (0.5 g) were adsorbed with NH₃ at 50 °C, then reacted with simulated flue gas containing 600 ppm NO/Ar and 3% O₂ from 50 °C to 300 °C at a rate of 5 °C·min⁻¹, and the total gas-flow rate was 350 mL·min⁻¹. Finally, the gas concentrations of NO were measured by a flue-gas analyzer (Testo350-XL, Testo SE & Co. KGaA, Germany).

4. Conclusions

In this study, the deactivation of MnO_x/PG catalysts by SO₂ for low-temperature SCR was investigated. The catalyst was pretreated in an SO₂-containing ammonia-free gas to avoid a reaction between SO₂ and NH₃, which helps explore the reasons for the deactivation of the catalyst by SO₂. Interestingly, the catalyst was determined to be considerably deactivated only by SO₂ and can be easily recovered by simple washing with water. The results suggest that the poisoning species is neither ammonia sulfates nor manganese sulfates but something similar to polysulfuric acid that coats active manganese oxides. The effects of Mn₈/PG SO₂-pretreated poisoning and water-washed regeneration on the SCR reaction were studied. In addition, fresh SO₂-pretreated poisoning and water-washed regeneration catalysts were characterized in detail. A possible SO₂-poisoning and regeneration mechanism was proposed. The obtained conclusions are as follows.

• The stratification of the ammonium sulfate salt on the surface and the sulfation of the active component (MnSO₄ formed) is not the main cause of catalyst deactivation. The main reason for the catalyst deactivation caused by SO₂-pretreated poisoning is the SO₄²⁻ polymer. This polymer coats the active component and affects the adsorption and activation of NH₃ and NO.
• The recovery mechanism of water-washed regeneration is as follows. The SO₄²⁻ polymer that coats the active component can be removed by the water-washing process to restore the catalytic activity of the active component; the water-washing process does not cause a loss of the active component.
• MnSO₄ produced during vulcanization cannot be removed by washing. Therefore, MnSO₄ is not the main cause of catalyst inactivation. The loading of a small amount of MnSO₄ on the PG does not result in a decrease in the catalytic activity. Instead, the abovementioned process can increase the denitrification activity of PG in the low-temperature range.