Effect of Potassium on the Performance of a CuSO₄/TiO₂ Catalyst Used in the Selective Catalytic Reduction of NO_x by NH₃

1.   Introduction

Nitrogen oxides (NO_x, main NO and NO₂), which come from the stationary sources like coal-fired and biofuel-fired power plant, can cause the formation of haze and photochemical smog ^{1, 2}. Therefore, various methods have been developed to abate NO_x emission from stationary sources. Among them, selective catalytic reduction of NO_x by NH₃ (NH₃-SCR) is the most effective technology to abate NO_x and has been successfully applied in coal-fired power plants ^{3, 4}. The catalyst is the key in NH₃-SCR method. Vanadium-based materials which was active in the temperature range from 300 to 400 ℃ and exhibited high stability in the presence of SO₂, have been the main commercial SCR catalysts and are widely used around the world ^{5, 6}. However, due to the high bio-toxicity of vanadium and facile oxidation of SO₂ to SO₃, a lot of vanadium-free SCR catalysts have been researched ^{7, 8}. Metal oxides catalysts (MnO_x, FeO_x, etc.) have drawn much attention of researchers, because they could exhibit excellent activity in NH₃-SCR reaction under temperature below 300 ℃ ^{4, 9}. Nevertheless, SO₂ usually exists in the flue gas and the resistance to SO₂ of these metal oxides catalysts is very poor. Owing to the poison of active sites by SO₂, metal oxides catalysts are easy to lose activity when SO₂ exists ¹⁰⁻¹². SCR catalysts with excellent resistance to SO₂ are highly desirable.

One of the strategies to obtain catalysts with high SO₂ resistance is using materials which were stable when SO₂ exists and could not be sulfated by SO₂ ¹³. Consequently, metal sulfates have drawn attention of researchers. Supported CuSO₄, Fe₂(SO₄)₃, MnSO₄ and Ce(SO₄)₂ catalysts were investigated in NH₃-SCR, and the result indicated that they was active in the reaction ¹⁴. Among them, due to its abundant acid sites and surface absorbed oxygen species, CuSO₄-based catalysts were more active than other metal sulfate catalysts in NH₃-SCR reaction ^{15, 16}. NO_x conversion of CuSO₄/TiO₂ catalysts was higher than 90% under 280–420 ℃, similar to that over commercial vanadium-based catalysts. Significantly, CuSO₄/TiO₂ catalysts performed an excellent SO₂ tolerance. NO_x conversion over CuSO₄/TiO₂ catalysts was around 95% when 0.15% SO₂ existed for 48 h at 350 ℃ ¹⁶. Thus, CuSO₄/TiO₂ catalyst was considered to be a promising candidate vanadium-free catalyst used in NH₃-SCR.

Due to renewable and widely available, biomass was considered to be an environmental-friendly fuel for power plants. Noticeably, the application of biomass was thought to be one of effective methods to achieve carbon emission peak ¹⁷. As a result, more and more biofuel-fired power plants have been built in the world to reduce carbon emission. However, there are much higher alkali metals especially potassium (K) element existed in the biomass, which is a major difference between conventional fossil fuel and biomass, so higher K content could be found in the flue gas from biomass-fired power ¹⁸. K would react with acid sites on commercial vanadium-based catalysts and reduced surface chemisorbed oxygen, then caused the deactivation of the catalyst ¹⁹. Chen et al. compared the influence of K, Na, Ca and Mg on V₂O₅-WO₃/TiO₂ (VWTi) catalysts. The adverse effect of K was the highest among four elements. Only 0.5% (w, mass fraction) K doping would cause the highest NO_x conversion (350 ℃) decreased from 100% to 80% over VWTi catalysts ²⁰. The decrease of Brønsted acid sites and surface chemisorbed oxygen were the main reasons for the deactivation of the catalyst ²⁰. CeMoTiO_x catalysts used in NH₃-SCR under low temperature also could be poisoned by K. K would suppress the oxidation of NO and NO₂, inhibiting the "fast NH₃-SCR" under low temperature. Thus, NO_x conversion (225 ℃) decreased from 97% to 70% on 0.3% (w) K doping CeMoTiO_x catalysts ²¹. Up to now, the studies focused on the effect of K on metal sulfate catalysts were still absent, which limited the further application of the catalyst.

In order to promote the application of CuSO₄/TiO₂ catalysts in NH₃-SCR, the effect of K on CuSO₄/TiO₂ catalysts was investigated and compared with commercial VWTi catalysts. Samples before and after K doping were characterized by N₂ adsorption-desorption, X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), temperature programmed desorption of NH₃ (NH₃-TPD), temperature programmed reduction of H₂ (H₂-TPR), in situ diffuse reflectance infrared Fourier transform spectroscopy (in situ DRIFTS). Based on the result of characterization, the deactivation mechanism of K over CuSO₄/TiO₂ catalysts was proposed.

2.   Experimental section

2.1   Catalyst preparation

CuSO₄/TiO₂ catalyst (CuK-0) with CuSO₄ loading as 10% (w) was first prepared. Dissolve 0.25 g CuSO₄∙5H₂O (AR, Aladdin, China) in 30 mL deionized water and then 1.8 g TiO₂ (T299213-100g, Aladdin, China) was added into the solution, stirred for 30 min and then dried at 90 ℃ for 12 h. Finally, the obtained sample was calcined at 500 ℃ in the presence of air for 3 h (named as CuK-0). K doping catalyst was also prepared by a wet impregnation method. CH₃COOK (AR, Aladdin, China) was dissolved in 25 mL ethanol and then corresponding CuSO₄/TiO₂ catalysts were added, stirred for 30 min and then dried at 60 ℃ for 12 h. Then the sample was calcined at 500 ℃ for 3 h. K amount of the prepared sample was 1%, 2% and 4% (w), named as CuK-1, CuK-2 and CuK-4, respectively.

According to the result of X-Ray Fluorescence Spectrometer, commercial VWTi catalyst contained 1.4% (w) V₂O₅, 5.7% (w) WO₃, 90.1% (w) TiO₂ and other substances (Al, S and so on). The K doping amount was 1.0% (w) on the catalyst (named as VWTi-K) and the prepared method was the same as K doping CuSO₄/TiO₂ sample.

2.2   Activity test

NO_x conversion and N₂ selectivity were measured at a fixed-bed reactor (Φ 8 mm × 600 mm). 0.8 mL sample (40–60 mush) was used and the total mixture gas was set at 800 mL∙min⁻¹ and gas hourly space velocity (GHSV) was 60000 h⁻¹. The mixture gas contained 0.05% (volume fraction) NO, 0.05% NH₃, 4% O₂, 4% H₂O, and was balanced with N₂. The concentration of NO, NO₂, N₂O and NH₃ in the inlet and outlet of the reactor was measured by an Antaris IGS flue gas analyzer (Thermo Scientific, USA). NO_x conversion (x) and N₂ selectivity (y) were obtained by the equation below:

2.3   Catalyst characterizations

SSA-6000 physical adsorption analyzer (Beijing Builder Electronic Technology, China) was used to collect N₂ adsorption-desorption isotherm of each sample at 77 K. Brunauer-Emmett-Teller (BET) method was used to calculate specific surface area. XRD was recorded at an X'Pert Pro XRD diffractometer (Panalytical, The Netherlands) with Cu K_α radiation under 40 kV and 40 mA. XPS was collected on an ESCALAB 250 spectrometer (Thermo Fisher Scientific, USA).

H₂-TPR was tested on a PCA 1200 chemisorption analyzer (Beijing Builder Electronic Technology, China). In order to remove H₂O and other gas impurities on the sample, the sample (100 mg) was first treated under a helium flow at 400 ℃ for 60 min. Then the temperature was decreased to 50 ℃ and switching to 5.0% H₂/Ar. After the baseline of TCD was stable, the temperature was heated to 800 ℃ at 10 ℃∙min⁻¹.

NH₃-TPD was performed on a ChemBET-3000 TPR-TPD chemisorption analyzer (Quantachrome, USA). The sample (100 mg) was treated under a helium flow at 400 ℃ for 60 min. Then the temperature was decreased to 50 ℃ and switching into 5% NH₃/Ar for 30 min. To remove any NH₃ weakly adsorbed on the surface, the sample was purged by helium at 100 ℃ for 60 min. Finally, the temperature was heated to 600 ℃ at 10 ℃∙min⁻¹. In order to avoid the influence of H₂O, the signal of NH₃ (m/e = 16) was recorded on an online mass spectrum (DYCORLC-D100, Ametek Company, USA).

In situ DRIFTS were measured on a Tensor 37 FTIR (Bruker, German). Catalyst powder was treated under N₂ at 400 ℃ for 60 min and then decreased to 300 ℃. The background spectra were recorded at 300 ℃. Finally, mixture gas was inlet into the cell. The spectrum was recorded under 4 cm⁻¹ and 100 scans.

3.   Results and discussion

3.1   NH₃-SCR performance

Catalytic activity of each sample can be found in Fig. 1a. Fresh CuSO₄/TiO₂ catalyst (CuK-0) showed an excellent activity in NH₃-SCR reaction. More than 90% NO_x was removed on the catalyst in 300–400 ℃. After doping with 1.0% (w) K, NO_x conversion of CuK-1 sample in the temperature range from 300 to 400 ℃ was close to that of fresh CuSO₄/TiO₂ catalyst. Nevertheless, catalytic activity below 300 ℃ and above 400 ℃ decreased to some degree. With the increase of K content in the catalyst, it could be found that the activity of CuSO₄/TiO₂ catalyst decreased largely. When the K content was raised to 4.0% (w), the highest of NO_x conversion for CuK-4 sample was only 70.3% (at 350 ℃) and the temperature window obviously narrowed, indicating that K in the fuel gas could cause the deactivation of CuSO₄/TiO₂ catalyst. The effect of K on activity of commercial VWTi catalyst can be found in Fig. 1b. Commercial VWTi catalyst also performed a high activity in NH₃-SCR reaction and NO_x conversion was also higher than 90% in 300–400 ℃. However, 1.0% (w) K caused an obvious decrease of NO_x conversion for commercial VWTi catalyst. The highest NO_x conversion over VWTi-K sample in the test temperature range decreased to 75.1% (350 ℃). K could lead to the deactivation of commercial VWTi catalyst, in accordance with other researchers ^{20, 22}. Compared with CuK-1 sample, it could be found that the adverse effect of 1.0% (w) K doping on VWTi-K sample was more serious, indicating that CuSO₄/TiO₂ catalyst owned a better resistance to K poisoning than commercial VWTi catalyst. CuSO₄/TiO₂ catalyst should be a potential SCR catalyst used in the flue gas containing high concentration of K.

Figure 1

Figure 1.  NO_x conversion (a, b) and N₂ selectivity (c, d) of catalyst sample before and after doping with K.

Test conditions: 0.05% (volume fraction) NO, 0.05% NH₃, 4% O₂, 4% H₂O, and balanced with N₂; GHSV = 60000 h⁻¹.

下载: 全尺寸图片 幻灯片

N₂O is the main byproduct of NH₃-SCR reaction and it is a main greenhouse gas with a 310-time greenhouse effect of CO₂ ^{23, 24}. Therefore, the effect of K on N₂ selectivity of CuSO₄/TiO₂ catalyst and commercial VWTi catalyst was investigated, as shown in Fig. 1c and 1d. Almost no N₂O could be detected in the temperature below 325 ℃ for all samples, indicating that N₂O was difficult to be formed under low temperature. N₂O would be produced under temperature higher than 325 ℃ and N₂ selectivity of CuSO₄/TiO₂ catalyst decreased gradually with temperature increasing. It could be found that N₂ selectivity of the catalyst decreased to 92.1% at 450 ℃. After K was doped on CuSO₄/TiO₂ catalyst, N₂ selectivity of the poisoned catalyst decreased largely, indicating that K would benefit the formation of N₂O over CuSO₄/TiO₂ catalyst. Commercial VWTi catalyst also exhibited the similar trend of N₂ selectivity to CuSO₄/TiO₂ catalyst. However, it could be found that the influence of K on N₂ selectivity of commercial VWTi catalyst was not the same as that of CuSO₄/TiO₂ catalyst. K obviously inhibited the formation of N₂O over commercial VWTi catalysts, which was consistent with other researches ^{17, 25}. The adverse effect of K on N₂ selectivity of CuSO₄/TiO₂ catalyst suggested that the deactivation mechanism on CuSO₄/TiO₂ and commercial VWTi catalyst might be different.

3.2   Structural and chemical properties

High specific surface area can provide more active sites and thus benefit activity of the catalyst. Specific surface area (S_BET), pore volume (V_p) and average pore diameter (D_p) of each sample were measured by N₂ adsorption-desorption isotherm and analyzed by BET-BJH method. As shown in Table 1, S_BET of CuSO₄/TiO₂ catalyst was 57.4 m²·g⁻¹. After doping with 1.0% (w) K, it could be found that S_BET, V_p and D_p did not have an obviously change, suggesting that 1.0 % (w) K would not have a significant effect on the pore structure. Nevertheless, it could be found 1.0% (w) K caused an obviously change of pore structure over commercial VWTi catalyst. After doping with 1.0% (w) K, S_BET and V_p of commercial VWTi catalyst decreased to 52.1 m²∙g⁻¹ and 0.242 cm³∙g⁻¹, with D_p increased to 9.29 nm, indicating CuSO₄/TiO₂ catalyst had a better resistance to the negative effect of K on pore structure than commercial VWTi catalyst. With the increase of K content, S_BET of CuSO₄/TiO₂ catalyst decreased obviously. When K content increased to 4.0% (w), S_BET and V_p decreased to 46.7 m²∙g⁻¹ and 0.245 cm³∙g⁻¹, about 81% and 93% of fresh CuSO₄/TiO₂ catalyst, respectively. Meanwhile, D_p increased from 9.12 to 10.49 nm, implying that high concentration of K would destroy partial pore structure of CuSO₄/TiO₂ catalyst. The harmful effect of K on pore structure should be one of the main reasons for the deactivation of SCR catalysts.

Table 1

Table 1.  Physical information and total acid sites of catalyst sample.

下载: 导出CSV

Sample	SBET/(m2∙g−1)	Vp/(cm3∙g−1)	Dp/nm	Tacid/(μmol∙g−1)	Oβ/Ot (%)
CuK-0	57.4	0.262	9.12	3673	43.6
CuK-1	56.1	0.259	9.23	2378	36.1
CuK-2	54.6	0.257	9.41	1286	31.2
CuK-4	46.7	0.245	10.49	679	22.9
VWTi	55.5	0.249	8.97	925	19.4
VWTi-K	52.1	0.242	9.29	222	15.1
SBET: specific surface area, Vp: pore volume, Dp: average pore diameter, Tacid : total amount of acid sites. Oβ/Ot: the content of surface adsorbed oxygen in total oxygen species.

XRD pattern of all samples can be found in Fig. 2. Several peaks were detected for CuSO₄/TiO₂ catalyst. Peaks at 2θ = 25.3°, 36.9°, 37.8°, 38.6°, 48.0°, 53.9°, 55.0°, 62.7°, 68.8°, 70.3° and 75.1° were the typical peaks of anatase TiO₂ (PDF number: 01-071-1166). In addition, two small peaks at 2θ = 21.2° and 25.1° which could be associated with CuSO₄ (PDF number: 01-072-1248) were also detected. After K doping, the intensity of the peak assigned to anatase TiO₂ changed little, indicating that K would not have a large influence on TiO₂ carrier. However, it could be found the intensity of peaks associated with CuSO₄ decreased with K content increasing, implying that K would significantly affect CuSO₄ of the catalyst. Besides the peaks associated with TiO₂ and CuSO₄, two peaks at 2θ = 29.8° and 30.9° assigned to K₂SO₄ (PDF number: 00-001-0944) were detected on CuK-4 sample. K₂SO₄ might come from the reaction of K₂O and CuSO₄. K₂O could react with CuSO₄ and thus formed K₂SO₄ and CuO. Nevertheless, there was no peak assigned to CuO could be detected on CuK-4 sample, implying that CuO should be amorphous or the amount of CuO was lower than the detection limit of XRD. The formation of K₂SO₄ and CuO might block up the pore structure of materials, and thus caused the obvious reduction of specific surface area. Commercial VWTi catalyst only showed the typical peaks of antase TiO₂. Due to the high dispersity of V₂O₅ and WO₃ on the carrier, and no peaks assigned to V₂O₅ or WO₃ could be detected, in accordance with other researches ²⁶. Moreover, the doping of 1.0% (w) K did not have a significant effect on the XRD pattern of the catalyst, in accordance with other researches ^{17, 22}.

Figure 2

Figure 2.  XRD patterns of catalyst sample.

下载: 全尺寸图片 幻灯片

Previous researches have affirmed that O species played a vital role in NH₃-SCR reaction ^{8, 27}. XPS was used to detect O species of each sample. As shown in the Fig. 3a, two fitting peaks at 529.5 and 531.4 eV were found on fresh CuSO₄/TiO₂ catalyst. The former peak could be associated with lattice oxygen (O_α) and the latter peak could be assigned to surface adsorbed oxygen (O_β) ²⁸. Based on the integral area of the corresponding peak, O_β was 43.6% of total O species for the catalyst. Both O_α and O_β were also found on CuSO₄/TiO₂ catalyst poisoned by K. However, the content of O_β obviously decreased with the increase of K content. After K content increased to 4.0% (w), O_β of the catalyst decreased to 22.9%, about 53% of that for fresh catalyst, indicating that K could lead to the decrease of surface adsorbed oxygen over CuSO₄/TiO₂ catalyst. The O 1s spectrum of commercial VWTi catalyst also could be fitted into two peaks assigned to O_α and O_β, respectively. The doping of K also caused the decrease of O_β content (from 19.4% to 15.1%) over commercial VWTi catalyst. Thus, the decrease of surface adsorbed oxygen caused by K should be one of the major factors for deactivation of CuSO₄/TiO₂ and commercial VWTi catalyst. Moreover, higher content of surface adsorbed oxygen should be a main factor for the higher resistance to K poisoning of CuSO₄/TiO₂ catalyst.

Figure 3

Figure 3.  XPS result of O 1s (a), Cu 2p_3/2 (b), S 2p (c) and Ti 2p_3/2 (d) for each catalyst sample.

下载: 全尺寸图片 幻灯片

Then, XPS was also used to detect Cu species on fresh and K-poisoned CuSO₄/TiO₂ catalyst and the result could be found in Fig. 3b. Only one fitting peak at 934.2 eV associated with Cu²⁺ was found in CuK-0 sample, indicating that all Cu species should exist as Cu²⁺ on CuSO₄/TiO₂ catalyst ¹⁶. After doping of 1.0% (w) K, the position of the peak was still at 934.2 eV. However, when K content increased to 2.0% (w), the peak of Cu²⁺ moved to lower binding energy (934.0 eV). The Cu peak of CuK-4 sample further moved to 933.8 eV. As found by previous researchers ¹⁶, the position of the peak assigned to Cu²⁺ in CuO was lower than that in CuSO₄. Thus, the movement of the peak assigned to Cu²⁺ indicated that CuO should be formed on the K-poisoned CuSO₄/TiO₂ catalyst, consistent with the result of XRD. S species on the catalyst were also detected and the result can be found in Fig. 3c. Only one peak at 168.5 eV could be found on CuSO₄/TiO₂ catalysts before and after K doping. The peak could be assigned to S⁶⁺ in SO₄^{2− 16}, indicating that K should also exist as metal sulfate (K₂SO₄). Ti spectra of all samples could be found in Fig. 3d. It could be found that K doping would not have an obvious effect on Ti species on CuSO₄/TiO₂ catalyst, implying that K could not react with the carrier of the catalyst, was consistent with the result of XRD. XPS result further indicated that K doping would cause the formation of CuO and K₂SO₄ on CuSO₄/TiO₂ catalyst.

3.3   Surface reducibility and acidity

Surface reducibility is an important factor to determine the activity of SCR catalyst ^{29, 30}. H₂-TPR was used to characterize surface reducibility of the catalyst before and after K doping and the result was shown in Fig. 4. CuSO₄/TiO₂ catalyst performed H₂ consumption in a broad temperature range from 200 to 550 ℃, with four peaks at 281, 353, 452 and 492 ℃. According to our previous research, these peaks could be assigned to the reduction of surface CuSO₄ by H₂. After 1.0% (w) K was doped on the catalyst, it could be found that the signal of H₂ consumption changed obviously. The sharp peak at 452 ℃ disappeared. Meanwhile, a small peak appeared at 426 ℃. In addition, the peak at 281 ℃ moved to low temperature (271 ℃). The disappearance of the sharp peak at 452 ℃ indicated that K should react with partial CuSO₄ of the catalyst. With the increase of K doping, it could be found that the peak at 281 ℃ move to lower temperature and the intensity increased largely. When 4.0% (w) K was doped on the catalyst, the peak moved to 248 ℃, indicating K could promote surface reducibility of the catalyst. CuO was easier to be reduced by H₂ than CuSO₄, and the formation of CuO should be the main reason for the movement of the peak. According to the result of XRD and XPS, K could react with CuSO₄ and thus formed CuO. The formation of CuO should change the surface reducibility of K-poisoned catalyst. CuO was active in the oxidation of NH₃ and would cause the reaction between NH₃ and O₂ became prior in the process ³¹, thus the formation of CuO should be the major factor for the deactivation of CuSO₄/TiO₂ catalyst. The effect of K on surface reducibility of commercial VWTi catalyst also could be found in Fig. 4b. Commercial VWTi catalyst presented only one reduction peak (548 ℃) associated with reduction of surface vanadium species. When 1.0 % (w) K was doped on the catalyst, it could be found that the peak moved to a higher temperature (641 ℃), implying that K performed a negative impact on the surface reducibility, and thus resulted in the deactivation of commercial VWTi catalyst.

Figure 4

Figure 4.  H₂-TPR results of (a) CuSO₄/TiO₂ and (b) commercial VWTi catalysts before and after K doping.

下载: 全尺寸图片 幻灯片

Adsorption of NH₃ on acid sites of catalysts was a vital step of NH₃-SCR reaction ^{8, 32}. According to previous research, NH₃ was first absorbed on acid sites of CuSO₄/TiO₂ catalyst and then reacted with NO and O₂ ¹⁶. The adsorption of NH₃ on the catalyst was characterized by NH₃-TPD and the result was shown in Fig. 5a. CuSO₄/TiO₂ sample presented desorption signal of NH₃ in 150–450 ℃ and three peaks could be found in the result. The first peak at 175 ℃ could be associated with NH₃ adsorbed on weak acid sites. Meanwhile, other two peaks at 251 and 332 ℃ could be associated with NH₃ adsorbed on medium strong and strong acid sites, respectively. According to the desorption amount of NH₃, the total acid sites (T_acid) on the sample was 3673 μmol·g⁻¹. CuK-1 sample also presented a similar NH₃ desorption signal to that on CuSO₄/TiO₂ catalyst. However, the intensity of the signal decreased to a certain extent and T_acid was 2378 μmol·g⁻¹, about 0.64-time of that for CuSO₄/TiO₂ catalyst, implying that K would poison acid sites on CuSO₄/TiO₂ catalyst. With K content increasing, the number of acid sites on the catalyst decreased significantly. When K content increased to 4.0% (w), T_acid on CuK-4 sample decreased to 679 μmol·g⁻¹, only 18.4% of that on CuSO₄/TiO₂ catalyst, suggesting that most acid sites of the catalyst could be poisoned with K doping. The poison of acid sites was one of the major factors for the deactivation of CuSO₄/TiO₂ catalyst. As shown in Fig. 5a, K also poisoned acid sites of commercial VWTi catalyst. Only 1.0% (w) K caused the T_acid decreased to 222 μmol·g⁻¹, about 24% of that on fresh VWTi catalyst (925 μmol·g⁻¹). In conclusion, acid sites on CuSO₄/TiO₂ catalyst had a stronger resistant to K than that on commercial VWTi catalysts, which was one of the major factors for the high resistant to K of CuSO₄/TiO₂ catalyst.

Figure 5

Figure 5.  (a) NH₃-TPD and (b) in situ DRIFTS of NH₃ adsorption on each catalyst sample.

下载: 全尺寸图片 幻灯片

In order to investigate the effect of K on acid sites further, in situ DRIFTS of NH₃ adsorption on each sample was performed. And the result was shown in Fig. 5b. CuSO₄/TiO₂ sample was first treated by 0.25% NH₃/N₂ for 15 min under 300 ℃. Then the sample was purged by N₂ for 15 min. A number of bands at 1270, 1315, 1386, 1442 and 1601 cm⁻¹ could be detected. Moreover, a broad band in 1650–1800 cm⁻¹ also could be found. The band at 1386 cm⁻¹ could be associated with ―NH₂ wagging. Two bands at 1270 and 1601 cm⁻¹ were the typical bands of NH₃ adsorbed on Lewis (L) acid sites ³³. The band at 1442 cm⁻¹ and the broad band (1650–1800 cm⁻¹) could be associated with chemisorbed NH₄⁺ on Brønsted (B) acid sites ³². According to previous researches, for CuSO₄/TiO₂ catalyst, surface S-OH supplied most of Brønsted acid sites and Cu atom mainly supplied Lewis acid sites ¹⁶. After doping with 1.0% (w) K, the intensity of bands associated with chemisorbed NH₄⁺ on Brønsted acid sites reduced significantly, indicating that K mainly affected S-OH on the catalyst. K should react with S-OH and formed S-OK, and thus caused the elimination of Brønsted acid sites on CuSO₄/TiO₂ catalyst. In situ DRIFTS was also used to investigate the influence of K on NH₃ adsorption over commercial VWTi catalyst. The bands associated with NH₄⁺ on Brønsted acid sites almost disappeared after 1.0% (w) K doping, indicating that Brønsted acid sites of commercial VWTi catalyst should be destroyed by K. K was easy to react with Brønsted acid sites (V-OH) on commercial VWTi catalysts, in accordance with other researches ^{20, 22}. From the Fig. 5b, it could be concluded that Brønsted acid sites on CuSO₄/TiO₂ catalyst had a higher resistant to K than commercial VWTi catalyst. K could poison Brønsted acid sites and then hindered NH₃ adsorption on SCR catalysts, which could cause the deactivation of catalysts.

According to previous researches, NH₃-SCR reaction on CuSO₄/TiO₂ catalyst mainly obeyed Eley-Rideal (E-R) mechanism ¹⁶. NH₃ was first adsorbed on Lewis and Brønsted acid sites, then gaseous NO and O₂ would react with NH₃/NH₄⁺ on the acid sites, N₂ and H₂O were formed finally ¹⁶. Reaction mechanism on fresh and deactivated catalyst might be different. In situ DRIFTS experiment was used at 300 ℃ to investigate the effect of K on reaction mechanism over CuSO₄/TiO₂ catalyst. The catalyst was treated first by 0.25% NH₃/N₂ for 15 min. Then the mixture gas was switched into pure N₂ for 15 min. After that, 0.25% NO + 4.0% O₂/N₂ was inlet into the reactor cell for 20 min. The obtained spectra can be found in Fig. 6. After the catalyst successively pretreated by NH₃/N₂ and N₂, bands associated with NH₃ and NH₄⁺ on Lewis and Brønsted acid sites appeared, consistent with the spectra in Fig. 5b. When the mixture was changed to NO + O₂/N₂, all bands assigned to NH₃/NH₄⁺ adsorbed on the surface of CuSO₄/TiO₂ catalyst decreased largely, implying that both NH₃ on Lewis acid sites and NH₄⁺ on Brønsted acid sites would react with gaseous NO and O₂. Thus, reaction mechanism on K-poisoned CuSO₄/TiO₂ catalysts should also obeyed E-R mechanism. It could be concluded that 1.0% (w) K doping would not change reaction mechanism of NH₃-SCR over CuSO₄/TiO₂ catalysts.

Figure 6

Figure 6.  In situ DRIFTS experiment of CuK-0 sample treated by NH₃ + N₂, N₂ and N₂ + NO + O₂ in proper order at 300 ℃.

下载: 全尺寸图片 幻灯片

3.4   Deactivation mechanism of K for CuSO₄/TiO₂ catalysts

K could cause the deactivation of CuSO₄/TiO₂ and commercial VWTi catalysts used in NH₃-SCR reaction. As shown in Fig. 1a and 1b, K caused the decrease of NO_x conversion for both catalysts. Nevertheless, an obvious difference could be found for the effect of K on N₂ selectivity over two catalysts. As shown in Fig. 1c and 1d, the doping of K would promote of N₂O formation over CuSO₄/TiO₂ catalyst but inhibited that over commercial VWTi catalyst. Thus, the deactivation mechanism should be different between CuSO₄/TiO₂ and commercial VWTi catalysts. The deactivation mechanism of K over V-based catalyst has been investigated by a number of researchers ^{17, 20, 22, 25}. They concluded that K would neutralize Brønsted acid sites (V-OH) with going against the reduction of surface V species over V-base catalysts, hindering the adsorption/activation of NH₃ on the catalyst, causing the deactivation of V-based catalyst in NH₃-SCR reaction. As shown in the results of NH₃-TPD, in situ DRIFTS of NH₃ adsorption, K could also poison Brønsted acid sites (S-OH) over CuSO₄/TiO₂ catalyst. Notably, K would benefit the surface reducibility of CuSO₄/TiO₂ catalyst (Fig. 4). In conclusion, the different effect of K on surface reducibility should be the major factors for the opposite change of N₂ selectivity for CuSO₄/TiO₂ and commercial VWTi catalyst.

As shown in the result of XRD, XPS and H₂-TPR, it could be concluded that CuO and K₂SO₄ was formed on K-poisoned CuSO₄/TiO₂ catalysts. K should react with CuSO₄ over the catalyst, as followed the chemical equation below:

The formation of CuO should be the major factor for promoting surface reducibility of CuSO₄/TiO₂ catalyst after K doping. In addition, CuO was more active in selective catalytic oxidation of NH₃ (NH₃-SCO). Previous researchers have confirmed that NH₃ was easy to be oxidized to N₂O, NO and NO₂ over CuO-based catalyst in the temperature above 300 ℃ ^{31, 34, 35}. As Fig. 1c showed, K doping caused the fall of N₂ selectivity for CuSO₄/TiO₂ catalyst. Therefore, the formation of CuO should be the major factor for the deactivation of CuSO₄/TiO₂ catalysts. In addition, as shown in Fig. 5, K could poison Brønsted acid sites by reacting with surface S-OH on the catalyst, which then hindered the adsorption of NH₃ on K-poisoned catalyst. The disappearance of Brønsted acid sites also should be one of the major reasons for the deactivation of the catalyst. Moreover, the formation of CuO and K₂SO₄ also harmed the pore structure, causing the decrease of specific surface area, which should be another reason for the catalyst deactivation. Proposed deactivation mechanism of K over CuSO₄/TiO₂ catalysts could be found in Fig. 7.

Figure 7

Figure 7.  Proposed deactivation mechanism of K over CuSO₄/TiO₂ catalysts.

下载: 全尺寸图片 幻灯片

4.   Conclusions

K in the flue gas would lead to the deactivation of CuSO₄/TiO₂ NH₃-SCR catalyst. The formation of CuO and K₂SO₄ through the reaction between K and CuSO₄ was the main factor for the catalyst deactivation. NH₃ would be oxidized to N₂O, NO and NO₂ over CuO of K-poisoned CuSO₄/TiO₂ catalysts, causing the decrease of both NO_x conversion and N₂ selectivity. Moreover, the disappearance of Brønsted acid sites caused by the reaction of K and S-OH was also one of the main factors for the deactivation of the catalyst. The decrease of specific surface area after K doping also contributed the deactivation of the catalyst. Notably, due to the abundant acid sites and surface adsorbed oxygen species of CuSO₄/TiO₂ catalyst, the catalyst exhibited a higher resistance to K than commercial VWTi catalyst. CuSO₄/TiO₂ catalyst might a candidate NH₃-SCR catalyst used in the flue gas containing high concentration of K.

References

1. [1]

   Fang, X.; Qu, W.; Qin, T.; Hu, X.; Chen, L.; Ma, Z.; Liu, X.; Tang, X. Environ. Sci. Technol. 2022, 56, 6631. doi: 10.1021/acs.est.2c00482  doi: 10.1021/acs.est.2c00482
   

2. [2]

   Ding, X. M.; Liang, Y. L.; Zhang, H. L.; Zhao, M.; Wang, J. L.; Chen, Y. Q. Acta Phys. -Chim. Sin. 2022, 38, 2005009.  doi: 10.3866/PKU.WHXB202005009
   

3. [3]

   Mesilov, V.; Dahlin, S.; Bergman, S. L.; Xi, S.; Han, J.; Olsson, L. Appl. Catal. B 2021, 299, 120626. doi: 10.1016/j.apcatb.2021.120626  doi: 10.1016/j.apcatb.2021.120626
   

4. [4]

   Han, L.; Cai, S.; Gao, M.; Hasegawa, J.; Wang, P.; Zhang, J.; Shi, L.; Zhang, D. Chem. Rev. 2019, 119, 10916. doi: 10.1021/acs.chemrev.9b00202  doi: 10.1021/acs.chemrev.9b00202
   

5. [5]

   Yang, X. Q.; Yang, H. L.; Lu, H.; Ding, H. X.; Tong, Y. X.; Rao, F.; Zhang, X.; Shen, Q.; Gao, J.; Zhu, G. Q. Acta Phys. -Chim. Sin. 2021, 37, 2005008.  doi: 10.3866/PKU.WHXB202005008
   

6. [6]

   Huang, X.; Liu, Z.; Wang, D.; Peng, Y.; Li, J. Catal. Sci. Technol. 2020, 10, 323. doi: 10.1039/C9CY02188D  doi: 10.1039/C9CY02188D
   

7. [7]

   Yu, Y.; Yi, X.; Zhang, J.; Tong, Z.; Chen, C.; Ma, M.; He, C.; Wang, J.; Chen, J.; Chen, B. Catal. Sci. Technol. 2021, 11, 5125. doi: 10.1039/D1CY00467K  doi: 10.1039/D1CY00467K
   

8. [8]

   Ni, S.; Tang, X.; Yi, H.; Gao, F.; Wang, C.; Shi, Y.; Zhang, R.; Zhu, W. J. Rare Earths 2022, 40, 268. doi: 10.1016/j.jre.2020.12.015  doi: 10.1016/j.jre.2020.12.015
   

9. [9]

   Li, L.; Ge, C.; Ji, J.; Tan, W.; Wang, X.; Wei, X.; Guo, K.; Tang, C.; Dong, L. Chin. J. Catal. 2021, 42, 1488. doi: 10.1016/S1872-2067(20)63778-0  doi: 10.1016/S1872-2067(20)63778-0
   

10. [10]

    Guo, K.; Ji, J.; Song, W.; Sun, J.; Tang, C.; Dong, L. Appl. Catal. B 2021, 297, 120388. doi: 10.1016/j.apcatb.2021.120388  doi: 10.1016/j.apcatb.2021.120388
    

11. [11]

    Zhang, Y.; Zhang, J.; Wang, H.; Yang, W.; Wang, C.; Peng, Y.; Chen, J.; Li, J.; Gao, F. J. Phys. Chem. C 2022, 126, 8720. doi: 10.1021/acs.jpcc.2c01268  doi: 10.1021/acs.jpcc.2c01268
    

12. [12]

    Hu, W.; He, J.; Liu, X.; Yu, H.; Jia, X.; Yan, T.; Han, L.; Zhang, D. Environ. Sci. Technol. 2022, 56, 5170. doi: 10.1021/acs.est.1c08715  doi: 10.1021/acs.est.1c08715
    

13. [13]

    Kang, L.; Han, L.; Wang, P.; Feng, C.; Zhang, J.; Yan, T.; Deng, J.; Shi, L.; Zhang, D. Environ. Sci. Technol. 2020, 54, 14066. doi: 10.1021/acs.est.0c05038  doi: 10.1021/acs.est.0c05038
    

14. [14]

    Yu, Y.; Zhang, J.; Chen, C.; Ma, M.; He, C.; Miao, J.; Li, H.; Chen, J. New J. Chem. 2020, 44, 13598. doi: 10.1039/D0NJ02647F  doi: 10.1039/D0NJ02647F
    

15. [15]

    Qi, K.; Xie, J.; Fang, D.; Gong, P.; Li, F.; He, F. J. Cleaner Prod. 2020, 256, 120411. doi: 10.1016/j.jclepro.2020.120411  doi: 10.1016/j.jclepro.2020.120411
    

16. [16]

    Yu, Y.; Miao, J.; Wang, J.; He, C.; Chen, J. Catal. Sci. Technol. 2017, 7, 1590. doi: 10.1039/c6cy02626e  doi: 10.1039/c6cy02626e
    

17. [17]

    Kong, M.; Liu, Q.; Zhou, J.; Jiang, L.; Tian, Y.; Yang, J.; Ren, S.; Li, J. Chem. Eng. J. 2018, 348, 637. doi: 10.1016/j.cej.2018.05.045  doi: 10.1016/j.cej.2018.05.045
    

18. [18]

    Gong, P.; He, F.; Fang, D.; Xie, J. J. Environ. Chem. Eng. 2022, 10, 107895. doi: 10.1016/j.jece.2022.107895.  doi: 10.1016/j.jece.2022.107895
    

19. [19]

    Li, J.; Peng, Y.; Chang, H.; Li, X.; Crittenden, J. C.; Hao, J. Front. Environ. Sci. Eng. 2016, 10, 413. doi: 10.1007/s11783-016-0832-3  doi: 10.1007/s11783-016-0832-3
    

20. [20]

    Chen, L.; Li, J.; Ge, M. Chem. Eng. J. 2011, 170, 531. doi: 10.1016/j.cej.2010.11.020  doi: 10.1016/j.cej.2010.11.020
    

21. [21]

    Chen, X.; Geng, Y.; Shan, W.; Liu, F. Ind. Eng. Chem. Res. 2018, 57, 1399. doi: 10.1021/acs.iecr.7b04444  doi: 10.1021/acs.iecr.7b04444
    

22. [22]

    Kong, M.; Liu, Q.; Jiang, L.; Tong, W.; Yang, J.; Ren, S.; Li, J.; Tian, Y. Chem. Eng. J. 2019, 370, 518. doi: 10.1016/j.cej.2019.03.156  doi: 10.1016/j.cej.2019.03.156
    

23. [23]

    Wu, H.; Li, W.; Guo, L.; Pan, Y.; Xu, X. J. Fuel Chem. Technol. 2011, 39, 550. doi: 10.1016/S1872-5813(11)60034-0  doi: 10.1016/S1872-5813(11)60034-0
    

24. [24]

    Byun, S. W.; Lee, S. J.; Kim, M.; Bae, W. B.; Shin, H.; Hazlett, M. J.; Kang, D.; Tesfaye, B.; Park, P. W.; Kang, S. B. Chem. Eng. J. 2022, 444, 136517. doi: 10.1016/j.cej.2022.136517  doi: 10.1016/j.cej.2022.136517
    

25. [25]

    Du, X.; Gao, X.; Qiu, K.; Luo, Z.; Cen, K. J. Phys. Chem. C 2015, 119, 1905. doi: 10.1021/jp511475b  doi: 10.1021/jp511475b
    

26. [26]

    Madia, G.; Elsener, M.; Koebel, M.; Raimondi, F.; Wokaun, A. Appl. Catal. B 2002, 39, 181. doi: 10.1016/S0926-3373(02)00099-1  doi: 10.1016/S0926-3373(02)00099-1
    

27. [27]

    Qi, X.; Han, L.; Deng, J.; Lan, T.; Wang, F.; Shi, L.; Zhang, D. Environ. Sci. Technol. 2022, 56, 5840. doi: 10.1021/acs.est.2c00944  doi: 10.1021/acs.est.2c00944
    

28. [28]

    Guo, M.; Mosevitzky, L. B.; Ford, M. E.; Wachs, I. E. Appl. Catal. B 2022, 306, 121108. doi: 10.1016/j.apcatb.2022.121108  doi: 10.1016/j.apcatb.2022.121108
    

29. [29]

    Zhang, X.; Xuan, Y.; Wang, B.; Gao, C.; Niu, S.; Zhao, G.; Wang, D.; Li, J.; Lu, C.; Crittenden, J. C. Front. Environ. Sci. Eng. 2021, 16, 88. doi: 10.1007/s11783-021-1447-x  doi: 10.1007/s11783-021-1447-x
    

30. [30]

    Ji, J.; Tang, Y.; Han, L.; Ran, P.; Song, W.; Cai, Y.; Tan, W.; Sun, J.; Tang, C.; Dong, L. Chem. Eng. J. 2022, 445, 136530. doi: 10.1016/j.cej.2022.136530  doi: 10.1016/j.cej.2022.136530
    

31. [31]

    Han, F.; Yuan, M.; Mine, S.; Sun, H.; Chen, H.; Toyao, T.; Matsuoka, M.; Zhu, K.; Zhang, J.; Wang, W.; et al. ACS Catal. 2019, 9, 10398. doi: 10.1021/acscatal.9b02975  doi: 10.1021/acscatal.9b02975
    

32. [32]

    Wang, Y.; Yi, W.; Yu, J.; Zeng, J.; Chang, H. Environ. Sci. Technol. 2020, 54, 12612. doi: 10.1021/acs.est.0c02840  doi: 10.1021/acs.est.0c02840
    

33. [33]

    Yu, Y.; Wei, D.; Tong, Z.; Wang, J.; Chen, J.; He, C. Chem. Eng. J. 2022, 442, 136356. doi: 10.1016/j.cej.2022.136356  doi: 10.1016/j.cej.2022.136356
    

34. [34]

    Qu, Z.; Wang, Z.; Zhang, X.; Wang, H. Catal. Sci. Technol. 2016, 6, 4491. doi: 10.1039/C5CY02125A  doi: 10.1039/C5CY02125A
    

35. [35]

    Zhang, X.; Wang, H.; Wang, Z.; Qu. Z. Appl. Surf. Sci. 2018, 447, 40. doi: 10.1016/j.apsusc.2018.03.220  doi: 10.1016/j.apsusc.2018.03.220