Nitrogen oxides (NO_x), which are emitted from combustion of fossil fuels, are a major cause of air pollution, and can cause phenomena such as photochemical smog, acid rain, and ozone depletion ^[1]. Selective catalytic reduction with NH₃ (NH₃-SCR) is an efficient method for abating NO_x from stationary sources and is widely used in coal- or biomass-fired power plants. The catalyst is important in SCR and V₂O₅-WO₃/TiO₂ is a popular commercial SCR catalyst. V₂O₅-WO₃/TiO₂ is excellent for NO_x abatement, but V₂O₅ is toxic, and the catalyst causes extensive oxidation of SO₂ to SO₃ ^[2]. Much effort has therefore been made to develop new SCR catalysts ^{[3, 4]}.

SO₂ is present in the flue gas from coal-fired power plants, therefore sulfur tolerance is an important index for SCR catalysts. Under the SCR conditions, SO₂ and O₂ react with oxides to form thermodynamically stable sulfate phases. These cover the catalyst surface and reduce the number of active sites for the SCR reaction, causing deactivation of the SCR catalyst ^[5]. For example, Ce/TiO₂ has excellent activity in NH₃-SCR, but when SO₂ is added to the reaction, Ce(SO₄)₂ and Ce₂(SO₄)₃ formation significantly decreases the activity ^[6]. MnO_x/Al₂O₃ shows high activity and selectivity in NH₃-SCR at 110-350 °C, but MnO_x is converted to MnSO₄ in the presence of SO₂, resulting in catalyst deactivation ^[7]. Al₂O₃, which has a high specific surface area, is a popular catalyst carrier, but it can react with SO₂ and O₂ to form Al₂(SO₄)₃; this greatly decreases the specific surface area and causes significant catalyst deactivation ^{[8, 9]}. In our recent research, we found that sulfation promoted the activity of commercial V₂O₅-WO₃/TiO₂, which has been deactivated by K, by increasing the number of acid sites on the deactivated catalyst ^[10]. The sulfation of a CuO/Al₂O₃ catalyst sorbent for a short time promotes SCR activity at high temperatures (> 300 °C) ^[11]. Sulfate catalysts may therefore be promising for use in NH₃-SCR and may have high sulfur tolerance and low toxicity. However, research on the use of sulfate catalysts in NH₃-SCR is limited. Ma et al. ^[12] found that an Fe₂(SO₄)₃/TiO₂ catalyst had excellent activity and low N₂O selectivity in NH₃-SCR at 350-450 °C, and displayed high sulfur and H₂O tolerance. Pietrogiacomi et al. ^[13] found that a CuSO₄/ZrO₂ catalyst showed good activity in NH₃-SCR, but the temperature window was narrow and the sulfur tolerance was unknown. It is still important to investigate sulfate catalysts for NH₃-SCR.

Because of its high specific surface area and high tolerance of SO₂, TiO₂ (anatase) is commonly used as a support in SCR catalysts ^{[14, 15]}. In this study, a series of CuSO₄/TiO₂ catalysts were prepared. Their activities and tolerance of SO₂ and H₂O were tested. The catalysts were characterized using N₂ adsorption-desorption, X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), temperature-programmed reduction of H₂ (H₂-TPR), and temperature-programmed desorption of NH₃ (NH₃-TPD). The NH₃-SCR reaction mechanism on CuSO₄/TiO₂ was investigated using in situ diffuse-reflectance infrared Fourier-transform spectroscopy (DRIFTS).

CuSO₄/TiO₂ catalysts were prepared using a wet impregnation method. TiO₂ (anatase, T104937-100g, Aladdin, China) was pretreated at 300 °C in air for 3 h and then cooled to room temperature in a desiccator. Solutions containing 1, 2, and 4 wt% CuSO₄ (AR; 100 mL) were prepared. TiO₂ (3.0 g) was added to the solution and the mixture was stirred at 40 °C for 1 h. The mixed solution was filtered and the residue was dried at 60 °C for 30 min and then at 120 °C for 2 h. The catalyst was calcined at 500 °C in air for 5 h. Samples are denoted by xCuTi, where x is the concentration of CuSO₄ solution.

The catalytic activity was evaluated using a fixed-bed quartz reactor (Φ10 mm × 600 mm). The catalyst sample (20-40 mesh) was placed in the middle of the reactor. The total gas flow rate was 1250 mL/min and the gas consisted of NO (0.05%), NH₃ (0.05%), SO₂ (0.15%, when used), H₂O (5%, when used), and O₂ (4%) in N₂. The concentrations of O₂, NO, and NO₂ were determined using a flue gas analyzer (T-350, Testo Company, Germany) and N₂O was detected using gas chromatography (7890A, Agilent Technologies, USA). The NO_x conversion (x) is defined by

N₂ adsorption-desorption was performed at -196 °C using a NOVA 2000e surface area and pore size analyzer (Quantachrome, USA). The specific surface area was calculated using the Brunauer-Emmett-Teller (BET) method and the pore size distribution was obtained using the Barrett-Joyner-Halenda method. Powdered samples were examined by XRD (X'Pert Pro diffractometer, PANalytical B.V., The Netherlands) using Cu K_α radiation. XPS was performed using an ESCALAB 250 spectrometer (Thermo Fisher Scientific Company, USA). H₂-TPR was performed using a Quantachrome ChemBET-3000 TPR-TPD chemisorption analyzer. The sample (50 mg) was preheated under a stream of He at 500 °C for 1 h and then cooled to 100 °C. The temperature was progressively increased to 600 °C at 10 °C/min in a stream of 5% H₂/Ar and the thermal conductivity detector signal was recoded. NH₃-TPD was performed using the same instrument; NH₃ was detected using mass spectrometry (MS; DYCOR LC-D100, Ametek Company, USA). A sample (50 mg) was preheated under a stream of He at 500 °C for 1 h and then cooled to 50 °C. The gas flow was changed to 5% NH₃/Ar for 30 min, and then the sample was purged with He at 100 °C for 1 h to remove weakly adsorbed NH₃. The temperature was progressively increased to 600 °C at 10 °C/min.

In situ DRIFTS was performed using a Bruker Vertex 70 infrared spectrometer with KBr optics and a mercury cadmium telluride detector cooled by liquid N₂. The DRIFTS reaction cell (Harrick Scientific) was fitted with KBr windows and a heating cartridge that enabled the sample to be heated to 500 °C. For NH₃ or NO + O₂ adsorption, the powdered sample was preheated in the reaction cell at 500 °C for 1 h under a N₂ flow (50 mL/min), and then cooled to 80, 180, 280, and 380 °C. The background spectrum was recorded at the corresponding temperature. The sample was exposed to 0.3% NH₃/N₂ or 0.3% NO + 4.0% O₂/N₂ for 15 min and then purged with N₂ for 15 min. DRIFTS was then performed. All spectra were recorded at a resolution of 4 cm^-1 by accumulating 100 scans.

In situ DRIFTS of NO + O₂ on a catalyst sample pretreated with NH₃ was also performed. The catalyst sample was first preheated at 500 °C for 1 h in N₂, and cooled to 380 °C; the background spectrum was then recorded. The sample was exposed to 0.3% NH₃/N₂ for 15 min and then purged with N₂ for 15 min. Then 0.3% NO + 4% O₂/N₂ was imported and spectra were recorded at 3, 5, 10, 15, and 20 min.

The N₂ adsorption-desorption isotherms and pore size distributions for all the samples are shown in Fig. 1. The isotherms show that pure TiO₂ is significantly mesoporous ^[16]. The isotherms did not change significantly after TiO₂ impregnation with various CuSO₄ solutions, suggesting that CuSO₄ did not noticeably affect the carrier pore structure. The pore size distributions support this conclusion. Fig. 1 shows that all the samples had similar pore size distributions, with a peak at around 4.7 nm. The specific surface areas (A_BET), total pore volumes (V_P), and average pore radii (r_A) of the samples are shown in Table 1. The specific surface area decreased slightly, and the average pore radius increased slightly, with increasing CuSO₄ concentration.

Fig. 1. N2 adsorption-desorption isotherms and pore size distributions (insert) for 0CuTi, 1CuTi, 2CuTi, and 4CuTi samples.

Table 1 Specific surface areas, total pore volumes, and average pore radii of samples Samples ABET (m2/g) VP* (cm3/g) rA (nm) 0CuTi 56.6 0.2311 8.17 1CuTi 54.3 0.2196 8.15 2CuTi 54.0 0.2201 8.16 4CuTi 52.9 0.2369 8.96 * Obtained at p/p0 = 0.99.

Table 1 Specific surface areas, total pore volumes, and average pore radii of samples

The XRD patterns of all the samples are shown in Fig. 2. Pure TiO₂ had diffraction peaks at 2θ = 25.3°, 36.9°, 37.9°, 38.6°, 48.0°, 53.9°, 55.0°, 62.8°, 69.0°, 70.3°, and 75.1°, corresponding to anatase (PDF 03-065-5714); anatase was therefore the main phase of the carrier. The anatase diffraction peaks were retained after TiO₂ impregnation with CuSO₄ solutions, suggesting that CuSO₄ was well dispersed on the carrier or the CuSO₄ concentration was lower than the limit of detection.

Fig. 2. XRD patterns for 0CuTi, 1CuTi, 2CuTi, and 4CuTi samples.

The O 1s, Cu 2p, and S 2p XPS profiles for the 0CuTi and 4CuTi samples are shown in Fig. 3. Two main oxygen species were detected in pure TiO₂, i.e., lattice oxygen, at 529.6 eV, and adsorbed oxygen, at 530.8 eV ^[17]. Lattice oxygen predominated and the ratio of adsorbed oxygen to lattice oxygen was about 0.41:1. Lattice oxygen and adsorbed oxygen were also present in the 4CuTi sample. The amount of adsorbed oxygen was higher in the 4CuTi sample than in pure TiO₂. The ratio of adsorbed oxygen to lattice oxygen increased to 1.50:1, suggesting that the presence of CuSO₄ greatly increased the amount of adsorbed oxygen in the catalyst. Fig. 3 shows that no distinct peak from Cu atoms was observed at 945.0-925.0 eV for the 0CuTi sample. However, a peak at 932.0 eV was observed for the 4CuTi sample, suggesting that Cu atoms were loaded on the support and mainly present as Cu^{2+ [18]}. The same trend was observed for S atoms. A peak at 168.2 eV, assigned to SO₄^2-, was detected for the 4CuTi sample ^[19].

Fig. 3. O 1s, Cu 2p, and S 2p XPS spectra for 0CuTi and 4CuTi samples.

The catalyst surface reducibility has an important effect on the activity in the NH₃-SCR reaction ^[20]. The surface reducibility of each catalyst sample was determined using H₂-TPR; the results are shown in Fig. 4(a). No significant peak was observed for pure TiO₂. Large reduction peaks were observed for the catalysts obtained by TiO₂ impregnation with CuSO₄ solutions. The reduction peaks were located at about 520 °C for 1CuTi and 500 °C for 2CuTi. This suggests that the surface reducibility increased with increasing CuSO₄ concentration. The reduction peak is assigned to sulfate reduction ^[12], i.e., CuSO₄ + 2H₂ → Cu + 2H₂O + SO₂. The amount of SO₂ produced during H₂-TPR was determined using MS (m/e = 64); the results are shown in Fig. 4(b). As the figure shows, the SO₂ peak was consistent with the reduction peak. However, two reduction peaks were observed for the 4CuTi sample, at about 350 and 490 °C. Fig. 4(b) shows that the latter peak can be assigned to reduction of surface sulfate. The former peak can be assigned to reduction of surface-adsorbed oxygen. It is concluded that CuSO₄ increases the amount of oxygen adsorbed on the catalyst; this is consistent with the XPS results. H₂S was also detected by MS (m/e = 34) during H₂-TPR, but the amount was few; this is consistent with the results of previous research ^[21].

Fig. 4. H2-TPR profiles (a) and release of SO2 (b) during H2-TPR for different samples.

It has been reported that NH₃ adsorption on acid sites of catalysts is an important step in the NH₃-SCR reaction ^{[10, 22]}. NH₃ is first adsorbed on Brønsted or Lewis acid sites to form NH₄⁺ or coordinated NH₃, and then reacts with gaseous or adsorbed nitric oxides to form N₂ and H₂O ^{[15, 22, 23]}. The effect of CuSO₄ on the catalyst acid sites was investigated using NH₃-TPD; the results are shown in Fig. 5. A broad peak from about 200 to 500 °C was observed for pure TiO₂. After CuSO₄ was loaded on the support, the onset temperature for NH₃ desorption dropped to about 150 °C and the intensity of the NH₃ desorption peak clearly increased, suggesting that CuSO₄ increased the amount of acid sites on the catalyst.

Fig. 5. NH3-TPD profiles for various catalysts.

The catalytic performances of the samples are shown in Fig. 6(a). Pure TiO₂ had the poorest activity. The NO_x conversion increased from 11.8% to 32.7% when the temperature was increased from 300 to 450 °C for the 0CuTi sample. The activities of the CuSO₄/TiO₂ samples were significantly higher than that of pure TiO₂. The catalytic activity increased with increasing CuSO₄ concentration, i.e., the 4CuTi sample had the highest activity. The activity of a commercial SCR catalyst was tested under the same conditions; the results are also shown in Fig. 6(a). The activity of the 4CuTi sample was similar to that of the commercial catalyst, and the NO_x conversion was higher than 94% when the temperature was higher than 340 °C. However, the 4CuTi activity was lower than that of the commercial catalyst below 340 °C. The 4CuTi activity therefore still needs to be improved, especially at temperatures below 340 °C. The influence of the gas hourly space velocity (GHSV) on the 4CuTi activity is shown in Fig. 6(b). The GHSV had little effect on the sample activity at temperatures above 340 °C. However, the activity decreased significantly at temperatures lower than 340 °C. N₂O is the main byproduct of NH₃-SCR; data on N₂O production are also shown in Fig. 6(a). N₂O production with 4CuTi was lower than that with VWTi. The effect of SO₂ or H₂O on the 4CuTi activity was investigated by adding 0.15% SO₂ or 5.0% H₂O to the reactor for about 24 h at 340 °C; the results are shown in Fig. 7. 4CuTi showed a high tolerance of SO₂, and SO₂ had little influence on the sample activity during the experimental period. After addition of 5.0% H₂O, the 4CuTi activity decreased slightly; the NO_x conversion was still higher than 93% when the sample was exposed to H₂O for about 24 h. The combined effect of H₂O and SO₂ on the 4CuTi activity was also examined; the results are shown in Fig. 7. A combination of H₂O and SO₂ had little influence on the activity of the catalyst. It is concluded that the samples had a high tolerance of SO₂ and H₂O, in accordance with previous results ^[12].

Fig. 6. (a) Activities of CuSO4/TiO2 and VWTi catalysts (GHSV = 50000 h-1) and (b) effect of GHSV on 4CuTi activity.

Fig. 7. Effects of SO2 and/or H2O on activity of 4CuTi sample. Reaction conditions: 0.05% NO, 0.05% NH3, 4% O2, 0.15% SO2 (when used) and 5% H2O (when used) in N2, 340 °C, 50000 h-1.

DRIFT spectra obtained after adsorption of NH₃ on 4CuTi at different temperatures are shown in Fig. 8(a). Bands at 1243, 1307, 1357, 1593, 1665, 3160, 3263, and 3350 cm^-1 and a broad band at 1390-1529 cm^-1 were detected after treatment of the catalyst with NH₃/N₂ for 15 min and purging with N₂ for 15 min at 80 °C. The band at 1665 cm^-1 and the broad band at 1390-1529 cm^-1 are assigned to NH₄⁺ formed by NH₃ protonation on Brönsted acid sites ^{[10, 24]}. The bands at 1243, 1307, and 1593 cm^-1 are assigned to NH₃ adsorbed on Lewis acid sites ^[24]. The band at 1357 cm^-1 is assigned to -NH₂ wagging ^[10]. The bands at 3350, 3263, and 3160 cm^-1 are assigned to coordinated NH₃ adsorbed on Lewis acid sites ^[25]. The bands moved to slightly higher wavenumbers with increasing temperature. The intensities of the band at 1665 cm^-1 and the broad band at 1390-1529 cm^-1 decreased significantly at 280 °C, and they almost disappeared at 380 °C, suggesting that NH₄⁺ on Brönsted acid sites was unstable. However, the bands assigned to NH₃ adsorbed on Lewis acid sites were still present at 380 °C, suggesting that NH₃ adsorbed on Lewis acid sites was more stable. It is concluded that at high temperatures NH₃ was mainly adsorbed on Lewis acid sites in 4CuTi.

Fig. 8. In situ DRIFT spectra of 4CuTi samples (a) treated with 0.3% NH3/N2 for 15 min and purged with N2 for 15 min at 80, 180, 280, and 380 °C; (b) Treated with 0.3% NO and 4% O2 for 15 min and purged with N2 for 15 min at 80, 180, 280, and 380 °C; (c) Pretreated with 0.3% NH3 and exposed to 0.3% NO + 4% O2 for various times at 380 °C.

The spectra for NO + O₂ adsorption on 4CuTi at different temperatures are shown in Fig. 8(b). Bands at 1243, 1285, 1358, 1514, 1549, 1585, and 1613 cm^-1 were observed at 80 °C. Bands at 1650-1500 cm^-1 are widely accepted to arise from surface nitrate or nitro species adsorbed primarily on transition-metal active sites of catalysts ^[26]. The bands at 1613 and 1243 cm^-1 can be assigned to bridging nitrates ^[27], and those at 1585 and 1549 cm^-1 can be assigned to bidentate nitrate ^{[25, 27]}. The bands at 1358 cm^-1 can be assigned to ν_asym(NO₂) of chelated nitrito surface species ^[28] and those at 1514 and 1285 cm^-1 can be assigned to monodentate nitrate ^[27]. When the temperature was increased to 180 °C, the bands at 1285 and 1514 cm^-1 disappeared, suggesting that surface monodentate nitrate was unstable. The other bands were still detected at 180 and 280 °C. However, all the bands almost disappeared at 380 °C, suggesting that little NO was adsorbed on the surface at high temperatures. This suggests that NO cannot form stable intermediates that could participate in the SCR reaction at high temperatures.

In situ DRIFT spectra of NO + O₂ on a catalyst sample pretreated with NH₃ at 380 °C are shown in Fig. 8(c). After NO + O₂ was introduced, the intensities of the bands assigned to NH₃ adsorbed on Lewis acid sites decreased greatly, suggesting that NH₃ adsorbed on the catalyst reacted with NO + O₂. The changes in the -NH region support this conclusion. It seems that some nitrates were formed on the catalyst surface, but the adsorption was unstable. When the supply of NO + O₂ was stopped, the bands disappeared. It can be concluded that the reaction mainly occurred between NH₃ adsorbed on Lewis acid sites and gaseous NO; this is different from the mechanism involved for the V₂O₅-WO₃/TiO₂ catalyst. Most researchers agree that with the V₂O₅-WO₃/TiO₂ catalyst, NH₃ is first adsorbed on Brönsted acid sites as NH₄⁺ and then takes part in the reaction ^[22]. The SCR reaction on CuSO₄/TiO₂ could follow the Eley-Rideal mechanism, and reaction of gaseous NO with NH₃ adsorbed on Lewis acid sites to form N₂ and H₂O could be the main reaction pathway. The XPS and H₂-TPR results show that CuSO₄ increased the amount of oxygen adsorbed on the catalyst surface. The adsorbed oxygen might favor the reaction of NH₃ adsorbed on Lewis acid sites with gaseous NO.

CuSO₄/TiO₂ catalysts showed excellent activities in the NH₃-SCR reaction at temperatures higher than 340 °C and a high tolerance of SO₂ or H₂O. These are promising NH₃-SCR catalysts for use in coal-fired power plants. The NH₃-SCR reaction on CuSO₄/TiO₂ could follow the Eley-Rideal mechanism. Unlike the case for a V₂O₅-WO₃/TiO₂ catalyst, the reaction of gaseous NO with NH₃ adsorbed on Lewis acid sites to form N₂ and H₂O could be the main reaction pathway on CuSO₄/TiO₂ catalysts. CuSO₄ increases the amounts of acid sites and adsorbed oxygen on the catalysts, and the adsorbed oxygen could promote the NH₃-SCR reaction on Lewis acid sites.