The TiO₂ powder (Industrial grade, 96% purity, 15 - 25 nm, 93.9 m²・g⁻¹) was dried at 105˚C for 1 h and milled for 150 mesh. Manganese nitrate, cerium nitrate and nitrate of transition metal (M) (Sinopharm Chemical Reagent Co.) at a certain ratio are solved into Deionized water (DW). Then, the obtained solution was placed into the ultrasonic reaction pool. The processed TiO₂ powder was added into the solution when the metal nitrate dissolved fully under the stirring condition with 160 W ultrasonic at 50˚C for 24 h. Then the immersed TiO₂ powder was dried at 105˚C for 24 h, calcined at 450˚C for 6 h and milled for 150 mesh to obtain catalysts powder. The obtained catalyst was expressed as Mn-Ce-M_0.025, and M represent W, La, Pr and Zr. The mole ratio of Ti:Mn:Ce:M is 1:0.4:0.07:0.025.

The catalyst’s catalytic activity was measured in a quartz reactor (35 cm i.d.) using 6.75 g catalyst, see in Figure 1. The feed gas mixture in a N₂ stream contained 1000 ppm NO, 1100 ppm NH₃, and 5%O₂. The total flow rate of the feed gas was 90 L/h (GHSV = 10,000 h⁻¹). The reaction temperature was scanned from 100˚C to 400˚C.

The concentration of the NO_x in the inlet and outlet of the reactor were measured on-line by a flue gas analyzer (MRU, Germany). H₂-temperature-programmed reduction (H₂-TPR, ChemBET Pulsar) experiments were carried out using every catalyst (40 mg) under a mixed gas flow of H₂ and Ar.Morphologies of the samples were characterized by scanning electron microscopy (SEM, quanta FEG 250). Powder XRD was characterized with Cu-Ka radiation at the range of 10˚ - 90˚. Fourier infrared spectrum (FT-IR, Prestige-21) was used to analyze functional group change of catalyst after operating for a time.

XRD patterns of the Mn-Ce-M catalysts are shown in Figure 2. Only TiO₂ peaks in all catalysts suggests a low loading and a high dispersion of Mn,Te and M elements on the surface of TiO₂. At the same time, compared with pure TiO₂ (industrial grade), the TiO₂ peaks become weak for all the catalyst, which further shows uniform active ingredients with small size for the catalyst can be obtained by impregnation technology. Figure 3 gives the SEM pictures of Mn-Ce-M catalyst. W-doped catalyst reveals uniform particle size from 500 nm to 800 nm as same as Mn-Ce catalyst. Adding Pr into Mn-Ce catalyst lead to slightly agglomeration of the catalyst particle, see the Figure 3(c). The prepared Mn-Ce-Zr and Mn-Ce-Y catalysts show obvious particle agglomeration, in the Figure 3(d) and Figure 3(e). At the same time, metal doping lead to bulk density increase or the porosity decrease of the catalyst. The catalyst morphology may possibly influence its performance, which need be investigated in the following.

The catalytic performance of Mn-Ce-M_0.025 catalysts, expressed as the percent conversion of NO as function of temperature, is shown in Figure 4. The Mn-Ce catalyst behaves above 80% NO conversion at 150˚C - 300˚C and achieved the maximum value with 91% conversion at 200˚C. Compared with Mn-Ce catalyst, Mn-Ce-M catalyst presents the lower NO conversion at 100˚C - 150˚C. Above 200˚C, the Mn-Ce-W catalyst shows the most excellent performance and obtains the 93.2% and 98.4% NO conversion at 200˚C and 250˚C respectively. Pr elements doped Mn-Ce catalysts decrease the catalytic performance relative to Mn-Ce catalyst at all test temperature. At the same time, Zr and Y elements addition into the Mn-Ce catalyst reduce the lower-tempera- ture (100˚C - 200˚C) performance and enhanced NO conversion at the high temperature (300˚C - 400˚C). At 250˚C, the Mn-Ce-Zr catalyst with 92.2% NO conversion is higher than that of Mn-Ce catalyst. The performance of Mn-Ce-Ycatalyst is near to that of Mn-Ce catalyst. These results demonstrate W element efficiently improves the catalyst’s NO conversion and Pr element presents poor action. Zr and Y elements promote Mn-Ce

catalyst activity only at the high temperature. However, at high temperature range (above 250˚C), the NO conversions for Mn-Ce-Pr and Mn-Ce-Zr are near to that of Mn-Ce. According to SEM picture, it is possible that uniform particle size and good porosity tends to present good catalyst, especially at low temperature. The reason of the transition metal effect on catalyst performance will be further investigated in the following.

The doped metal effect on the redox property of Mn-Ce-M_0.025 catalysts was observed by H₂-TPR, see in Figure 5. Use of Mn-Ce catalyst resulted in a broad reduction profile peak in the temperature range of 163˚C - 486˚C,

which was attributed to the continuous reduction steps of MnO₂ → Mn₂O₃ → Mn₃O₄ → MnO [7] . The reduction peak at 553˚C was attributed to a reduction of Ce⁴⁺ to Ce³⁺ on the surface. Above 850˚C, the reduction profile peak was from the reduction of Ce⁴⁺ to Ce³⁺ in the bulk. When W was introduced to the Mn-Ce catalyst, all reduction peaks were connected together and not back to the baseline. Comparing to Mn-Ce catalyst, the initial reduction temperature of Mn-Ce-W catalyst increased, which may be the reason of its lower catalytic performance at low temperature. On the other hand, the reduction peak area of Mn-Ce-W catalyst at 163˚C - 486˚C is larger than that for Mn-Ce catalyst. The larger reduction peak area presents the higher oxygen capacity, so the catalytic performance of Mn-Ce-W catalyst begins to exceed that of Mn-Ce catalyst from 200˚C. At the same time, the new reduction peaks at 639˚C and 804˚C can be attributed to tungsten reduction, which result in a greatly increase of reactive oxygen species on the catalyst surface and an improvement of redox ability of the catalyst. Hence, comparing to the Mn-Ce catalyst, the Mn-Ce-W catalyst presents the more excellent catalytic performance. When Pr and Zr element were introduced to the Mn-Ce catalyst, the reduction peak of manganese oxide was divided to two peaks of MnO₂ or Mn₂O₃ reduced to Mn₃O₄ below 398˚C and Mn₃O₄ to MnO above 398˚C. Two peaks area is lower than that of the Mn-Ce catalyst, which led to lower oxygen capacity. At the same time, the initial reduction temperature of Mn-Ce-Pr and Mn-Ce-Zr catalysts also increased, which indicate low reduction ability at low temperature. These may be two reasons of lower catalytic performance at lower temperature after doping Pr and Zr into the catalyst. At higher temperature, the H₂-TPR profiles of Mn-Ce, Mn-Ce-Pr and Mn-Ce-Zr is similar, so the NO conversion of Mn-Ce-Pr is small lower to that of Mn-Ce and only a little promotion after doping Zr element to Mn-Ce catalyst. For the Mn-Ce-Y catalyst, two independent peaks present at 431˚C and 501˚C attributed to manganese oxide, which obviously shift to high temperature and the peaks area is reduced. The reduction peak area of Ce⁴⁺ to Ce³⁺ is also reduced that may be due to the action between Y and Ce element. Hence, in Mn-Ce-M catalysts, Mn-Ce-Y behaves the lowest NO conversion at the low temperature (below 250˚C). The reduction peak at 748˚C can be attributed to Y₂O₃, which promote catalytic performance of Mn-Ce-Y catalyst at high temperature. At 350˚C and 400˚C, the NO conversion of Mn-Ce-Y is second to that of Mn-Ce-W.

The tolerance to the SO₂ for these prepared SCR catalysts is further investigated. When SO₂ concentration is 0.01%, activities of Mn-Ce-M were performed at 250˚C, as shown in Figure 6. For all the catalysts, after importing SO₂ into the reactant gas, the NO conversion is kept or slightly increases within 50 min and then begins to decrease. The performance improvement for the time being can be attributed to the enhancement of NH3 adsorption on the acid site formed from the SO₂ oxidation. After a period of time, a lot of sulfate were formed and adhered to the catalyst surface which leads to catalytic active decrease. Use of Mn-Ce-Pr and Mn-Ce-Y catalysts, the NO conversion are lower than that of Mn-Ce catalyst at initial stage, but decreased slowly and begin to ex-

ceed that of Mn-Ce catalyst after 150 min in SO₂ condition. After 275 min in SO₂ condition, the NO_x conversion decrease from 88.8% to 73.5% over Mn-Ce catalyst, 87.1% to 77.1% over Mn-Ce-Pr catalyst and 88.7% to 75% over Mn-Ce-Y. The results demonstrated that Pr presents better tolerance to the SO₂ than Y element. Doping W and Zr into the Mn-Ce catalyst not only improve the catalytic performance but also enhance the catalyst’s sulfur tolerance. In all catalysts, Mn-Ce-W catalyst presented the best performance with 88.1% NO conversion in SO₂ condition after 275 min.

In order to study the catalyst sulfur tolerance, the fresh catalyst and deactivated Mn-Ce-M catalysts with SO₂ was carried out by FT-IR spectroscopy to detect the adsorbed species on the Mn-Ce-M catalyst. The FTIR spectra for the fresh Mn-Ce catalyst, deactivated Mn-Ce and Mn-Ce-W catalysts with SO₂ deactivated catalysts with SO₂ are shown in Figure 7. The adsorption peak at 1634 cm⁻¹ band was commonly considered as O-H stretching vibration peak [8] . Comparing to fresh Mn-Ce catalyst, four new bands appeared when the catalysts were exposed to flue gas with SO₂. NH⁴⁺ was adsorbed on catalyst acid site at 1400 cm⁻¹ and three bands between 1048 - 1129 cm⁻¹ were identified as adsorption peaks [9] [10] . The results from the FTIR spectra indicated sulfate formation on the catalysts.

There into, the adsorption peak of NH⁴⁺ was weakest on Mn-Ce catalyst, but the strongest peaks were formed. That is to say, over Mn-Ce catalyst, a little ammonium sulfate or ammonium bisulfate was formed, but a lot of metal sulphates deposited. So, the maximum sulfate formation was observed on Mn-Ce catalyst, which may be the main reason of the catalyst inactivation. Metal doping into the Mn-Ce catalyst could inhibit the sulfate formation, especially the suppression to the deposition of metal sulfate, which could lead to the block of catalyst pores or channels. The weakest peaks presented over Mn-Ce-W catalyst, which proved that W addition could greatly inhibit the sulfate formation. Then Mn-Ce-W catalyst presented the optimal performance in the aspect of sulfur tolerance.