Byproduct Analysis of SO 2 Poisoning on NH 3-SCR over MnFe / TiO 2 Catalysts at Medium to Low Temperatures

The byproducts of ammonia-selective catalytic reduction (NH3-SCR) process over MnFe/TiO2 catalysts under the conditions of both with and without SO2 poisoning were analyzed. In addition to the NH3-SCR reaction, the NH3 oxidation and the NO oxidation reactions were also evaluated at temperatures of 100–300 ◦C to clarify the reactions occurred during the SCR process. The results indicated that major byproducts for the NH3 oxidation and NO oxidation tests were N2O and NO2, respectively, and their concentrations increased as the reaction temperature increased. For the NH3-SCR test without the presence of SO2, it revealed that N2O was majorly from the NH3-SCR reaction instead of from NH3 oxidation reaction. The byproducts of N2O and NO2 for the NH3-SCR reaction also increased after increasing the reaction temperature, which caused the decreasing of N2-selectivity and NO consumption. For the NH3-SCR test with SO2 at 150 ◦C, there were two decay stages during SO2 poisoning. The first decay was due to a certain amount of NH3 preferably reacted with SO2 instead of with NO or O2. Then the catalysts were accumulated with metal sulfates and ammonium salts, which caused the second decay of NO conversion. The effluent N2O increased as poisoning time increased, which was majorly from oxidation of unreacted NH3. On the other hand, for the NH3-SCR test with SO2 at 300 ◦C, the NO conversion was not decreased after increasing the poisoning time, but the N2O byproduct concentration was high. However, the SO2 led to the formation of metal sulfates, which might inhibit NO oxidation reactions and cause the concentration of N2O gradually decreased as well as the N2-selectivity increased.


Introduction
Nitrogen oxides (NO x , NO and NO 2 ) produced from stationary sources are major air pollutants that lead to environmental concerns such as photochemical smog and acid rain [1].The most effective technology for the removal of NO x emission from coal-fired power plants is ammonia-selective catalytic reduction (NH 3 -SCR; SCR hereafter) [2].The traditional SCR catalysts are active within the temperature window of 300-400 • C [3,4].Even though some of the traditional catalyst compositions such as V 2 O 5 -WO 3 / TiO 2 or Fe-zeolite-based catalysts can lower down their working temperature window to be as low as 250 or even 200 • C [5][6][7][8][9][10], there is still a strong demand in developing SCR catalysts to be active at less than 200 • C and placing them downstream of the electrostatic precipitator and desulfurizer [11][12][13].
Literature data showed that Mn-based catalysts have good activity for low-temperature SCR [14 -17].Moreover, in iron containing SCR catalysts, the introduction of Mn could obviously enhance the low-temperature activity, probably due to the fact that synergistic effect between iron and manganese species [18].It was reported that the MnFe/TiO 2 could improve the activity, stability and SO 2 durability of the SCR catalysts using NH 3 as the reducing agent [19][20][21][22][23].
In the past, there have been extensive studies usingFourier-transform infrared spectroscopy (FTIR) for understanding the mechanism of SCR reaction on the surface of MnFe catalysts [24][25][26][27][28].In addition, several different types of reaction mechanisms have been proposed including the typical Eley-Rideal mechanism and Langmuir-Hinshelwood mechanism.For the Eley-Rideal mechanism, it is assumed that the gaseous NO directly reacts with an activated ammonia surface complex [29].On the other hand, the Langmuir-Hinshelwood reaction mechanism involves that a surface NO complex reacts with an activated ammonia [3,30].Moreover, Yang et al. [31] used in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS-FTIR) to reveal the mechanism of low-temperature SCR reaction over the MnFe spinel.The results indicated that the contribution of Eley-Rideal mechanism to NO conversion increased after increasing the reaction temperature.
Although many studies have been done on the SCR mechanism, to the authors' best knowledge, there has been no work on clarifying reactions occurred in the medium to low-temperature SCR system with and without the presence of SO 2 .Therefore, this study employed MnFe/TiO 2 catalyst to study the oxidation of NH 3 , the oxidation of NO and the NH 3 -SCR reaction with and without the presence of SO 2 at 100-300 • C. The reactants and byproducts of gaseous NO, NH 3 , N 2 O, and NO 2 in the effluent streams as well as solid byproducts of ammonium salts and metal sulfates on the catalysts were analyzed.The results can offer useful information to understand the reaction pathway at different operation conditions for the application of medium to low-temperature SCR catalysts.

Oxidation Reactions of NH 3 and NO without SO 2
To understand the products and byproducts of oxidation reactions, the NH 3 and NO oxidation reactions over the MnFe/TiO 2 catalyst were studied within the temperature range of 100-300 • C, which was the most active temperature region for MnFe/TiO 2 catalyst in SCR reaction [32].In addition, the N-balance of NH 3 and NO oxidation tests were calculated by Equations ( 1) and (2), respectively.

N-balance in NH
N-balance in NO oxidation = The results of N-balance and outlet gas concentrations in the NH 3 oxidation reaction without SO 2 over MnFe/TiO 2 catalyst at different reaction temperatures are shown in Figure 1.It can be seen that the major oxidation product of NH 3 was N 2 O (i.e., Equation (3) shown later in the Materials and Method section) instead of NO (Equation ( 4)).In addition, N 2 O increased from 31 ppm to 219 ppm after increasing the reaction temperature from 100 • C to 300 • C. On the other hand, the effluent NO concentration was only 36 ppm at most, which occurred at reaction temperature of 300 • C; but no NO 2 was found in the effluent at all tested temperatures.
The results of N-balance and outlet gas concentrations in the NO oxidation reaction over MnFe/TiO 2 catalyst at different reaction temperatures are shown in Figure 2, it can be seen that NO could be oxidized to NO 2 , and the concentration of NO 2 was significantly increased from 10 to 309 ppmv for reaction temperatures from 100 to 300 • C.
The above results indicated that NH 3 and NO oxidations (Equations (3)-( 5)) increased after increasing the reaction temperature, and the major products of NH 3 oxidation and NO oxidation were N 2 O and NO 2 , respectively.In addition, it can be observed that both the N-balance results shown in Figures 1 and 2 were very high at 97~100% in the 100-300 • C range.This indicates that we can detect almost all the reaction species.

SCR Reactions without SO 2
Figure 3 shows the NO consumption (in ppmv) and outlet gas concentrations of N 2 O, NO and NH 3 over the MnFe/TiO 2 catalyst for reaction temperatures of 100-300 • C. It can be seen that the MnFe/TiO 2 catalyst maintained high NO consumption in the SCR reaction.When raising reaction temperature from 100 to 200 • C, the NO consumption in the SCR reaction increased slightly from 450 ppm to 495 ppm, and the outlet concentration of NH 3 was slightly decreased from 63 to 0 ppm.
At the reaction temperature of 200 • C, the concentrations of N 2 O and NO 2 were 175 and 0 ppm, respectively, with 98% of NO consumption in the SCR reactions (500 ppm NO reacted with 500 ppm NH 3 ).On the other hand, the N 2 O from NH 3 oxidation (500 ppm NH 3 reacted with O ) at the reaction temperature of 200 • C was 156 ppm (as indicated by Figure 1).Therefore, it can be seen that in the SCR system, lower N 2 O percentage ( 175ppm N 2 O 500 ppm NO + 500 ppm NH 3 = 17.5%) was being produced as compared to that during only NH 3 oxidation ( 156 ppm N 2 O 500 ppm NH 3 = 31.2%).
temperature from 100 to 200 °C, the NO consumption in the SCR reaction increased slightly from 450 ppm to 495 ppm, and the outlet concentration of NH3 was slightly decreased from 63 to 0 ppm.
At the reaction temperature of 200 °C, the concentrations of N2O and NO2 were 175 and 0 ppm, respectively, with 98% of NO consumption in the SCR reactions (500 ppm NO reacted with 500 ppm NH3).On the other hand, the N2O from NH3 oxidation (500 ppm NH3 reacted with O2) at the reaction temperature of 200 °C was 156 ppm (as indicated by Figure 1).Therefore, it can be seen that in the SCR system, lower N2O percentage ( The "fast SCR" (Equation ( 8)), first proposed in 1986 [33], proceeds at a much higher reaction rate than "standard SCR" reactions (Equations ( 6) and ( 7)) was developed to improve deNOx efficiency at lower temperatures [34,35].As indicated by Figure 2 for the NO oxidation reaction, NO could be oxidized to NO2 at temperatures higher than 150 °C, but NO2 was not detected at temperature below 200 °C in the SCR test as seen in Figure 3.This might be due to the fact that NO2 would react with NO and NH3 according to the fast SCR reaction (Equation (8)).The "fast SCR" (Equation ( 8)), first proposed in 1986 [33], proceeds at a much higher reaction rate than "standard SCR" reactions (Equations ( 6) and ( 7)) was developed to improve deNO x efficiency at lower temperatures [34,35].As indicated by Figure 2 for the NO oxidation reaction, NO could be oxidized to NO 2 at temperatures higher than 150 • C, but NO 2 was not detected at temperature below 200 • C in the SCR test as seen in Figure 3.This might be due to the fact that NO 2 would react with NO and NH 3 according to the fast SCR reaction (Equation ( 8)).
After further raising the reaction temperature from 200 to 300 • C, NO consumption in the SCR reaction was slightly decreased from 495 ppm to 455 ppm, and concentrations of N 2 O and NO 2 increased from 175 to 270 ppm and 0 to 36 ppm, respectively.When reaction temperature increased, oxidation reactions occurred more quickly as indicated by Figures 1 and 2. Hence the reason for slight decreases in the NO consumption at above 200 • C might be due to the fact that NO was oxidized (Equation ( 5)) as also demonstrated in the literature [26,[36][37][38].Moreover, the results of Figure 3 also indicated that products of the SCR reaction gradually changed from N 2 to N 2 O when raising the temperature from 100 to 300 • C. Therefore, the N 2 -selectivity of SCR reaction was gradually decreased from 93% to 41% after increasing the reaction temperature from 100 • C to 300 • C.

SCR Reactions with SO 2
To clarify the byproducts of SCR reactions with SO 2 , NO consumption in the SCR reaction and outlet gas concentrations in the NH 3 -SCR reaction over MnFe/TiO 2 catalyst at 150 and 300 • C are shown in Figure 4a,b, respectively.It can be seen in Figure 4a that there were four stages during the SCR test period.At stage I where no SO 2 was introduced, the NO consumption in the SCR reaction was very high (495 ppm out of 500 ppm).After 150 ppm SO 2 was introduced, the NO consumption in the SCR reaction decreased rapidly within 60 min SO 2 poisoning (stage II), and they remained roughly stable for another 60 min (stage III), then decreased gradually with time again (stage IV).
form ammonium salts.However, the MnFe/TiO2 catalyst still had good activity at stage III.Thus, during this stage, there was no sufficient NH3 to react with NO, which caused NO consumption remained relatively lower but stable at around 425 ppm from 60 to 120 min SO2 poisoning time as compared to stages I and II.At stage IV, NO consumption in the SCR reaction decreased significantly from 420 ppm to 153 ppm for poisoning time from 120 to 360 min, which was due to accumulation of metal sulfates and ammonium salts which blocked the active sites of catalyst.On the other hand, no effluent NH3 was detected from stage I to stage III.Then NH3 slip occurred at stage IV as seen in Figure 4a, which was due to less active sites and decreasing NO consumption in the SCR reaction.One can also see that concentrations of N2O slightly decreased from 76 to 63 ppm within 120 min SO2 poisoning time (stage II and III), then increased gradually to 86 ppm after 360 min SO2 poisoning time (stage IV).This result indicated that SO2 would react with NH3 and form ammonium salts, which cause concentrations of N2O slightly decreased at stages II and III.Then NH3 slip occurred at stage IV, which caused increasing N2O concentrations.
For the SCR test with SO2 at 300 °C as shown in Figure 4b, it can be seen that the SO2 poisoning effect was negligible as compared to the poisoning test at low temperature.When the reaction temperature was at 300 °C, the NO consumption in the SCR reaction remained around 460 ppm, which was almost the same as those without SO2 poisoning.Moreover, during the SCR activity tests no SO2 concentration was detected, which indicated that all the gas phase SO2 molecules might be adsorbed and/or reacted with metal catalysts.
The main reason for the inhibition of SO2 poisoning might be attributed to different SCR reaction mechanisms at different temperatures.Literature results showed that at lower temperatures (<200 °C), SCR reactions would follow the Langmuir−Hinshelwood mechanism.In the Langmuir−Hinshelwood mechanism, SO2 would compete with NO to be adsorbed on the active sites, which cause the decreases of SCR efficiencies.However, when reaction temperature increased, the SCR reaction mechanism would transform from Langmuir−Hinshelwood mechanism to Eley-Rideal mechanism.In our previous work [32], it was found that the gradual decrease (stage II) of NO consumption in the SCR reaction was probably due to the fact that SO 2 competed with NO to react with NH 3 and form ammonium salts.However, the MnFe/TiO 2 catalyst still had good activity at stage III.Thus, during this stage, there was no sufficient NH 3 to react with NO, which caused NO consumption remained relatively lower but stable at around 425 ppm from 60 to 120 min SO 2 poisoning time as compared to stages I and II.At stage IV, NO consumption in the SCR reaction decreased significantly from 420 ppm to 153 ppm for poisoning time from 120 to 360 min, which was due to accumulation of metal sulfates and ammonium salts which blocked the active sites of catalyst.
On the other hand, no effluent NH 3 was detected from stage I to stage III.Then NH 3 slip occurred at stage IV as seen in Figure 4a, which was due to less active sites and decreasing NO consumption in the SCR reaction.One can also see that concentrations of N 2 O slightly decreased from 76 to 63 ppm within 120 min SO 2 poisoning time (stage II and III), then increased gradually to 86 ppm after 360 min SO 2 poisoning time (stage IV).This result indicated that SO 2 would react with NH 3 and form ammonium salts, which cause concentrations of N 2 O slightly decreased at stages II and III.Then NH 3 slip occurred at stage IV, which caused increasing N 2 O concentrations.
For the SCR test with SO 2 at 300 • C as shown in Figure 4b, it can be seen that the SO 2 poisoning effect was negligible as compared to the poisoning test at low temperature.When the reaction temperature was at 300 • C, the NO consumption in the SCR reaction remained around 460 ppm, which was almost the same as those without SO 2 poisoning.Moreover, during the SCR activity tests no SO 2 concentration was detected, which indicated that all the gas phase SO 2 molecules might be adsorbed and/or reacted with metal catalysts.
The main reason for the inhibition of SO 2 poisoning might be attributed to different SCR reaction mechanisms at different temperatures.Literature results showed that at lower temperatures (<200 • C), SCR reactions would follow the Langmuir−Hinshelwood mechanism.In the Langmuir−Hinshelwood mechanism, SO 2 would compete with NO to be adsorbed on the active sites, which cause the decreases of SCR efficiencies.However, when reaction temperature increased, the SCR reaction mechanism would transform from Langmuir−Hinshelwood mechanism to Eley-Rideal mechanism.In the high temperature range (>200 • C), the SCR reaction mainly followed the Eley-Rideal mechanism, over which the gaseous NO could directly react with an activated ammonia [30,39,40].
One can also see from Figure 4b that at reaction temperature of 300 • C, the SCR system did not have NH 3 slip.The concentrations of NO 2 and N 2 O decreased from 30 to 0 ppm and from 270 to 172 ppm, respectively; and the N 2 -selectivity increased from 41% to 62% for the SCR reaction from stage II to stage IV.The result indicated that by increasing the reaction temperature to 300 • C the SO 2 poisoning effect could be inhibited, and the N 2 selectivity can be enhanced due to the decreased N 2 O concentrations.
The amount of sulfate species on the catalysts was estimated by thermo-gravimetric analysis (TGA) analysis and the results are shown in Figure 5 in terms of differential thermogram (DTG) spectra.The weight loss profiles of all samples showed three distinct decomposition steps: (1). the weight loss at low temperature (<200 • C) was assigned to the water desorption on the catalyst surface.(2). the weight loss at 200-400 • C could be attributed to decomposition of ammonium salts [41][42][43].(3). the weight loss at high temperature (>670 • C) was originated from metal sulfates [44,45].
Catalysts 2018, 8, x FOR PEER REVIEW 7 of 17 poisoning effect could be inhibited, and the N2 selectivity can be enhanced due to the decreased N2O concentrations.
The amount of sulfate species on the catalysts was estimated by thermo-gravimetric analysis (TGA) analysis and the results are shown in Figure 5 in terms of differential thermogram (DTG) spectra.The weight loss profiles of all samples showed three distinct decomposition steps: (1). the weight loss at low temperature (<200 °C) was assigned to the water desorption on the catalyst surface.
(3). the weight loss at high temperature (>670 °C) was originated from metal sulfates [44,45].It can be seen from Figure 5 that the fresh catalyst had only one major weight loss peak, which appeared at 50-150 °C and corresponded to H2O desorption on the catalyst surface.On the other hand, decomposition peaks of both ammonium salts (200-400 °C) and metal sulfates (670-900 °C) were observed on all MnFe/TiO2 catalysts poisoned at temperature ranges of 100-200 °C.When reaction temperatures were 250 °C and 300 °C, there were no decomposition peak of ammonium salts.This reveals that ammonium salts were not formed on the catalyst surface at temperatures above It can be seen from Figure 5 that the fresh catalyst had only one major weight loss peak, which appeared at 50-150 • C and corresponded to H 2 O desorption on the catalyst surface.On the other hand, decomposition peaks of both ammonium salts (200-400 • C) and metal sulfates (670-900 • C) were observed on all MnFe/TiO 2 catalysts poisoned at temperature ranges of 100-200 • C. When reaction temperatures were 250 • C and 300 • C, there were no decomposition peak of ammonium salts.This reveals that ammonium salts were not formed on the catalyst surface at temperatures above 250 • C. As also noted in Figure 5, there was another peak in the range of temperature of 400-600 • C for the fresh catalyst and poisoned catalyst at 100 • C.However, since we cannot find related literature discussed on this, so the reason for this peak is not clear.
The amounts of sulfate species of fresh and poisoned catalysts are listed in Table 1.It is observed that the amounts of ammonium salts on the catalysts decreased from 2.3 wt.% to negligible amounts by increasing the SCR reaction temperature from 100 to 300 • C.This indicated that the reaction temperature would directly affect the formation of ammonium salts.It is noted that rigid quantification of metal sulfates accumulated on the poisoned catalyst was not possible via the DTG data because some ammonium salts could also be transformed into metal sulfates during the continuous TGA heating process [46].Hence we can only confirm that all the poisoned catalysts had roughly similar amounts of metal sulfates during the temperature from 100 • C to 250 • C (within experimental error of TGA instrument, ±0.3 wt.%), except for the case at 300 • C. a The amount of ammonium salts was calculated by weight difference between the fresh and poisoned catalysts from the TGA spectrum of 200-400 • C. b The amount of metal sulfates was calculated by weight difference between the fresh and poisoned catalysts from the TGA spectrum of 670-900 • C. c Rigid quantification of the amounts of metal sulfates is not possible via the DTG data because it was possible that some of the ammonium salts could transform into metal sulfates during the heating process of TGA.
At high temperature of 300 • C, the NO consumption was not affected by SO 2 .Thus, it is easy to predict that adding an excessive amount of NH 3 tends to be oxidized and forming N 2 O and NO as observed in Figure 1.However, at low temperature of 150 • C, the NO consumption was significantly affected by SO 2 .Thus, to ensure the gradual decrease of NO consumption at stage II of Figure 4a was due to the competition between SO 2 and NO to react with NH 3 , different inlet amounts of NH 3 were tested to study the SO 2 poisoning mechanism at temperature of 150 • C. The results are shown in Figure 6.One can see that adding different amounts of NH 3 had a similar effect on NO consumption at stages I and IV; but it had different NO consumptions at stages II and III.On the other hand, the NH 3 outlet concentrations were different at all stages.At stage I, NH 3 only reacted with a certain amount of NO and thus extra NH 3 slip was detected in the outlet gas.When SO 2 was introduced, NH 3 not only reacted with NO but also reacted with SO 2 .Therefore, both the outlet concentrations of NH 3 and NO consumption decreased at stages II and III.Besides, when increasing NH 3 amount to above 550 ppm (i.e., NH 3 /NO molar ratio of 1.1), the first decay of NO consumption at stage II could be inhibited.This indicated that NH 3 concentration was sufficient to react with NO and SO 2 .The result shown in Figure 6 indicated that if a sufficient amount of NH 3 can be provided to react with both SO 2 and NO, then the first decay of NO consumption could be inhibited at early time of stage II.However, after 120 min of SO 2 poisoning (stage IV), the NO consumption could not be affected by different amounts of NH 3 .Thus, the more injection amount of NH 3 led to eventually the more amount of NH 3 slip to the atmosphere.

Product and Byproduct Analysis
Since different reactions occurred during the SCR process, i.e., Equations (3)-( 12), were considered both with and without SO2, thus the percentages of product and byproducts can be calculated.Table 2 lists formulas for calculating the percentages of N-containing product (N2) as well as gaseous and solid byproducts (NO, N2O and salts) formed during the SCR process.The byproducts could be formed by the SCR reaction as well as by the NH3 oxidation or the NO oxidation reactions.Because it is difficult to clarify that the outlet NO was from NH3 oxidation or from the unreacted NO, therefore the NO formation from NH3 oxidation (Equation ( 4)) is neglected and all the effluent NO was assumed to be only from the unreacted NO.This is an acceptable assumption since the formation of NO from NH3 oxidation was very minor (0~7%) at reaction temperatures of 100~300 °C as observed from Figure 1.Besides, it was assumed that the fast SCR reaction (Equation ( 8)) only served as the intermediate reaction at the low temperature SCR process [35].Thus, the fast SCR reaction was not considered in the calculation of the percentages of all N-containing product and byproducts.Table 2. Formulas for calculating the N-containing product and byproduct percentages during the SCR process.

NO2
NO oxidation to NO2 (Equation ( 5)): The percentage of NO2 from NO oxidation

Product and Byproduct Analysis
Since different reactions occurred during the SCR process, i.e., Equations (3)-( 12), were considered both with and without SO 2 , thus the percentages of product and byproducts can be calculated.Table 2 lists formulas for calculating the percentages of N-containing product (N 2 ) as well as gaseous and solid byproducts (NO, N 2 O and salts) formed during the SCR process.The byproducts could be formed by the SCR reaction as well as by the NH 3 oxidation or the NO oxidation reactions.Because it is difficult to clarify that the outlet NO was from NH 3 oxidation or from the unreacted NO, therefore the NO formation from NH 3 oxidation (Equation ( 4)) is neglected and all the effluent NO was assumed to be only from the unreacted NO.This is an acceptable assumption since the formation of NO from NH 3 oxidation was very minor (0~7%) at reaction temperatures of 100~300 • C as observed from Figure 1.Besides, it was assumed that the fast SCR reaction (Equation ( 8)) only served as the intermediate reaction at the low temperature SCR process [35].Thus, the fast SCR reaction was not considered in the calculation of the percentages of all N-containing product and byproducts.
Based on formulas shown in Table 3, the results on percentages of all N-containing species during the NH 3 -SCR process tested at 150 • C are shown in Figure 7a; and those tested at 300 • C are shown in Figure 7b.It can be seen from Figure 7a that at SCR operation temperature of 150 • C and without the presence of SO 2 , the major product of SCR process appeared to be N 2 .This indicated that MnFe/TiO 2 catalyst can serve as a good catalyst and achieve high N 2 selectivity at low temperature of 150 • C when SO 2 was not presented in the system.The minor presence of N 2 O byproduct in the exhaust was majorly from the SCR reaction rather than from the NH 3 oxidation reaction.
Table 2. Formulas for calculating the N-containing product and byproduct percentages during the SCR process.

Product and Byproduct Reaction Equation Percentage Calculation Equation
NO 2 NO oxidation to NO 2 (Equation ( 5)): The percentage of NO 2 from NO oxidation Ammonium salts Equations ( 10) and ( 11): The percentage of salts were indicated by result of Figure 6 N 2 O The percentage of N 2 O from NH 3 slip oxidation [NH3] : from results of NH 3 oxidation test (Figure 1) SCR reaction to form N 2 O Equation ( 7):

SCR reaction to form N 2
Equation ( 6): The percentage of N 2 from SCR reaction On the other hand, when SO 2 was introduced at 150 • C, the percentages of N 2 and N 2 O decreased after increasing the poisoning time as observed in Figure 7a.Because NH 3 could not be reacted with NO due to decreased availability of active sites, it tends to increase the NO and NH 3 slip to the atmosphere.In addition, the portion of N 2 O from NH 3 oxidation was also gradually increased as increasing the SO 2 poisoning time.After 360 min of SO 2 poisoning, the exhaust N 2 O from NH 3 oxidation was more than from SCR reaction.One can also see that percentages of ammonium salts were similar at different SO 2 poisoning times.This indicated that certain amounts of NH 3 would preferentially be reacted with SO 2 .In addition, the remaining NH 3 would then be reacted with NO or O 2 .Moreover, it can be seen that percentages of NO slip were higher than percentages of NH 3 slip at different SO 2 poisoning times, which is due to the fact that NH 3 not only reacted with NO but also reacted with SO 2 and O 2 .
For results at 300 • C as seen in Figure 7b, it is observed that NO 2 from NO oxidation disappeared during the SO 2 poisoning.However, the unreacted NO (NO slip) appeared during the SO 2 poisoning.This indicated that the formation of metal sulfates at 300 • C might inhibit the NO oxidation reaction.At high temperature, the SCR reaction mainly followed the Eley-Rideal mechanism, so the continuous decreasing of N 2 O concentration as poisoning time increases and no SO 2 concentration in the outlet gases might be related to the reaction between SO 2 and activated NH 3 instead of the NH 3 oxidation reaction at 300 • C.Moreover, the total percentages of N 2 and N 2 O from SCR reactions remained almost the same no matter SO 2 was presented in the system or not.The low N 2 selectivity revealed that MnFe/TiO 2 catalyst may not be a good candidate for SCR process at 300 • C unless the space velocity can be reduced to further enhance the more complete reduction of NO to N 2 instead of forming the N 2 O byproduct.However, it is interested to note that the N 2 selectivity gradually increased and the percentages of N 2 O from SCR reaction gradually decreased after increasing SO 2 poisoning time.This indicated that SO 2 promotion phenomenon might exist at 300 • C, which was attributed to the formation of SO shown in Figure 7b.It can be seen from Figure 7a that at SCR operation temperature of 150 °C and without the presence of SO2, the major product of SCR process appeared to be N2.This indicated that MnFe/TiO2 catalyst can serve as a good catalyst and achieve high N2 selectivity at low temperature of 150 °C when SO2 was not presented in the system.The minor presence of N2O byproduct in the exhaust was majorly from the SCR reaction rather than from the NH3 oxidation reaction.3.
On the other hand, when SO2 was introduced at 150 °C, the percentages of N2 and N2O decreased after increasing the poisoning time as observed in Figure 7a.Because NH3 could not be reacted with NO due to decreased availability of active sites, it tends to increase the NO and NH3 slip to the atmosphere.In addition, the portion of N2O from NH3 oxidation was also gradually increased as increasing the SO2 poisoning time.After 360 min of SO2 poisoning, the exhaust N2O from NH3 oxidation was more than from SCR reaction.One can also see that percentages of ammonium salts were similar at different SO2 poisoning times.This indicated that certain amounts of NH3 would  3.

Reaction Pathways
From the above results, one could surmise the reaction pathway in the NH 3 -SCR system with/without SO 2 at 150 • C and 300 • C as shown in Figure 8.When the SCR system was at 150 • C without SO 2 , it can be seen from the top left plot that a fraction of NH 3 and NO would be oxidized to N 2 O and NO 2 , respectively.In addition, the major reaction product of SCR reaction was N 2 instead of N 2 O. On the other hand, when increasing temperature to 300 • C, it can be seen from the bottom left plot that NH 3 would be oxidized to both N 2 O and NO.Moreover, the major reaction product of SCR reaction was N 2 O instead of N 2 , which was revealed by the low N 2 -selectivity as seen in Figure 7b.
promoted NH3 reaction with NO via Eley−Rideal mechanism.

Reaction Pathways
From the above results, one could surmise the reaction pathway in the NH3-SCR system with/without SO2 at 150 °C and 300 °C as shown in Figure 8.When the SCR system was at 150 °C without SO2, it can be seen from the top left plot that a fraction of NH3 and NO would be oxidized to N2O and NO2, respectively.In addition, the major reaction product of SCR reaction was N2 instead of N2O.On the other hand, when increasing temperature to 300 °C, it can be seen from the bottom left plot that NH3 would be oxidized to both N2O and NO.Moreover, the major reaction product of SCR reaction was N2O instead of N2, which was revealed by the low N2-selectivity as seen in Figure 7b.When SO2 was introduced at 150 °C, it can be seen from the top right plot of Figure 8 that SO2 would be reacted with both NH3 and metal catalyst, which resulted in the formation of ammonium salts and metal sulfates, respectively.In addition, because NO could not be adsorbed on metal sulfates, therefore it could not be reacted with NH3 at low temperature.As a result, unreacted NH3 (NH3 slip) turned out to be the major N-containing species in addition to the unreacted NO.
When increasing temperature to 300 °C, it can be seen from the bottom right plot that ammonium salts would not be formed in the presence of SO2, but SO2 would react with metal catalyst to form metal sulfates.In addition, gaseous NO could directly react with adsorbed ammonia via Eley−Rideal mechanism [29].The major reaction product of SCR reaction gradually changed from N2O to N2 after increasing poisoning time.

NH3 oxidation:
2NH3 + 2O2 → N2O + 3H2O (3) When SO 2 was introduced at 150 • C, it can be seen from the top right plot of Figure 8 that SO 2 would be reacted with both NH 3 and metal catalyst, which resulted in the formation of ammonium salts and metal sulfates, respectively.In addition, because NO could not be adsorbed on metal sulfates, therefore it could not be reacted with NH 3 at low temperature.As a result, unreacted NH 3 (NH 3 slip) turned out to be the major N-containing species in addition to the unreacted NO.
When increasing temperature to 300 • C, it can be seen from the bottom right plot that ammonium salts would not be formed in the presence of SO 2 , but SO 2 would react with metal catalyst to form metal sulfates.In addition, gaseous NO could directly react with adsorbed ammonia via Eley−Rideal mechanism [29].The major reaction product of SCR reaction gradually changed from N 2 O to N 2 after increasing poisoning time.

Reactions in SCR System
In the SCR system, it may contain oxidation reactions, SCR reactions and SO 2 poisoning reactions [46][47][48][49][50][51][52] In this study, experimental tests were designed to clarify the products and byproducts of the above reactions.

Synthesis of MnFe/TiO 2 Catalysts
Mn and Fe metal oxides were supported on TiO 2 (in the form of TiO(OH) 2 ) by the co-precipitation method.In a typical procedure, 8 g of TiO 2 (China Steel Corp., Kaohsiung, Taiwan), 11.57g of ferric nitrate 9-hydrate (99%,J.T. Baker, Radnor, PA, USA), 7.13 g of manganese (II) acetate tetrahydrate (99%, Merck, Kenilworth, NJ, USA) and D.I. water (76 g) were mixed then adjusted to pH = 10 with 25 wt.% ammonia solution to form a precipitate.It was filtered and washed thoroughly with D.I. water, then dried at 120 • C for 12 h.Finally, the material was calcined at 350 • C for 6 h in air.

Catalyst Reaction
The NH 3 oxidation, NO oxidation and SCR activity tests were carried out at atmospheric pressure in a fixed-bed reactor loaded with sieved pelletized (16-30 mesh) catalysts.The operation conditions of inlet gas concentrations, reaction temperatures, and gas hourly space velocity for different tests are shown in Table 3.Under typical NH 3 oxidation, NO oxidation, and SCR reaction tests, the concentrations of NH 3 and/or NO were the same at 500 ppmv, and the SO 2 concentration was 150 ppmv if SO 2 poisoning effect was considered.The feed gases were mixed in a gas mixer.Then the catalysts were preheated in the reactor for 30 min to ensure that an isothermal reaction temperature was reached.During the oxidation and SCR test, the NO and SO 2 concentrations at the inlet and outlet of the reactor were monitored by a NO/SO 2 analyzer (Ultramat 23, SIEMENS, Munich, Germany).In addition, the concentrations of NH 3 , N 2 O, and NO 2 at the inlet and outlet of the reactor were monitored by a FTIR Spectrophotometer (Bomem MB 104,San Jose, CA, USA and ITRI, Hsinchu, Taiwan).The NO consumption due to SCR reactions (Equations ( 6)-( 8)) must be subtracted by the NO oxidation to NO 2 (Equation ( 5)).Thus, the NO consumption of SCR, [NO] SCR is defined by: [NO] SCR = [NO] in − ([NO] out − [NO 2 ] out ) (13)

FreshFigure 5 .
Figure 5. DTG spectra of fresh catalyst and catalysts poisoned at different reaction temperatures after 6 h poisoning time.

Figure 5 .
Figure 5. DTG spectra of fresh catalyst and catalysts poisoned at different reaction temperatures after 6 h poisoning time.

Figure 7 .
Figure 7. Product and byproduct percentages of N-containing species during the SCR process over MnFe/TiO2 catalyst with SO2 poisoning at (a) 150 °C and (b) 300 °C.The inlet NH3 and NO molar concentration ratio was 1:1; and the percentage calculation formulas were based on those listed in Table3.

Figure 7 .
Figure 7. Product and byproduct percentages of N-containing species during the SCR process over MnFe/TiO 2 catalyst with SO 2 poisoning at (a) 150 • C and (b) 300 • C. The inlet NH 3 and NO molar concentration ratio was 1:1; and the percentage calculation formulas were based on those listed in Table3.

Figure 8 .
Figure 8. Proposed reaction pathway in the NH3-SCR system with/without SO2 at 150 °C and 300 °C.The reaction pathway was based on inlet NH3 and NO molar concentration ratio of 1:1.

Figure 8 .
Figure 8. Proposed reaction pathway in the NH 3 -SCR system with/without SO 2 at 150 • C and 300 • C. The reaction pathway was based on inlet NH 3 and NO molar concentration ratio of 1:1.

Table 1 .
Amounts of ammonium salts and metal sulfates deposited on the catalyst surfaces.
4  2−on the catalyst surface.This increased NH 3 adsorption and promoted NH 3 reaction with NO via Eley−Rideal mechanism.

Table 3 .
Operation conditions of experimental tests in this study.