The Study of C3H6 Impact on Selective Catalytic Reduction by Ammonia (NH3-SCR) Performance over Cu-SAPO-34 Catalysts

In present work, the catalytic performance of Cu-SAPO-34 catalysts with or without propylene during the NH3-SCR process was conducted, and it was found that the de-NOx activity decreased during low temperature ranges (<350 °C), but obviously improved within the range of high temperatures (>350 °C) in the presence of propylene. The XRD, BET, TG, NH3-TPD, NOx-TPD, in situ DRIFTS and gas-switch experiments were performed to explore the propylene effect on the structure and performance of Cu-SAPO-34 catalysts. The bulk characterization and TG results revealed that neither coke deposition nor the variation of structure and physical properties of catalysts were observed after C3H6 treatment. Generally speaking, at the low temperatures (<350 °C), active Cu2+ species could be occupied by propylene, which inhibited the adsorption and oxidation of NOx species, confining the SCR reaction rate and causing the deactivation of Cu-SAPO-34 catalysts. However, with the increase of reaction temperatures, the occupied Cu2+ sites would be recovered and sequentially participate into the NH3-SCR reaction. Additionally, C3H6-SCR reaction also showed the synergetic contribution to the improvement of NOx conversion at high temperature (>350 °C).


Introduction
Regardless of diesel or lean-burn gasoline engine exhaust, NO x fouling has induced severe environmental problems [1][2][3] and the elimination of NO x is still a big challenge under stringent emission regulations [4]. Selective catalytic reduction of NO x with NH 3 (NH 3 -SCR) is considered as the most potential de-NO x technology for engine exhaust emission control [5,6]. In the past decades, V 2 O 5 -WO 3 /TiO 2 SCR catalyst has been widely used in the commercial de-NO x application [7,8]. However, the main drawbacks of these catalysts for mobile source domains are their insufficient low temperature SCR activity, poor thermal stability, and latent toxicity [9]. Compared with V 2 O 5 -WO 3 /TiO 2 , ZSM-5 and Beta zeolite-based catalysts, generally promoted by Cu or Fe ions, have been intensively studied for NO x elimination [10,11] and they were considered as a potential substitute for vanadium-based catalysts due to their competitive activity and thermal stability [12]. However, their durability under hydrothermal aging condition is less than optimal [13][14][15]. In recent years, the Cu supported Chabazite catalyst (e.g., Cu-SAPO-34 or Cu-SSZ-13) has been reported as a leading candidate of the NH 3 -SCR catalyst, owing to its high catalytic activity [16,17] and superior stability [18]. Furthermore, extensive researches have focused on the following topics, such as active Cu species. Especially based on the paper published by Xue et al., it was shown that four copper species in Cu/SAPO-34 samples were identified, including external surface copper oxide clusters, isolated Cu 2+ , nanosized copper oxide crystallites, and Cu + species. It was indicated that the isolated Cu 2+ species were responsible for the active sites for the NH 3 -SCR reaction over the Cu/SAPO-34 catalyst (100-200 • C) [19,20], acidity [21], reaction intermediates [22], reaction mechanism [23][24][25][26], formation and decomposition of the by-products N 2 O [27][28][29][30][31], hydrothermal stability [32,33], and SO x poisoning over Cu-SAPO-34 [34,35].
Although Cu-SAPO-34 has been widely reported in previous studies as a powerful potential application, another issue that needs to be addressed is the impact of hydrocarbons on SCR activity during cold startup or inactivation of upstream TWC catalysts, especially during the passive-SCR process [36], since the hydrocarbons in exhaust could slip and interact with downstream SCR catalyst. Even if the HCs-SCR could be a possible pathway for NO x elimination, the hydrocarbons (e.g., C 3 H 6 ) were still considered as poisons for NH 3 -SCR catalyst [37][38][39]. Li et al. illustrated that SCR performance of Fe-zeolites catalysts (Fe-BEA, Fe-ZSM-5, and Fe-MOR) was obviously inhibited in the presence of propylene [40,41]. The origin of C 3 H 6 inhibition on activity over Cu-ZSM-5 and Fe-SZM-5 catalysts was the competitive adsorption of NH 3 and C 3 H 6 on catalyst sites [42]. Meanwhile, another deactivation mechanism resulted from hydrocarbons was the blocking of active sites by carbonaceous species, which limited the overall SCR reaction rate [41]. Furthermore, Ma et al. stated that competitive adsorption of NO x and C 3 H 6 in the low temperature ranges and the coke formation on active sites during medium temperature ranges were critical for the deactivation of the Cu-SSZ-13 catalyst [36]. Conversely, Wang et al. reported that the improved NO x removal efficiency resulted from the presence of hydrocarbons such as C 3 H 6 in NH 3 -SCR gas stream for the commercial Cu-CHA catalyst in LNT-SCR configuration [43,44]. Meanwhile, Z. Gao et al. illustrated that the presence of CuO in Cu-SAPO-34 catalysts would accelerate the C 3 H 6 oxidation process and improve the C 3 H 6 conversions [45]. Furthermore, D. Yang et al. reported that Fe-SAPO-34 catalyst exhibited attractive C 3 H 6 -SCR catalytic performance at 250-350 • C in the presence of 5% H 2 O and 10% O 2 [46]. In a nutshell, the propylene impact on NH 3 -SCR performance during a different temperature range and the mechanism behind the catalyst's deactivation [47][48][49][50][51] is still controversial.
In the present work, the impact of C 3 H 6 on NO conversion of Cu-SAPO-34 catalyst and the C 3 H 6 influence mechanism were investigated at different temperature ranges. XRD, TG, gas-switch tests, and the temperature programmed desorption (TPD) of NH 3 (NO x ) with or without propylene were conducted to identify the influence of propylene on Bronsted acid sites, active sites, and reaction intermediates. Finally, in situ diffuse reflectance infrared Fourier transform spectroscopy (in situ DRIFTS) was employed to probe into the C 3 H 6 effect mechanism on NH 3 -SCR performance.

Structure and Physical Properties of Cu-SAPO-34 Catalysts
For XRD patterns of fresh Cu-SAPO-34 and HCs-Cu-SAPO-34 in the Figure 1, no significant change of CHA crystal structure was observed before and after propylene treatment and any detected crystal variation was not found through the above XRD patterns. Namely, it implied that the CHA structure of Cu-SAPO-34 barely changed after 1000 ppm, 6h C 3 H 6 treatment at 300 • C.
The results of BET results of the fresh Cu-SAPO-34 and HCs-Cu-SAPO-34 were listed in Table 1. Compared to the fresh Cu-SAPO-34, the BET surface area of HCs-Cu-SAPO-34 still remained at 415 m 2 /g, which was comparable with the fresh catalyst. It also suggested that the C 3 H 6 treatment scarcely damaged the porous structure of catalysts. The results of BET results of the fresh Cu-SAPO-34 and HCs-Cu-SAPO-34 were listed in Table 1. Compared to the fresh Cu-SAPO-34, the BET surface area of HCs-Cu-SAPO-34 still remained at 415 m 2 /g, which was comparable with the fresh catalyst. It also suggested that the C3H6 treatment scarcely damaged the porous structure of catalysts.

Catalysts
Fresh Cu-SAPO-34 HCs-Cu-SAPO-34 BET specific surface area (m 2 /g) 417 415 Figure 2a,b demonstrated the TG differential profiles and TG mass loss profiles of fresh Cu-SAPO-34 and HCs-Cu-SAPO-34 catalysts. As shown in the Figure 2a, two obvious mass losses were observed around 100 °C and 250 °C. Specifically, the main mass variation at 100 °C and the minor mass loss at 250 °C was deemed to derive from the dehydration process of adsorbed water on Cu-SAPO-34 samples [27]. Furthermore, it was calculated from Figure 2b that the H2O mass losses were 13%, 5% for fresh Cu-SAPO-34, and 12%, 4% for HCs-Cu-SAPO-34, respectively. Importantly, no distinct mass loss was discovered at high temperature above 400 °C, which usually was considered as the essential temperature for coke removal on catalysts. Consequently, it indicated that the possibility of coke formation on Cu-SAPO-34 catalyst was small.   Figure 2a,b demonstrated the TG differential profiles and TG mass loss profiles of fresh Cu-SAPO-34 and HCs-Cu-SAPO-34 catalysts. As shown in the Figure 2a, two obvious mass losses were observed around 100 • C and 250 • C. Specifically, the main mass variation at 100 • C and the minor mass loss at 250 • C was deemed to derive from the dehydration process of adsorbed water on Cu-SAPO-34 samples [27]. Furthermore, it was calculated from Figure 2b that the H 2 O mass losses were 13%, 5% for fresh Cu-SAPO-34, and 12%, 4% for HCs-Cu-SAPO-34, respectively. Importantly, no distinct mass loss was discovered at high temperature above 400 • C, which usually was considered as the essential temperature for coke removal on catalysts. Consequently, it indicated that the possibility of coke formation on Cu-SAPO-34 catalyst was small.

NH 3 -SCR Activity of Cu-SAPO-34 Catalysts
For NO conversion in Figure 3a, the catalytic performance of fresh Cu-SAPO-34 catalyst presented a volcanic-type profile. After the propylene was introduced into the reaction system, the de-NO x activity of Cu-SAPO-34 catalyst was obviously inhibited below 350 • C. However, the NO x conversion was rapidly improved when temperatures were higher than 350 • C and C 3 H 6 existed in the feed. The above results revealed that the propylene played an opposite role in NO x removal during the low-and high-temperature ranges, implying that the influence of propylene on the NH 3 -SCR reaction could be related to the reaction temperatures. Meanwhile, as shown in Figure 3b, the NH 3 conversion rate curve exhibited the same change pattern as the NO conversion in the low-temperature stage, but the NH 3 conversion kept increasing until 100% in the high-temperature stage. Simultaneously, it can be seen from the results in Figure 3c that the Cu/SAPO-34 catalysts shown excellent nitrogen selectivity with or without C 3 H 6 . Finally, the by-products NO 2 and N 2 O were both less than 18 ppm.

NH3-SCR Activity of Cu-SAPO-34 Catalysts
For NO conversion in Figure 3a, the catalytic performance of fresh Cu-SAPO-34 catalyst presented a volcanic-type profile. After the propylene was introduced into the reaction system, the de-NOx activity of Cu-SAPO-34 catalyst was obviously inhibited below 350 °C. However, the NOx conversion was rapidly improved when temperatures were higher than 350 °C and C3H6 existed in the feed. The above results revealed that the propylene played an opposite role in NOx removal during the low-and high-temperature ranges, implying that the influence of propylene on the NH3-SCR reaction could be related to the reaction temperatures. Meanwhile, as shown in Figure 3b, the NH3 conversion rate curve exhibited the same change pattern as the NO conversion in the low-temperature stage, but the NH3 conversion kept increasing until 100% in the high-temperature stage. Simultaneously, it can be seen from the results in Figure 3c that the Cu/SAPO-34 catalysts shown excellent nitrogen selectivity with or without C3H6. Finally, the by-products NO2 and N2O were both less than 18 ppm.  Figure 4 displayed the results of long-term NH 3 -SCR activity changes with propylene over Cu-SAPO-34 catalyst through the gas-switch experiments. It was seen that the NO conversion achieved equilibrium after 50 min. Subsequently, propylene was introduced into the system with identical total flow rate for 100 mL·min −1 . Then the C 3 H 6 was cut off again at 150 min. Finally, the NH 3 -SCR reaction achieved stability after another 80 min. As shown in Figure 4, the NO conversion presented two types of trends at different temperatures after the propylene was introduced. The NO conversion sharply decreased at 200 • C and 300 • C, but distinctly increased under 400 • C and 500 • C. Specifically, the catalytic activity changed from 89.2% to 81.2% at 200 • C, 88% to 83% at 300 • C, but increased from 77.6% to 80.4% at 400 • C and 68% to 74.4% at 500 • C, respectively. It is worthy of mention that after propylene was removed from the infeed, the NO conversion could recover completely to the initial value at 400 • C and 500 • C. Nevertheless, de-NO x activity could merely be regained to some extent at 200 • C and 300 • C. peratures after the propylene was introduced. The NO conversion sharply decreased at 200 °C and 300 °C, but distinctly increased under 400 °C and 500 °C. Specifically, the catalytic activity changed from 89.2% to 81.2% at 200 °C, 88% to 83% at 300 °C, but increased from 77.6% to 80.4% at 400 °C and 68% to 74.4% at 500 °C, respectively. It is worthy of mention that after propylene was removed from the infeed, the NO conversion could recover completely to the initial value at 400 °C and 500 °C. Nevertheless, de-NOx activity could merely be regained to some extent at 200 °C and 300 °C.  Figure 5 displayed the NH3-TPD profiles with or without propylene over H-SAPO-34 support at 100 °C. It was seen that three broad ammonia desorption peaks were detected over the whole temperature range for both NH3-TPD curves. According to previous studies [25], the peak at low-temperature area around 150 °C was ascribed to NH3 desorption from weak Brønsted acid sites at surface hydroxyl. The peaks around 246 °C and 335 °C were assigned to the moderate and strong structural Brønsted acid sites. Compared NH3-TPD results without C3H6, only the peak at low temperature slightly decreased, while the peaks at moderate-and high-temperature ranges almost remained identical level when C3H6 existed in the gas composition. This meant that impact of C3H6 to NH3 adsorption was mainly embodied in weak Brønsted acid sites, which could become inapparent with the increase of temperatures.

The Adsorption
Performance of Reactants 2.4.1. NH 3 -TPD Figure 5 displayed the NH 3 -TPD profiles with or without propylene over H-SAPO-34 support at 100 • C. It was seen that three broad ammonia desorption peaks were detected over the whole temperature range for both NH 3 -TPD curves. According to previous studies [25], the peak at low-temperature area around 150 • C was ascribed to NH 3 desorption from weak Brønsted acid sites at surface hydroxyl. The peaks around 246 • C and 335 • C were assigned to the moderate and strong structural Brønsted acid sites. Compared NH 3 -TPD results without C 3 H 6 , only the peak at low temperature slightly decreased, while the peaks at moderate-and high-temperature ranges almost remained identical level when C 3 H 6 existed in the gas composition. This meant that impact of C 3 H 6 to NH 3 adsorption was mainly embodied in weak Brønsted acid sites, which could become inapparent with the increase of temperatures.  Figure 6 presented the NH3-TPD profiles with or without propylene over the Cu-SAPO-34 catalyst at 100 °C. Specifically, four ammonia desorption peaks were observed over the range from 100 °C to 600 °C. The peak around 150 °C was ascribed to NH3 desorption from weak Brønsted sites and Lewis acid sites related to Cu species. And the  Figure 6 presented the NH 3 -TPD profiles with or without propylene over the Cu-SAPO-34 catalyst at 100 • C. Specifically, four ammonia desorption peaks were observed over the range from 100 • C to 600 • C. The peak around 150 • C was ascribed to NH 3 desorption from weak Brønsted sites and Lewis acid sites related to Cu species. And the peak at medium temperature range around 246 • C was assigned to residual structural Brønsted acid sites. Additionally, peaks above 300 • C were considered as strong Brønsted acid sites and the new Lewis acid sites created by the Cu 2+ species. Compared with results without C 3 H 6 , the tendency of NH 3 -TPD test with C 3 H 6 showed that just ammonia adsorption on weak acid site around 150 • C mildly decreased. And the propylene barely affected ammonia adsorption over medium and strong acid sites. Meanwhile, due to the huge difference of desorption amount, almost no C 3 H 6 species was detected during the desorption process of the NH 3 -TPD test.  Figure 6 presented the NH3-TPD profiles with or without propylene over the Cu-SAPO-34 catalyst at 100 °C. Specifically, four ammonia desorption peaks were observed over the range from 100 °C to 600 °C. The peak around 150 °C was ascribed to NH3 desorption from weak Brønsted sites and Lewis acid sites related to Cu species. And the peak at medium temperature range around 246 °C was assigned to residual structural Brønsted acid sites. Additionally, peaks above 300 °C were considered as strong Brønsted acid sites and the new Lewis acid sites created by the Cu 2+ species. Compared with results without C3H6, the tendency of NH3-TPD test with C3H6 showed that just ammonia adsorption on weak acid site around 150 °C mildly decreased. And the propylene barely affected ammonia adsorption over medium and strong acid sites. Meanwhile, due to the huge difference of desorption amount, almost no C3H6 species was detected during the desorption process of the NH3-TPD test.  Figure 7a,b exhibited the NO x -TPD profiles over the Cu-SAPO-34 catalyst with or without propylene at 80 • C. For NO+O 2 adsorption behavior on the Cu-SAPO-34 catalyst, there was only one main NO 2 desorption peak and slight NO desorption peak around 250 • C, as shown in Figure 7a. Concretely, the NO 2 desorption species could be related to the decomposition of nitrates on the Cu-SAPO-34 catalyst. However, when C 3 H 6 was introduced into the NO + O 2 adsorption process, all desorption peaks obviously shifted to the lower temperature around 150 • C and the desorption species mainly turned into NO. Only a slight amount of C 3 H 6 and little N 2 O were observed during the NO x -TPD experiments. Accordingly, this implied that the adsorption and oxidation of NO x on catalyst surface could apparently be weakened in the presence of C 3 H 6 species, and the NO x desorption peaks located around 150 • C were attributed to the weak adsorbed NO x species on the catalyst's surface.

NO x -TPD Experiments
duced into the NO + O2 adsorption process, all desorption peaks obviously shifted to the lower temperature around 150 °C and the desorption species mainly turned into NO. Only a slight amount of C3H6 and little N2O were observed during the NOx-TPD experiments. Accordingly, this implied that the adsorption and oxidation of NOx on catalyst surface could apparently be weakened in the presence of C3H6 species, and the NOx desorption peaks located around 150 °C were attributed to the weak adsorbed NOx species on the catalyst's surface.  Figure 8 shows the C3H6-TPD profiles after C3H6 adsorption saturation over H-SAPO-34 support and Cu-SAPO-34 catalyst at 100 °C. For C3H6 adsorptions on the H-SAPO-34 support, one main desorption peak at 220 °C was attributed to the weak adsorbed C3H6 species, yet the slight shoulder peak around 370 °C was assigned to the strong C3H6 adsorption species. Specifically, they were both related to C3H6 species adsorption on the  Figure 8 shows the C 3 H 6 -TPD profiles after C 3 H 6 adsorption saturation over H-SAPO-34 support and Cu-SAPO-34 catalyst at 100 • C. For C 3 H 6 adsorptions on the H-SAPO-34 support, one main desorption peak at 220 • C was attributed to the weak adsorbed C 3 H 6 species, yet the slight shoulder peak around 370 • C was assigned to the strong C 3 H 6 adsorption species. Specifically, they were both related to C 3 H 6 species adsorption on the Brønsted acid sites. Compared with the C 3 H 6 -TPD results over the H-SAPO-34 sample, two obvious desorption peaks were detected over the Cu-SAPO-34 catalyst. The larger C 3 H 6 desorption peak around 370 • C should be ascribed to adsorption of C 3 H 6 on Cu species, since only Cu species over the Cu-SAPO-34 catalyst could supply extra adsorption sites. Furthermore, it was suggested that the C 3 H 6 adsorption strength on Cu species was stronger than that on the Brønsted acid sites, and the Cu 2+ species played an important role in adsorption of HCs species, as shown in Figure 8.

C 3 H 6 -TPD Experiments
Brønsted acid sites. Compared with the C3H6-TPD results over the H-SAPO-34 sample, two obvious desorption peaks were detected over the Cu-SAPO-34 catalyst. The larger C3H6 desorption peak around 370 °C should be ascribed to adsorption of C3H6 on Cu species, since only Cu species over the Cu-SAPO-34 catalyst could supply extra adsorption sites. Furthermore, it was suggested that the C3H6 adsorption strength on Cu species was stronger than that on the Brønsted acid sites, and the Cu 2+ species played an important role in adsorption of HCs species, as shown in Figure 8.

Co-Adsorption of NH3 and C3H6 Species
The DRIFTS spectra of NH3, C3H6, and NH3+C3H6 adsorption over the Cu-SAPO-34 catalysts at 200 °C are shown in Figure 9. For NH3 adsorption over Cu-SAPO-34 reported previously [25], the doublet bands at 3624 cm −1 and 3590 cm −1 were assigned to the stretching model of bridged Brønsted OH groups Al-(OH)-Si, which was used to represent the Brønsted acid sites on the catalyst. Meanwhile, the bands at 885 cm −1 and 858 cm −1 were related to Lewis acid sites of copper cations. For C3H6 adsorption on the Cu-SAPO-34 catalysts, obvious IR adsorption peaks were only observed at 885 and 858 cm −1 , which could be related to the adsorption of C3H6 on copper species over Cu-SAPO-34. However, during the co-adsorption process of NH3 and C3H6, the peak intensity of 3624 cm −1 was almost consistent with the individual NH3 adsorption test, which revealed that the ammonia adsorption on the Brønsted acid site was barely affected by the presence of C3H6. Nevertheless, IR adsorption peaks strength at 885 cm −1 was relatively decreased in Figure 9b compared with the results of the NH3 adsorption test. Based on the above results, it was concluded that the adsorption of NH3 on the Brønsted acid sites of Cu-SAPO-34 catalysts would not be affected by the existence of propylene, while the ammonia adsorption on the Lewis acid sites was partially influenced by C3H6.

Co-Adsorption of NH 3 and C 3 H 6 Species
The DRIFTS spectra of NH 3 , C 3 H 6 , and NH 3 +C 3 H 6 adsorption over the Cu-SAPO-34 catalysts at 200 • C are shown in Figure 9. For NH 3 adsorption over Cu-SAPO-34 reported previously [25], the doublet bands at 3624 cm −1 and 3590 cm −1 were assigned to the stretching model of bridged Brønsted OH groups Al-(OH)-Si, which was used to represent the Brønsted acid sites on the catalyst. Meanwhile, the bands at 885 cm −1 and 858 cm −1 were related to Lewis acid sites of copper cations. For C 3 H 6 adsorption on the Cu-SAPO-34 catalysts, obvious IR adsorption peaks were only observed at 885 and 858 cm −1 , which could be related to the adsorption of C 3 H 6 on copper species over Cu-SAPO-34. However, during the co-adsorption process of NH 3 and C 3 H 6 , the peak intensity of 3624 cm −1 was almost consistent with the individual NH 3 adsorption test, which revealed that the ammonia adsorption on the Brønsted acid site was barely affected by the presence of C 3 H 6 . Nevertheless, IR adsorption peaks strength at 885 cm −1 was relatively decreased in Figure 9b compared with the results of the NH 3 adsorption test. Based on the above results, it was concluded that the adsorption of NH 3 on the Brønsted acid sites of Cu-SAPO-34 catalysts would not be affected by the existence of propylene, while the ammonia adsorption on the Lewis acid sites was partially influenced by C 3 H 6 .  Figure 10 displayed the DRIFTS spectra of NO, C3H6, and NO + C3H6 adsorption over the Cu-SAPO-34 catalysts at 200 °C. When 500 ppm NO was introduced into the reaction cell at 200 °C, IR adsorption peak was scarcely observed, which suggested the NO species was hardly adsorbed on Cu-SAPO-34 catalyst, which was consistent with previous results of NO reaction orders of 1 [21]. For C3H6 adsorption over Cu-SAPO-34 at 200 °C, various IR bands of carbonaceous species appeared as shown in Figure 10. The IR bands at 1656 cm −1 and 1591 cm −1 were assigned to the C=O vibrations and C-H stretching vibrations of hydrocarbons, respectively. The band at 1430 cm −1 was related to the surface carboxylate species. Nevertheless, when the NO and C3H6 were both present during the adsorption competition, the IR results were almost identical with C3H6 adsorption experiment results. Consequently, this illustrated that the adsorption strength of C3H6 was obviously stronger than NO on the Cu-SAPO-34, which could lead into the inhibition of nitrate/nitrite formation on the catalyst.  Figure 10 displayed the DRIFTS spectra of NO, C 3 H 6 , and NO + C 3 H 6 adsorption over the Cu-SAPO-34 catalysts at 200 • C. When 500 ppm NO was introduced into the reaction cell at 200 • C, IR adsorption peak was scarcely observed, which suggested the NO species was hardly adsorbed on Cu-SAPO-34 catalyst, which was consistent with previous results of NO reaction orders of 1 [21]. For C 3 H 6 adsorption over Cu-SAPO-34 at 200 • C, various IR bands of carbonaceous species appeared as shown in Figure 10. The IR bands at 1656 cm −1 and 1591 cm −1 were assigned to the C=O vibrations and C-H stretching vibrations of hydrocarbons, respectively. The band at 1430 cm −1 was related to the surface carboxylate species. Nevertheless, when the NO and C 3 H 6 were both present during the adsorption competition, the IR results were almost identical with C 3 H 6 adsorption experiment results. Consequently, this illustrated that the adsorption strength of C 3 H 6 was obviously stronger than NO on the Cu-SAPO-34, which could lead into the inhibition of nitrate/nitrite formation on the catalyst.

The in Situ DRIFTS Experiments
The DRIFTS spectra of saturated NH3 titration by NO + O2 and NO + C3H6 + O2 at 200 °C were shown in Figure 11a,b. After the Cu-SAPO-34 catalyst was purged by NH3 until equilibrium, the NO + O2 (Figure 11a) and NO + C3H6 + O2 (Figure 11b) were introduced into the reaction cell. The spectra were recorded with the interval of 20 min until reaction equilibrium was attained. Herein, the doublet bands at 3624 cm −1 and 3590 cm −1 were ascribed to the Brønsted acid sites, while the bands at 885 cm −1 and 858 cm −1 were related to the Lewis acid sites of Cu 2+ species on the Cu-SAPO-34. The peak intensity at 3624 cm −1 and 885 cm −1 in the present DRIFTS spectra were chosen as the representative of the Brønsted acid sites and Lewis acid sites to evaluate the behaviors of adsorbed ammonia species on the Cu-SAPO-34 catalyst. As shown in Figure 11a, the adsorbed ammonia species at 3624 cm −1 and 885 cm −1 had been consumed completely around 60 min. However, the pre-adsorbed NH3 on the Lewis acid sites still distinctly existed after 60 min in the presence of C3H6 in Figure 11b. Consequently, it suggested that the consumption rate of adsorbed ammonia was apparently inhibited when the C3H6 existed in the mixture gas.

The In Situ DRIFTS Experiments
The DRIFTS spectra of saturated NH 3 titration by NO + O 2 and NO + C 3 H 6 + O 2 at 200 • C were shown in Figure 11a,b. After the Cu-SAPO-34 catalyst was purged by NH 3 until equilibrium, the NO + O 2 ( Figure 11a) and NO + C 3 H 6 + O 2 (Figure 11b) were introduced into the reaction cell. The spectra were recorded with the interval of 20 min until reaction equilibrium was attained. Herein, the doublet bands at 3624 cm −1 and 3590 cm −1 were ascribed to the Brønsted acid sites, while the bands at 885 cm −1 and 858 cm −1 were related to the Lewis acid sites of Cu 2+ species on the Cu-SAPO-34. The peak intensity at 3624 cm −1 and 885 cm −1 in the present DRIFTS spectra were chosen as the representative of the Brønsted acid sites and Lewis acid sites to evaluate the behaviors of adsorbed ammonia species on the Cu-SAPO-34 catalyst. As shown in Figure 11a, the adsorbed ammonia species at 3624 cm −1 and 885 cm −1 had been consumed completely around 60 min. However, the pre-adsorbed NH 3 on the Lewis acid sites still distinctly existed after 60 min in the presence of C 3 H 6 in Figure 11b. Consequently, it suggested that the consumption rate of adsorbed ammonia was apparently inhibited when the C 3 H 6 existed in the mixture gas. Figure 12a showed the SCR catalytic activity of NO x removal with C 3 H 6 over the Cu-SAPO-34 catalyst. Figure 12b displayed the NO x conversion performance with a variation of the C 3 H 6 concentration at 500 • C. As shown in Figure 12a, the NO x conversion increased with an increase of reaction temperature. Concretely, the de-NO x efficiency was very low when the temperature was below 300 • C, while NO x conversion obviously improved up to 50% at 550 • C. Subsequently, in order to evaluate the effect of C 3 H 6 -SCR on the catalytic performance of NO x , the NH 3 -SCR reactions with different C 3 H 6 concentration were performed at 500 • C. It was presented in Figure 12b that the NO x conversion continuously increased with C 3 H 6 concentration added from 100 ppm to 900 ppm. Accordingly, this implied that C 3 H 6 -SCR exhibited the evident contribution to the NO x removal efficiency at high temperatures.  Figure 12a showed the SCR catalytic activity of NOx removal with C3H6 over the Cu-SAPO-34 catalyst. Figure 12b displayed the NOx conversion performance with a variation of the C3H6 concentration at 500 °C. As shown in Figure 12a, the NOx conversion increased with an increase of reaction temperature. Concretely, the de-NOx efficiency was very low when the temperature was below 300 °C, while NOx conversion obviously improved up to 50% at 550 °C. Subsequently, in order to evaluate the effect of C3H6-SCR on the catalytic performance of NOx, the NH3-SCR reactions with different C3H6 concentration were performed at 500 °C. It was presented in Figure 12b that the NOx conversion continuously increased with C3H6 concentration added from 100 ppm to 900 ppm. Accordingly, this implied that C3H6-SCR exhibited the evident contribution to the NOx removal efficiency at high temperatures.   Figure 12a showed the SCR catalytic activity of NOx removal with C3H6 over the Cu-SAPO-34 catalyst. Figure 12b displayed the NOx conversion performance with a variation of the C3H6 concentration at 500 °C. As shown in Figure 12a, the NOx conversion increased with an increase of reaction temperature. Concretely, the de-NOx efficiency was very low when the temperature was below 300 °C, while NOx conversion obviously improved up to 50% at 550 °C. Subsequently, in order to evaluate the effect of C3H6-SCR on the catalytic performance of NOx, the NH3-SCR reactions with different C3H6 concentration were performed at 500 °C. It was presented in Figure 12b that the NOx conversion continuously increased with C3H6 concentration added from 100 ppm to 900 ppm. Accordingly, this implied that C3H6-SCR exhibited the evident contribution to the NOx removal efficiency at high temperatures.

Catalyst Preparation
H-SAPO-34 was synthesized via the hydrothermal method with mole composition of 0.2 morpholine (MA): 0.1, Al2O3: 0.1, P2O5: 0.06, and SiO2: 6.4 H2O. The sources of Si, P, and Al were silica sol, 85% phosphoric acid and pseudoboehmite, respectively. Firstly, the pseudoboehmite and phosphoric acid were mixed with water. The mixture was fiercely stirred for 1 h. Secondly, the MA and silica sol were mingled well and dropwise added into the former mixture. Subsequently, the blend was sealed in 200 mL Teflon-lined stainless-steel pressure vessels and heated in an oven at 200 °C under autogenic pressure for 48 h. Finally, the H-SAPO-34 was obtained through centrifugal separation, washed, dried at 100 °C overnight, and calcined at 650 °C in air for 6 h.
The Cu-SAPO-34 catalyst was prepared by the ion-exchange method over H-SAPO-34 molecular sieve as following. Firstly, Cu-SAPO-34 was obtained by exchanging H-SAPO-34 in ammonium nitrate solution at 80 °C for 3 h. Additionally, NH4-SAPO-34 was added into the copper sulfate solution then stirred at 70 °C for 4 h. The slurry was filtered, washed, and dried at 90~100 °C for 16 h. Lastly, the dried Cu-SAPO-34 was calcined at 500 °C for 5 h and denoted as fresh Cu-SAPO-34 catalyst. The Cu content of catalysts was 2.5% (wt. %), which was calculated on account of the ICP test result. Furthermore, after 1000 ppm C3H6 treatment at 300 °C for 6 h, the obtained Cu-SAPO-34 sample was called as HCs-Cu-SAPO-34.

Catalytic Performance Measurement
The SCR activity with or without C3H6 was tested in a quartz reactor (20 mm inner diameter), in which 0.1 g sample (60~80 mesh) was mixed with 0.9 g quartz (60~80 mesh) at atmospheric pressure. The sample was sealed in a reactor tube with quartz wool. The K-type thermocouple was inserted into the center of catalyst to control the temperature. A Fourier Transform Infrared (FTIR) spectrometer (MKS-2030, Andover, MA, United States) equipped with 5.11 m gas cell was used to measure the concentration of NH3, NO, NO2, N2O, and C3H6. The flow rates in all experiments were controlled at 500 mL min −1 . After pretreated with 5% O2/N2 at 500 °C for 30 min, SCR tests were conducted with feed gas composition of 500 ppm NO, 500 ppm NH3, 500 ppm C3H6 (when used), and 5% O2.

Catalyst Preparation
H-SAPO-34 was synthesized via the hydrothermal method with mole composition of 0.2 morpholine (MA): 0.1, Al 2 O 3 : 0.1, P 2 O 5 : 0.06, and SiO 2 : 6.4 H 2 O. The sources of Si, P, and Al were silica sol, 85% phosphoric acid and pseudoboehmite, respectively. Firstly, the pseudoboehmite and phosphoric acid were mixed with water. The mixture was fiercely stirred for 1 h. Secondly, the MA and silica sol were mingled well and dropwise added into the former mixture. Subsequently, the blend was sealed in 200 mL Teflon-lined stainlesssteel pressure vessels and heated in an oven at 200 • C under autogenic pressure for 48 h. Finally, the H-SAPO-34 was obtained through centrifugal separation, washed, dried at 100 • C overnight, and calcined at 650 • C in air for 6 h.
The Cu-SAPO-34 catalyst was prepared by the ion-exchange method over H-SAPO-34 molecular sieve as following. Firstly, Cu-SAPO-34 was obtained by exchanging H-SAPO-34 in ammonium nitrate solution at 80 • C for 3 h. Additionally, NH 4 -SAPO-34 was added into the copper sulfate solution then stirred at 70 • C for 4 h. The slurry was filtered, washed, and dried at 90~100 • C for 16 h. Lastly, the dried Cu-SAPO-34 was calcined at 500 • C for 5 h and denoted as fresh Cu-SAPO-34 catalyst. The Cu content of catalysts was 2.5% (wt. %), which was calculated on account of the ICP test result. Furthermore, after 1000 ppm C 3 H 6 treatment at 300 • C for 6 h, the obtained Cu-SAPO-34 sample was called as HCs-Cu-SAPO-34.

Catalytic Performance Measurement
The SCR activity with or without C 3 H 6 was tested in a quartz reactor (20 mm inner diameter), in which 0.1 g sample (60~80 mesh) was mixed with 0.9 g quartz (60~80 mesh) at atmospheric pressure. The sample was sealed in a reactor tube with quartz wool. The K-type thermocouple was inserted into the center of catalyst to control the temperature. A Fourier Transform Infrared (FTIR) spectrometer (MKS-2030, Andover, MA, United States) equipped with 5.11 m gas cell was used to measure the concentration of NH 3 , NO, NO 2 , N 2 O, and C 3 H 6 . The flow rates in all experiments were controlled at 500 mL min −1 . After pretreated with 5% O 2 /N 2 at 500 • C for 30 min, SCR tests were conducted with feed gas composition of 500 ppm NO, 500 ppm NH 3 , 500 ppm C 3 H 6 (when used), and 5% O 2 .
The test temperature range was from 100 • C to 550 • C with an interval of 50 • C. The NO conversion and NH 3 conversion were calculated using the following equations:

Catalyst Characterization
The XRD patterns were recorded using an X-ray diffraction (XRD, Bruker D8 Focus, Cu Kα radiation, Karlsruhe, Germany). The XRD pattern was collected from 5 • to 50 • with the step size of 0.02 • . The BET surface area (m 2 ·g −1 ) was determined from the linear portion of the BET plot by measuring the N 2 isotherm of the samples at 77 K using F-Sorb 3400 volumetric adsorption-desorption apparatus. Prior to the measurement, the zeolites were degassed at 150 • C under vacuum for 3 h.
Temperature programmed desorption (TPD, Tianjin, China) experiments were performed to evaluate the adsorption amount of NH 3 and NO x in the presence or absence of C 3 H 6 . Prior to each of the following experiments, the catalysts were pretreated at 500 • C for 30 min in 5% O 2 /N 2 , then cooled to the test temperatures in N 2 . At test temperatures, NH 3 , NO x , and C 3 H 6 (when used) were introduced in 500 ppm NH 3 /N 2 until the outlet concentration of NH 3 was stable. The catalysts were purged with N 2 to remove any weakly adsorbed NH 3 . Finally, the catalysts were heated from test temperature to 600 • C at a ramp rate of 10 • C min −1 .
Thermo Gravimetric (TG) analysis was conducted on the STAR System TGA/DSC1 (Mettler, Zurich, Switzerland) to probe the mass variation during the temperature programing process. The sensitivity of the TG instruments is 0.1 µg, which ensures the veracity of the TG experiments. After the blank experiment, the samples were heated from 50 • C to 950 • C at a ramp rate of 10 • C min −1 .

In Situ DRIFTS Experiments
In situ Diffuse Reflectance Infrared Fourier transform spectra (in situ DRIFTS, Madison, Wisconsin, United States) were performed on Nicolet 6700 FTIR equipped with MCT detector at a resolution of 1 cm −1 . Four scans were operated for each spectrum. The total flow rate was 150 mL min −1 . Prior to each experiment, the samples were pretreated with 2% O 2 /He at 500 • C for 30 min and the background spectra were collected under He gas at test temperature.
For the adsorption experiments, the NH 3 + C 3 H 6 or NO + C 3 H 6 were introduced into sample at 200 • C until the spectra were stable. And the spectra were collected after purged by He gas.
For the DRIFTS titration experiments of adsorbed NH 3 , the 500 ppm NH 3 was firstly introduced into the catalyst at 200 • C until the spectra were stable. Then the catalyst was purged by He gas to eliminate the physically adsorbed NH 3 . Finally, the sample was exposed to 500 ppm NO, 500 ppm C 3 H 6 (when used), and 5% O 2 until equilibrium was attained.

The Effect of C 3 H 6 Treatment on Catalyst Structure
Although some researchers reported that the rapid deactivation during the SCR of NO x with HCs (HCs-SCR) could result from coke formation, which could be due to the blocking of the pore, and reducing the pore volume and surface area of the catalyst [40]. However, based on the obtained results of BET, XRD, and TG of fresh and HCs treated Cu-SAPO-34 catalysts, it showed that the BET surface area of Cu-SAPO-34 was scarcely affected by 6 h C 3 H 6 treatment on its microporous structure. Meanwhile, the CHA structure of Cu-SAPO-34 catalyst still remained intact after C 3 H 6 treatment and no new crystalline phase was detected as exhibited in Figure 1, which was consistent with previous results that the zeolite crystal structure did not change before and after propylene poisoning over Fe-zeolite catalyst [40]. Furthermore, according to the results of TG experiments in Figure 2, no obvious coke removal signals were observed except two mass loss peaks which could be assigned to the dehydration process at 100 • C and 250 • C, respectively. Consequently, it could be concluded that the physical properties of Cu-SAPO-34 catalysts were not affected by the long-term HCs treatment compared with the fresh catalysts, and the deactivation phenomenon caused by coke deposition would be less likely to appear on the Cu-SAPO-34 catalysts than other mesoporous molecular sieves.

The C 3 H 6 Effect on the Adsorption Performance of NH 3 , NO Species
Based on the results of NH 3 -TPD and NO x -TPD with C 3 H 6 in Figures 5-7 and C 3 H 6 -TPD in Figure 8, it was shown that the ammonia adsorption had not been obviously affected by C 3 H 6 on the Brønsted acid sites, but competitive adsorption could occur between C 3 H 6 and NH 3 on the Cu 2+ sites. Although ammonia adsorption plays a dominant role in the adsorption process, the denitration performance at low temperature still could be reduced due to the reduction of ammonia adsorption capacity. In addition, the propylene adsorption site mainly ascribed to the copper divalent species, and the adsorption strength of propylene is stronger than that of nitrogen oxides. This could result in that nitrogen oxides cannot be oxidized on copper divalent at low temperature, on which the hydrocarbon SCR reaction was unable to be completed smoothly. Therefore, it would finally cause the decline of de-NO x reaction activity at low temperature.

The C 3 H 6 Effect on the SCR Performance over Cu-SAPO-34 Catalyst
Based on the obtained results and the published paper, it was shown that the isolated Cu 2+ cations were proved as both of the active sites for NH 3 -SCR reaction and C 3 H 6 -SCR reaction [45,52]. The standard NH 3 -SCR utilizes NH 3 as a reductant to remove NO x as follows: 4NH 3 + 4NO + O 2 = 4N 2 + 6H 2 O. The main products are N 2 and H 2 O, while the by-products are NO 2 and N 2 O. Herein, at low temperatures (<350 • C), the standard NH 3 -SCR is the dominant reaction, whether or not propylene exists, the amount of byproduct remains below 15 ppm as shown in Figure 3c. Yet the conversion of C 3 H 6 -SCR is just under 10% at this time. It is concluded that the competitive adsorption of propylene could play an important role in reducing the probability of adsorbed NH 3 participating in NH 3 -SCR reaction on the Cu 2+ cations, without affecting the product distribution of the NH 3 -SCR, while during high temperatures (>350 • C), it was known that the NH 3 oxidation and NH 3 -SCR competed with each other during the SCR process. The NH 3 oxidation is one of the main causes of the decline of the NO conversion at high temperature [26]. But NO conversions are surging rather than falling with propylene at the present time, resulting from the increased activity of C 3 H 6 -SCR on the Cu 2+ cations which compensates for the loss of de-NO x efficiency due to the restraint of ammonia oxidation at high temperatures. In other words, C 3 H 6 -SCR shows the synergistic enhancement effect. Importantly, the amount of by-product still remains below 18 ppm, showing good nitrogen selectivity.
The mechanism of propylene effect on Selective Catalytic Reduction by Ammonia over Cu-SAPO-34 catalyst was shown in Scheme 1. For Scheme 1a, it was essentially shown that the sequence of adsorption strength on the Cu 2+ sites of the catalysts during the low temperatures were as follows: ammonia > propylene > nitrogen oxides. Importantly, the presence of propylene not only affects the adsorption capacity of ammonia but also hinders the further oxidation of nitrogen oxides on the Cu 2+ sites, while the above two aspects are of great significance for SCR reaction on the Cu-SAPO-34 catalyst. In short, the influence of propylene at low temperature is mainly reflected in the inhibition of the above adsorption or oxidation process, which ultimately reduces the performance of SCR reaction. For Scheme 1b, the possibility of competitive adsorption of propylene with ammonia and nitrogen oxides was weakened with the increase of reaction temperature. The C 3 H 6 -SCR under high temperature could synergically improve the removal efficiency of nitrogen oxides and inhibit the ammonia oxidation reaction which is unfavorable to the SCR reaction. In brief, the effect of propylene during high temperature is mainly embodied in the enhancement, which through restraining of the ammonia oxidation reaction and providing alternative de-NO x reaction pathways, i.e., C 3 H 6 -SCR. These, finally, could give rise to the de-NO x efficiency of the Cu-SAPO-34. reaction. For Scheme 1b, the possibility of competitive adsorption of propylene with ammonia and nitrogen oxides was weakened with the increase of reaction temperature. The C3H6-SCR under high temperature could synergically improve the removal efficiency of nitrogen oxides and inhibit the ammonia oxidation reaction which is unfavorable to the SCR reaction. In brief, the effect of propylene during high temperature is mainly embodied in the enhancement, which through restraining of the ammonia oxidation reaction and providing alternative de-NOx reaction pathways, i.e., C3H6-SCR. These, finally, could give rise to the de-NOx efficiency of the Cu-SAPO-34.

Scheme 1.
The mechanism of propylene effect on NH3-SCR over the Cu/SAPO-34 catalyst at low temperature range (a) and high temperature range (b).

Conclusions
The effect of propylene on NH3-SCR of NOx over the Cu-SAPO-34 catalysts has been systematically studied in the present work. Firstly, the XRD and BET results showed that the structure and physical properties of Cu-SAPO-34 were almost not influenced by the propylene treatment. Meanwhile, the TG results proved that coke deposition might not formed on the surface of Cu-SAPO-34 in spite of undergoing a long-term C3H6+O2 treatment. Secondly, the adsorption intensity of C3H6 on Cu 2+ species was stronger than that on the Brønsted sites, and the C3H6 mainly adsorbed on the Cu 2+ sites of Cu-SAPO-34 catalyst. Furthermore, the oxidation state of Cu 2+ species on the catalyst's surface could be affected by the propylene in the feed gas and it would also influence the adsorption performance of NOx on Cu species and the formation ratio of nitrite/nitrate. Hence, the competitive effect of C3H6 on the adsorption amount and strength of ammonia could not be the dominating reason for the activity decrease of low temperate. Thirdly, the propylene effect on NH3-SCR of NOx over Cu-SAPO-34 catalyst could follow two different mechanisms with changes of reaction temperatures. Specifically, the activity decrease of NH3-SCR with propylene over Cu-SAPO-34 at low temperatures (<350 °C) could result from occupation of the Cu 2+ sites by C3H6 and inhibition to the proceeding of Cu 2+ -Cu + -Cu 2+ redox cycle. Nevertheless, the activity of the occupied Cu species could be recovered and sequentially participate in the NH3-SCR reaction with an increase of temperature. Meanwhile, the NO conversion efficiency was also benefited by the contribution of C3H6-SCR, which could restrain the ammonia oxidation reaction at high temperatures. Consequently, the propylene played a beneficial role in NH3-SCR process over Cu-SAPO-34 at high temperatures (>350 °C) due to the synergetic contribution of C3H6-SCR reaction. Scheme 1. The mechanism of propylene effect on NH 3 -SCR over the Cu/SAPO-34 catalyst at low temperature range (a) and high temperature range (b).

Conclusions
The effect of propylene on NH 3 -SCR of NO x over the Cu-SAPO-34 catalysts has been systematically studied in the present work. Firstly, the XRD and BET results showed that the structure and physical properties of Cu-SAPO-34 were almost not influenced by the propylene treatment. Meanwhile, the TG results proved that coke deposition might not formed on the surface of Cu-SAPO-34 in spite of undergoing a long-term C 3 H 6 +O 2 treatment. Secondly, the adsorption intensity of C 3 H 6 on Cu 2+ species was stronger than that on the Brønsted sites, and the C 3 H 6 mainly adsorbed on the Cu 2+ sites of Cu-SAPO-34 catalyst. Furthermore, the oxidation state of Cu 2+ species on the catalyst's surface could be affected by the propylene in the feed gas and it would also influence the adsorption performance of NO x on Cu species and the formation ratio of nitrite/nitrate. Hence, the competitive effect of C 3 H 6 on the adsorption amount and strength of ammonia could not be the dominating reason for the activity decrease of low temperate. Thirdly, the propylene effect on NH 3 -SCR of NO x over Cu-SAPO-34 catalyst could follow two different mechanisms with changes of reaction temperatures. Specifically, the activity decrease of NH 3 -SCR with propylene over Cu-SAPO-34 at low temperatures (<350 • C) could result from occupation of the Cu 2+ sites by C 3 H 6 and inhibition to the proceeding of Cu 2+ -Cu + -Cu 2+ redox cycle. Nevertheless, the activity of the occupied Cu species could be recovered and sequentially participate in the NH 3 -SCR reaction with an increase of temperature. Meanwhile, the NO conversion efficiency was also benefited by the contribution of C 3 H 6 -SCR, which could restrain the ammonia oxidation reaction at high temperatures. Consequently, the propylene played a beneficial role in NH 3 -SCR process over Cu-SAPO-34 at high temperatures (>350 • C) due to the synergetic contribution of C 3 H 6 -SCR reaction.