Co-Exchange of Mn: A Simple Method to Improve Both the Hydrothermal Stability and Activity of Cu–SSZ-13 NH 3 –SCR Catalysts

: A series of Cu–Mn–SSZ-13 catalysts were obtained by co-exchange of Mn and Cu into SSZ-13 together (ion exchange under a mixed solution of Cu(NO 3 ) 2 and Mn(NO 3 ) 2 ) and compared with Cu–SSZ-13 catalysts on the selective catalytic reduction (SCR) of nitric oxide (NO) by ammonia. The e ﬀ ects of total ion exchange degree and the e ﬀ ect of Mn species on the structure and performance of catalysts before and after hydrothermal aging were studied. All fresh and aged catalysts were characterized with several methods including temperature-programmed desorption with NH 3 (NH 3 -TPD), X-ray di ﬀ raction (XRD), 27 Al and 29 Si solid-state nuclear magnetic resonance (NMR), scanning electron microscopy-energy dispersive spectroscopy (SEM-EDS) and low-temperature N 2 adsorption–desorption techniques. The results showed that the increase of the total ion exchange degree can reduce the content of residual Brønsted acid sites of catalysts, thus relieved the dealumination and the decrease of crystallinity of the catalyst during hydrothermal aging. The moderate addition of a Mn component in Cu–Mn–SSZ-13 catalysts signiﬁcantly increased the activity of NO conversion at low temperature range. The selected Cu(0.2)Mn(0.1)–SSZ-13 catalyst achieved a high NO conversion of > 90% in the wide and low temperature range of 175–525 ◦ C and also exhibited good N 2 selectivity and excellent hydrothermal stability, which was related to the inhibition of the Mn component on the aggregation of Cu species and the pore destruction of the catalyst during hydrothermal aging.


Introduction
Nitrogen oxides (NO x ) emitted from mobile or stationary sources are among the major sources of air pollution, leading to a number of environmental problems such as photochemical smog and acid rain, which have great negative impacts on human health [1][2][3][4][5]. Owing to their negative environmental and health impacts, the removal of NO x is becoming increasingly urgent. In addition to modifications made to combustion systems to reduce NO x emissions, exhaust after-treatment technologies are required to meet stringent pollution control standards. In recent years, the selective catalytic reduction (SCR) of NO x with NH 3 was proven to be an efficient method, regarded as one of the most promising candidates for NO x reduction in flue gas [6][7][8][9][10].

Activity Test
The XRD curve and SEM images of synthetic SSZ-13 zeolite are shown in Figures S1 and S2. The result indicates the zeolite presents obvious CHA structural features [30,31]. Figure 1 shows the NH 3 -SCR activity of these Cu-Mn catalysts. As the Mn/Cu ratio (in the mother liquid) rose from 0/10 to 4/10, the activity of the catalyst at low temperature increased and the temperature at NO conversion of 90% decreased from 225 to 175 • C. This indicates that the Cu-Mn catalysts obtained by co-exchange method can achieve good activity at both low and high temperature. However, the Mn x O y -based, Mn x O y -modified or Mn x O y -SSZ-13 catalysts can only improve low-temperature NH 3 -SCR properties [24,25,[32][33][34]. This is due to the fact that the Mn element here was introduced into the SSZ-13 zeolite in the form of stable and well-dispersed metal ions instead of big Mn x O y particles. The XRD results ( Figure S3) shows that the Cu-Mn catalysts had the feature of the CHA structure. It was difficult to form the Mn x O y particles on the surface of Cu-Mn catalysts, since the concentration of the ionic solution used for exchange was quite dilute and the catalysts were rinsed by deionized water after exchange. However, the activity decreased when the Mn/Cu ratio reached 6/10 and the feature presented a seagull-shape showing an interesting phenomenon that from~150 to~250 • C the NO conversion decreased. This occurs when the gas hourly space velocity (GHSV) is relatively high [35,36], therefore, when the Mn/Cu ratio increased to 6/10,Mn ions competed with the Cu ions violently then the content of Cu ions decreased as shown in Table 1, which leads to an increase of relative GHSV. element here was introduced into the SSZ-13 zeolite in the form of stable and well-dispersed metal ions instead of big MnxOy particles. The XRD results ( Figure S3) shows that the Cu-Mn catalysts had the feature of the CHA structure. It was difficult to form the MnxOy particles on the surface of Cu-Mn catalysts, since the concentration of the ionic solution used for exchange was quite dilute and the catalysts were rinsed by deionized water after exchange. However, the activity decreased when the Mn/Cu ratio reached 6/10 and the feature presented a seagull-shape showing an interesting phenomenon that from ~150 to ~250 °C the NO conversion decreased. This occurs when the gas hourly space velocity (GHSV) is relatively high [35,36], therefore, when the Mn/Cu ratio increased to 6/10,Mn ions competed with the Cu ions violently then the content of Cu ions decreased as shown in Table 1, which leads to an increase of relative GHSV.   In order to understand the effect of Mn species, we used Cu(0.2) and Cu(0.3) samples (with the same Cu content and the same total ion exchange degree as the Cu(0.2)Mn(0.1) catalyst) for comparison. It should be noted that it is difficult to obtain the exact same ion exchange degree or the same Cu content by the ion exchange method, so we used approximate processing.
The temperature programmed reduction with hydrogen (H2-TPR) curves given in Figure S4 show that the addition of Mn can make the reduction peak at ca. 220 °C enlargen and broaden due to the occurrence of shoulder peak at the side of high temperature. It indicates the addition of Mn species can interact with Cu ions and weak the interaction between Cu ions and the CHA framework. It is beneficial to the low-temperature SCR activity [22]. The formation of NO2 from NO is a key step in standard SCR reactions [37]. The increase in the oxidizing ability of the catalyst made the reaction easier to occur and thereby increased the low temperature activity of the Cu-Mn catalyst. On the other hand, the introduction of Mn species by the co-exchange method caused an increase in the total ion exchange degree, leading to a decrease of the residual B acid sites, which are activity-silent [37][38][39][40] and vulnerable during hydrothermal aging. Another important conclusion of this result is that the decrease of the residual B acid sites also can improve the stability of Cu-Mn catalyst [22,41].  In order to understand the effect of Mn species, we used Cu(0.2) and Cu(0.3) samples (with the same Cu content and the same total ion exchange degree as the Cu(0.2)Mn(0.1) catalyst) for comparison. It should be noted that it is difficult to obtain the exact same ion exchange degree or the same Cu content by the ion exchange method, so we used approximate processing.
The temperature programmed reduction with hydrogen (H 2 -TPR) curves given in Figure S4 show that the addition of Mn can make the reduction peak at ca. 220 • C enlargen and broaden due to the occurrence of shoulder peak at the side of high temperature. It indicates the addition of Mn species can interact with Cu ions and weak the interaction between Cu ions and the CHA framework. It is beneficial to the low-temperature SCR activity [22]. The formation of NO 2 from NO is a key step in standard SCR reactions [37]. The increase in the oxidizing ability of the catalyst made the reaction easier to occur and thereby increased the low temperature activity of the Cu-Mn catalyst. On the other hand, the introduction of Mn species by the co-exchange method caused an increase in the total ion exchange degree, leading to a decrease of the residual B acid sites, which are activity-silent [37][38][39][40] Catalysts 2019, 9, 455 4 of 15 and vulnerable during hydrothermal aging. Another important conclusion of this result is that the decrease of the residual B acid sites also can improve the stability of Cu-Mn catalyst [22,41].
From Figure 2a, we can see the Cu(0.2)Mn(0.1) catalyst presents the widest active temperature window. For the Cu(0.2) and Cu(0.3) samples, they were similar in activity at high temperatures and the Cu(0.3) catalyst had a higher activity in the low temperature range since it contained more active components. However, their activity in the low-temperature range was much lower than that of Cu(0.2)Mn(0.1), indicating that the increase of activity was not caused by the increase in ion exchange degree but was mainly caused by the promotion of Mn species. Compared with Figure 1 and the literature, we found that the temperature at which the seagull-like curve inflection appeared was different, meaning that the interaction between Cu and Mn changed the transition temperature [35,36]. This indicates that Cu 2+ and Mn species interact with each other. The N 2 selectivity of all catalysts exceeded 95% throughout all of the temperature range, showing that this method not only increased activity, but also maintained high N 2 selectivity. conversion in the range of 180-475 °C. This indicates that it exhibits not only remarkable low temperature activity but also excellent hydrothermal stability, which is quite beneficial for the practical application of commercial SCR catalysts [42,43]. The Cu(0.3) catalyst had the most serious activity decrease, maintaining 90% NO conversion only between 250-425 °C. Furthermore, when the temperature was higher than 400 °C, all the catalysts NO conversion rate dropped sharply, while for fresh sample it needed 500 °C. The Cu(0.2) catalyst had a similar activity like the Cu(0.3) catalyst between 100 and 425 °C while it had a better performance in high-temperature range. This is due to the difference of NH3 oxidation ability of catalysts and will be discussed in detail below. The N2 selectivity of the three aged catalysts was slightly lower than that of the fresh samples, and the selectivity of the Cu(0.2)Mn(0.1) sample was the best which was consistent with the activity.
After hydrothermal aging the deterioration of the catalysts performance in high temperature was more serious than that in the low temperature section, which indicated that the catalyst activity in the high temperature section was more sensitive to hydrothermal aging. This is because during hydrothermal aging parts of active species translate from Cu 2+ or Cu(OH) + (two forms of Cu ions in the zeolite [44,45]) to CuxOy [20,23]. These species have stronger oxidation than Cu ions and could oxidize the reducing agent NH3 through the non-selective oxidation as shown in Formula 1, which is the main side reaction in SCR [31].
Since in the standard SCR process the NH3 molecules and NO molecules are introduced at 1:1, the loss of reducing agent leads to a decrease in the NO conversion efficiency.  Hydrothermal aging is one of the common reasons for the deactivation of NH 3 -SCR catalysts for diesel vehicles and it is important to evaluate the catalytic activity of catalysts after hydrothermal aging. Figure 2b is the activity curve of the catalysts after hydrothermal aging. As it shows, all the catalysts had varying degrees of deterioration compared with the fresh sample and the Cu(0.2)Mn(0.1) still possessed the best catalytic activity and maintained more than 90% NO conversion in the range of 180-475 • C. This indicates that it exhibits not only remarkable low temperature activity but also excellent hydrothermal stability, which is quite beneficial for the practical application of commercial SCR catalysts [42,43]. The Cu(0.3) catalyst had the most serious activity decrease, maintaining 90% NO conversion only between 250-425 • C. Furthermore, when the temperature was higher than 400 • C, all the catalysts NO conversion rate dropped sharply, while for fresh sample it needed 500 • C. The Cu(0.2) catalyst had a similar activity like the Cu(0.3) catalyst between 100 and 425 • C while it had a better performance in high-temperature range. This is due to the difference of NH 3 oxidation ability of catalysts and will be discussed in detail below. The N 2 selectivity of the three aged catalysts was slightly lower than that of the fresh samples, and the selectivity of the Cu(0.2)Mn(0.1) sample was the best which was consistent with the activity.
After hydrothermal aging the deterioration of the catalysts performance in high temperature was more serious than that in the low temperature section, which indicated that the catalyst activity in the high temperature section was more sensitive to hydrothermal aging. This is because during hydrothermal aging parts of active species translate from Cu 2+ or Cu(OH) + (two forms of Cu ions in the zeolite [44,45]) to Cu x O y [20,23]. These species have stronger oxidation than Cu ions and could oxidize the reducing agent NH 3 through the non-selective oxidation as shown in Formula (1), which is the main side reaction in SCR [31].
Since in the standard SCR process the NH 3 molecules and NO molecules are introduced at 1:1, the loss of reducing agent leads to a decrease in the NO conversion efficiency.

NH 3 and NO Oxidation Test
NH 3 oxidation capacity is an important indicator of SCR catalysts and the suitable oxidizability of the catalyst is pivotal to the SCR reaction [31,37,43,46]. Figure 3 shows the NH 3 oxidation performance curve of the three catalysts before and after hydrothermal aging. It can be seen from Figure 3a that all the catalysts had weak NH 3 oxidation ability below 200 • C and when the temperature was higher than 250 • C the oxidation capacity started to rise rapidly, both Cu 2+ and Cu(OH) + had a weak ability to oxidize NH 3 to N 2 at low temperatures [11]. As the temperature increased, the oxidation rate of NH 3 rose continuously and reached 100% at 450 • C. The Cu(0.2)Mn(0.1) and Cu(0.2) catalysts had very similar oxidation activities, meaning that the Mn introduced by the co-exchange method did not cause excessive oxidation of NH 3 . Comparing Figure 3a,b, we can see that the catalysts' NH 3 oxidation ability increased after hydrothermal aging. Too high of an NH 3 oxidation capacity leads to non-selective oxidation of reducing NH 3 , which is a major cause of reduced activity [20]. Among the samples, the Cu(0.3) sample had the highest NH 3 oxidation promotion, indicating that samples with a high Cu content generated the most Cu x O y species during hydrothermal aging, leading to a higher oxidation ability [20,23,32,47]. This is the reason why the Cu(0.3) catalyst performed the worst after hydrothermal aging, especially in the high temperature range.

NH3 and NO Oxidation Test
NH3 oxidation capacity is an important indicator of SCR catalysts and the suitable oxidizability of the catalyst is pivotal to the SCR reaction [31,37,43,46]. Figure 3 shows the NH3 oxidation performance curve of the three catalysts before and after hydrothermal aging. It can be seen from Figure 3a that all the catalysts had weak NH3 oxidation ability below 200 °C and when the temperature was higher than 250 °C the oxidation capacity started to rise rapidly, both Cu 2+ and Cu(OH) + had a weak ability to oxidize NH3 to N2 at low temperatures [11]. As the temperature increased, the oxidation rate of NH3 rose continuously and reached 100% at 450 °C. The Cu(0.2)Mn(0.1) and Cu(0.2) catalysts had very similar oxidation activities, meaning that the Mn introduced by the co-exchange method did not cause excessive oxidation of NH3. Comparing Figure  3a,b, we can see that the catalysts' NH3 oxidation ability increased after hydrothermal aging. Too high of an NH3 oxidation capacity leads to non-selective oxidation of reducing NH3, which is a major cause of reduced activity [20]. Among the samples, the Cu(0.3) sample had the highest NH3 oxidation promotion, indicating that samples with a high Cu content generated the most CuxOy species during hydrothermal aging, leading to a higher oxidation ability [20,23,32,47]. This is the reason why the Cu(0.3) catalyst performed the worst after hydrothermal aging, especially in the high temperature range. NO oxidation is also relevant to better understand the SCR process [48,49]. As previously reported, isolated Cu species do not catalyze NO oxidation at low temperatures and NO oxidation only occurs when relatively stable O-bridged Cu-dimers are formed [35,49]. Therefore, the oxidation activity of NO reflects the difficulty of forming Cu-dimers in the catalyst. Figure 4a,b shows the performance of NO oxidation of the three catalysts before and after hydrothermal aging. All the activity curves exhibited a volcanic-like shape [20,35]. With the increase of the temperature, the NO conversion rate rose to reach the highest at around 400 °C and then decreased. This is because when the temperature is higher than 400 °C, Cu ions lose their complex small molecules and reduce their migration ability, which is not conducive to the formation of Cu dimers [44,50]. Furthermore, Cu dimers are more readily formed when there are more Cu ions, thus the Cu(0.3) sample had the highest NO oxidation capacity. Unlike for the NH3 oxidation process, the Cu(0.2) and Cu(0.2)Mn(0.1) catalysts exhibited different catalytic activities in the NO oxidation reaction, though they had the same Cu content. This indicates that the addition of Mn species has an inhibitory effect on the migration of Cu ions to generate Cu dimers.
One possible way to enhance the SCR activity at low temperatures is to oxidize some of the NO to NO2 so that a fast SCR occurs [47]. The reaction equation is presented below: NO oxidation is also relevant to better understand the SCR process [48,49]. As previously reported, isolated Cu species do not catalyze NO oxidation at low temperatures and NO oxidation only occurs when relatively stable O-bridged Cu-dimers are formed [35,49]. Therefore, the oxidation activity of NO reflects the difficulty of forming Cu-dimers in the catalyst. Figure 4a,b shows the performance of NO oxidation of the three catalysts before and after hydrothermal aging. All the activity curves exhibited a volcanic-like shape [20,35]. With the increase of the temperature, the NO conversion rate rose to reach the highest at around 400 • C and then decreased. This is because when the temperature is higher than 400 • C, Cu ions lose their complex small molecules and reduce their migration ability, which is not conducive to the formation of Cu dimers [44,50]. Furthermore, Cu dimers are more readily formed when there are more Cu ions, thus the Cu(0.3) sample had the highest NO oxidation capacity. Unlike for the NH 3 oxidation process, the Cu(0.2) and Cu(0.2)Mn(0.1) catalysts exhibited different catalytic activities in the NO oxidation reaction, though they had the same Cu content. This indicates that the addition of Mn species has an inhibitory effect on the migration of Cu ions to generate Cu dimers. was mainly caused by the promotion of Mn species. Figure 4b shows the ability of the catalysts to catalyze the oxidation of NO was greatly decreased after hydrothermal aging. As discussed above, some of the isolated Cu 2+ or Cu(OH) + species translate into CuxOy, thus reducing the ability of catalysts to catalyze the oxidation of NO. Furthermore, the change of the NO oxidation capacity of the Cu(0.2)Mn(0.1) catalyst after hydrothermal aging was minimal, which also indicates that this catalyst had the best resistance to hydrothermal aging.

Temperature-Programmed Desorption with NH3 (NH3-TPD)
Hydrothermal aging can seriously affect the acidity of the catalysts, which is an important factor affecting catalytic performance [48]. NH3-TPD experiments are often used to study the acidity of SSZ-13-based catalysts [38,44]. Figure 5 shows the NH3 desorption profiles of the samples before and after hydrothermal aging. The fresh samples show the presence of three desorption peaks, centered at about 150 °C (low temperature, LT), 300 °C (intermediate temperature, IT), and 475 °C (high temperature, HT). The LT peak corresponding to the weak adsorption of NH3 is assigned to physically adsorbed NH3 and NH3 adsorbed on weak acid sites [31,40,51]. The IT peak is attributed to NH3 desorption from Lewis (L) acid sites and the HT peak is due to strongly bound NH3, arising from protonated NH3 formed over the B acid sites [31,44,52]. Notably, the Cu(0.3) and the Cu(0.2)Mn(0.1) samples presented similar NH3-TPD curves and the areas of every peak were also substantially uniform. This is because they have the same total ion exchange degree, so they have the same L acid sites. (the L sites are generated by metal elements entering the zeolite via the co-exchange method [44]) This also indicates that the adsorption strength of NH3 on Cu sites and Mn sites are similar. As for the Cu(0.2) sample, it was clear that the IT peak was smaller than that of others while the HT peak was a little bigger, as this sample had the lowest ion exchange degree. The LT peak areas of the three fresh samples were almost identical since they all had the same structure.
After hydrothermal aging, the NH3 adsorption capacity was greatly decreased and the desorption peaks at different temperatures were various. All the LT peaks had a reduced area, which indicates that the structure of the catalysts was damaged during the hydrothermal aging. The area reduction of the Cu(0.2) sample was the most obvious, indicating that its damage was the most serious. The peaks in the HT range almost completely disappeared, revealing the fact that the B acid sites are the most vulnerable during the hydrothermal aging [22]. Meanwhile, the destruction of B acid sites causes a chain reaction that leads to the damage of the catalysts [23,53]. Thus, this influence can be effectively improved by increasing the amount of ion exchange to reduce the amount of residual B acid sites. It is interesting to note that the amount of NH3 adsorbed at the IT segment One possible way to enhance the SCR activity at low temperatures is to oxidize some of the NO to NO 2 so that a fast SCR occurs [47]. The reaction equation is presented below: Our experiment proves that the introduction of Mn species by the co-exchange method in increasing the low-temperature activity of the catalyst does not follow the above mechanism and it was mainly caused by the promotion of Mn species. Figure 4b shows the ability of the catalysts to catalyze the oxidation of NO was greatly decreased after hydrothermal aging. As discussed above, some of the isolated Cu 2+ or Cu(OH) + species translate into Cu x O y , thus reducing the ability of catalysts to catalyze the oxidation of NO. Furthermore, the change of the NO oxidation capacity of the Cu(0.2)Mn(0.1) catalyst after hydrothermal aging was minimal, which also indicates that this catalyst had the best resistance to hydrothermal aging.

Temperature-Programmed Desorption with NH 3 (NH 3 -TPD)
Hydrothermal aging can seriously affect the acidity of the catalysts, which is an important factor affecting catalytic performance [48]. NH 3 -TPD experiments are often used to study the acidity of SSZ-13-based catalysts [38,44]. Figure 5 shows the NH 3 desorption profiles of the samples before and after hydrothermal aging. The fresh samples show the presence of three desorption peaks, centered at about 150 • C (low temperature, LT), 300 • C (intermediate temperature, IT), and 475 • C (high temperature, HT). The LT peak corresponding to the weak adsorption of NH 3 is assigned to physically adsorbed NH 3 and NH 3 adsorbed on weak acid sites [31,40,51]. The IT peak is attributed to NH 3 desorption from Lewis (L) acid sites and the HT peak is due to strongly bound NH 3 , arising from protonated NH 3 formed over the B acid sites [31,44,52]. Notably, the Cu(0.3) and the Cu(0.2)Mn(0.1) samples presented similar NH 3 -TPD curves and the areas of every peak were also substantially uniform. This is because they have the same total ion exchange degree, so they have the same L acid sites. (the L sites are generated by metal elements entering the zeolite via the co-exchange method [44]) This also indicates that the adsorption strength of NH 3 on Cu sites and Mn sites are similar. As for the Cu(0.2) sample, it was clear that the IT peak was smaller than that of others while the HT peak was a little bigger, as this sample had the lowest ion exchange degree. The LT peak areas of the three fresh samples were almost identical since they all had the same structure. This is because Cu species move to more thermodynamically stable sites at high temperatures [44,50,52]. At the same time, the curve of the Cu(0.2)Mn(0.1) sample was similar to that of the Cu(0.3) sample, which suggests that the Mn species may also move to more thermodynamically stable sites at high temperatures, although more experiments are needed to clearly illustrate this problem. In summary, by the co-exchange method, increasing the total ion exchange degree can slow down the destruction of the acidity and structure.

XRD
In order to illustrate the changes of phase of the three catalysts before and after hydrothermal aging, their XRD patterns are shown in Figure 6. From Figure 6a, it can be seen that the XRD patterns of these fresh samples are similar to those reported in the literature for SSZ-13 materials [54]. The CHA phase (JCPDS file No.47-0762) was observed as the only crystalline phase in all samples. In each case, the diffraction intensities of the CHA phase of all samples had no difference. This indicates that the introduction of a small number of Cu and Mn elements does not influence the host structure. However, there was an obvious decrease of the diffraction intensities of the CHA phase for aged samples. This indicates that the crystal phase of the catalysts was damaged, resulting in a decrease in crystallinity. The maintained CHA phase can be used to explain the stable catalytic performance of NH3-SCR over these catalysts. It should be note that the degree of damage of the Cu(0.2) sample was more serious than that of Cu(0.2)Mn(0.1) and Cu(0.3) samples. This means that the total ion exchange degree has a great influence on the stability. As mentioned above this is because the B acid sites are susceptible to damage during hydrothermal aging, while the higher exchange degree means that more H + sites in SSZ-13 can be replaced by metal ions, which effectively reduce the amount of residual B acid sites and enhance the stability of the catalysts [22,23,55]. After hydrothermal aging, the NH 3 adsorption capacity was greatly decreased and the desorption peaks at different temperatures were various. All the LT peaks had a reduced area, which indicates that the structure of the catalysts was damaged during the hydrothermal aging. The area reduction of the Cu(0.2) sample was the most obvious, indicating that its damage was the most serious. The peaks in the HT range almost completely disappeared, revealing the fact that the B acid sites are the most vulnerable during the hydrothermal aging [22]. Meanwhile, the destruction of B acid sites causes a chain reaction that leads to the damage of the catalysts [23,53]. Thus, this influence can be effectively improved by increasing the amount of ion exchange to reduce the amount of residual B acid sites. It is interesting to note that the amount of NH 3 adsorbed at the IT segment increased. As mentioned above, Cu species change during hydrothermal aging. In addition to some of the isolate Cu species being converted to Cu x O y clusters, some of the Cu(OH) + specie were converted to Cu 2+ specie by the following equation: This is because Cu species move to more thermodynamically stable sites at high temperatures [44,50,52]. At the same time, the curve of the Cu(0.2)Mn(0.1) sample was similar to that of the Cu(0.3) sample, which suggests that the Mn species may also move to more thermodynamically stable sites at high temperatures, although more experiments are needed to clearly illustrate this problem. In summary, by the co-exchange method, increasing the total ion exchange degree can slow down the destruction of the acidity and structure.

XRD
In order to illustrate the changes of phase of the three catalysts before and after hydrothermal aging, their XRD patterns are shown in Figure 6. From Figure 6a, it can be seen that the XRD patterns of these fresh samples are similar to those reported in the literature for SSZ-13 materials [54]. The CHA phase (JCPDS file No.47-0762) was observed as the only crystalline phase in all samples. In each case, the diffraction intensities of the CHA phase of all samples had no difference. This indicates that the introduction of a small number of Cu and Mn elements does not influence the host structure. However, there was an obvious decrease of the diffraction intensities of the CHA phase for aged samples. This indicates that the crystal phase of the catalysts was damaged, resulting in a decrease in crystallinity. The maintained CHA phase can be used to explain the stable catalytic performance of NH 3 -SCR over these catalysts. It should be note that the degree of damage of the Cu(0.2) sample was more serious than that of Cu(0.2)Mn(0.1) and Cu(0.3) samples. This means that the total ion exchange degree has a great influence on the stability. As mentioned above this is because the B acid sites are susceptible to damage during hydrothermal aging, while the higher exchange degree means that more H + sites in SSZ-13 can be replaced by metal ions, which effectively reduce the amount of residual B acid sites and enhance the stability of the catalysts [22,23,55].

NMR
NMR is a commonly used method to characterize the changes in a zeolite framework structure. Figure 7a. displays the 27 Al spectra acquired from hydrated, ambient samples before and after hydrothermal aging; the corresponding 29 Si NMR spectra are presented in Figure 7b. The features appearing at 58 ppm are attributed to the framework Al (Alf) and the features appearing at 0 ppm are attributed to the extra framework Al (EFAl) [53,56,57]. EFAl also has a certain catalytic activity and it can be seen from the spectrum that all three fresh samples contained almost no EFA1, thus eliminating the effect of the EFAl on the catalyst [40]. This is because the zeolite used in our experiments had a relatively high Si/Al ratio (20), which is less prone to generate EFAl [48,56]. After hydrothermal aging, the Alf of every sample greatly decreased and the corresponding EFAL slightly increased, indicating that the dealumination of the zeolite skeleton occurred during hydrothermal aging-one of the main reasons leading to the destruction of the zeolite. Note that the reduction of the Alf was much larger than the increase of the EFAl. This is because some of Al can be converted to NMR-silent species [23,45,58]. Since the acidity of the zeolite is caused by the coordination imbalance of Al and O atoms, the loss of this portion of Al can result in a decrease in the total acidity of the zeolite. This is consistent with the NH3-TPD results. Among the three aged samples, the Alf of the Cu(0.2) was the most severely decreased while the changes of the Cu(0.3) and the Cu(0.2)Mn(0.1) samples were much slighter. This is similar to the result of XRD and it means that increasing the total ion exchange degree of the zeolite can reduce the dealumination of the zeolite. The 29 Si NMR spectra show the same result. The features appearing at −111 ppm are attributed to Si(4Si, 0Al) and the features appearing at −105 ppm are attributed to Si(3Si, 1Al) [53,56,57]. After hydrothermal aging, the −105 ppm feature declined while the −111 ppm feature rose, consistent with catalyst dealumination. It is clear that the Cu(0.2) sample had the most obvious change in peak intensity, also indicating that the degree of total ion exchange has an important influence on the structural stability of the catalysts.
From the discussion above, it can be seen that the total ion exchange plays an important role in the hydrothermal aging process of the catalyst. The content of residual B acid sites in the catalyst can be decreased, thus the acidity and structural damage of the catalyst during hydrothermal aging can be effectively reduced by increasing the total ion exchange degree of the catalysts. However, simply increasing the content of Cu 2+ to increase the total ion exchange degree makes the catalyst prone to creating CuxOy species, as shown in the NH3 oxidation experiment, which leads to non-selective reduction reactions. Meanwhile, simply increasing the content of Cu 2+ leads to the damage of the catalyst's pores during hydrothermal aging and this will be discussed in detail below.

NMR
NMR is a commonly used method to characterize the changes in a zeolite framework structure. Figure 7a. displays the 27 Al spectra acquired from hydrated, ambient samples before and after hydrothermal aging; the corresponding 29 Si NMR spectra are presented in Figure 7b. The features appearing at 58 ppm are attributed to the framework Al (Alf) and the features appearing at 0 ppm are attributed to the extra framework Al (EFAl) [53,56,57]. EFAl also has a certain catalytic activity and it can be seen from the spectrum that all three fresh samples contained almost no EFA1, thus eliminating the effect of the EFAl on the catalyst [40]. This is because the zeolite used in our experiments had a relatively high Si/Al ratio (20), which is less prone to generate EFAl [48,56]. After hydrothermal aging, the Alf of every sample greatly decreased and the corresponding EFAL slightly increased, indicating that the dealumination of the zeolite skeleton occurred during hydrothermal aging-one of the main reasons leading to the destruction of the zeolite. Note that the reduction of the Alf was much larger than the increase of the EFAl. This is because some of Al can be converted to NMR-silent species [23,45,58]. Since the acidity of the zeolite is caused by the coordination imbalance of Al and O atoms, the loss of this portion of Al can result in a decrease in the total acidity of the zeolite. This is consistent with the NH 3 -TPD results. Among the three aged samples, the Alf of the Cu(0.2) was the most severely decreased while the changes of the Cu(0.3) and the Cu(0.2)Mn(0.1) samples were much slighter. This is similar to the result of XRD and it means that increasing the total ion exchange degree of the zeolite can reduce the dealumination of the zeolite. The 29 Si NMR spectra show the same result. The features appearing at −111 ppm are attributed to Si(4Si, 0Al) and the features appearing at −105 ppm are attributed to Si(3Si, 1Al) [53,56,57]. After hydrothermal aging, the −105 ppm feature declined while the −111 ppm feature rose, consistent with catalyst dealumination. It is clear that the Cu(0.2) sample had the most obvious change in peak intensity, also indicating that the degree of total ion exchange has an important influence on the structural stability of the catalysts. Catalysts 2019, 9, x FOR PEER REVIEW 9 of 15 Figure 7. Solid-state 27 Al-NMR spectra of the samples of catalysts before and after hydrothermal aging (a) and solid-state 29 Si-NMR spectra of the samples of catalysts before and after hydrothermal aging (b). Figure 8 shows the SEM images of the fresh and aged samples. The fresh samples exhibited a cube shape, which is the typical morphology of SSZ-13 and the average particle size of the catalyst was about 200 nm. In addition, the images indicate that the catalysts showed a perfect degree of crystallization which is in accordance with the XRD results. After hydrothermal aging, most of the particles still retained the structure of the cube. That means that the structure of the catalysts was partially destroyed after hydrothermal aging while its basic structure ca=ould be maintained and this consistent with our findings above. It is clear from the mapping pictures shown in Figure S5 that the Cu element and Mn element of the Cu(0.2)Mn(0.1) sample were quite evenly distributed. Figure 9 depicts the curves of the micropores and mesopores before and after the hydrothermal aging of the catalysts. In the micropores (diameter <2 nm), as Figure 9a,c,e show, the pores of the samples were concentrated at about 0.38 nm, which is a perfect reflection of the theoretical pore size of the CHA zeolite. It can be seen that after hydrothermal aging, the Cu(0.3) sample suffered the largest reduction of microporous pores, while the Cu(0.2)Mn(0.1) sample exhibited the least reduction. Meanwhile, as shown in the mesoporous portion, the amount of mesopores (2~50 nm as shown in Figure 9b,d,f) increased after hydrothermal aging. Consistent with the change of micropores, the increase of the mesopores of the Cu(0.3) sample was the largest while there was almost no difference in the mesopores of the Cu(0.2)Mn(0.1) sample. This is also important evidence for the pore destruction of the catalysts.

Physical Properties
The destruction of the pore structure of the catalyst is due to the formation and migration of CuxOy, with a larger than the primary pore size (0.38 nm) of SSZ-13 [23]. The results indicate that addition of Mn species can slow down the formation of CuxOy and alleviate the damage of the pore structure of the catalysts. The Brunauer-Emmett-Teller (BET) specific surface area embedded in Figure 9 also reflects the extent to which the catalysts were affected by hydrothermal aging. All the samples before hydrothermal aging had similar specific surface areas and their slight differences may have been due to subtle differences in the preparation process. After hydrothermal aging, the specific surface area of the Cu(0.2)Mn(0.1) was the largest, and that of the Cu(0.3) was the smallest also revealing that the addition of Mn species greatly slowed down the damage of the catalyst during hydrothermal aging. From the discussion above, it can be seen that the total ion exchange plays an important role in the hydrothermal aging process of the catalyst. The content of residual B acid sites in the catalyst can be decreased, thus the acidity and structural damage of the catalyst during hydrothermal aging can be effectively reduced by increasing the total ion exchange degree of the catalysts. However, simply increasing the content of Cu 2+ to increase the total ion exchange degree makes the catalyst prone to creating Cu x O y species, as shown in the NH 3 oxidation experiment, which leads to non-selective reduction reactions. Meanwhile, simply increasing the content of Cu 2+ leads to the damage of the catalyst's pores during hydrothermal aging and this will be discussed in detail below. Figure 8 shows the SEM images of the fresh and aged samples. The fresh samples exhibited a cube shape, which is the typical morphology of SSZ-13 and the average particle size of the catalyst was about 200 nm. In addition, the images indicate that the catalysts showed a perfect degree of crystallization which is in accordance with the XRD results. After hydrothermal aging, most of the particles still retained the structure of the cube. That means that the structure of the catalysts was partially destroyed after hydrothermal aging while its basic structure ca=ould be maintained and this consistent with our findings above. It is clear from the mapping pictures shown in Figure S5 that the Cu element and Mn element of the Cu(0.2)Mn(0.1) sample were quite evenly distributed. Figure 9 depicts the curves of the micropores and mesopores before and after the hydrothermal aging of the catalysts. In the micropores (diameter <2 nm), as Figure 9a,c,e show, the pores of the samples were concentrated at about 0.38 nm, which is a perfect reflection of the theoretical pore size of the CHA zeolite. It can be seen that after hydrothermal aging, the Cu(0.3) sample suffered the largest reduction of microporous pores, while the Cu(0.2)Mn(0.1) sample exhibited the least reduction. Meanwhile, as shown in the mesoporous portion, the amount of mesopores (2~50 nm as shown in Figure 9b,d,f) increased after hydrothermal aging. Consistent with the change of micropores, the increase of the mesopores of the Cu(0.3) sample was the largest while there was almost no difference in the mesopores of the Cu(0.2)Mn(0.1) sample. This is also important evidence for the pore destruction of the catalysts.

Physical Properties
The destruction of the pore structure of the catalyst is due to the formation and migration of Cu x O y , with a larger than the primary pore size (0.38 nm) of SSZ-13 [23]. The results indicate that addition of Mn species can slow down the formation of Cu x O y and alleviate the damage of the pore structure of the catalysts. The Brunauer-Emmett-Teller (BET) specific surface area embedded in Figure 9 also reflects the extent to which the catalysts were affected by hydrothermal aging. All the samples before hydrothermal aging had similar specific surface areas and their slight differences may have been due to subtle differences in the preparation process. After hydrothermal aging, the specific surface area of the Cu(0.2)Mn(0.1) was the largest, and that of the Cu(0.3) was the smallest also revealing that the addition of Mn species greatly slowed down the damage of the catalyst during hydrothermal aging.     Figure 9. Micropore size distributions of catalysts before and after hydrothermal aging (a,c,e) and mesopore size distributions before and after hydrothermal aging (b,d,f).

Catalyst Preparation
The SSZ-13 zeolite was prepared using a hydrothermal method with the N, N, N-trimethyl-1-1-adamantammonium hydroxide (TMAdaOH) serving as a structure-directing agent (SDA). Firstly, 4 g of NaAlO 2 and 0.2 g NaOH were added to 20 g of deionized water and the mixture was stirred at room temperature for 30 min. Then 6 g SDA was added to the solution and 20 g of silica sol (40%) was added dropwise with constant stirring. After stirring for 4 h, the sol was transferred into Teflon-lined autoclaves and reacted at 160 • C for 96 h. After that, the product was separated from the mother liquid via centrifugation and washing with deionized water three times. Finally, the solid powder was dried overnight at 120 • C, then calcined at 390 • C for 1 hour and then heated at 2 • C/min to 550 • C and held for 6 h.
The Cu-SSZ-13 catalysts were prepared using a standard two-step solution ion exchange method. First, the obtained Na-SSZ-13 was exchanged three times with a 0.1 M solution of NH 4 NO 3 at 80 • C for 12 h to generate the NH 4 -SSZ-13. Then the Cu-SSZ-13 was obtained by exchanging with a 0.1 M solution of Cu(NO 3 ) and the pH of the solution was adjusted to 3 by adding dilute HNO 3 solution (10%). Similarly, Cu-Mn-SSZ-13 samples with different Mn/Cu ratios were prepared by ion exchange under a mixed solution of Cu(NO 3 ) 2 and Mn(NO 3 ) 2 (co-exchange method). By changing the Mn/Cu ratio in the solution, Cu-Mn-SSZ-13 catalysts with different ion exchange degrees were obtained and the content of Cu and Mn was measured by inductively coupled plasma atomic emission spectroscopy (ICP-AES). The ion exchange degree was calculated by Formula 5 and the results are shown in Table 1, at the same time, the Si/Al molar ratio of all samples was 20. Finally, all the solid powders were calcined in air at 550 • C for 6 h.
The Cu-Mn-SSZ-13 samples with different molar ratio of Mn/Cu catalysts were named as Cu(x)Mn(y)-F(A), in which x and y represents the ion exchange degree of Cu and Mn elements, and fresh or hydrothermally aged samples were distinguished by the suffix letter F or A
The element contents were determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES) (Massachusetts, USA).
Temperature programmed reduction with hydrogen (H 2 -TPR) experiments were performed on a Micromeritics Auto Chem II chemisorption analyzer (Norcross, USA). In each experiment, 100 mg of sample was loaded into a quartz reactor and then pretreated in He (50 mL/min) at 400 • C for 2 h. The reduction of the sample was carried out from 50 • C to 600 • C under a flow of 10% H 2 /Ar (50 mL/min) at 10 • C/min. Low-temperature N 2 adsorption-desorption analysis was carried out at 77 K using a Micromeritics ASAP 2460 instrument (Norcross, USA). Before the measurement, all samples were dried overnight at 120 • C, then treated under vacuum at 90 • C for 3 h and finally treated under vacuum at 250 • C for 6 h. The specific surface area, mesoporous and micropore size were determined by the Brunauer-Emmett-Teller (BET) method, Barrett-Joyner-Halenda (BJH) method and Horvath-Kawazoe (H-K) method, respectively.
Temperature programmed desorption with NH 3 (NH 3 -TPD) tests were conducted on a Micromeritics Auto Chem II chemisorption analyzer (Norcross, USA). The powder samples were about 100 mg and before testing the powder samples needed to be pre-treated with He for 2 h at 400 • C, then cooled to 50 • C. After that, the samples were exposed to 2% NH 3 /He for 2 h at 50 • C, followed by a purge in He at 50 • C for 2 h. Finally, the TPD data were obtained by heating the samples from 50 • C to 750 • C with a heating rate of 10 • C/min under pure He (100 mL/min).
Solid-state nuclear magnetic resonance (NMR) analysis of the samples was conducted on an Infinityplus 300 instrument (California, USA). About 0.1 g of the samples were transferred into a gastight rotor, and the rotor mounted into the NMR spectrometer. 27 Al chemical shifts were reported relative to a 0.1 M aqueous AlCl 3 solution and 29 Si chemical shifts were reported relative to a Si(CH 3 ) 4 standard.
Scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS) was carried out on Hitachi S-4800 instrument (Tokyo, Japan) using a 3.0 kV electron beam and before test, all samples were sprayed with gold for 50 s.

Catalytic Activity and Hydrothermal Stability Measurements
NH 3 -SCR experiments were performed in a fixed bed reactor with a 12 mm inner diameter. The catalysts were crushed and sieved into 40-60 mesh before the catalytic measurements. The reaction conditions were 500 ppm NO, 500 ppm NH 3 , 5 vol% O 2 , 2.5 vol% H 2 O, and the gas hourly space velocity (GHSV) = 30,000 h −1 . The concentrations of NH 3 , NO, NO 2 , and N 2 O were measured by a Fourier-transform infrared (FTIR) spectrometer (Thermo Nicolet IS 10).
The NH 3 and NO oxidation reactions were also conducted under the same conditions except that, in NH 3 oxidation, the NO feed was stopped while in the NO oxidation, the NH 3 feed was stopped.
The hydrothermal aging experiment was carried out in a vacuum tube furnace. The aging gas components included 75% N 2 , 15% O 2 and 10% H 2 O. The aging temperature was 800 • C and the GHSV = 30,000 h −1 .
The conversion of NO and NH 3 and the selectivity of N 2 were calculated by the following equations: N 2 selectivity = NH 3inlet + NO inlet − NH 3outlet − NO outlet − 2N 2 O outlet − NO 2 outlet NH 3inlet + NO inlet − NH 3outlet − NO outlet × 100%. (8)

Conclusions
In our study, it was found that Mn was a promising promoter and changed the structure, stability and performance of Cu-SSZ-13 catalysts. Therefore, some points can be put forward as follows: (1) Several Cu-Mn-SSZ-13 catalysts with different Mn/Cu molar ratio and well-crystallized SSZ-13 CHA structure were prepared by ion-exchange of Mn and Cu into SSZ-13 together.
(2) The Cu(0.2)Mn(0.1)-SSZ-13 catalyst has high reactivity at low-temperature and strong resistance to hydrothermal aging. It presents a more than 90% NO conversion in a wide temperature range (175-525 • C) and can still maintain more than 90% NO conversion in a temperature range of 180-475 • C after rigorous hydrothermal aging.
(3) The high total ion exchange degree is beneficial in reducing the residual B acid sites and thus decreasing the destruction of SSZ-13 phase and inhibiting the dealumination of the skeleton during the hydrothermal process.
(4) The moderate addition of Mn to Cu-SSZ-13 contribute to the high NH 3 -SCR activity and inhibits the aggregation of Cu species and the pore destruction of the catalyst during hydrothermal aging thus achieving excellent hydrothermal stability.
Author Contributions: C.S. and Y.L. conceived and designed the experiments; C.S., Z.L. and K.L. performed the experiments and contributed reagents and materials; C.S. wrote the manuscript; L.Z. and Y.L. checked and corrected the manuscript.
Funding: This research was funded by National Natural Science Foundation of China grant number [21776214].

Conflicts of Interest:
The authors declare no conflict of interest.