Analysis of Ion-Exchanged ZSM-5, BEA, and SSZ-13 Zeolite Trapping Materials under Realistic Exhaust Conditions

: An industry-deﬁned evaluation protocol was used to evaluate the hydrocarbon trapping (HCT) and passive NOx adsorption (PNA) potential for BEA, ZSM-5, and SSZ-13 zeolites with ion-exchanged Pd or Ag. All materials underwent 700 ◦ C degreening prior to exposure to an industry-derived protocol gas stream, which included NOx, ethylene, toluene, and decane as measured trapping species as well as common exhaust gasses CO, H 2 O, O 2 , CO 2 , and H 2 . Evaluation showed that BEA and ZSM-5 zeolites were effective at trapping hydrocarbons (HCs), as saturation was not achieved after 30 min of exposure. SSZ-13 also stored HCs but was only able to adsorb 20–25% compared to BEA and ZSM-5. The presence of Ag or Pd did not impact the overall HC uptake, particularly in the ﬁrst three minutes. Pd/zeolites had signiﬁcantly lower THC release temperature, and it aided in the conversion of the released HCs; Ag only had a moderate effect in both areas. With respect to NOx adsorption, the level of uptake was much lower than HCs on all samples, and Ag or Pd was necessary with Pd being notably more effective. Additionally, only Pd/ZSM-5 and Pd/SSZ-13 continue to store a portion of the NOx above 200 ◦ C, which is critical for downstream selective catalytic NOx reduction (SCR). Hydrothermal aging (800 ◦ C for 50 h) of a subset of the samples were performed: BEA, Pd/BEA, ZSM-5, Pd/ZSM-5, and Pd/SSZ-13. There was a minimal effect on the HC storage, ~10% reduction in capacity with no effect on release temperature; however, only Pd/SSZ-13 showed signiﬁcant NOx storage after aging.


Introduction
Since the 1970s automotive emissions control has been a continual challenge due to increasing understanding of the hazards of pollutants and subsequently increasing regulatory efforts to lower harmful emissions. Initially, much of the lean-burn automotive emissions reduction was handled through emissions-conscious engine design and the use of a diesel oxidation catalyst (DOC) [1][2][3][4][5][6][7]. However, modern lean-burn emissions control requires a complex approach involving a series of specialized materials [8][9][10][11][12][13][14] to manage the wide variety of pollutants and exhaust conditions due to the increasingly strict emissions standards. These include diesel particulate filters (DPF) and selective catalytic NOx reduction catalysts (SCR), in addition to the DOC. This emissions control system requires a minimum exhaust temperature to become fully effective, and thus, a significant percentage of the overall emissions occur in the first 1-2 min after starting the engine [15][16][17]. Additionally, improved engine efficiency results in cooler exhaust [18][19][20][21], such that under low load operation the exhaust and catalyst temperatures could temporarily decrease to the point the catalysts are not active in the emissions control system [20]. In most cases temperatures less than 200 • C would be considered a significant challenge for emissions control catalysts [20]. This temperature is typical for DOCs to be able to convert greater than 90% of the exhaust CO and total hydrocarbons (THCs), and above this temperature urea can be injected, which will hydrolyze to NH 3 and is used to reduce NOx over the SCR. Development of a solution to these low temperature emissions control challenges is crucial to meeting future emissions requirements [21,22].
One frequently proposed approach to reducing low temperature emissions is the introduction of trapping materials within the emissions control system that would be capable of adsorbing HC and NOx until the catalysts have warmed up enough to become reactive [16,[23][24][25][26][27]. Development of effective trapping materials often focuses on three primary factors: (i) a significant adsorption capacity at low temperatures (<200 • C), (ii) a desorption temperature compatible with emissions control active windows (greater than 200 • C), and (iii) suitable durability against hydrothermal aging. With these factors in mind, a variety of zeolites have shown great promise in studies for the adsorption of common exhaust components such as CO, light and heavy HCs [26][27][28][29][30], and NOx species [31][32][33][34]. Unfortunately, many fundamental studies are often conducted in simplified gas streams which result in many questions as to their applicability to automotive systems which can include high concentrations of water and CO 2 , multiple types of HC species, NOx, CO, and high space velocity. On the other end of the research spectrum, industrially relevant studies are performed with full exhaust gas from engines [35][36][37]. These realistic evaluations show how the materials would work in the application, but the hundreds of different HCs in the exhaust make it difficult to track the individual components. Thus, evaluation of potential trapping materials under complex, yet manageable flow constituents, is necessary to prove feasibility, highlight challenges, and identify which family of exhaust components are most challenging.
With that in mind, we present an evaluation of nine zeolite-based trapping materials utilizing a flow reactor protocol developed by a collaboration of industry, national laboratories, and academic researchers [38,39] to mimic realistic lean-burn automotive exhaust. As described in detail below, this protocol includes the opportunity to simultaneously trap CO, NOx, and three HCs representative of those found in automotive exhaust: ethylene (shortchain alkenes), toluene (aromatics), and decane (long-chain alkanes). ZSM-5, BEA, and SSZ-13 zeolites and two ion-exchange metals, Pd or Ag, were chosen for this study based on previous reports of high storage efficiency and suitable release temperatures [40][41][42][43][44][45][46]. The representative structure of each of these zeolites are shown in Figure 1. Evaluation of both the storage and release characteristics of these materials was performed to elucidate the attributes of each material.
Catalysts 2021, 11, x FOR PEER REVIEW 2 of 15 emissions control system [20]. In most cases temperatures less than 200 °C would be considered a significant challenge for emissions control catalysts [20]. This temperature is typical for DOCs to be able to convert greater than 90% of the exhaust CO and total hydrocarbons (THCs), and above this temperature urea can be injected, which will hydrolyze to NH3 and is used to reduce NOx over the SCR. Development of a solution to these low temperature emissions control challenges is crucial to meeting future emissions requirements [21,22]. One frequently proposed approach to reducing low temperature emissions is the introduction of trapping materials within the emissions control system that would be capable of adsorbing HC and NOx until the catalysts have warmed up enough to become reactive [16,[23][24][25][26][27]. Development of effective trapping materials often focuses on three primary factors: (i) a significant adsorption capacity at low temperatures (<200 °C), (ii) a desorption temperature compatible with emissions control active windows (greater than 200 °C), and (iii) suitable durability against hydrothermal aging. With these factors in mind, a variety of zeolites have shown great promise in studies for the adsorption of common exhaust components such as CO, light and heavy HCs [26][27][28][29][30], and NOx species [31][32][33][34]. Unfortunately, many fundamental studies are often conducted in simplified gas streams which result in many questions as to their applicability to automotive systems which can include high concentrations of water and CO2, multiple types of HC species, NOx, CO, and high space velocity. On the other end of the research spectrum, industrially relevant studies are performed with full exhaust gas from engines [35][36][37]. These realistic evaluations show how the materials would work in the application, but the hundreds of different HCs in the exhaust make it difficult to track the individual components. Thus, evaluation of potential trapping materials under complex, yet manageable flow constituents, is necessary to prove feasibility, highlight challenges, and identify which family of exhaust components are most challenging.
With that in mind, we present an evaluation of nine zeolite-based trapping materials utilizing a flow reactor protocol developed by a collaboration of industry, national laboratories, and academic researchers [38,39] to mimic realistic lean-burn automotive exhaust. As described in detail below, this protocol includes the opportunity to simultaneously trap CO, NOx, and three HCs representative of those found in automotive exhaust: ethylene (short-chain alkenes), toluene (aromatics), and decane (long-chain alkanes). ZSM-5, BEA, and SSZ-13 zeolites and two ion-exchange metals, Pd or Ag, were chosen for this study based on previous reports of high storage efficiency and suitable release temperatures [40][41][42][43][44][45][46]. The representative structure of each of these zeolites are shown in Figure 1. Evaluation of both the storage and release characteristics of these materials was performed to elucidate the attributes of each material.

Hydrocarbon Trapping
An ideal hydrocarbon trap (HCT) would have high storage below 200 • C and release HCs above 200 • C where a DOC could oxidize them to CO 2 and H 2 O. To understand how the evaluated samples meet these HCT criteria, a systematic analysis routine was employed. Figure 2 shows a characteristic storage and release profile for each of the three trials on Pd/ZSM-5. The difference between the reactor effluent and the bypass gas concentration indicates a large amount of storage on this sample, and we have analyzed this storage in 2 parts. The first portion is 3 min and is related to how much storage would be expected under a typical startup condition [47,48]. It is represented by the blue shading with white dots ( , in Figure 2a,b), but as can be seen here, the sample was not fully saturated as the bypass concentration was not reached. After the 30-min storage, the reactant gases are turned off, and the sample is quickly heated to 600 • C. During this portion of the experiment the HCs desorb from the sample, and some of them react to form CO and CO 2 . As suggested by the protocol [38,39], the release profile is described by three characteristic temperatures. The first is the temperature of 10% release (T10), i.e., the temperature where 10% of the observed desorbed HCs were released. This is also shown in Figure 2a with a dark blue shading ( ) beginning at 30 min (0% release) and continuing to~33 min (10% release); the temperature is also indicated with a dark blue arrow. Figure 2c indicates this value with the dark blue ( ) portion of the bar. The next segment indicates the temperature of 50% release (T50). In Figure 2a this is represented with the medium blue shading ( ) that reaches from 10% to 50% and the temperature is indicated with medium blue arrow and by the second section of the bar in Figure 2c ( ). The next segment of interest is for 90% release of the HCs, which is represented by a light blue shading ( ) for the 50-90% release in Figure 2a. The T90 is represented by a light blue arrow in Figure 2a  This approach is used for each of the samples evaluated, and the full bar charts for both HC storage values (mmols C 1 /g cat ) and release temperatures for the degreened samples are shown in Figure 3 with the values listed in Table 1. All the BEA and ZSM-5 samples have similar uptake during the first 3 min with an uptake of 0.9-1.0 mmols C 1 /g cat . The SSZ-13 samples have notably less uptake with only 0.6-0.7 mmols C 1 /g cat occurring in the first 3 min. After 30 min the BEA samples store between 4.8 and 5.1 mmols C 1 /g cat while the ZSM-5 samples have slightly less uptake of 4.3-5.0 mmols C 1 /g cat . The SSZ-13 samples reach saturation well before 30 min and only uptake a total of 0.9-1.0 mmols C 1 /g cat after 30 min. These results clearly indicate that the larger pore structures associated with BEA (0.7-0.8 nm) and ZSM-5 (0.5-0.6 nm) allow more uptake of the HCs studied here compared to SSZ-13 (~0.4 nm), which has smaller pores [49,50]. An additional observation is that the presence of Ag or Pd does not generally result in higher uptake of the HCs in any of these samples. The one exception to this observation is that Ag/ZSM-5 uptakes 5.0 mmols C 1 /g cat compared to 4.3-4.4 for the other ZSM-5 samples. Enhanced uptake capacity in the presence of Ag has been reported elsewhere [27,40,51], so it is interesting that we do not observe it in this study. A likely explanation is the large size and concentration of the decane molecule used here overwhelming the adsorption sites and zeolite pores; most prior work reporting these benefits of metals focused on shorter hydrocarbon lengths [27,40,51,52].  ) minutes of adsorption during uptake and the zones/temperatures associated with 10% (T10, ), 50% (T50, ), or 90% (T90, ) of THC released during desorption, which are also indicated with arrows. These characteristics are then represented by bar charts for both (b) adsorption and (c) desorption.   ) minutes for each of the materials studied. (b) The temperature where 10% (T10, ), 50% (T50, ), or 90% (T90, ) of THC is released. Table 1. THC storage values (mmol/g cat ) for ion-exchanged zeolites during lean trapping evaluation after 3 and 30 min; THCs are reported on a C 1 basis. The total amount of observed THC released during the first 15 min of the ramp, and the temperature where 10% (T10), 50% (T50), or 90% (T90) of THC is released is also shown. At the end of the storage phase, a portion of the HCs is released, but most of them stay adsorbed until higher temperatures. The temperature of 10% release (T10) is shown in Figure 3 and listed in Table 1, and there are no distinct trends across the nine samples other than that the Pd/zeolites have the lowest T10. Comparing T50s and T90s, the Pd/zeolites continue to show lower release temperatures for the BEA and ZSM-5 samples. Since a small quantity of HCs are adsorbed on the SSZ-13 samples direct comparisons with BEA and ZSM-5 are difficult, but the trend of T90 being the lowest on the Pd sample is consistent. One of the confounding factors in evaluating the samples as we describe here, is that Pd is an excellent oxidation catalyst and the released HCs are readily oxidized over these zeolites in the presence of excess O 2 [40]. Thus, even though 90% of the HCs are released below the previously mentioned goal of 200 • C, the ability to convert the HCs directly on the HCT is a desirable feature. The released HC quantity is listed in Table 1, and in each Pd containing sample the released HC is less than the quantity stored at 30 min. This is most apparent in the Pd/BEA where nearly 60% of the released HC is converted. This feature could possibly allow for a smaller DOC, or at least less PGM content to balance the PGM being used in the HCT; this is the focus of a future study of ours. It should be noted that although Ag is not an excellent oxidation catalyst, it is expected to catalyze a portion of the stored HCs, and there is evidence of this in the Ag/BEA sample with 12% less HCs observed during the release compared to the uptake. Another key observation in the release temperatures is that both BEA and Ag/BEA release 50% of the HC above 200 • C, and thus many DOCs would be able to oxidize the HCs upon their release. ZSM-5 and Ag/ZSM-5 have similar stability, but their T50s are 190 and 185 • C, respectively, and thus not as desirable as the BEA samples. The SSZ-13 zeolites not only store less HCs, but they release nearly all of the HCs by 150 • C, and thus continue to show undesirable HCT properties.

NO Storage
As discussed above an ideal passive NOx adsorber (PNA) would have high storage below 200 • C and release NOx above 200 • C where a downstream SCR could be used in conjunction with ammonia from hydrolyzed urea. This hydrolysis occurs above 133 • C [53], and the SCR reactivity typically reaches 90% by 180 • C [54,55]. The same systematic analysis procedure described above in the HC section was also used with NOx to understand how the evaluated samples meet the PNA criteria. Figure 4 shows a characteristic NOx storage and desorption profile; Pd/ZSM-5 is shown here, which is the same sample and experiment shown in Figure 2. Notably less NOx is adsorbed than HCs, but the adsorbed NOx is more strongly bound and releases at higher temperatures. Similar to THC, the uptake is divided into two sections, 3 ( At the end of the storage phase, a portion of the HCs is released, but most of them stay adsorbed until higher temperatures. The temperature of 10% release (T10) is shown in Figure 3 and listed in Table 1, and there are no distinct trends across the nine samples other than that the Pd/zeolites have the lowest T10. Comparing T50s and T90s, the Pd/zeolites continue to show lower release temperatures for the BEA and ZSM-5 samples. Since a small quantity of HCs are adsorbed on the SSZ-13 samples direct comparisons with BEA and ZSM-5 are difficult, but the trend of T90 being the lowest on the Pd sample is consistent. One of the confounding factors in evaluating the samples as we describe here, is that Pd is an excellent oxidation catalyst and the released HCs are readily oxidized over these zeolites in the presence of excess O2 [40]. Thus, even though 90% of the HCs are released below the previously mentioned goal of 200 °C, the ability to convert the HCs directly on the HCT is a desirable feature. The released HC quantity is listed in Table 1, and in each Pd containing sample the released HC is less than the quantity stored at 30 min. This is most apparent in the Pd/BEA where nearly 60% of the released HC is converted. This feature could possibly allow for a smaller DOC, or at least less PGM content to balance the PGM being used in the HCT; this is the focus of a future study of ours. It should be noted that although Ag is not an excellent oxidation catalyst, it is expected to catalyze a portion of the stored HCs, and there is evidence of this in the Ag/BEA sample with 12% less HCs observed during the release compared to the uptake. Another key observation in the release temperatures is that both BEA and Ag/BEA release 50% of the HC above 200 °C, and thus many DOCs would be able to oxidize the HCs upon their release. ZSM-5 and Ag/ZSM-5 have similar stability, but their T50s are 190 and 185 °C, respectively, and thus not as desirable as the BEA samples. The SSZ-13 zeolites not only store less HCs, but they release nearly all of the HCs by 150 °C, and thus continue to show undesirable HCT properties.

NO Storage
As discussed above an ideal passive NOx adsorber (PNA) would have high storage below 200 °C and release NOx above 200 °C where a downstream SCR could be used in conjunction with ammonia from hydrolyzed urea. This hydrolysis occurs above 133 °C [53], and the SCR reactivity typically reaches 90% by 180 °C [54,55]. The same systematic analysis procedure described above in the HC section was also used with NOx to understand how the evaluated samples meet the PNA criteria. Figure 4 shows a characteristic NOx storage and desorption profile; Pd/ZSM-5 is shown here, which is the same sample and experiment shown in Figure 2. Notably less NOx is adsorbed than HCs, but the adsorbed NOx is more strongly bound and releases at higher temperatures. Similar to THC, the uptake is divided into two sections, 3 ( ) and 30 ( ) minutes (Figure 4b), and the temperatures of 10%, 50% and 90% NOx release are determined ( Figure 4c). The data for all the samples are shown in Figure 5 with the specific values listed in Table 2. From these data two important observations are apparent. First, the Pd-zeolites are the only samples to have significant uptake, and second, only Pd/ZSM-5 and Pd/SSZ-13 maintains 50% of the NOx above 150 °C; in fact, over 50% of the NOx is still on Pd Pd/SSZ-13 at 224 °C. This finding is consistent with other reports of Pd/SSZ-13 being one of the best PNAs [56,57] and that NO storage only occurs on ion-exchanged Pd sites [41]. Neither catalyst closed the NOx balance during release, suggesting there is some direct reduction in the stored NOx from HCs. This typically yields some N2O formation, but this reactor was not equipped with an analyzer that would allow N2O to be measured accurately.

Hydrothermal Aging
After fully evaluating all the samples, five of the most promising zeolites were hydrothermally aged to investigate their durability. BEA, Pd/BEA, ZSM-5, Pd/ZSM-5, and Pd/SSZ-13 were heated to 800 °C for 50 h and re-evaluated following the same protocols as before. Figure 6 shows the HC metrics for both storage and release with the values listed in Table   ) and 30 (  At the end of the storage phase, a portion of the HCs is released, but most of them stay adsorbed until higher temperatures. The temperature of 10% release (T10) is shown in Figure 3 and listed in Table 1, and there are no distinct trends across the nine samples other than that the Pd/zeolites have the lowest T10. Comparing T50s and T90s, the Pd/zeolites continue to show lower release temperatures for the BEA and ZSM-5 samples. Since a small quantity of HCs are adsorbed on the SSZ-13 samples direct comparisons with BEA and ZSM-5 are difficult, but the trend of T90 being the lowest on the Pd sample is consistent. One of the confounding factors in evaluating the samples as we describe here, is that Pd is an excellent oxidation catalyst and the released HCs are readily oxidized over these zeolites in the presence of excess O2 [40]. Thus, even though 90% of the HCs are released below the previously mentioned goal of 200 °C, the ability to convert the HCs directly on the HCT is a desirable feature. The released HC quantity is listed in Table 1, and in each Pd containing sample the released HC is less than the quantity stored at 30 min. This is most apparent in the Pd/BEA where nearly 60% of the released HC is converted. This feature could possibly allow for a smaller DOC, or at least less PGM content to balance the PGM being used in the HCT; this is the focus of a future study of ours. It should be noted that although Ag is not an excellent oxidation catalyst, it is expected to catalyze a portion of the stored HCs, and there is evidence of this in the Ag/BEA sample with 12% less HCs observed during the release compared to the uptake. Another key observation in the release temperatures is that both BEA and Ag/BEA release 50% of the HC above 200 °C, and thus many DOCs would be able to oxidize the HCs upon their release. ZSM-5 and Ag/ZSM-5 have similar stability, but their T50s are 190 and 185 °C, respectively, and thus not as desirable as the BEA samples. The SSZ-13 zeolites not only store less HCs, but they release nearly all of the HCs by 150 °C, and thus continue to show undesirable HCT properties.

NO Storage
As discussed above an ideal passive NOx adsorber (PNA) would have high storage below 200 °C and release NOx above 200 °C where a downstream SCR could be used in conjunction with ammonia from hydrolyzed urea. This hydrolysis occurs above 133 °C [53], and the SCR reactivity typically reaches 90% by 180 °C [54,55]. The same systematic analysis procedure described above in the HC section was also used with NOx to understand how the evaluated samples meet the PNA criteria. Figure 4 shows a characteristic NOx storage and desorption profile; Pd/ZSM-5 is shown here, which is the same sample and experiment shown in Figure 2. Notably less NOx is adsorbed than HCs, but the adsorbed NOx is more strongly bound and releases at higher temperatures. Similar to THC, the uptake is divided into two sections, 3 ( ) and 30 ( ) minutes (Figure 4b), and the temperatures of 10%, 50% and 90% NOx release are determined ( Figure 4c). The data for all the samples are shown in Figure 5 with the specific values listed in Table 2. From these data two important observations are apparent. First, the Pd-zeolites are the only samples to have significant uptake, and second, only Pd/ZSM-5 and Pd/SSZ-13 maintains 50% of the NOx above 150 °C; in fact, over 50% of the NOx is still on Pd Pd/SSZ-13 at 224 °C. This finding is consistent with other reports of Pd/SSZ-13 being one of the best PNAs [56,57] and that NO storage only occurs on ion-exchanged Pd sites [41]. Neither catalyst closed the NOx balance during release, suggesting there is some direct reduction in the stored NOx from HCs. This typically yields some N2O formation, but this reactor was not equipped with an analyzer that would allow N2O to be measured accurately.

Hydrothermal Aging
After fully evaluating all the samples, five of the most promising zeolites were hydrothermally aged to investigate their durability. BEA, Pd/BEA, ZSM-5, Pd/ZSM-5, and Pd/SSZ-13 were heated to 800 °C for 50 h and re-evaluated following the same protocols as before. Figure 6 shows the HC metrics for both storage and release with the values listed in Table   ) minutes (Figure 4b), and the temperatures of 10%, 50% and 90% NOx release are determined (Figure 4c). The data for all the samples are shown in Figure 5 with the specific values listed in Table 2. From these data two important observations are apparent. First, the Pd-zeolites are the only samples to have significant uptake, and second, only Pd/ZSM-5 and Pd/SSZ-13 maintains 50% of the NOx above 150 • C; in fact, over 50% of the NOx is still on Pd Pd/SSZ-13 at 224 • C. This finding is consistent with other reports of Pd/SSZ-13 being one of the best PNAs [56,57] and that NO storage only occurs on ion-exchanged Pd sites [41]. Neither catalyst closed the NOx balance during release, suggesting there is some direct reduction in the stored NOx from HCs. This typically yields some N 2 O formation, but this reactor was not equipped with an analyzer that would allow N 2 O to be measured accurately.
functionality, it is clear that each of these hydrothermally aged samples would offer significant HCT capacity. The release profile of unexchanged BEA is slightly more favorable, since ~50% of the stored HCs are released above 200 °C, compared to 185 °C for ZSM-5. If Pd is going to be used for HCTs, Pd/BEA offers significant durability, and although Pd does not aid in the storage capacity, its oxidation capability is valuable with the ability to oxidize over 50% of the stored HCs even after hydrothermal aging.   of the storage phase, a portion of the HCs is released, but most of them until higher temperatures. The temperature of 10% release (T10) is shown listed in Table 1, and there are no distinct trends across the nine samples the Pd/zeolites have the lowest T10. Comparing T50s and T90s, the Pd/zeto show lower release temperatures for the BEA and ZSM-5 samples. Since y of HCs are adsorbed on the SSZ-13 samples direct comparisons with BEA difficult, but the trend of T90 being the lowest on the Pd sample is conthe confounding factors in evaluating the samples as we describe here, is cellent oxidation catalyst and the released HCs are readily oxidized over n the presence of excess O2 [40]. Thus, even though 90% of the HCs are the previously mentioned goal of 200 °C, the ability to convert the HCs HCT is a desirable feature. The released HC quantity is listed in Table 1, containing sample the released HC is less than the quantity stored at 30 ost apparent in the Pd/BEA where nearly 60% of the released HC is conture could possibly allow for a smaller DOC, or at least less PGM content PGM being used in the HCT; this is the focus of a future study of ours. It d that although Ag is not an excellent oxidation catalyst, it is expected to ion of the stored HCs, and there is evidence of this in the Ag/BEA sample Cs observed during the release compared to the uptake. Another key obrelease temperatures is that both BEA and Ag/BEA release 50% of the HC nd thus many DOCs would be able to oxidize the HCs upon their release. /ZSM-5 have similar stability, but their T50s are 190 and 185 °C, respecs not as desirable as the BEA samples. The SSZ-13 zeolites not only store they release nearly all of the HCs by 150 °C, and thus continue to show T properties. ed above an ideal passive NOx adsorber (PNA) would have high storage nd release NOx above 200 °C where a downstream SCR could be used in th ammonia from hydrolyzed urea. This hydrolysis occurs above 133 °C R reactivity typically reaches 90% by 180 °C [54,55]. The same systematic ure described above in the HC section was also used with NOx to underevaluated samples meet the PNA criteria. Figure 4 shows a characteristic nd desorption profile; Pd/ZSM-5 is shown here, which is the same sample t shown in Figure 2. Notably less NOx is adsorbed than HCs, but the admore strongly bound and releases at higher temperatures. Similar to THC, ivided into two sections, 3 ( ) and 30 ( ) minutes (Figure 4b), and the f 10%, 50% and 90% NOx release are determined (Figure 4c). The data for are shown in Figure 5 with the specific values listed in Table 2. From these rtant observations are apparent. First, the Pd-zeolites are the only samples ant uptake, and second, only Pd/ZSM-5 and Pd/SSZ-13 maintains 50% of 150 °C; in fact, over 50% of the NOx is still on Pd Pd/SSZ-13 at 224 °C. This istent with other reports of Pd/SSZ-13 being one of the best PNAs [56,57] orage only occurs on ion-exchanged Pd sites [41]. Neither catalyst closed ce during release, suggesting there is some direct reduction in the stored s. This typically yields some N2O formation, but this reactor was not an analyzer that would allow N2O to be measured accurately.  Table 1, and there are no distinct trends across the nine samples e Pd/zeolites have the lowest T10. Comparing T50s and T90s, the Pd/zeshow lower release temperatures for the BEA and ZSM-5 samples. Since of HCs are adsorbed on the SSZ-13 samples direct comparisons with BEA ifficult, but the trend of T90 being the lowest on the Pd sample is cone confounding factors in evaluating the samples as we describe here, is ellent oxidation catalyst and the released HCs are readily oxidized over the presence of excess O2 [40]. Thus, even though 90% of the HCs are he previously mentioned goal of 200 °C, the ability to convert the HCs CT is a desirable feature. The released HC quantity is listed in Table 1, ontaining sample the released HC is less than the quantity stored at 30 t apparent in the Pd/BEA where nearly 60% of the released HC is conre could possibly allow for a smaller DOC, or at least less PGM content M being used in the HCT; this is the focus of a future study of ours. It that although Ag is not an excellent oxidation catalyst, it is expected to n of the stored HCs, and there is evidence of this in the Ag/BEA sample s observed during the release compared to the uptake. Another key obelease temperatures is that both BEA [54,55]. The same systematic re described above in the HC section was also used with NOx to underaluated samples meet the PNA criteria. Figure 4 shows a characteristic desorption profile; Pd/ZSM-5 is shown here, which is the same sample shown in Figure 2. Notably less NOx is adsorbed than HCs, but the adore strongly bound and releases at higher temperatures. Similar to THC, ided into two sections, 3 ( ) and 30 ( ) minutes (Figure 4b), and the 10%, 50% and 90% NOx release are determined (Figure 4c). The data for re shown in Figure 5 with the specific values listed in Table 2. From these ant observations are apparent. First, the Pd-zeolites are the only samples nt uptake, and second, only Pd/ZSM-5 and Pd/SSZ-13 maintains 50% of 50 °C; in fact, over 50% of the NOx is still on Pd Pd/SSZ-13 at 224 °C. This tent with other reports of Pd/SSZ-13 being one of the best PNAs [56,57] rage only occurs on ion-exchanged Pd sites [41]. Neither catalyst closed during release, suggesting there is some direct reduction in the stored . This typically yields some N2O formation, but this reactor was not n analyzer that would allow N2O to be measured accurately.      At the end of the storage phase, a portion of the HCs is released, but most of them stay adsorbed until higher temperatures. The temperature of 10% release (T10) is shown in Figure 3 and listed in Table 1, and there are no distinct trends across the nine samples other than that the Pd/zeolites have the lowest T10. Comparing T50s and T90s, the Pd/zeolites continue to show lower release temperatures for the BEA and ZSM-5 samples. Since a small quantity of HCs are adsorbed on the SSZ-13 samples direct comparisons with BEA and ZSM-5 are difficult, but the trend of T90 being the lowest on the Pd sample is consistent. One of the confounding factors in evaluating the samples as we describe here, is that Pd is an excellent oxidation catalyst and the released HCs are readily oxidized over these zeolites in the presence of excess O2 [40]. Thus, even though 90% of the HCs are released below the previously mentioned goal of 200 °C, the ability to convert the HCs directly on the HCT is a desirable feature. The released HC quantity is listed in Table 1, and in each Pd containing sample the released HC is less than the quantity stored at 30 min. This is most apparent in the Pd/BEA where nearly 60% of the released HC is converted. This feature could possibly allow for a smaller DOC, or at least less PGM content to balance the PGM being used in the HCT; this is the focus of a future study of ours. It should be noted that although Ag is not an excellent oxidation catalyst, it is expected to catalyze a portion of the stored HCs, and there is evidence of this in the Ag/BEA sample with 12% less HCs observed during the release compared to the uptake. Another key observation in the release temperatures is that both BEA and Ag/BEA release 50% of the HC above 200 °C, and thus many DOCs would be able to oxidize the HCs upon their release. ZSM-5 and Ag/ZSM-5 have similar stability, but their T50s are 190 and 185 °C, respectively, and thus not as desirable as the BEA samples. The SSZ-13 zeolites not only store less HCs, but they release nearly all of the HCs by 150 °C, and thus continue to show undesirable HCT properties.

NO Storage
As discussed above an ideal passive NOx adsorber (PNA) would have high storage below 200 °C and release NOx above 200 °C where a downstream SCR could be used in conjunction with ammonia from hydrolyzed urea. This hydrolysis occurs above 133 °C [53], and the SCR reactivity typically reaches 90% by 180 °C [54,55]. The same systematic analysis procedure described above in the HC section was also used with NOx to understand how the evaluated samples meet the PNA criteria. Figure 4 shows a characteristic NOx storage and desorption profile; Pd/ZSM-5 is shown here, which is the same sample and experiment shown in Figure 2. Notably less NOx is adsorbed than HCs, but the adsorbed NOx is more strongly bound and releases at higher temperatures. Similar to THC, the uptake is divided into two sections, 3 ( ) and 30 ( ) minutes (Figure 4b), and the temperatures of 10%, 50% and 90% NOx release are determined (Figure 4c). The data for all the samples are shown in Figure 5 with the specific values listed in Table 2. From these data two important observations are apparent. First, the Pd-zeolites are the only samples to have significant uptake, and second, only Pd/ZSM-5 and Pd/SSZ-13 maintains 50% of the NOx above 150 °C; in fact, over 50% of the NOx is still on Pd Pd/SSZ-13 at 224 °C. This finding is consistent with other reports of Pd/SSZ-13 being one of the best PNAs [56,57] and that NO storage only occurs on ion-exchanged Pd sites [41]. Neither catalyst closed the NOx balance during release, suggesting there is some direct reduction in the stored NOx from HCs. This typically yields some N2O formation, but this reactor was not equipped with an analyzer that would allow N2O to be measured accurately.

Hydrothermal Aging
After fully evaluating all the samples, five of the most promising zeolites were hydrothermally aged to investigate their durability. BEA, Pd/BEA, ZSM-5, Pd/ZSM-5, and Pd/SSZ-13 were heated to 800 °C for 50 h and re-evaluated following the same protocols as before. Figure 6 shows the HC metrics for both storage and release with the values listed in Table   ) and 30 ( At the end of the storage phase, a portion of the HCs is released, but most of them stay adsorbed until higher temperatures. The temperature of 10% release (T10) is shown in Figure 3 and listed in Table 1, and there are no distinct trends across the nine samples other than that the Pd/zeolites have the lowest T10. Comparing T50s and T90s, the Pd/zeolites continue to show lower release temperatures for the BEA and ZSM-5 samples. Since a small quantity of HCs are adsorbed on the SSZ-13 samples direct comparisons with BEA and ZSM-5 are difficult, but the trend of T90 being the lowest on the Pd sample is consistent. One of the confounding factors in evaluating the samples as we describe here, is that Pd is an excellent oxidation catalyst and the released HCs are readily oxidized over these zeolites in the presence of excess O2 [40]. Thus, even though 90% of the HCs are released below the previously mentioned goal of 200 °C, the ability to convert the HCs directly on the HCT is a desirable feature. The released HC quantity is listed in Table 1, and in each Pd containing sample the released HC is less than the quantity stored at 30 min. This is most apparent in the Pd/BEA where nearly 60% of the released HC is converted. This feature could possibly allow for a smaller DOC, or at least less PGM content to balance the PGM being used in the HCT; this is the focus of a future study of ours. It should be noted that although Ag is not an excellent oxidation catalyst, it is expected to catalyze a portion of the stored HCs, and there is evidence of this in the Ag/BEA sample with 12% less HCs observed during the release compared to the uptake. Another key observation in the release temperatures is that both BEA and Ag/BEA release 50% of the HC above 200 °C, and thus many DOCs would be able to oxidize the HCs upon their release. ZSM-5 and Ag/ZSM-5 have similar stability, but their T50s are 190 and 185 °C, respectively, and thus not as desirable as the BEA samples. The SSZ-13 zeolites not only store less HCs, but they release nearly all of the HCs by 150 °C, and thus continue to show undesirable HCT properties.

NO Storage
As discussed above an ideal passive NOx adsorber (PNA) would have high storage below 200 °C and release NOx above 200 °C where a downstream SCR could be used in conjunction with ammonia from hydrolyzed urea. This hydrolysis occurs above 133 °C [53], and the SCR reactivity typically reaches 90% by 180 °C [54,55]. The same systematic analysis procedure described above in the HC section was also used with NOx to understand how the evaluated samples meet the PNA criteria. Figure 4 shows a characteristic NOx storage and desorption profile; Pd/ZSM-5 is shown here, which is the same sample and experiment shown in Figure 2. Notably less NOx is adsorbed than HCs, but the adsorbed NOx is more strongly bound and releases at higher temperatures. Similar to THC, the uptake is divided into two sections, 3 ( ) and 30 ( ) minutes (Figure 4b), and the temperatures of 10%, 50% and 90% NOx release are determined (Figure 4c). The data for all the samples are shown in Figure 5 with the specific values listed in Table 2. From these data two important observations are apparent. First, the Pd-zeolites are the only samples to have significant uptake, and second, only Pd/ZSM-5 and Pd/SSZ-13 maintains 50% of the NOx above 150 °C; in fact, over 50% of the NOx is still on Pd Pd/SSZ-13 at 224 °C. This finding is consistent with other reports of Pd/SSZ-13 being one of the best PNAs [56,57] and that NO storage only occurs on ion-exchanged Pd sites [41]. Neither catalyst closed the NOx balance during release, suggesting there is some direct reduction in the stored NOx from HCs. This typically yields some N2O formation, but this reactor was not equipped with an analyzer that would allow N2O to be measured accurately.

Hydrothermal Aging
After fully evaluating all the samples, five of the most promising zeolites were hydrothermally aged to investigate their durability. BEA, Pd/BEA, ZSM-5, Pd/ZSM-5, and Pd/SSZ-13 were heated to 800 °C for 50 h and re-evaluated following the same protocols as before. Figure 6 shows the HC metrics for both storage and release with the values listed in Table   ) minutes for each of the materials studied. (b) The temperature where 10% (T10, ), 50% (T50, ), or 90% (T90, ) of the NOx is released. Table 2. NOx storage values (mmol/g cat ) for ion-exchanged zeolites during lean trapping evaluation after 3 and 30 min. The total amount of observed NOx released during the first 15 min of the ramp, and the temperature where 10% (T10), 50% (T50), or 90% (T90) of NOx is released is also shown.

Hydrothermal Aging
After fully evaluating all the samples, five of the most promising zeolites were hydrothermally aged to investigate their durability. BEA, Pd/BEA, ZSM-5, Pd/ZSM-5, and Pd/SSZ-13 were heated to 800 • C for 50 h and re-evaluated following the same protocols as before. Figure 6 shows the HC metrics for both storage and release with the values listed in Table 3. The amount of deactivation is minimal on storage, as the change is less than 10%. The release temperatures for HCs are also similarly unchanged, as the T10s, T50s, and T90s are generally within a couple degrees of each other. The Pd/ZSM-5 sample did result in a higher release temperature, as the T90 increased from 195 to 216 • C, suggesting its oxidation capability diminished after aging. This is supported by the closer match of storage and release quantities, 4.7 and 4.5 (mmols C 1 /g cat ). However, Pd/BEA maintained a large discrepancy in storage and release values as greater than 50% of the stored HCs are converted during release. In deciding which one of these samples would offer the best HCT functionality, it is clear that each of these hydrothermally aged samples would offer significant HCT capacity. The release profile of unexchanged BEA is slightly more favorable, since~50% of the stored HCs are released above 200 • C, compared to 185 • C for ZSM-5. If Pd is going to be used for HCTs, Pd/BEA offers significant durability, and although Pd does not aid in the storage capacity, its oxidation capability is valuable with the ability to oxidize over 50% of the stored HCs even after hydrothermal aging.      The NOx storage behavior was notably different after hydrothermally aging. Figure 7 shows the NOx storage and release metrics for the five aged samples with the values shown in Table 4. From these results it is clear only the Pd/SSZ-13 maintains significant NOx storage capacity; in fact, both the storage capacity and release characteristics improve after aging. The storage increased from 0.028 to 0.036 mmols NOx/g cat , and the T50 increased from 224 to 262 • C. Pd/BEA is the only other sample to show storage, but all of the NOx is released before 150 • C. All these results continue to reinforce that Pd/SSZ-13 is the only viable option for PNA amongst the samples studied here.
(b) Figure 6. Following hydrothermal aging at 800 °C for 50 h (a) THC adsorption after 3 ( ) and 30 ( ) minutes for each of the materials studied, and (b) the temperature where 10% (T10, ■), 50% (T50, ■), or 90% (T90, ■) of THC is released. The NOx storage behavior was notably different after hydrothermally aging. Figure  7 shows the NOx storage and release metrics for the five aged samples with the values shown in Table 4. From these results it is clear only the Pd/SSZ-13 maintains significant NOx storage capacity; in fact, both the storage capacity and release characteristics improve after aging. The storage increased from 0.028 to 0.036 mmols NOx/gcat, and the T50 increased from 224 to 262 °C. Pd/BEA is the only other sample to show storage, but all of the NOx is released before 150 °C. All these results continue to reinforce that Pd/SSZ-13 is the only viable option for PNA amongst the samples studied here.   At the end of the storage phase, a portion of the HCs is released, but most of them stay adsorbed until higher temperatures. The temperature of 10% release (T10) is shown in Figure 3 and listed in Table 1, and there are no distinct trends across the nine sample other than that the Pd/zeolites have the lowest T10. Comparing T50s and T90s, the Pd/ze olites continue to show lower release temperatures for the BEA and ZSM-5 samples. Since a small quantity of HCs are adsorbed on the SSZ-13 samples direct comparisons with BEA and ZSM-5 are difficult, but the trend of T90 being the lowest on the Pd sample is con sistent. One of the confounding factors in evaluating the samples as we describe here, i that Pd is an excellent oxidation catalyst and the released HCs are readily oxidized ove these zeolites in the presence of excess O2 [40]. Thus, even though 90% of the HCs are released below the previously mentioned goal of 200 °C, the ability to convert the HCs directly on the HCT is a desirable feature. The released HC quantity is listed in Table 1 and in each Pd containing sample the released HC is less than the quantity stored at 30 min. This is most apparent in the Pd/BEA where nearly 60% of the released HC is con verted. This feature could possibly allow for a smaller DOC, or at least less PGM conten to balance the PGM being used in the HCT; this is the focus of a future study of ours. I should be noted that although Ag is not an excellent oxidation catalyst, it is expected to catalyze a portion of the stored HCs, and there is evidence of this in the Ag/BEA sample with 12% less HCs observed during the release compared to the uptake. Another key ob servation in the release temperatures is that both BEA and Ag/BEA release 50% of the HC above 200 °C, and thus many DOCs would be able to oxidize the HCs upon their release ZSM-5 and Ag/ZSM-5 have similar stability, but their T50s are 190 and 185 °C, respec tively, and thus not as desirable as the BEA samples. The SSZ-13 zeolites not only store less HCs, but they release nearly all of the HCs by 150 °C, and thus continue to show undesirable HCT properties.

NO Storage
As discussed above an ideal passive NOx adsorber (PNA) would have high storage below 200 °C and release NOx above 200 °C where a downstream SCR could be used in conjunction with ammonia from hydrolyzed urea. This hydrolysis occurs above 133 °C [53], and the SCR reactivity typically reaches 90% by 180 °C [54,55]. The same systemati analysis procedure described above in the HC section was also used with NOx to under stand how the evaluated samples meet the PNA criteria. Figure 4 shows a characteristi NOx storage and desorption profile; Pd/ZSM-5 is shown here, which is the same sample and experiment shown in Figure 2. Notably less NOx is adsorbed than HCs, but the ad sorbed NOx is more strongly bound and releases at higher temperatures. Similar to THC the uptake is divided into two sections, 3 ( ) and 30 ( ) minutes (Figure 4b), and the temperatures of 10%, 50% and 90% NOx release are determined (Figure 4c). The data fo all the samples are shown in Figure 5 with the specific values listed in Table 2. From these data two important observations are apparent. First, the Pd-zeolites are the only sample to have significant uptake, and second, only Pd/ZSM- 5  At the end of the storage phase, a portion of the HCs is released, but most of them stay adsorbed until higher temperatures. The temperature of 10% release (T10) is shown in Figure 3 and listed in Table 1, and there are no distinct trends across the nine samples other than that the Pd/zeolites have the lowest T10. Comparing T50s and T90s, the Pd/zeolites continue to show lower release temperatures for the BEA and ZSM-5 samples. Since a small quantity of HCs are adsorbed on the SSZ-13 samples direct comparisons with BEA and ZSM-5 are difficult, but the trend of T90 being the lowest on the Pd sample is consistent. One of the confounding factors in evaluating the samples as we describe here, is that Pd is an excellent oxidation catalyst and the released HCs are readily oxidized over these zeolites in the presence of excess O2 [40]. Thus, even though 90% of the HCs are released below the previously mentioned goal of 200 °C, the ability to convert the HCs directly on the HCT is a desirable feature. The released HC quantity is listed in Table 1, and in each Pd containing sample the released HC is less than the quantity stored at 30 min. This is most apparent in the Pd/BEA where nearly 60% of the released HC is converted. This feature could possibly allow for a smaller DOC, or at least less PGM content to balance the PGM being used in the HCT; this is the focus of a future study of ours. It should be noted that although Ag is not an excellent oxidation catalyst, it is expected to catalyze a portion of the stored HCs, and there is evidence of this in the Ag/BEA sample with 12% less HCs observed during the release compared to the uptake. Another key observation in the release temperatures is that both BEA and Ag/BEA release 50% of the HC above 200 °C, and thus many DOCs would be able to oxidize the HCs upon their release. ZSM-5 and Ag/ZSM-5 have similar stability, but their T50s are 190 and 185 °C, respectively, and thus not as desirable as the BEA samples. The SSZ-13 zeolites not only store less HCs, but they release nearly all of the HCs by 150 °C, and thus continue to show undesirable HCT properties.

NO Storage
As discussed above an ideal passive NOx adsorber (PNA) would have high storage below 200 °C and release NOx above 200 °C where a downstream SCR could be used in conjunction with ammonia from hydrolyzed urea. This hydrolysis occurs above 133 °C [53], and the SCR reactivity typically reaches 90% by 180 °C [54,55]. The same systematic analysis procedure described above in the HC section was also used with NOx to understand how the evaluated samples meet the PNA criteria. Figure 4 shows a characteristic NOx storage and desorption profile; Pd/ZSM-5 is shown here, which is the same sample and experiment shown in Figure 2. Notably less NOx is adsorbed than HCs, but the adsorbed NOx is more strongly bound and releases at higher temperatures. Similar to THC, the uptake is divided into two sections, 3 ( ) and 30 ( ) minutes (Figure 4b), and the temperatures of 10%, 50% and 90% NOx release are determined (Figure 4c). The data for all the samples are shown in Figure 5 with the specific values listed in Table 2. From these data two important observations are apparent. First, the Pd-zeolites are the only samples to have significant uptake, and second, only Pd/ZSM-5 and Pd/SSZ-13 maintains 50% of the NOx above 150 °C; in fact, over 50% of the NOx is still on Pd Pd/SSZ-13 at 224 °C. This Table 4. NOx storage values (mmol/g cat ) for hydrothermally aged samples. The total amount of observed NOx released during the first 15 min of the ramp, and the temperature where 10% (T10), 50% (T50), or 90% (T90) of NOx is released is also shown. sample and then averaged. Following the degreened evaluations of the samples, the most promising materials were hydrothermally aged at 800 °C for 50 h in [O2] = 12%, [CO2] = 6%, and [H2O] = 6% with Ar balance. These samples were then re-evaluated according to the procedure outlined above. According to the low-temperature storage protocol [38,39], this time and temperature is intended to stress the materials to a similar level that would be expected at the end of their full-useful-life.  Following the degreened evaluations of the samples, the most promising materials were hydrothermally aged at 800 • C for 50 h in [O 2 ] = 12%, [CO 2 ] = 6%, and [H 2 O] = 6% with Ar balance. These samples were then re-evaluated according to the procedure outlined above. According to the low-temperature storage protocol [38,39], this time and temperature is intended to stress the materials to a similar level that would be expected at the end of their full-useful-life.

Conclusions
Overall, these samples show significant capability to perform both HCT and PNA functionalities. The presence of Ag or Pd does not universally result in increased HC storage, but Pd and its excellent oxidation reactivity aids in the conversion of the released HCs directly on the HCT. The pores of SSZ-13 are too small to store significant quantities of HCs, and thus BEA and ZSM-5 are preferred for HCTs. Pd is crucial for the adsorption of NOx, and under these gas mixtures, Pd/SSZ-13 provides the best storage and release characteristics. Hydrothermal aging largely supports these results with minimal storage losses observed for BEA, Pd/BEA, ZSM-5, Pd/ZSM-5, and Pd/SSZ-13. Only Pd/BEA maintained its HC oxidation reactivity after aging. Pd/SSZ-13 is clearly the only sample evaluated to illustrate reasonable PNA behavior after aging, and in fact its storage capacity increased while its release temperature showed improved characteristics with greater than 50% of the stored NOx being released above 250 • C.