Identifying the Factors Governing the Early-Stage Degradation of Cu-Chabazite Zeolite for NH3-SCR.

To understand the degradation mechanism of the copper-ion-exchanged SSZ-13 (Cu-SSZ-13) is of high significance for rationally designing a zeolitic catalyst for ammonia-selective catalytic reduction of NO x (NH3-SCR). In this work, we focused on an Al-rich Cu-SSZ-13 and studied its structural degradation under hydrothermal conditions through a set of characterization techniques, including in situ X-ray diffraction (XRD), pair distribution function analysis and transmission electron microscopy-energy dispersive X-ray analysis (TEM-EDX). The results indicated that the chabazite structure tends to contract in the c direction upon hydrothermal treatment and consequently leads to the collapse of the four-membered ring. Such a structure change then results in the movement of isolated Cu2+ species from the face of the double six-membered ring to its center, which damages the structure further. However, the larger rings (6MRs and 8MRs) partially remain during the structure degradation, which possibly explains that some of the isolated Cu2+ species are alive even when the XRD-detectable crystallinity completely loses. The particle-by-particle observations through TEM-EDX analysis suggested that the occurrence of structural degradation differs remarkably from one individual particle to another. In general, particles with smaller size, having a lower Si/Al ratio and a higher Cu/Al ratio, tend to degrade easily. These results offer a thorough understanding of the structural degradation of Cu-SSZ-13 from the microscopic point of view and point out that the uniformity in composition and particle size of the zeolites plays a critical role in the early-stage degradation.

Introduction

Zeolites are crystalline porous materials that are composed of tetrahedral aluminosilicates.[1,2] The presence of tetrahedral AlO4 units brings about negative charges, which offer ion-exchange sites to accommodate cations. When the cation is a proton, Brønsted acidity is generated, which is an important aspect that determines the catalytic performance of zeolites in certain reactions.[3−5] Various metallic cations can also be ion-exchanged into zeolites, creating potential active centers that broaden the catalytic applications of zeolites.[6,7] One prominent example of commercial applications of zeolites is ammonia-selective catalytic reduction (NH3-SCR) of NO.[8−10] Efforts in applying metal-ion-exchanged zeolites on the SCR of NO date back to 1980s when Iwamoto et al. discovered the catalytic activities of copper-ion-exchanged FAU and MFI zeolites for the decomposition of NO.[11,12] Due to structural degradation of medium- and large-pore zeolites (e.g., MFI and FAU) under severe hydrothermal aging conditions, however, such copper-ion-exchanged zeolites had not been utilized in real engines.[13] In recent years, copper-ion-exchange SSZ-13, a small-pore zeolite with chabazite structure, has been found to possess excellent hydrothermal stability (stable against aging temperatures above 700 °C).[14−18] In addition, compared to conventional catalysts like TiO2–V2O5–WO3, Cu-SSZ-13 possesses a wider temperature range for NO conversion. Therefore, Cu-SSZ-13 has already been commercialized as an NH3-SCR catalyst for NO abatement in diesel engines.[19]

Hydrothermal stability is an important aspect to the application of zeolites as it dictates their catalytic performances. Typically, amorphization of zeolites occurs in hydrothermal aging conditions, causing structural degradation.[20,21] In the previous work by Maugé et al., it was supposed that H2O tends to attack the aluminum sites, which results in dealumination and consequently structural amorphization.[22] Therefore, increasing the silica-to-aluminum ratio (Si/Al ratio) of a zeolite is one means of improving its hydrothermal stability because the dealumination could be minimized. However, the high-silica zeolites offer fewer framework aluminums that can create ion-exchange sites to accommodate the cations for catalytic reactions. Specific to Cu-SSZ-13 for NH3-SCR application, an SSZ-13 zeolite with a too high Si/Al ratio is not favored because its performance in the low temperature range is not ideal.[23] Rather, Al-rich Cu-SSZ-13 zeolites are on the trend because they enable a high copper loading together with abundant Brønsted acidity that is beneficial to the SCR activity over a broad temperature.[24−26] Nevertheless, a thorough understanding of the degradation mechanism of Cu-SSZ-13 from a microscopic point of view is of high significance yet still lacking. In this work, we attempt to clarify the degradation mechanism of the Al-rich Cu-SSZ-13 catalyst from a microscopic point of view through various characterization techniques such as X-ray diffraction (XRD)-Rietveld refinement, pair distribution function (PDF) analysis, and transmission electron microscopy (TEM). These results, identifying the factors that are responsible for structural degradation, demonstrate that nonuniformity is the major reason accounting for the early-stage degradation of the Cu-SSZ-13 catalyst.

Results and Discussion

The Cu-SSZ-13 sample was prepared according to the recipe described elsewhere, and the procedure can also be found in the Supporting Information. The Si/Al ratio of Cu-SSZ-13 was 6.5, and the copper loading was ca. 3.6 wt %. This Cu-SSZ-13 was proved to rival the mainstream zeolitic NH3-SCR catalyst on the market as it possesses high hydrothermal stability and good NH3-SCR performance;[26] yet, it is still susceptible to severe hydrothermal treatments. We first investigated, from the microscopic point of view, how the amorphization of the Cu-SSZ-13 catalyst evolves under hydrothermal aging conditions. Figure shows the changes in relative crystallinity of Cu-SSZ-13 with respect to hydrothermal temperatures. It can be seen that the crystallinity changed very slightly (decreased by less than 6%) when the temperature was lower than 750 °C, indicating that the sample was quite stable against moderate aging treatments. Although a sharp drop in the crystallinity was observed when the hydrothermal temperature was 800 °C and above, only a crystallinity of 1.8% was observed after the hydrothermal treatment at 830 °C. This observation is consistent with the tendency revealed in previous studies on Cu-SSZ-13.[26] We further studied the dependence of the amount of isolated Cu2+ species on the crystallinity. As seen from Figure S1, the decrease of the isolated Cu2+ species was initially in line with the crystallinity (for example, ca. 75% amount of Cu2+ species remained when the crystallinity decreased to 73%). However, the amount of Cu2+ species remained stable at ca. 55–58% when the crystallinity further decreased (to even 0%). This result indicates that the presence of isolated Cu2+ species does not necessarily rely on the preservation of crystallinity. However, crystallinity is of high importance to the catalytic performance, as microporosity decreased with crystallinity, which resulted in a considerable loss in catalytic activity. In addition, the 27Al MAS NMR spectra in Figure S2 show the coordination state of the aluminum species. The peak centering at ca. 50 ppm, representing the tetrahedral aluminum, can be observed on the samples hydrothermally treated at a temperature below 750 °C. A higher hydrothermal temperature could result in the disappearance of the peak for tetrahedral aluminum, while no other peaks were observed, probably because NMR invisible aluminum species were generated along with the progressing of dealumination.

Figure 1

Cu/SSZ-13 degradation curve by XRD. The hydrothermal stability test was conducted at several temperatures for 5 h in flowing air containing 10% H2O.

To probe the structural evolution taking place under the hydrothermal aging condition, in situ XRD measurement was carried out. A homemade setup was designed (Figure S3), which allowed us to accurately collect the XRD data under controlled hydrothermal conditions (see temperature program in Figure S4). Figure shows degradation curves reflected by relative intensities for three representative peaks at 9.55° (hkl = 101), 20.90° (hkl = 211), and 13.09° (hkl = 110), which are denoted hereafter as the 1st peak, 2nd peak, and 3rd peak, respectively. A full range of the time-resolved XRD patterns can be found in Figure S5. Almost no degradation was observed from the intensity of these peaks after 1 h of hydrothermal aging. The intensity of the 2nd peak decreased by 9% after 2 h and further decreased to 27% after 8 h of hydrothermal aging, whereas the intensities of both the 1st and 3rd peaks remained at ca. 97 and 40% after aging periods of 2 and 8 h, respectively. This result indicates that the degradation did not proceed isotropically but exhibited different rates in different directions. To further interpret the in situ XRD data, Rietveld refinement of the XRD patterns collected at 0, 3.3, 6.6, and 14.4 h of hydrothermal aging was performed. Figure S6 shows the changes in the unit cell refined after different time intervals, which suggests that the chabazite structure tended to contract in the c direction while remain almost unaffected in the a and b directions. The refinement results further revealed that the structure degradation is an anisotropic process, which, not beyond expectation, is consistent with the peak intensity results in Figure .

Figure 2

Relative crystallinity retention of Cu-SSZ-13 based on different peak intensities measured by in situ XRD.

In addition, the maximum entropy method (MEM) analysis of the time-resolved XRD patterns (also collected at 0, 3.3, 6.6, and 14.4 of hydrothermal aging) was also performed.[27,28] The electron density distribution maps in Figure indicate that the copper cations migrated alongside the degradation of the chabazite structure. After the aging of 3.3 h, the crystallinity was about 90%, where the location of the copper cations did not change too much, and only a small deviation of the unit cell parameter in the c direction was observed. After the aging of 6.6 h, only a crystallinity of 50% remained. In such a case, the location of the copper cations was observed to move considerably, from the face of the D6R to its center. While after the aging of 14.4 h, only a crystallinity of 30% remained, and the copper cations were found to locate in the center of the double six-membered ring (D6R).

Figure 3

Rietveld refinement of the time-resolved XRD patterns collected at different periods and the corresponding electron density distribution maps determined by the maximum entropy method. (An illustration of the CHA structure is shown on top of this figure. Color of the elements: Cu, yellow; Si, blue; O, red. The green color denotes the electron density information calculated from MEM).

In a pioneer work by Fickel and Lobo,[29] the location of the Cu cations as well as the thermal stability of a Cu-SSZ-13 sample was studied by variable-temperature XRD. According to their results, the chabazite structure tends to contract in the a direction in the temperature range of 200–800 °C (from ca. 13.53 Å to ca. 13.43 Å), whereas in the c direction, the unit cell undergoes a slight expansion from ca. 15.00 Å at 200 °C to ca. 15.02 Å at 800 °C, with a peak reaching ca. 15.05 Å at 350 °C. While in our case, the refinement of the in situ XRD results (Figure S6) demonstrated a different scenario that the unit cell of the Cu-SSZ-13 exhibits a considerable shrinkage in the c direction (from ca. 14.93 Å to ca. 14.84 Å) while maintaining its dimensions in the a and b directions. It is worth noting that the two Cu-SSZ-13 samples bear little difference in terms of composition. Nevertheless, the conditions for the in situ XRD measurement remarkably differed. In the work by Fickel and Lobo, the temperature-variable XRD measurements were carried out in an atmosphere of 5% O2/He mixture, without the addition of any water. In contrast, we performed the in situ XRD measurement in humid air containing 10 vol % H2O, a common atmosphere used for hydrothermal stability tests, which is considered to be responsible for the different structure degradation observed. Considering the above results, it could be inferred that the chabazite structure tends to contract in the c direction under hydrothermal conditions, which drives the copper cations to move from the face of D6R to its center and causes the structure degradation further.

X-ray total scattering together with PDF analysis was performed to understand the structure of the amorphization taking place in the hydrothermal aging. Figure S7 compares the pair distribution function, G(r) (obtained by a Fourier transformation of the total structure factors, S(Q)), for the fully crystalline and completely degraded samples. From the PDF, various atomic distances concerning framework connectivity can be identified, which allowed us to analyze the structural difference between these two types of materials.[30−33] From this comparison, it can be seen that the peak centering at 3.8 Å (representing the correlation in the four-membered ring (4MR)) for the degraded sample became less pronounced with respect to that for the fully crystalline sample, probably because the 4MR in the degraded sample had collapsed. In contrast, the peaks centering at 4.1 Å, which represent the correlation in the six-membered ring (6MR) and eight-membered ring (8MR), maintained the same intensity after amorphization, indicating the preservation of the larger ring structures. These results suggest that the degradation of the chabazite structure could probably start from the collapse of the 4MRs.

On the basis of the characterizations at the atomic level, a degradation mechanism was proposed as shown in Figure . The contraction of the structure starts from the c direction, which drives the copper cations to move from the face of the D6R to the center. The copper cations remained to some extent to be the isolated Cu2+ species, which minimizes the direct contact between two coppers and thus reduces the possibility of forming CuO clusters (or even CuO phase), which was supposed to be a major reason for structural degradation of Cu-SSZ-13. With the progress of the hydrothermal aging, the larger rings (6MRs and 8MRs) maintained their connection, although considerable amorphization occurs until the crystallinity reaches 0%. Because of the preservation of the larger rings, some of the isolated copper species could be protected even though the sample becomes completely degraded.

Figure 4

Proposed degradation scheme for Cu-SSZ-13.

The above mechanism, proposed on the basis of the XRD and PDF results, provides an explanation at the microscopic level. To identify the factors governing the degradation at the early stage from other microscopic points of view, TEM observation together with EDX analysis was carried out to study the size, composition, and crystallinity of each particle. Statistical analysis was then conducted on the basis of the data collected on individual particles. The TEM observation as well as EDX analysis was performed using powder samples. As the particle size of the Cu-SSZ-13 was in the range of several hundred nanometers (up to 500 nm), the measurement was feasible. The accuracy of the statistical analysis could be reflected by the error bars, which are associated with the averaged values. To describe the diameter of the particles, diagonal diameter was used, as the Cu-SSZ-13 samples had a cubic morphology. Selected area electron diffraction (SAED) analysis could provide the level of crystallinity of the particle observed, as a halo was observed in the case of amorphous particles but spot patterns were observed for the crystalline ones. Representative TEM images and the corresponding SAED analysis results of a set of samples are provided in Figure S8.

Table shows the statistic data for three representative samples with different levels of crystallinity (100, 86, and 0%). In the fully crystalline sample (100% in crystallinity), there was no particle on which an amorphous halo was observed. In the sample with a crystallinity of 86%, 29% of the particles were considered to be amorphous from the TEM–SAED analysis as a halo was observed, and 71% of the particles were confirmed to be crystalline. In the case of the fully degraded sample (0% in crystallinity), all particles were considered to be amorphous from the TEM–SAED analysis. From these results, we can see that the tendency observed from the TEM–SAED analysis is consistent with the results of powdered XRD. Interesting conclusion, however, could be reached if we give a closer look. From the TEM–SAED analysis, no transitional state, which means a mixture of crystalline and amorphous parts, was observed in any single particle (Figure ). This result suggests that the amorphization of the Cu-SSZ-13 could occur suddenly when a critical condition was met and that the amorphization did not spread uniformly but happened first to those that were vulnerable. To better understand such a mechanism, the nature of the particle (crystalline vs amorphous) was plotted in two parameter settings (particle size vs Cu/Al ratio and particle size vs Si/Al ratio) in Figure . In the synthesis of zeolites, a uniform particle size distribution is generally difficult to achieve. The final average particle size and particle size distribution of the zeolite product are dependent upon the competition between the timing of nucleation and crystal growth. However, the timing of nucleation of zeolite crystallization is very complicated and challenging to control.[34,35] In the case of nucleation taking place at an early stage, the crystal growth of each nucleus has abundant time to consume the nutrients, which may result in large particle size, whereas in the case of nucleation taking place in a later time, small particle size may be obtained. In addition, the Cu/Al ratio as well as the Si/Al ratio within a particle also matters. The Si/Al ratio is largely determined by the synthesis parameters, such as type of the structure-directing agent, raw materials, synthesis temperature, etc., and in most cases, the distribution of aluminum atoms and thereof the local Si/Al ratio are difficult to control.[36] On the contrary, the Cu/Al ratio can be adjusted through postsynthesis treatment, while the local Cu/Al ratio could vary depending on the nature of the parent zeolite as well as the parameters of the postsynthesis treatment.[37] Such aspects suggest that the Si/Al ratio as well as the Cu/Al ratio could also be nonuniform among the individual particles. From Figure , it can be clearly seen that amorphization first happened to the particles with smaller particle size.

Table 1

Number of Particles Observed (No Intermediate State, Only Dead or Alive) with TEM Photo Example (Several Positions Are Observed)

condition		TEM diffraction pattern
temp. /°C at hydrothermal durability	crystallinity by XRD (%)a	CHA pattern (%)	amorphous halo (%)
fresh	100	100 (N = 36/36)	0 (N = 0)
800	86	71 (N = 20/28)	29 (N = 8/28)
850	0	0 (N = 0/1)	100 (N = 1/1)

The crystallinity was calculated taking the fresh sample as a reference.

Figure 5

Representative scanning electron microscopy and TEM images and the corresponding selected area electron diffraction patterns of the individual particles for the fresh Cu-SSZ-13 sample and the samples subjected to different hydrothermal treatment conditions.

Figure 6

Scatter plot of the nature of the Cu-SSZ-13 particles (crystalline vs amorphous). (a) Cu/Al ratio vs particle size. (b) Si/Al ratio vs particle size. Note that the Cu/Al and Si/Al ratios were measured by TEM–EDX. Five spots from one particle were selected to obtain the compositional information, and therefore the Cu/Al and Si/Al ratios were presented with an average bar.

The nature of the particle was further plotted in terms of the Cu/Al ratio vs the Si/Al ratio (Figure ). The result indicates that amorphization tended to occur on the particles that had a lower Si/Al ratio and a higher Cu/Al ratio. It is generally considered that a sample with a higher Si/Al ratio should have higher hydrothermal stability, which was again confirmed by our result. In addition, the hydrothermal stability of the Cu-SSZ-13 catalyst is also supposed to depend on the Cu/Al ratio.[38] While our result in Figure , obtained from the statistical analysis based on a particle-by-particle measurement, confirms that a high Cu/Al ratio is not favored for the hydrothermal stability. Since the particle size was relatively small (which ranged from 250 to 500 nm), a gradual amorphization was not observed. It was speculated that the particle started to contract in the c direction under the hydrothermal aging condition, which suddenly drove the crystalline structure to collapse. Therefore, the uniformity of the sample plays an important role as a sample that is homogeneous in particle size, and Si/Al and Cu/Al ratios could be beneficial for the parameter screening toward the enhancement of the overall hydrothermal stability.

Figure 7

Scatter plot of the nature of the Cu-SSZ-13 particles (crystalline vs amorphous) in function of Cu/Al vs Si/Al ratio. Note that the average bars were given for both the Cu/Al and Si/Al ratios.

Conclusions

In summary, we have focused on an Al-rich Cu-SSZ-13 zeolite and clarified the factors governing its structural degradation under hydrothermal treatment. A collective effort by in situ XRD (together with Rietveld refinement) and PDF analysis revealed that the hydrothermal treatment tends to cause the zeolite structure to contract in the c direction and then leads to the collapse of the 4MRs. Such a structural change drives the isolated Cu2+ species to move from the face of D6R to its center, resulting in further structural degradation. The larger rings (6MRs and 8MRs) partially remain during the structural degradation, which could possibly explain the facts that no octahedral aluminum can be observed from 27Al MAS NMR and that some of the isolated Cu2+ species remain even when the XRD-detectable crystallinity completely loses. The results from the particle-by-particle observations through TEM–EDX analysis suggested that the occurrence of structural degradation was not evenly distributed among the particles. Rather, the particles with a smaller size, having a lower Si/Al ratio and a higher Cu/Al ratio, tended to degrade easily. These results demonstrate a study of the structural degradation of Cu-SSZ-13 from a microscopic point of view and provide insights into the design of NH3-SCR catalysts. The findings point out that controlling the chemical composition as well as narrowing the particle size distribution, by either bottom-up approach[39] or top-down approach,[40] is an important aspect in the synthesis of zeolites, as both the composition and morphology play critical roles in determining the hydrothermal stability of the zeolites and thereof the performance in practical applications.