Skip to main content
Chemical Science logoLink to Chemical Science
. 2026 Sep 21. Online ahead of print. doi: 10.1039/d6sc04592h

Transcrystallization of X zeolite into SSZ-39 for enhanced NH3-SCR catalysis through tailored Brønsted acid sites and Cu species

Jiachen Wang a,†, Hui Wang a,†, Xiao Jiang b,†, Gang Liu a, Hongwei Su a, Ke Zhang b, Michael J Forte b, Chunyan Liu a, Shaoyun Chen a, Jikai Zhang a, Ronghe Lin d,✉, Limin Ren c,e,✉, Jiaxu Liu a,✉
PMCID: PMC13629304  PMID: 42825086

Abstract

Traditional synthesis of SSZ-39 zeolites via Y zeolite precursors faces challenges of high cost, complex dealumination, and limited scalability. Here, we present a pioneering interzeolite conversion (IZC) strategy using low-cost X and A zeolites as precursors, bypassing the need for pretreatments. A new crystallization mechanism centered on sodalite cages is proposed, supported by theoretical calculations. The synthesized SSZ-39 zeolites exhibit excellent crystallinity and well-defined pore structures. The strong Brønsted acid sites and [Cu(OH)]+ species in Cu-X-SSZ-39 both contribute to its outstanding NOx conversion performance, maintaining a remarkable 97% conversion even at the high temperature of 600 °C. In situ dual-beam Fourier transform infrared (DB-FTIR) studies confirm the rapid depletion of Cu–NO3− on Cu-X-SSZ-39, highlighting the superior redox dynamics. This work not only advances the mechanistic understanding of zeolite transformations but also establishes a scalable, eco-friendly route for designing high-performance NH3-SCR catalysts with industrial potential.


Low-cost X zeolite is rapidly transcrystallized into SSZ-39 through a SOD-cage-mediated pathway, producing Cu-X-SSZ-39 with optimized Brønsted acid sites and reactive Cu species for wide-temperature-window NH3-SCR catalysis.graphic file with name d6sc04592h-ga.webp

Introduction

SSZ-39 is an aluminosilicate zeolite with AEI topology, characterized by its three-dimensional eight member ring (8MR) pore system (3.8 Å × 3.8 Å) and AEI cage (8.5 Å × 9.9 Å).1,2 Due to its extensive micropore volume, high specific surface area, and remarkable structural robustness, SSZ-39 has emerged as a high-performance catalyst for a diverse range of industrial transformations. These include the methanol-to-olefins (MTO) process,3,4 ammonia-selective catalytic reduction (NH3-SCR) of nitrogen oxides,5–7 and the partial oxidation of methane.8 In the context of NH3-SCR catalysis, Cu-SSZ-39 is characterized by the coexistence of two distinct populations of active copper sites: Z2-Cu2+ species located within six-membered rings (6MRs) and Z-[Cu(OH)]+ species situated in 8MRs. The Z2-Cu2+ species, anchored by paired framework Al sites, serve as the fundamental determinant of the catalyst's exceptional hydrothermal stability. Conversely, Z-[Cu(OH)]+ species, associated with isolated Al sites, are recognized for their superior intrinsic activity at low temperatures.9–11 While the exceptional hydrothermal durability conferred by paired Al sites is a recognized paradigm in the field, it frequently imposes an intrinsic activity-stability dichotomy.5 Foundational investigations into Cu-CHA systems have established that low-temperature NH3-SCR is governed by a dynamic multinuclear mechanism, wherein NH3 solvation mobilizes framework-coordinated Cu ions, facilitating the formation of transient dimeric intermediates necessary for the rate-limiting O2 activation. Within this framework, the accessibility and lability of Z-[Cu(OH)]+ species, associated with isolated Al sites, are of paramount importance. These species function as highly reactive precursors that readily transform into mobile Cu-ammine complexes under specific reaction conditions. In contrast, Z2-Cu2+ species, anchored at paired Al sites, primarily serve to reinforce framework-stabilized Cu speciation, thereby safeguarding hydrothermal integrity.12–14 Consequently, the rational design of a superior SSZ-39 catalyst necessitates a meticulous orchestration of Al-siting to achieve a synergistic balance between rapid redox kinetics and long-term structural robustness.5

Interzeolite conversion (IZC) represents a robust methodology for zeolite synthesis, synergistically integrating structural inheritance with precise compositional regulation.2 On a structural level, locally ordered fragments derived from the parent framework function as pre-organized motifs that effectively lower the kinetic barriers to nucleation and facilitate the assembly of the target lattice. This transformative efficiency is particularly pronounced when the parent and daughter zeolites share common secondary building units (SBUs).15,16 A quintessential example is the FAU-to-AEI transition employed for SSZ-39 synthesis,17–19 which benefits from shared double six-ring (D6R) subunits and a thermodynamically favorable evolution from a low-density to a high-density framework.20–22 From a compositional standpoint, the parent zeolite imparts a “structural memory” of its intrinsic Si/Al ratio, providing a versatile handle to modulate the framework composition and Al-siting within the crystallized product. Despite these advantages, conventional IZC strategies for SSZ-39 have predominantly relied on high-silica Y-type precursors, typically USY.22–25 However, this reliance entails laborious dealumination and auxiliary silica supplementation, which significantly convolutes the synthetic protocol and undermines the cost-effectiveness inherent to the IZC approach. While D6R units are traditionally regarded as indispensable for the FAU-to-AEI transition,26,27 their absolute necessity remains a subject of active debate. Notably, SSZ-39 has been successfully derived from D6R-free precursors, such as Beta and ZSM-5, highlighting the flexibility of the IZC transformation route.16,28 Although Beta and ZSM-5 zeolites circumvent the high-temperature dealumination limitations of USY via IZC routes, their high cost remains a bottleneck for industrial scale-up. Thus, a more cost-effective IZC route is highly desirable for synthesizing highly crystalline SSZ-39 with robust thermal stability.

Recognizing the decisive role of SBUs in steering the IZC process, we propose the implementation of X zeolite as a strategic precursor. X zeolite shares the FAU topology with Y-type analogues, but its inherently low silica-to-alumina ratio (SAR ≈ 2) eliminates the requirement for intensive high-temperature dealumination, providing an atom-economical and environmentally benign alternative for SSZ-39 production. Nevertheless, this transition introduces distinct challenges; unlike USY-based routes where the parent framework acts as the dominant silicon source, X zeolite contributes only a portion of the framework silica, demanding precise modulation of synthetic parameters to synchronize dissolution kinetics with crystallization efficiency. Simultaneously, A zeolite (LTA topology) presents a conceptually attractive, highly aluminous platform for SSZ-39 synthesis. Although LTA and FAU frameworks differ in their sodalite (SOD) cage connectivity, linked via double four-rings (D4Rs) in the former and D6Rs in the latter, both share the underlying single 4-ring (S4R) and single 6-ring (S6R) motifs required for AEI framework assembly. The LTA-to-AEI transition is energetically favored by the evolution toward increased framework density (14.2 to 15.1 T/1000 Å3). However, the absence of intrinsic D6R units in LTA, coupled with its rapid dissolution and potential for excessive Al release, renders this route synthetically nontrivial and prone to competitive phase crystallization. The successful synthesis of SSZ-39 from these low-silica precursors would not only broaden the available feedstock for low-cost preparation but also challenge the conventional D6R-centric mechanism. Such findings would substantiate an alternative crystallization pathway centered on the reorganization of SOD-cage-derived aluminosilicate species, rather than the direct inheritance of D6R motifs. Furthermore, the use of X and A zeolites offers a unique opportunity to manipulate framework Al-siting (paired vs. isolated sites), a critical factor for optimizing Cu speciation and acid site density to enhance NH3-SCR performance. Resolving these synthetic and mechanistic complexities is essential to harness the full potential of SBU-mediated zeolite transformations while advancing fundamental insights into interzeolite transformation mechanisms.

By optimizing the synthesis conditions, we successfully prepared SSZ-39 via an X-zeolite-assisted route that exhibits significantly faster crystallization kinetics than traditional Y-derived methods. Density functional theory (DFT) calculations, supported by the successful conversion of D6R-free A-zeolite, reveal that SOD cage units, rather than conventional D6R motifs, govern this FAU-to-AEI transformation. The resulting Cu-X-SSZ-39 catalyst demonstrates superior NH3-SCR performance across the entire temperature window (250–600 °C). In situ dual-beam Fourier transform infrared (DB-FTIR) spectroscopy reveals that this enhancement stems from a more favorable reaction mechanism, characterized by rapid NH3–Cu2+ dynamic coordination and the efficient turnover of reactive bridged nitrate intermediates. Furthermore, the enhanced coordination stability of Z2-Cu2+ species in Cu-X-SSZ-39 ensures a robust balance between low-temperature activity and hydrothermal durability, providing a scalable and mechanistic framework for the design of high-performance zeolite catalysts.

Results and discussion

Preparation of X-SSZ-39 zeolite

Under the proposed synthetic protocol, X zeolite (Fig. S1) serves as the aluminum source and a portion of the silicon source, with the rest of the silicon derived from colloidal silica. The synthesis of SSZ-39 zeolite was successfully achieved over a wide range of starting silica-to-alumina ratios (SARs, 25–90, Table S1). By fixing the ratio of the organic structure directing agent (OSDA) to SiO2 at 0.29, we found that the key factor that governs the transformation process is the interplay between the SARs and the NaOH/SiO2 ratios. A higher SAR requires a greater amount of NaOH to obtain the pure-phase SSZ-39, but the yield is lower compared to the case obtained from a lower SAR in the feed (Table S1). All resulting pure SSZ-39 zeolites with different SARs show good crystallinities and uniform rectangular morphology (Fig. S2). The crystal sizes reduce as the SAR ratio increases in the feed, due to a high SAR-induced requirement of the NaOH amount in the synthesis that leads to the formation of more nuclei.29,30

The crystallization process of SSZ-39 with a SAR of 14 (run 8, Table S1) was monitored through time-resolved characterization studies (Fig. 1). The X-ray diffraction (XRD) patterns of the un-treated sample present characteristic peaks of the FAU structure (Fig. 1a), confirming the preservation of the X framework. This observation agrees with the octahedral morphology observed from the scanning electron microscopy (SEM) image at this stage (Fig. 1b). Crucially, the complete absence of detectable AEI-related crystalline reflections or structural framework vibrations in the 0 h sample provides direct, concrete evidence of a sacrificial seed dissolution pathway in a highly alkaline synthesis environment. Instead of persisting as passive crystalline centers from the outset, the introduced seeds rapidly disaggregate during the initial stage of hydrothermal treatment, thereby liberating critical molecular-level secondary building units into the amorphous gel to direct the framework transformation. This rapid dissolution behavior aligns seamlessly with established mechanistic precedents where even higher initial seed loadings remain fully transparent to XRD and spectroscopy during the early induction period.2,31 After 12 h crystallization, a distinct new diffraction peak at 9.6° emerges. Meanwhile, the surface of the octahedral crystals shows significant coarsening and breakage, with small rectangular and irregularly shaped particles attached (Fig. 1b). These changes signify the onset of the crystallization process toward SSZ-39 zeolites. Between 12 and 24 h, the characteristic AEI XRD peaks (9.6, 10.6, 13, and 16.8°) intensify, while the typical FAU peaks, particularly the first peak at 6.3°, begin to diminish (Fig. 1a). During this period, the previously aggregated fuzzy small particles evolve into cubic particles with well-defined edges, signalling a gradual transformation into SSZ-39. By 28 h, the FAU peaks completely disappear, and the SEM image (Fig. 1b) reveals only uniformly stacked cubic crystals, indicating a pure AEI phase at this point. The crystallization process completes after 30 h.

Fig. 1. Time-resolved (a) XRD patterns, (b) SEM images of the crystallization process from X zeolite to X-SSZ-39 zeolite, and (c) FT-IR spectra and (d) UV-Raman spectra of the solid products collected at different crystallization times: 0, 12, 18, 24, and 30 h during the IZC process from X zeolite to X-SSZ-39 zeolite.

Fig. 1

Fourier transform infrared (FT-IR) and UV-Raman spectroscopy were conducted for probing the growth process at the building unit level. In the FT-IR spectra shown in Fig. 1c, the main peaks observed between 0 and 18 h are attributed to vibration signals from X zeolite, indicating its predominance. The peaks at 558 and 950 cm−1 correspond to D6R32 and T–O asymmetric stretching vibrations in X zeolite, respectively.33 The other peaks (747, 667, and 446 cm−1) are assigned to T–O related signals.34 The 24 h spectrum features a broadened peak at around 1000 cm−1, which encompasses both the 950 and 1050 cm−1 peaks, marking a transitional point where both structures co-exist. Beyond 24 h, the 1050 cm−1 peak stabilizes. The blue shift from 950 cm−1 to 1050 cm−1 signifies an increase in the SAR within the zeolite framework, agreeing with the transformation from X zeolite (SAR = 2) to SSZ-39 zeolite (SAR = 14).35 Meanwhile, typical framework vibration peaks corresponding to the SSZ-39 structure are observed, including the characteristic D6R peak at 618 cm−1, and 4R peaks at 450 and 542 cm−1 in SSZ-39.36

The UV-Raman spectra (Fig. 1d) reveal a transition stage at 12 h, consistent with the XRD results. At 0 h, characteristic vibration peaks corresponding to X zeolite are observed, including a main peak at ca. 506 cm−1 assigned to the vibration of the S4R, and the bending modes of the D6R at ca. 288 and ca. 375 cm−1.37 At the 12 h mark, the main peak broadens and shifts to 500 cm−1, likely due to a partial decomposition of the X zeolite framework (particularly the S4R-rich cages) and the emerging S4R units belonging to the AEI structure. Starting at 18 h of crystallization, a further shift of the 500 cm−1 peak toward 475 cm−1 is evidenced; and the latter corresponds to the S4R unit in SSZ-39.28 Meanwhile, a new peak emerges at ca. 335 cm−1, attributed to the bending mode of D6R in the SSZ-39 structure.38 This indicates that the formation of D6R follows the occurrence of S4Rs.28 From 18 to 30 h, the Raman peaks of SSZ-39 sharpen, suggesting the development of the AEI structure over time.

To provide a comprehensive comparison, the inter-zeolite crystallization kinetics of the Y-zeolite-derived process were systematically monitored (Fig. S3 and Table S2). At the onset of the reaction (0 h), the XRD pattern (Fig. S3a) and SEM analysis (Fig. S3b) exhibit characteristic FAU reflections and pristine octahedral morphology, confirming that the parent framework remains intact prior to hydrothermal treatment. Between 12 and 66 h, a progressive attenuation of FAU reflections is observed alongside a concomitant intensification of AEI diffraction peaks. Correspondingly, SEM images capture the initial emergence of cubic particles at 12 h, with their population steadily increasing over time, signifying the continuous growth of SSZ-39. By 72 h, the AEI reflections plateau in intensity and the diffraction profiles stabilize, indicating the achievement of peak crystallinity. At this stage, the solid product consists predominantly of well-defined cubic crystals with a representative particle size of approximately 500 nm. Notably, a discernible decrease in relative crystallinity is observed upon extending the crystallization to 96 h. This phenomenon is attributed to partial framework degradation resulting from prolonged exposure to the caustic alkaline environment under high-temperature and high-pressure hydrothermal conditions. These observations collectively demonstrate that the phase transformation from Y-zeolite to SSZ-39 reaches completion at 72 h.

The evolution of framework vibrational signatures during the Y-to-SSZ-39 transformation was monitored via FT-IR spectroscopy (Fig. S3c). The parent Y zeolite exhibits characteristic bands at 1000, 792, 727, 578, and 461 cm−1, whereas the fully crystallized SSZ-39 features shifted bands at 1050, 797, 728, 618, 542, and 430 cm−1.36 Throughout the crystallization process, the asymmetric stretching vibration (Peak 1) undergoes a distinct blue shift from 1000 to 1050 cm−1, while the T–O bending vibration (Peak 2) shifts toward lower frequencies. Notably, the absorption band associated with D6R units remains persistent, shifting from 578 cm−1 in the parent Y to 618 cm−1 in the AEI framework. This continuous presence suggests that the D6R building units are largely preserved rather than completely dissolved during the interzeolite conversion. Consistent with XRD findings, the disappearance of Y-zeolite characteristic reflections by 72 h confirms the completion of the phase transformation.

UV-Raman spectroscopy was further employed to characterize the evolution of structural units during the crystallization process (Fig. S3d). At 12 h, the vibrational profile is dominated by the parent Y-zeolite, featuring a primary band at approximately 500 cm−1 associated with S4R units and a D6R bending mode at approximately 300 cm−1.28 Between 12 and 24 h, the intensity of the S4R signal increases significantly while that of the D6R mode increases marginally, suggesting that D6R units are formed by combining two S6R segments generated from additional S4R motifs.39 This behavior explicitly diverges from the X-zeolite-based interzeolite conversion, where D6R units assemble directly from S4R intermediates. This difference strongly implies that X and Y zeolites undergo the topological transformation via entirely different crystallization mechanisms. By 48 h, the broadening of the main peak signifies the onset of structural transformation, resulting from the partial depolymerization of the parent framework and the incipient emergence of S4R units within the AEI topology.

To quantitatively evaluate the impact of the parent framework topology and composition on the crystallization kinetics, the transformation efficiencies of the X- and Y-zeolite-derived processes were compared in terms of relative crystallinity (RC) as a function of time (Fig. 2). This comparative analysis reveals a distinct disparity in the induction periods and growth rates between the two systems. Specifically, a rapid RC drop of X zeolite is evidenced between 20 and 28 h and reaches 0% after 28 h, indicating its rapid dissolution. In contrast, the RC of Y zeolite exhibits a much slower decline throughout the IZC process (0–72 h). This significant difference in crystallization rates is attributed to the fact that the Y zeolite-based IZC process relies entirely on nutrients from Y zeolite, while the X zeolite-based process contributes only a small portion, with most silicon present as amorphous silica sol. The dissolution rate of the amorphous silicon source in the case of the X zeolite-based process is faster than that of the crystal-to-structural units in the case of the Y zeolite-based one, resulting in a faster transformation rate of X zeolite to SSZ-39.1,40

Fig. 2. Crystallinity profiles of parent X and Y zeolites versus their transformed X-SSZ-39 and Y-SSZ-39 phases as a function of crystallization time.

Fig. 2

While these macroscopic kinetic profiles clearly illustrate the superior transformation efficiency of the X-zeolite system, the underlying molecular mechanism warrants deeper investigation. Given that the addition of external colloidal silica reduces the relative abundance of pre-organized parent motifs in the X-zeolite mixture, its accelerated crystallization stands in contrast to the conventional view that structural inheritance primarily dictates the transformation rate. This distinct kinetic behavior suggests that the microscopic pathways governing framework dismantling and topological reorganization in the X-zeolite system possess inherently lower energetic barriers.

To validate this hypothesis, DFT calculations were performed to evaluate the hydrolysis energetics of S6R and S4R units derived from X and Y zeolites, respectively (Fig. 3). The results demonstrate that the hydrolysis steps for the X-zeolite framework are significantly more exothermic than those for Y-zeolite. Consequently, X-zeolite exhibits a higher thermodynamic propensity to release SOD cage units via localized hydrolysis. It can be further deduced that these liberated SOD cages undergo more facile depolymerization into constituent S6R and S4R building blocks compared to Y-zeolite, thereby accelerating the subsequent assembly of D6R motifs. Furthermore, X-zeolite can directly depolymerize into D6R units (Fig. 3b) through an energetically highly favorable pathway (−17.54 kcal mol−1), contrasting sharply with the endothermic nature of the corresponding process in Y-zeolite (3.89 kcal mol−1). Based on these collective findings, a distinct mechanistic divergence between the X- and Y-zeolite-assisted transformation pathways can be established. X-zeolite exhibits a pronounced thermodynamic propensity to depolymerize into S4R subunits via highly exothermic localized hydrolysis steps. This rapid liberation and subsequent preferential assembly of S4R building blocks into the D6R motifs of the target AEI framework perfectly rationalize the initial consumption of S4R intensity observed from 0 to 12 h in the UV-Raman spectra. This streamlined, S4R-mediated assembly under hydrothermal equilibrium effectively accelerates nucleation and enhances overall crystallization efficiency. In sharp contrast, the traditional Y-zeolite-derived process initially relies on the reorganization of S6R units to generate D6R motifs (0–24 h). DFT energetics reveal that this pathway is thermodynamically unfavorable, which inherently retards early-stage conversion efficiency and slows down crystallization kinetics, as confirmed by the UV-Raman results. Crucially, the X-zeolite system uniquely benefits from an alternative pathway involving the direct dissociation of intact D6R segments from the parent lattice for seamless integration into the growing AEI structure. DFT calculations confirm that this direct incorporation process is energetically highly favorable (−17.54 kcal mol−1), contrasting sharply with the endothermic nature of the corresponding process in Y-zeolite (3.89 kcal mol−1). The viability of this spontaneous direct integration, coupled with the facile S4R-mediated exothermic assembly, fundamentally underpins the kinetic superiority of X-zeolite over Y-zeolite.

Fig. 3. Shedding of (a) SOD cage units and (b) D6R units from X and Y zeolites, respectively. The optimized structures for the formation of the D6R from (c) two S6Rs and (d) three S4Rs, respectively.

Fig. 3

Given the potential for SOD cages in X zeolite to decompose into S4R and S6R for further assembly into the AEI structure, we further explored the IZC process using A zeolite (LTA structure, Fig. S4), consisting of SOD cages linked with D4R (Table S2). The crystallization evolution of the A-zeolite-derived transformation into SSZ-39 was systematically characterized via XRD and FT-IR spectroscopy (Fig. S5). At the initial stages (0 and 4 h), the XRD profiles (Fig. S5a) exhibit well-defined reflections at 7.1, 10.1, 12.3, 16.0, and 23.9°, characteristic of pristine A zeolite. Concurrently, the FT-IR spectra (Fig. S5b) display intense vibrational bands at 970, 550, and 463 cm−1, corresponding to the asymmetric T–O stretching, D4R cage vibrations, and T–O bending modes of the LTA parent lattice, respectively. By 8 h, the emergence of AEI reflections at 9.6, 16.8, and 21.0° indicates a mixed-phase regime where LTA and AEI topologies coexist. This topological reconstruction is clearly mirrored in the corresponding FT-IR profiles, which reveal the appearance of characteristic SSZ-39 skeletal vibrations at 1050, 797, 678, 609, 533, and 450 cm−1, accompanied by a sharp diminution of the LTA-specific D4R vibrational signature. Notably, the asymmetric T–O stretching band undergoes a pronounced hypsochromic shift from 970 to 1050 cm−1, directly fingerprinting the progressive framework silicification as the SAR evolves from the highly aluminous parent matrix (SAR = 2) to the crystallized SSZ-39 architecture (SAR = 9). Upon extending the hydrothermal treatment to 10 h, the primary LTA reflections are entirely extinguished, yielding an embryonic AEI phase that undergoes continuous long-range ordering and structural maturation, as evidenced by the steady sharpening of diffraction peaks and the systematic enhancement of relative crystallinity with prolonged reaction time.

The morphological evolution of the samples during the LTA-to-AEI transformation was systematically monitored via SEM analysis (Fig. S5c). At 0 h, the parent A-zeolite displays its characteristic well-defined cubic morphology with an average particle size of approximately 2.3 µm. Upon hydrothermal treatment for 8 h, the pristine cubic matrix begins to undergo structural erosion and dissolution, accompanied by the incipient precipitation of nascent cuboidal SSZ-39 particles in the surrounding media. By 10 h, the population of these cuboidal SSZ-39 crystals increases dramatically, though residual fragments of the decaying LTA cubes remain partially discernible. Extending the reaction time to 12 h leads to the complete eradication of the original LTA cubic morphology, signifying the thorough dissolution of the parent framework and its quantitative rearrangement into the target AEI phase. With further prolonged hydrothermal aging, the SSZ-39 zeolite enters a stage of structural maturation and deep crystallization, yielding well-crystallized particles with dimensions stabilized in the range of 0.5 to 1.5 µm.

Comprehensive correlation of our kinetic profiles, spectroscopic fingerprinting, and DFT calculations establishes a unified mechanistic framework governing the interzeolite conversion (IZC) to the AEI topology (Scheme 1). The overall transformation efficiency is likely governed by a synergistic combination of specific synthesis compositions and the intrinsic structural disassembly pathways of the parent matrix. Specifically, the observed faster crystallization of the X-route may arise from a combination of the lower Si/Al ratio of the precursor promoting faster dissolution, the presence of external colloidal silica increasing the nutrient supply, and the possibly more favorable structural disassembly pathway involving SOD-cage release. This manifests as a stark mechanistic divergence between the X and Y zeolite precursors, where both compositional variations and topological features play significant roles. The highly aluminous X zeolite framework (Si/Al = 1.5) undergoes rapid dissolution via highly exothermic localized hydrolysis. This process liberates a transient pool of reactive SOD cage intermediates that undergo facile depolymerization into S4R structural motifs, thereby accelerating nucleation kinetics under hydrothermal equilibrium. Crucially, DFT energetics confirm that X zeolite uniquely permits the direct cleavage of intact D6R segments from the parent lattice via a thermodynamically spontaneous pathway (−17.54 kcal mol−1). These favorable thermodynamic and kinetic features of the X-zeolite route drastically accelerate the crystallization rate, allowing full phase transformation within merely 30 h and effectively shielding the emerging AEI framework from severe alkaline corrosion. Consequently, this fast crystallization process fundamentally suppresses framework dealumination and defect generation, preserving highly intact tetrahedral aluminum coordination (vide infra) and yielding a high-crystallinity SSZ-39 zeolite matrix with a distinct isolated-Al-dominated framework structure. The confluence of this spontaneous direct structural integration with the efficient S4R-mediated exothermic assembly effectively bypasses the sluggish induction periods intrinsic to traditional Y zeolite routes, where framework reconstruction is constrained by the kinetically hindered and thermodynamically less favorable recombination of S6R fragments. Unequivocal validation for this SOD cage mediated pathway, which challenges the conventional D6R-centric inheritance model, is provided by the strategic deployment of D6R free A zeolite (LTA topology, Si/Al = 1). Under DMPOH-mediated hydrothermal conditions at 150 °C, the LTA framework undergoes localized topological dismantling to yield individual SOD cages. Despite the absolute absence of intrinsic D6R units in the parent matrix, these liberated SOD cages undergo controlled depolymerization into reactive S6R and S4R fragments, which cooperatively condense to assemble the constituent D6R motifs required for the embryonic AEI network (Si/Al = 6). This successful LTA-to-AEI transition firmly establishes the role of the SOD cage as a versatile, transferable intermediate, suggesting that the rigid inheritance of intact D6R motifs is not an absolute prerequisite for AEI framework construction. In essence, this unified scheme proposes that IZC efficiency depends on the thermodynamic and kinetic accessibility of framework disassembly into adaptable sub-structural motifs, rather than the rigid inheritance of identical building units. Bridging the mechanism of FAU-to-AEI and LTA-to-AEI transitions offers insights for designing small-pore zeolites via precursor engineering. While direct observation of intact SOD cages remains elusive, the thermodynamic favourability of their release and the structural feasibility of their recombination provide a compelling and self-consistent rationale for all experimental observations.

Scheme 1. Proposed mechanisms for the crystallization of SSZ-39 zeolite assisted by X zeolite and A zeolite, respectively.

Scheme 1

NH3-SCR performance

The NOx removal efficiency of the Cu-SSZ-39 catalysts prepared from different starting zeolites (X, Y, and A) was evaluated for the NH3-SCR reaction. The impact of varying gas hourly space velocity (GHSVs = 120 000, 300 000, and 600 000 h−1) on the denitrification performance was examined (Fig. 4 and S6). At lower GHSVs of 120 000 and 300 000 h−1, Cu-X-SSZ-39 and Cu-Y-SSZ-39 exhibit overlapping NOx conversion profiles within experimental error. However, Cu-A-SSZ-39 shows a clear decrease in activity when the GHSV is increased to 300 000 h−1. When evaluated at an ultra-high GHSV of 600 000 h−1, the superiority of Cu-X-SSZ-39 becomes highly pronounced, maintaining the highest NOx conversion across the entire investigated temperature range. As shown in Fig. 4a, both Cu-X-SSZ-39 and Cu-Y-SSZ-39 maintain an excellent and stable deNOx platform (approximately 97% conversion) in the mid-to-high temperature window of 250–600 °C. Nevertheless, Cu-Y-SSZ-39 displays lower efficiency in the low-temperature region (<250 °C). While Cu-A-SSZ-39 exhibits a low-temperature light-off behavior comparable to Cu-Y-SSZ-39 between 150 and 250 °C, its high-temperature performance is significantly impaired. Above 500 °C, NOx conversion over Cu-A-SSZ-39 decreases sharply, dropping to 65.2% at 600 °C, which severely narrows its effective operating temperature window (Fig. 4b) due to the unwanted competitive oxidation of ammonia. Overall, the Cu-X-SSZ-39 catalyst demonstrates the most robust NH3-SCR activity and widest temperature window among all evaluated samples.

Fig. 4. NOx conversion over the Cu-SSZ-39 catalysts (a) before and after hydrothermal treatment and (b) the temperature range of >90% of NOx conversion. Reaction conditions: 0.5 vol% of NO, 0.5 vol% of NH3, 20 vol% of O2 and balanced with N2 at GHSV of 600 000 h−1.

Fig. 4

To evaluate hydrothermal robustness, Cu-SSZ-39 catalysts were subjected to severe aging at 800 °C in the presence of 10 vol% water vapor for 16 h; the aged samples are denoted as Cu-SSZ-39-HTA. Following this harsh hydrothermal treatment, all catalysts exhibit varying degrees of deactivation (Fig. 4a). Remarkably, Cu-X-SSZ-39-HTA demonstrates superior hydrothermal durability compared to the other aged counterparts, maintaining a high NOx conversion of above 90% across a broad temperature window of 250–550 °C. In sharp contrast, Cu-A-SSZ-39-HTA suffers from severe performance degradation, achieving over 90% conversion only within a drastically restricted temperature range of 300 to 450 °C (Fig. 4b).

Importantly, the product selectivity remains uncompromised after aging (Fig. S7), i.e., all fresh and aged catalysts maintain an excellent N2 selectivity exceeding 97% over the entire investigated temperature trajectory. Concurrently, the formation of unwanted N2O byproduct emissions is effectively suppressed, with the maximum N2O concentration remaining below 8 ppm across all evaluated samples. This suppressed N2O generation indicates that the severe hydrothermal aging does not promote detrimental side reactions, such as the non-selective oxidation of ammonia to nitrous oxide. Collectively, these results demonstrate that the Cu-X-SSZ-39 catalyst exhibits exceptional hydrothermal stability alongside its superior fresh activity, showcasing outstanding potential for practical application.

Characterization of Cu-SSZ-39 derived from different precursors

Elemental analysis by the inductively coupled plasma atomic emission (ICP-OES) technique (Table 1) suggests that all three Cu-SSZ-39 catalysts have similar Cu contents (1.71–1.90 wt%). Cu-X-SSZ-39 and Cu-Y-SSZ-39 have a similar Si/Al ratio slightly higher than that of Cu-A-SSZ-39 (6.1–6.2 vs. 4.4).

Table 1. The elemental analysis, Si/Al ratio, content of Cu species, BET surface areas (SBET) and acid content of different Cu-SSZ-39 catalysts.

Samples Cua (wt%) Si/Ala Cu/Ala S BET b (m2 g−1) Acid contentc (mmol g−1)
Weak Cu ions Strong
Cu-X-SSZ-39 1.90 6.2 0.19 678 0.290 0.292 0.316
Cu-Y-SSZ-39 1.87 6.1 0.19 656 0.219 0.242 0.157
Cu-A-SSZ-39 1.72 4.4 0.12 629 0.213 0.170 0.155
a

Determined by ICP-OES.

b

BET method applied to the N2 isotherms.

c

Obtained from the peak fitting of the NH3-TPD results.

Hydrothermal stability is closely linked with the Si/Al ratios. Since Cu-A-SSZ-39 has the lowest Si/Al ratio, it is more prone to Al leaching during the aging process,41 resulting in inferior NH3-SCR performance. Scanning transmission electron microscopy and energy-dispersive X-ray spectroscopy (STEM-EDS) mapping (Fig. S8) results indicate the uniform dispersion of Cu and other elements (Al, Si and O). While Cu and framework elements (Al, Si, and O) are uniformly distributed in Cu-X-SSZ-39 (Fig. S8a) and Cu-Y-SSZ-39 (Fig. S8b), a distinct local agglomeration of Cu species is observed in Cu-A-SSZ-39 (Fig. S8c). This pristine-state segregation reveals a weak coordination capacity of the low-SAR framework to anchor isolated Cu centers, structurally suggesting its vulnerability to framework breakdown and de-alumination under severe hydrothermal aging.

The XRD patterns of the copper-loaded catalysts (Fig. 5a) display well-defined reflections characteristic of the AEI topology, confirming that the crystalline framework remains intact post-metallation. However, the relative crystallinity exhibits a strong dependence on the parent precursor, following the order of Cu-X-SSZ-39 (100%) > Cu-Y-SSZ-39 (92.7%) > Cu-A-SSZ-39 (59.8%). This structural trend directly mirrors the NH3-SCR performance. The highly crystalline framework of Cu-X-SSZ-39 implies a long-range ordered lattice with low defect density, which optimizes the stabilization of active isolated Cu ions and underpins its superior space-velocity tolerance. Conversely, the severely diminished crystallinity of Cu-A-SSZ-39 (59.8%) points to substantial structural degradation or disorder. This framework deficiency limits the effective coordination of active Cu centers and potentially promotes the formation of inactive CuO clusters, directly resulting in its truncated operating temperature window and inferior hydrothermal durability.

Fig. 5. Characterization studies of different Cu-SSZ-39 catalysts: (a) XRD patterns, (b) NH3-TPD profiles, (c) H2-TPR profiles and (d) UV-vis spectra.

Fig. 5

To quantify the acid site distribution, NH3 temperature-programmed desorption (NH3-TPD) profiles were recorded and deconvoluted into three distinct ammonia desorption regimes (Fig. 5b and Table 1). The low-temperature peak centered at around 200 °C corresponds to weakly bound NH3 on weak acid sites, such as terminal silanols or extra-framework Al species.42 The mid-temperature desorption (250–400 °C) is ascribed to ammonia coordinated to strong Lewis acid sites created by ion-exchanged isolated Cu2+ ions.43 The high-temperature feature near 500 °C originates from structural Brønsted acid sites associated with framework Si–O(H)–Al motifs.44 Notably, Cu-X-SSZ-39 features the highest density of both strong Lewis and Brønsted acid sites among all evaluated candidates. Intriguingly, while Cu-X-SSZ-39 and Cu-Y-SSZ-39 share similar bulk Si/Al ratios and copper loadings, the former possesses nearly twofold the Brønsted acid capacity of the latter. This striking disparity strongly implies that the X-zeolite-derived transformation pathway promotes a superior preservation of framework Al atoms within the crystallized lattice, which aligns seamlessly with its higher structural crystallinity and lower defect density. Consequently, the significantly enhanced NH3 adsorption and storage capacity provides a substantial reactant reservoir, facilitating low-temperature light-off and underpinning the outstanding NH3-SCR activity of the Cu-X-SSZ-39 catalyst.

Hydrogen temperature-programmed reduction (H2-TPR) was performed to elucidate the redox behaviors and structural coordination of the host copper species (Fig. 5c). Within the 150–400 °C regime, the reduction profiles resolve into a triad of well-defined hydrogen consumption features centered at approximately 230, 290, and 340 °C, consecutively assigned to the reduction of isolated Z-[Cu(OH)]+ cations to Cu+,45 Cu2+ inside highly dispersed amorphous CuOx aggregates,46 and framework-stabilized Z2-Cu2+ ions.41,47 Notably, Cu-Y-SSZ-39 displays an isolated high-temperature reduction feature at 420 °C, signifying the presence of a parasitic CuAl2O4 spinel phase.48 The generation of CuAl2O4 within the Y-derived framework is driven by the pronounced thermodynamic affinity between copper and mobile extra-framework aluminum species, which selectively consumes the active Z-[Cu(OH)]+ precursors while leaving the Z2-Cu2+ configuration unaffected.49 This specific structural redistribution rationalizes the compromised low-temperature SCR activity (below 200 °C) of Cu-Y-SSZ-39 relative to Cu-X-SSZ-39. Conversely, the thermal stability of this aluminate phase, coupled with the preserved Z2-Cu2+ population, enables Cu-Y-SSZ-39 to sustain a competitive high-temperature deNOx platform despite its markedly inferior Brønsted acidity. In the case of Cu-A-SSZ-39, the reduction envelope shifts systematically toward lower temperatures, disclosing an exceptionally labile coordination sphere between the copper ions and framework oxygen atoms, which directly correlates with the compromised structural integrity stemming from its low Si/Al ratio.7 Furthermore, Cu-A-SSZ-39 possesses the lowest concentration of framework-bound Z2-Cu2+ species, whereas Cu-X-SSZ-39 exhibits the highest density. Given that these rigidly coordinated Z2-Cu2+ ions are highly resilient and serve as the indispensable active centers for high-temperature denitration, their structural abundance fundamentally underpins the superior, wide-window NH3-SCR efficiency observed exclusively over the Cu-X-SSZ-39 catalyst.50

UV-vis diffuse reflectance spectroscopy (UV-vis DRS) was employed to fingerprint the copper speciation and coordination environments (Fig. 5d). All catalysts display a dominant absorption envelope within 200–280 nm, unambiguously assigned to the oxygen-to-metal charge transfer of isolated Cu2+ cations (comprising both Z-[Cu(OH)]+ and Z2-Cu2+ states).51 As the primary catalytic engines driving the dynamic NH3-SCR redox cycle, these monomeric Cu2+ centers facilitate the reversible interconversion between Cu+ and Cu2+ required for reactant activation.52,53 The secondary shoulder at 300–350 nm corresponds to the charge transfer of oligomeric, amorphous CuOx clusters. Notably, Cu-X-SSZ-39 exhibits the highest absorption intensity in the isolated Cu2+ region. Benefiting from its exceptional framework crystallinity and minimized defect density, the X-zeolite-derived matrix exhibits an enhanced metal-exchange capacity that acommodates a higher density of active monomeric cations alongside slightly elevated CuOx species. This electronic distribution seamlessly corroborates the H2-TPR quantification, demonstrating that Cu-X-SSZ-39 successfully maximizes the population of active, isolated Cu centers.

Co2+ ion-exchange experiments were conducted to investigate the framework Al distribution and elucidate its decisive role in directing the speciation of Z-[Cu(OH)]+ and Z2-Cu2+ monomers (Table S3). This diagnostic method leverages the preferential coordination of divalent Co2+ cations at paired Al sites to form stable charge-balanced complexes. The results propose that the parent zeolite topology deeply dictates the spatial proximity of framework Al atoms. Cu-A-SSZ-39, with the lowest Si/Al ratio, exhibits the highest proportion of paired Al sites (29%), whereas the X-zeolite-derived matrix stabilizes the largest fraction of isolated Al sites (82%), with Cu-Y-SSZ-39 falling in between.

To further validate the preservation of framework aluminum, solid-state 27Al MAS NMR spectroscopy was conducted on the synthesized samples (Fig. S9). The dominant resonance peak centered at approximately 56 ppm is unambiguously assigned to tetrahedral framework aluminum species within the zeolite lattice, while the minor peak emerging at around 0 ppm corresponds to octahedral extra-framework aluminum species.54,55 Quantitative analysis of the spectra reveals that the X-SSZ-39 zeolite possesses the highest concentration of framework aluminum and the lowest fraction of extra-framework aluminum among all evaluated candidates. This direct spectroscopic evidence perfectly corroborates our aforementioned characterization results, demonstrating that the unique S4R-mediated assembly pathway of the X-zeolite route effectively minimizes dealumination and preserves a highly intact aluminosilicate framework, which underpins the abundant Brønsted acid sites and the optimal coordination of active copper species.

Characteristically, monovalent Z-[Cu(OH)]+ species anchor to isolated Al sites and primarily govern low-temperature NH3-SCR activity, while divalent Z2-Cu2+ ions require paired Al configurations to impart high-temperature hydrothermal stability.56,57 The exceptional low-temperature light-off efficiency of Cu-X-SSZ-39 (below 200 °C) aligns seamlessly with its high isolated Al abundance, which drives the preferential siting of highly reactive Z-[Cu(OH)]+ species, a distribution independently validated by UV-vis and H2-TPR profiles.57,58 Furthermore, the anchoring of these copper cations partially neutralizes the structural Brønsted acid sites.59 Consequently, the superior preservation of framework Al within the Cu-X-SSZ-39 matrix, coupled with this unique isolated-Al-rich distribution, rationalizes the prominent acidity and active-site density disparities observed between the X- and Y-zeolite-derived catalysts.

The key features of the developed Cu-SSZ-39 synthesized from different precursors are summarized in Fig. 6. Accordingly, a comprehensive structure–activity relationship can be established to interconnect the parent precursor topologies, framework acid attributes, copper speciation, and the resultant NH3-SCR trajectories. The selection of the starting zeolite precursor fundamentally dictates the long-range structural integrity and localized crystallization pathways of the tailored AEI framework. Cu-X-SSZ-39, inheriting the structurally optimized lattice parameters from the X-zeolite precursor, demonstrates the absolute highest relative crystallinity (100%) and a minimized framework defect density. The distinct crystallization pathway involving SOD cage reorganization suggests the atomic-level aluminum siting and structural defect density in the synthesized zeolite framework. Specifically, the rapid and intact assembly of the AEI lattice in Cu-X-SSZ-39 preserves double the concentration of framework Brønsted acid sites compared to the Y-derived counterpart, while directing 82% of the framework aluminum into isolated Al sites. This synthesized framework structure directly dictates the subsequent copper coordination chemistry, driving the preferential formation of highly active mononuclear Z-[Cu(OH)]+ species alongside stable Z2-Cu2+ centers. This robust framework architecture ensures the maximum, intact preservation of tetrahedral framework Al atoms, which directly translates into a nearly twofold enrichment of structural Brønsted acid sites compared to the Y-zeolite-derived counterpart. Crucially, this unique Al distribution, dominated by an exceptional abundance of isolated Al sites (82%) as cross-validated by Co2+ titration, acts as a stereospecific blueprint that orchestrates the subsequent coordination chemistry of the copper ions during the metallation process. The rich isolated Al landscape in Cu-X-SSZ-39 drives the preferential siting of highly dispersed, mononuclear Z-[Cu(OH)]+ cations, while simultaneously maintaining a highly optimized density of framework-bound Z2-Cu2+ ions, effectively suppressing the formation of parasitic, inactive CuAl2O4 or bulk CuOx aggregates. This sophisticated orchestration of active sites underpins the outstanding, wide-window deNOx efficiency and hydrothermal robustness unique to the Cu-X-SSZ-39 catalyst. The high density of Brønsted acid sites provides an abundant reactant reservoir that dynamically enriches ammonia molecules near the active zones. Concurrently, the coexisting Z-[Cu(OH)]+ and Z2-Cu2+ single-copper centers operate as high-efficiency redox engines, accelerating low-temperature intermediate turnover and maintaining high-temperature structural stability, respectively. In sharp contrast, the framework degradation and mismatched Al positioning characteristic of Cu-A-SSZ-39 and Cu-Y-SSZ-39 result in a critical deficit of active centers and the intrusion of spectator phases, rationalizing their truncated operating windows. Consequently, this precursor-directed regulation of the Al and Cu spatial environments establishes Cu-X-SSZ-39 as a well-balanced catalytic platform for advanced deNOx applications.

Fig. 6. The distribution of Cu species and total acid content for Cu-SSZ-39 synthesized from different precursors.

Fig. 6

DB-FTIR study of the reaction intermediates

The preceding characterization studies have established a clear correlation between the zeolite precursor, the resultant Al distribution, and the speciation of Cu active sites. To bridge these structural properties to the observed catalytic performance, it is essential to probe the interactions between these specific Cu sites and the reactant molecules under working conditions. According to seminal studies, the low-temperature NH3-SCR mechanism over Cu-CHA zeolites relies on a dynamic catalytic cycle fundamentally different from the high-temperature pathway. In this regime, NH3 solvates and mobilizes Cu ions from their framework coordination sites, enabling the formation of transient, highly active dinuclear complexes that are essential for the rate-limiting O2 activation step. The lability and accessibility of the initial Cu sites, however, are critical for this process. Therefore, to further elucidate the structure–performance relationship connecting Al siting, Cu speciation, and catalytic reactivity, we employed in situ DB-FTIR spectroscopy to directly monitor the evolution of adsorbed species.

We first studied NO adsorption on Cu-SSZ-39 at −196 °C (Fig. 7). The absorption bands in the 1700–1800 cm−1 range (specifically resolving at 1733, 1778, and 1826 cm−1) represent NO molecules interacting with isolated Cu+ ions, while the peak at 1889 cm−1 is attributed to [Cu(OH)]+-NO complexes within 8MR.60 The distinct peaks observed between 1910 and 1950 cm−1 are associated with NO molecules adsorbed on isolated Cu2+ ions in 6MR,11 accompanied by a minor peak at 2329 cm−1 corresponding to physisorbed N2O molecules.61 This species is generated in situ through the low-temperature disproportionation of reactant NO catalyzed by poorly stabilized or defective copper sites, exhibiting a highly sharpened profile due to the intense steric confinement within the AEI cages.

Fig. 7. Series of DB-FTIR spectra of pre-oxidized Cu-X-SSZ-39, Cu-A-SSZ-39, and Cu-Y-SSZ-39 catalysts after exposure to saturated NO at −196 °C. The samples were annealed in a vacuum at 400 °C for 1 h prior to the measurements.

Fig. 7

All three catalysts exhibit characteristic Cu2+-NO bands near 1947 and 1911 cm−1, reflecting two types of isolated Cu2+ centers with different stabilities. Notably, Cu-Y-SSZ-39 shows a higher relative intensity for the less stable Cu2+-NO species at 1947 cm−1 compared to the more stable 1911 cm−1 peak.60 This framework instability in the Y-derived sample is further supported by the continuous increase in the N2O signal at 2329 cm−1 with longer NO exposure time, suggesting a side reaction catalyzed by these poorly stabilized copper centers. Furthermore, Cu-Y-SSZ-39 shows no distinct Cu+-NO absorption bands at 1733, 1778, and 1826 cm−1, indicating a lack of alternative active copper species. These chemisorption results demonstrate that despite sharing similar bulk copper loadings with Cu-X-SSZ-39, the framework-anchored Z2-Cu2+ active sites in Cu-Y-SSZ-39 exhibit lower structural stability. This deficiency explains its rapid high-temperature deactivation and narrowed operating window after hydrothermal aging. In contrast, Cu-X-SSZ-39 maintains a much higher abundance of well-coordinated, stable, and highly active isolated copper species, fundamentally underpinning its excellent fresh activity and exceptional hydrothermal durability.

NH3 adsorption was then conducted to resolve the surface acid distributions and copper coordination characteristics under working states (Fig. 8). As shown in Fig. 8a, the control H-X-SSZ-39 sample features well-defined bands at 3394, 3335, and 1458 cm−1, explicitly assigned to the vibrations of NH4+ ions chemisorbed on Brønsted acid sites (B-NH4+).62 Upon copper introduction, Cu-X-SSZ-39 exhibits new distinct features at 3272, 3178, and 1620 cm−1, which track NH3 molecules coordinated to Lewis acidic copper centers (L-NH3).63 Concurrently, the negative reflections appearing in the hydroxyl stretching region at 3725, 3654, 3606, and 3576 cm−1 monitor the consumption of Si–OH, Al–OH, Cu–OH, and Si–OH–Al groups, respectively, upon reaction with the incoming ammonia probes. Notably, the sharp negative band at 3606 cm−1 directly fingerprints the interaction with reactive [Cu(OH)]+ species.42,63 Among all evaluated candidates, Cu-X-SSZ-39 displays the most intense [Cu(OH)]+ depletion signal, while Cu-A-SSZ-39 exhibits the weakest response, a trend that coincides with its minimal L-NH3 intensity at 1620 cm−1 and points to a low density of accessible copper sites.53 The time-resolved evolution of ammonia uptake was sequentially tracked to evaluate the adsorption kinetics (Fig. 8b and c). For both Cu-X-SSZ-39 and Cu-Y-SSZ-39, the normalized peak areas for L-NH3 at 1620 cm−1 increase rapidly within the first 7 min of exposure, demonstrating an efficient dynamic coordination between the isolated Cu2+ sites and reactant ammonia molecules. In contrast, Cu-A-SSZ-39 displays a noticeable kinetic lag, with a significant signal increase observed only after 10 min, revealing that ammonia capture over its isolated copper sites is substantially less reactive. Meanwhile, the B-NH4+ feature at 1458 cm−1 follows a parallel kinetic trajectory to the L-NH3 band prior to saturation.61 These transient adsorption profiles clearly reveal that the X-zeolite-derived matrix successfully optimizes both the absolute density and the initial kinetic responsiveness of the active Brønsted acid and copper centers.

Fig. 8. DB-FTIR spectra collected at 200 °C in a 1000 ppm NH3/Ar atmosphere of Cu-SSZ-39 (a) after 30 min, and (b) at different intervals, and (c) normalized band areas associated with (b).

Fig. 8

To probe the surface catalytic pathways and intermediate reaction kinetics under working conditions, transient in situ DB-FTIR spectra were recorded at 200 °C (Fig. 9 and S10). Upon exposing the catalysts to a NO and O2 atmosphere (Fig. 9a), the primary adsorbed NOx intermediates resolve into bridged nitrates (∼1625 cm−1),64 monodentate nitrates bound to framework Al sites (∼1597 cm−1),65 and monodentate nitrates coordinated to Cu centers (∼1575 cm−1).66 Notably, Cu-X-SSZ-39 demonstrates the most prominent adsorption capacity for all nitrate configurations, particularly the highly reactive bridged nitrates at 1625 cm−1. Benefiting from the superior redox capability confirmed by H2-TPR, the X-zeolite-derived matrix effectively accelerates the initial oxidation of NO to NO2, thereby promoting the rapid formation of active nitrate pools required for low-temperature light-off.67,68 To evaluate the dynamic reactivity of these surface intermediates, reactant NH3 was subsequently introduced into the NOx-preadsorbed cell (Fig. 9b–d).69 Across all three catalysts, the intensities of the nitrate bands (1500–1700 cm−1) experience a swift attenuation concurrent with the growth of adsorbed NH3 species. This synchronous trend verifies that the reaction proceeds via the typical Langmuir–Hinshelwood (L–H) mechanism driven by the interaction between co-adsorbed nitrates and ammonia.53

Fig. 9. In situ DB-FTIR spectra of (a) Cu-SSZ-39 at 200 °C in a 1000 ppm NO + 5 vol% O2 atmosphere after 30 min, (b–d) Cu-SSZ-39 at 200 °C in a 1000 ppm NH3 and pre-adsorbed NO + 5 vol% O2 species atmosphere, and (e) normalized band areas associated with the results in Fig. 8d.

Fig. 9

To precisely compare the intermediate consumption kinetics without spectral overlap interference, the normalized area variations of the distinct features were systematically quantified over time (Fig. S11). In the 1620 cm−1 region (Fig. S11b), the absorption envelope drops instantly before rising again.42 This unique V-shaped trajectory originates from a sequential two-step process: the rapid chemical consumption of Cu-bound nitrates (∼1625 cm−1) by the incoming ammonia, followed by the immediate re-coordination of excess reactant ammonia onto the newly liberated Lewis acidic copper centers (L-NH3). Crucially, Cu-X-SSZ-39 demonstrates the deepest drop in intensity during the first minute, reflecting an outstanding active site responsiveness. This superior kinetic profile is unambiguously validated by the isolated Cu-nitrate band at 1575 cm−1 (Fig. S11c), where the signal over Cu-X-SSZ-39 drops sharply within 1 min and completely vanishes by 15 min. In contrast, Cu-A-SSZ-39 exhibits a slower consumption rate, while Cu-Y-SSZ-39 suffers from the most sluggish depletion kinetics, with its nitrate feature remaining visible up to 30 min. Concurrently, the accumulation rate of B-NH4+ at 1448 cm−1 confirms that Cu-X-SSZ-39 activates its Brønsted acid reservoir with optimal responsiveness (Fig. 9e). This superior reactivity profile clarifies a critical mechanistic trade-off regarding the binding energy of surface intermediates. Although the structurally optimized aluminum siting inherited from the X-zeolite precursor directs the highest adsorption capacity (or local concentration) of surface bridged nitrate species within the AEI cages, this relatively strong binding does not induce site poisoning or block the catalytic centers. Instead, as firmly corroborated by the H2-TPR profiles, the potential kinetic bottleneck typically imposed by strong adsorption is successfully bypassed by the outstanding low-temperature redox capability of Cu-X-SSZ-39. This enhanced electron-transfer proficiency dramatically lowers the activation barrier for the Langmuir–Hinshelwood pathway, prompting the highly concentrated surface nitrates to undergo the most rapid reaction with neighboring ammonia molecules. Collectively, these transient observations provide concrete molecular-level evidence that the X-zeolite-derived catalyst dramatically accelerates intermediate turnover and coordinates reactants with optimal efficiency, fundamentally rationalizing its unparalleled NH3-SCR efficiency at a reduced operating temperature of 200 °C.

Conclusions

To summarize, this study demonstrates a highly efficient interzeolite conversion (IZC) strategy for synthesizing high-performance SSZ-39 zeolites, highlighting the breakthrough deployment of cost-effective X zeolite as a strategic precursor to replace traditional Y zeolite. Our approach completely circumvents energy-intensive pretreatments and dealumination steps, providing critical, non-conventional insights into the crystallization mechanism. The combined experimental observations and DFT calculations suggest that the FAU-to-AEI topological transition is likely governed by the localized hydrolysis and reorganization of sodalite (SOD) cage units, providing a compelling alternative to the direct inheritance of traditional double six-ring (D6R) motifs. This proposed SOD-cage-mediated pathway is highly consistent with the successful transformation of D6R-free A zeolite into pure-phase SSZ-39. The resulting X-zeolite-derived SSZ-39 exhibits superior structural parameter integrity and uniform cubic morphology. Upon copper introduction, the Cu-X-SSZ-39 catalyst sustains outstanding NOx conversion across the entire evaluated temperature trajectory, maintaining an impressive 97% deNOx efficiency even at an elevated temperature of 600 °C and ultra-high space velocities. In situ DB-FTIR spectroscopy reveals that this exceptional performance stems from a synergistic orchestration between abundant framework Brønsted acid sites and highly accessible, active Z-[Cu(OH)]+ species, which dramatically accelerates intermediate turnover and drives the rapid depletion of reactive Cu-bound nitrate complexes. Furthermore, Cu-X-SSZ-39 exhibits exceptional hydrothermal durability, retaining over 90% NOx conversion over a broad temperature window of 250–550 °C even after severe aging at 800 °C. Based on fresh-state characterization, we tentatively attribute this robust stability to its superior crystallinity, enriched Brønsted acidity, and a higher proportion of framework-anchored Z2-Cu2+ species, which are recognized for their resistance to dealumination and copper migration. While these fresh-state correlations logically infer the structural origins of its durability, future investigations will prioritize the systematic characterization of aged catalysts to conclusively elucidate the underlying stabilization mechanism. Overall, this precursor engineering approach seamlessly bridges deep fundamental insights with scalable, eco-friendly catalyst design, positioning Cu-X-SSZ-39 as a promising platform for industrial NOx abatement.

Author contributions

The manuscript was completed through contributions of all authors. All authors have given approval to the final version of the manuscript. Jiachen Wang, Hui Wang and Gang Liu carried out the experiments. Limin Ren, Xiao Jiang, Ke Zhang, Hongwei Su, Michael J. Forte, and Ronghe Lin discussed the results and commented on the manuscript. Chunyan Liu, Shaoyun Chen, Jikai Zhang, and Jiachen Wang participated in some of the characterization studies. Jiaxu Liu guided the research project.

Conflicts of interest

There are no conflicts to declare.

Supplementary Material

SC-OLF-D6SC04592H-s001

Acknowledgments

This work was supported by the National Key Research and Development Program of China (2022YFA1504402), the National Energy R&D Center of Petroleum Refining Technology (RIPP, SINOPEC), the National Natural Science Foundation of China (U23B20169, 22472016), the Key R&D Program of Ningbo (No. 2023Z144), the Fundamental Research Funds for the Central Universities (DUT22LAB601, etc.), and the Research Fund of Key Laboratory of the Ministry of Education for Advanced Catalysis Materials, Zhejiang Key Laboratory of Advanced Catalysis and Adsorption Materials, Zhejiang Normal University (2025ZY01096).

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6sc04592h.

Notes and references

  1. Dusselier M. Davis M. E. Chem. Rev. 2018;118:5265–5329. doi: 10.1021/acs.chemrev.7b00738. [DOI] [PubMed] [Google Scholar]
  2. Xu H. Zhu J. Qiao J. Yu X. Sun N.-B. Bian C. Li J. Zhu L. Microporous Mesoporous Mater. 2021;312:110736–110742. [Google Scholar]
  3. Dusselier M. Deimund M. A. Schmidt J. E. Davis M. E. ACS Catal. 2015;5:6078–6085. [Google Scholar]
  4. Martín N. Li Z. Martínez-Triguero J. Yu J. Moliner M. Corma A. Chem. Commun. 2016;52:6072–6075. doi: 10.1039/c5cc09719c. [DOI] [PubMed] [Google Scholar]
  5. Shan Y. Du J. Yu Y. Shan W. Shi X. He H. Appl. Catal., B. 2020;266:118655–118667. [Google Scholar]
  6. Zhu N. Shan Y. Shan W. Sun Y. Liu K. Zhang Y. He H. Environ. Sci. Technol. 2020;54:15499–15506. doi: 10.1021/acs.est.0c06256. [DOI] [PubMed] [Google Scholar]
  7. Du J. Han S. Huang C. Shan Y. Zhang Y. Shan W. He H. Appl. Catal., B. 2023;338:123072–123082. [Google Scholar]
  8. Xiao P. Wang Y. Lu Y. De Baerdemaeker T. Parvulescu A.-N. Müller U. De Vos D. Meng X. Xiao F.-S. Zhang W. Marler B. Kolb U. Gies H. Yokoi T. Appl. Catal., B. 2023;325:122395–122407. [Google Scholar]
  9. Du J. Tang X. Huang C. Liu J. Shan Y. Zhang Y. Ren M. Wang Y. Shan W. Yu Y. Zheng A. He H. Appl. Catal., B. 2024;356:124258–124270. [Google Scholar]
  10. Paolucci C. Khurana I. Parekh A. A. Li S. Shih A. J. Li H. Di Iorio J. R. Albarracin-Caballero J. D. Yezerets A. Miller J. T. Delgass W. N. Ribeiro F. H. Schneider W. F. Gounder R. Science. 2017;357:898–903. doi: 10.1126/science.aan5630. [DOI] [PubMed] [Google Scholar]
  11. Liu W. Zhang J. Dong M. Ding Z. Wang Q. Ning F. He Q. Chen Y. Zhao Z. Li Z. Liu C. Sep. Purif. Technol. 2025;354:129201–129212. [Google Scholar]
  12. Paolucci C. Parekh A. A. Khurana I. Di Iorio J. R. Li H. Albarracin Caballero J. D. Shih A. J. Anggara T. Delgass W. N. Miller J. T. Ribeiro F. H. Gounder R. Schneider W. F. J. Am. Chem. Soc. 2016;138:6028–6048. doi: 10.1021/jacs.6b02651. [DOI] [PubMed] [Google Scholar]
  13. Sonoda T. Maruo T. Yamasaki Y. Tsunoji N. Takamitsu Y. Sadakane M. Sano T. J. Mater. Chem. A. 2015;3:857–865. [Google Scholar]
  14. Daya R. Trandal D. Dadi R. K. Li H. Joshi S. Y. Luo J. Kumar A. Yezerets A. Appl. Catal., B. 2021;297:120444–120460. [Google Scholar]
  15. Ransom R. Coote J. Moulton R. Gao F. Shantz D. F. Ind. Eng. Chem. Res. 2017;56:4350–4356. [Google Scholar]
  16. Tsunoji N. Shimono D. Tsuchiya K. Sadakane M. Sano T. Chem. Mater. 2019;32:60–74. [Google Scholar]
  17. Liu Z. Chokkalingam A. Miyagi S. Yoshioka M. Ishikawa T. Yamada H. Ohara K. Tsunoji N. Naraki Y. Sano T. Okubo T. Wakihara T. Phys. Chem. Chem. Phys. 2022;24:4136–4146. doi: 10.1039/d1cp03751j. [DOI] [PubMed] [Google Scholar]
  18. Li J. Gao M. Yan W. Yu J. Chem. Sci. 2023;14:1935–1959. doi: 10.1039/d2sc06010h. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Bae J. Dusselier M. Chem. Commun. 2023;59:852–867. doi: 10.1039/d2cc05370e. [DOI] [PubMed] [Google Scholar]
  20. Yoshioka T. Liu Z. Iyoki K. Chokkalingam A. Yonezawa Y. Hotta Y. Ohnishi R. Matsuo T. Yanaba Y. Ohara K. Takewaki T. Sano T. Okubo T. Wakihara T. React. Chem. Eng. 2021;6:74–81. [Google Scholar]
  21. Goel S. Zones S. I. Iglesia E. J. Am. Chem. Soc. 2014;136:15280–15290. doi: 10.1021/ja507956m. [DOI] [PubMed] [Google Scholar]
  22. Li C. Moliner M. Corma A. Angew. Chem., Int. Ed. 2018;57:15330–15353. doi: 10.1002/anie.201711422. [DOI] [PubMed] [Google Scholar]
  23. Goel S. Zones S. I. Iglesia E. Chem. Mater. 2015;27:2056–2066. [Google Scholar]
  24. Honda K. Itakura M. Matsuura Y. Onda A. Ide Y. Sadakane M. Sano T. J. Nanosci. Nanotechnol. 2013;13:3020–3026. doi: 10.1166/jnn.2013.7356. [DOI] [PubMed] [Google Scholar]
  25. Itabashi K. Kamimura Y. Iyoki K. Shimojima A. Okubo T. J. Am. Chem. Soc. 2012;134:11542–11549. doi: 10.1021/ja3022335. [DOI] [PubMed] [Google Scholar]
  26. Nakazawa N. Inagaki S. Kubota Y. Chem. Lett. 2016;45:919–921. [Google Scholar]
  27. Dusselier M. Schmidt J. E. Moulton R. Haymore B. Hellums M. Davis M. E. Chem. Mater. 2015;27:2695–2702. [Google Scholar]
  28. Zhang J. Chu Y. Deng F. Feng Z. Meng X. Xiao F.-S. Inorg. Chem. Front. 2020;7:2204–2211. [Google Scholar]
  29. Liu P. Yu C. Li W. Shen B. Microporous Mesoporous Mater. 2023;347:112341–112352. [Google Scholar]
  30. Chaves T. F. Pastore H. O. Cardoso D. Microporous Mesoporous Mater. 2012;161:67–75. [Google Scholar]
  31. Hu P. Iyoki K. Fujinuma H. Yu J. Yu S. Anand C. Yanaba Y. Okubo T. Wakihara T. Microporous Mesoporous Mater. 2022;330:111583–111592. [Google Scholar]
  32. Yao G. Lei J. Zhang X. Sun Z. Zheng S. Komarneni S. Mater. Res. Bull. 2018;107:132–138. [Google Scholar]
  33. Rodríguez-Fuentes G. Ruiz-Salvador A. R. Mir M. Picazo O. Quintana G. Delgado M. Microporous Mesoporous Mater. 1998;20:269–281. [Google Scholar]
  34. Purnomo C. W. Salim C. Hinode H. Microporous Mesoporous Mater. 2012;162:6–13. [Google Scholar]
  35. Fernández-Jiménez A. Palomo A. Microporous Mesoporous Mater. 2005;86:207–214. [Google Scholar]
  36. Tian Z. Li X. Cao Y. Wang Y. Peng X. J. Lumin. 2019;211:62–68. [Google Scholar]
  37. Xiong G. Yu Y. Feng Z.-c. Xin Q. Xiao F.-S. Li C. Microporous Mesoporous Mater. 2001;42:317–323. [Google Scholar]
  38. Zhang J. Chu Y. Liu X. Xu H. Meng X. Feng Z. Xiao F.-S. Chin. J. Catal. 2019;40:1854–1859. [Google Scholar]
  39. Chen C.-T. Iyoki K. Hu P. Yamada H. Ohara K. Sukenaga S. Ando M. Shibata H. Okubo T. Wakihara T. J. Am. Chem. Soc. 2021;143:10986–10997. doi: 10.1021/jacs.1c03351. [DOI] [PubMed] [Google Scholar]
  40. Asgar Pour Z. Alassmy Y. A. Sebakhy K. O. Crystals. 2023;13:959–972. [Google Scholar]
  41. Fu G. Yang R. Liang Y. Yi X. Li R. Yan N. Zheng A. Yu L. Yang X. Jiang J. Microporous Mesoporous Mater. 2021;320:111060–111073. [Google Scholar]
  42. Shen Y. Wang Z. Ge S. Wang L. Zhan W. Dai Q. Guo Y. Guo Y. Wang A. Appl. Catal., B. 2024;353:124094–124106. [Google Scholar]
  43. Wang A. Elena Azzoni M. Han J. Xie K. Olsson L. Chem. Eng. J. 2023;454:140040–1400052. [Google Scholar]
  44. Luo J. Gao F. Kamasamudram K. Currier N. Peden C. H. F. Yezerets A. J. Catal. 2017;348:291–299. [Google Scholar]
  45. Gao F. Washton N. M. Wang Y. Kollár M. Szanyi J. Peden C. H. F. J. Catal. 2015;331:25–38. [Google Scholar]
  46. Zhao Z. Yu R. Zhao R. Shi C. Gies H. Xiao F.-S. De Vos D. Yokoi T. Bao X. Kolb U. Feyen M. McGuire R. Maurer S. Moini A. Müller U. Zhang W. Appl. Catal., B. 2017;217:421–428. [Google Scholar]
  47. Kim J. Cho S. J. Kim D. H. ACS Catal. 2017;7:6070–6081. [Google Scholar]
  48. Chen Z. Fan C. Pang L. Ming S. Liu P. Li T. Appl. Catal., B. 2018;237:116–127. [Google Scholar]
  49. Ma Y. Cheng S. Wu X. Ma T. Liu L. Jin B. Liu M. Liu J. Ran R. Si Z. Weng D. J. Catal. 2022;405:199–211. [Google Scholar]
  50. Lin J. Hu X. Li Y. Shan W. Tan X. He H. Appl. Catal., B. 2023;331:122705–122715. [Google Scholar]
  51. Tarach K. A. Jabłońska M. Pyra K. Liebau M. Reiprich B. Gläser R. Góra-Marek K. Appl. Catal., B. 2021;284:119752–119768. [Google Scholar]
  52. Greenaway A. G. Marberger A. Thetford A. Lezcano-González I. Agote-Arán M. Nachtegaal M. Ferri D. Kröcher O. Catlow C. R. A. Beale A. M. Chem. Sci. 2020;11:447–455. doi: 10.1039/c9sc04905c. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Zhang J. Liang J. Peng H. Mi Y. Luo P. Xu H. He M. Wu P. Appl. Catal., B. 2021;292:120163–120175. [Google Scholar]
  54. Xie K. Woo J. Bernin D. Kumar A. Kamasamudram K. Olsson L. Appl. Catal., B. 2019;241:205–216. [Google Scholar]
  55. Zhao H. Zhao Y. Liu M. Li X. Ma Y. Yong X. Chen H. Li Y. Appl. Catal., B. 2019;252:230–239. [Google Scholar]
  56. Di Iorio J. R. Gounder R. Chem. Mater. 2016;28:2236–2247. [Google Scholar]
  57. Wang Y. Han J. Chen M. Lv W. Meng P. Gao W. Meng X. Fan W. Xu J. Yan W. Yu J. Angew. Chem., Int. Ed. 2023;62:e202306174. doi: 10.1002/anie.202306174. [DOI] [PubMed] [Google Scholar]
  58. Han J. Li J. Zhao W. Li L. Chen M. Ge X. Wang S. Liu Q. Mei D. Yu J. J. Am. Chem. Soc. 2024;146:7605–7615. doi: 10.1021/jacs.3c13855. [DOI] [PubMed] [Google Scholar]
  59. Zhang J. Liu W. Wang Q. Ning F. He Q. Li G. Liu C. Li Z. Peng H. Sep. Purif. Technol. 2024;346:127544–127555. [Google Scholar]
  60. Zhang R. McEwen J.-S. Kollár M. Gao F. Wang Y. Szanyi J. Peden C. H. F. ACS Catal. 2014;4:4093–4105. [Google Scholar]
  61. Szanyi J. Kwak J. H. Zhu H. Peden C. H. F. Phys. Chem. Chem. Phys. 2013;15:2368–2380. doi: 10.1039/c2cp43467a. [DOI] [PubMed] [Google Scholar]
  62. Lan T. Deng J. Zhang X. Wang F. Liu X. Cheng D. Zhang D. ACS Catal. 2022;12:3955–3964. [Google Scholar]
  63. Jiang B. Zhao S. Wang Y. Wenren Y. Zhu Z. Harding J. Zhang X. Tu X. Zhang X. Appl. Catal., B. 2021;286:119886–119897. [Google Scholar]
  64. Su W. Chang H. Peng Y. Zhang C. Li J. Environ. Sci. Technol. 2014;49:467–473. doi: 10.1021/es503430w. [DOI] [PubMed] [Google Scholar]
  65. Shen Y. Li T. Yang J. Wang A. Wang L. Zhan W. Guo Y. Guo Y. Chem. Eng. J. 2023;473:145275–145287. [Google Scholar]
  66. Shen Y. Dong W. Zhang L. Wang L. Chen B. Guo Y. Zhan W. Wang A. Ge C. Guo Y. Sep. Purif. Technol. 2024;330:125248–125262. [Google Scholar]
  67. Ming S. Pang L. Chen Z. Guo Y. Guo L. Liu Q. Liu P. Dong Y. Zhang S. Li T. Microporous Mesoporous Mater. 2020;303:110294–110305. [Google Scholar]
  68. Wang X. Xu Y. Qin M. Li Q. Liang X. J. Environ. Chem. Eng. 2023;11:110840–110850. [Google Scholar]
  69. Zhang Y. Peng Y. Li K. Liu S. Chen J. Li J. Gao F. Peden C. H. F. ACS Catal. 2019;9:6137–6145. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

SC-OLF-D6SC04592H-s001

Data Availability Statement

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6sc04592h.


Articles from Chemical Science are provided here courtesy of Royal Society of Chemistry

RESOURCES