Skip to main content
NCBI home page
Nature Communications logoLink to Nature Communications
. 2026 Aug 27;17:10288. doi: 10.1038/s41467-026-77159-y

Yolk-shell zeolite nanoreactors enable multi-poison resistance for NOx reduction

Yonglong Li 1,2,#, Guobo Li 1,#, Rongxing Li 1, Jian Ji 1, Xiaonan Hu 1, Fengbo Yu 1, Wenming Liu 3, Hongmei Li 4, Min Liu 4,5, Honggen Peng 1,3,✉
PMCID: PMC13619632  PMID: 42805984

Abstract

The practical application of selective catalytic reduction with ammonia (NH3-SCR) for NOx abatement from stationary sources is severely restricted by catalyst deactivation induced by SO2 and alkali metal poisons in flue gas. Although integrating transition metal oxides with zeolites is widely pursued to enhance catalyst poisoning tolerance, most reported strategies have inherent limitations. Conventional loading approaches deliver high active phase loading but leave active sites directly exposed to poisons, whereas confinement strategies offer shielding protection but typically suffer from limited active phase content. Herein, we demonstrate the concept of a zeolite-confined nanoreactor by constructing a yolk-shell structured catalyst, in which high-content (> 10 wt.%) MnFeOx nanoparticles are encapsulated within a hollow ZSM-5 zeolite shell. In situ extended X-ray absorption fine structure spectroscopy (EXAFS) and diffuse reflectance infrared Fourier-transform spectroscopy (DRIFTS), combined with density functional theory (DFT) calculations, reveal a synergistic dual-shielding mechanism: the ZSM-5 zeolite shell physically hinders SO2 from accessing and poisoning the inner active oxides core, while its intrinsic framework Brønsted acid sites chemically suppress SO2 adsorption and trap alkali metal species. Consequently, the catalyst maintains over 95% NOx conversion at 250 °C under co-poisoning by SO2 and potassium. This work establishes hollow zeolite-confined nanoreactors as an effective and versatile platform for developing durable NOx emission control catalysts applicable to industrially relevant multi-poisoning conditions.

Subject terms: Pollution remediation, Heterogeneous catalysis, Catalyst synthesis

Ammonia selective catalytic reduction for stationary NOx removal is limited by catalyst deactivation caused by SO2 and alkali metals in flue gas. Here, the authors develop a zeolite-confined yolk-shell nanoreactor that enables stable long-term operation under industrially relevant poisoning conditions.

Introduction

The abatement of nitrogen oxides (NOx) from stationary sources, including coal-fired power plants and waste incinerators, remains a critical environmental challenge owing to their adverse impacts on atmospheric quality and human health1–4. Selective catalytic reduction with ammonia (NH3-SCR) is currently one of the most efficient industrial technologies for NOx removal5–9. However, the service life of SCR catalysts is severely shortened in industrial flue gas, where NOx coexists with typical poisons such as sulfur dioxide (SO2) and alkali metal species (e.g., K-containing compounds)10,11. These poisons induce rapid and severe catalyst deactivation, and their combined poisoning effect is usually far more pronounced than that of any single poison, making the development of catalysts with simultaneous resistance to SO2 and alkali poisoning a paramount demand for sustainable industrial operation12–14.

Integrating transition metal oxides with zeolites has emerged as a promising strategy, as such composite systems combine the complementary functionalities of both phases15–18. Zeolites such as ZSM-5 offer a large specific surface area and abundant Brønsted acid sites capable of trapping alkali metal species19,20, whereas metal oxides (e.g., MnOx, FeOx, CeOx) provide the redox sites required for SCR21,22. However, conventional loading approaches (e.g., simple mixing, deposition, or impregnation) leave oxide nanoparticles exposed on the external surface of the zeolite supports17,18,23,24, rendering the active sites vulnerable to sulfation and alkali poisoning even with increased oxides loading. Even emerging confinement strategies, which encapsulate active oxides within zeolite shells, often face an inherent limitation25,26: high active phase loading tends to disrupt zeolite framework integrity and reduce shielding performance, whereas a well-preserved zeolite shell usually comes at the cost of limited active phase content.

To overcome this bottleneck, we propose a zeolite-confined nanoreactor design, in which a high-content (> 10 wt.%) active metal oxide core is encapsulated within a hollow porous acidic zeolite shell with a well-preserved zeolitic framework. In this architecture, the zeolite shell not only spatially confines the active metal oxide nanoparticles but also modulates their interfacial environment, permitting inward diffusion of reactants while intercepting poisons before they reach the inner active phase. This design thus simultaneously achieves high active phase content and robust resistance to multi-component poisoning. We demonstrate this concept by fabricating a yolk-shell structured MnFeOx@YS-ZSM-5 nanoreactor via an impregnation–dissolution–recrystallization (IDR) strategy, with MnFeOx nanoparticles encapsulated in a hollow ZSM-5 zeolite shell. Combined structural characterizations, operando spectroscopic studies, and theoretical calculations collectively uncover a synergistic dual-shielding mechanism: the zeolite shell provides physical isolation to hinder SO2 access to the inner active MnFeOx core, while its intrinsic framework Brønsted acid sites offer chemical protection by suppressing SO2 adsorption and trapping alkali metal species. The catalyst exhibits outstanding NH3-SCR activity and stability under simultaneous exposure to SO2 and alkali metal species, and the synthetic strategy is extendable to other mixed-metal oxide systems. This work establishes hollow zeolite nanoreactors as a promising platform for developing durable NOx emission control catalysts, offering a viable route toward long-life industrial NH3-SCR catalysts.

Results

Formation of the yolk-shell structured nanoreactor

The yolk-shell structured MnFeOx@YS-ZSM-5 nanoreactor was synthesized via an IDR strategy (Fig. 1a). In this architecture, crystalline ZSM-5 zeolite forms a hollow shell, while MnFeOx nanoparticles are confined as an inner yolk. Bulk ZSM-5 zeolite is first impregnated with Mn and Fe precursors and calcined to generate supported MnFeOx/ZSM-5, in which metal oxides are dispersed on the external surface of zeolite. Subsequent hydrothermal treatment in tetrapropylammonium hydroxide (TPAOH) induces selective internal dissolution of the zeolite and simultaneous recrystallization at the exterior, ultimately yielding a hollow ZSM-5 zeolite shell that fully encapsulates the MnFeOx nanoparticles. This selective etching originates from the intrinsic compositional gradient of ZSM-5 zeolite crystals, and bulky TPA+ cations direct the outward recrystallization to maintain zeolitic framework integrity27.

Fig. 1. Morphological and structural characterization.

Fig. 1

Open in a new tab

a Schematic illustration of the synthesis of the yolk-shell structured MnFeOx@YS-ZSM-5 nanoreactor via the IDR method. b, c TEM images, d HAADF-STEM image, e, f EDS elemental mappings, and iDPC-STEM images of g the yolk-shell structured MnFeOx@YS-ZSM-5 nanoreactor and h the hollow ZSM-5 zeolite shell. i N2 adsorption–desorption isotherms of ZSM-5, MnFeOx/ZSM-5, and MnFeOx@YS-ZSM-5. j Fe 2p and Mn 2p XPS spectra of MnFeOx/ZSM-5 and MnFeOx@YS-ZSM-5.

Electron microscopy reveals a clear structural evolution during this process. Pristine ZSM-5 zeolites consist of well-defined blocky crystals with sharp edges and an average particle size of ~145 nm (Figs. S1 and S2). After impregnation, MnFeOx/ZSM-5 retains the original morphology (Fig. S3), but transmission electron microscopy (TEM) and energy-dispersive X-ray spectroscopy (EDS; Fig. S4) mappings confirm that Mn and Fe species are distributed on the external surface of zeolite. In contrast, the final MnFeOx@YS-ZSM-5 nanoreactor exhibits a distinct yolk-shell architecture. Scanning electron microscopy (SEM) shows hollow cubic particles (Fig. S5), while TEM (Figs. 1b, c and  S6) and high-angle annular dark-field scanning TEM (HAADF-STEM) image (Fig. 1d) unambiguously demonstrate that MnFeOx nanoparticles are confined within the internal cavity of hollow ZSM-5 zeolite, and almost no metal oxide species are detectable on the external surface. The nanoreactors retain an average size of ~145 nm, with shell thicknesses ranging from ~10 to ~30 nm. Notably, the MnFeOx nanoparticles are preferentially located near the inner shell surface rather than at the geometric center. EDS elemental mappings (Fig. 1e, f) and line-scan profiles (Fig. S7) confirm that Mn and Fe are spatially confined inside the zeolite shell rather than on the external surface. The integrated differential phase-contrast STEM (iDPC-STEM) (Figs. 1g and  S8) results directly resolve the atomic-scale MFI framework of the ZSM-5 zeolite shell while simultaneously visualizing the encapsulated MnFeOx yolk. The corresponding Fourier-transform pattern (Fig. S8f) reveals a well-ordered diffraction spot array, confirming the single-crystalline nature of the zeolite shell. Given that the ZSM-5 zeolite micropores (~ 0.54 nm, Fig. 1h) are much smaller than the MnFeOx nanoparticles (~ 8.73 nm), migration of the active phase into the zeolite channels is effectively prohibited.

Nitrogen adsorption–desorption results (Figs. 1i and  S9) confirm the hollow structure of MnFeOx@YS-ZSM-5 nanoreactor, as evidenced by its distinctive H2-type hysteresis loop and increased total pore volume (Table S1)28. XRD analysis (Fig. S10) verifies that the ZSM-5 zeolite retains its MFI crystal structure after the encapsulation process. Meanwhile, the obvious attenuation of Fe 2p and Mn 2p signals in the X-ray photoelectron spectroscopy (XPS) spectra (Fig. 1j) of MnFeOx@YS-ZSM-5 provides further evidence for the successful confinement of the MnFeOx metal oxides core within the zeolite shell, as the Fe 2p and Mn 2p signal intensity of MnFeOx@YS-ZSM-5 nanoreactor is markedly reduced compared with that of the supported MnFeOx/ZSM-5 sample. Time-dependent hydrothermal treatments reveal the mechanism underlying yolk-shell formation. Progressive development of internal voids is observed with increasing treatment duration, as evidenced by microscopy (Figs. S11–S17) and N2 adsorption–desorption results (Fig. S18), and XRD patterns (Fig. S19) confirm that the MFI framework is preserved throughout the process. XPS results (Fig. S20) show pronounced Si enrichment and Al depletion on the external surface after TPAOH hydrothermal treatment, indicative of preferential internal Si dissolution and outward recrystallization of silicate species. Based on the characterizations described above, a formation mechanism of the yolk-shell structure is proposed in Fig. S21. The universality of the IDR synthetic strategy is further demonstrated through the successful synthesis of hollow ZSM-5 zeolite encapsulating MnCoOx or MnCeOx (Figs. S22 and S23). Both systems display well-defined yolk-shell architectures, as evidenced by N2 adsorption–desorption profiles (Figs. S24 and S25) and XRD patterns (Fig. S26), confirming that the zeolite crystallinity remains intact regardless of the metal oxide precursors.

NH3-SCR performance and multi-poison resistance

To clarify the catalytic advantages of the yolk-shell structured MnFeOx@YS-ZSM-5 nanoreactor in NH3-SCR (Fig. 2a), comparative catalysts (MnFeOx, MnFeOx/ZSM-5, and MnFeOx/S-1; the S-1 support is an aluminum-free zeolite with much weaker acidity than ZSM-5 zeolite) were prepared (Figs. S27–S32 and Table S1) and then evaluated. Pure MnFeOx shows high activity only below 200 °C, but suffers rapid deactivation because of excessive NH3 oxidation (Fig. S33) and poor N2 selectivity (Fig. S34). Coupling MnFeOx with ZSM-5 zeolite broadens the operational window. Notably, MnFeOx@YS-ZSM-5 nanoreactor delivers outstanding NH3-SCR performance, maintaining over 80% NOx conversion across a wide temperature range of 220–500 °C, while MnFeOx/S-1 and MnFeOx/ZSM-5 display lower medium-high-temperature performance. Samples treated hydrothermally in TPAOH for different durations (Fig. S35) all retain similarly active wide temperature windows. Compared with reported MnFeOx-based catalysts (Fig. S36 and Table S2), MnFeOx@YS-ZSM-5 nanoreactor exhibits a remarkably broader operational temperature window. Kinetic analysis (Fig. S37 and Tables S3–S5) shows that all oxide-zeolite composites possess substantially lower apparent activation energies than pure MnFeOx.

Fig. 2. Catalytic NH3-SCR performance.

Fig. 2

Open in a new tab

a NOx conversion as a function of temperature in NH3-SCR reaction over MnFeOx, MnFeOx/ZSM-5, MnFeOx/S-1, and MnFeOx@YS-ZSM-5 catalysts. b SO2 resistance performance over pure MnFeOx, supported MnFeOx/ZSM-5, and yolk-shell structured MnFeOx@YS-ZSM-5 nanoreactor at 246 °C. c, d NOx conversion as a function of temperature in NH3-SCR reaction over the K-poisoned catalysts. Reaction conditions: 500 ppm NO, 500 ppm NH3, 5 vol.% O2, 5 vol.% H2O, 50 ppm SO2 (when used), balanced with N2, WHSV = 60,000 mL gcat.−1 h−1.

The SO2 tolerance was evaluated at 246 °C with 50 ppm SO2 (Fig. 2b). The supported MnFeOx/ZSM-5 catalyst undergoes rapid deactivation, with NOₓ conversion dropping from 98% to 33% within 0.5 h and stabilizing at about 39% after SO2 is cut off. Pure MnFeOx shows only transient resistance before severe deterioration. In striking contrast, yolk-shell structured MnFeOx@YS-ZSM-5 nanoreactor stabilizes at ~83% NOx conversion and rapidly recovers to ~90% once SO2 is removed, demonstrating largely reversible deactivation. Pre-sulfidation experiments (Fig. S38) further confirm the better SO2 resistance of the yolk-shell structured nanoreactor. To distinguish the contributions of hollow morphology from encapsulation, a control catalyst consisting of MnFeOx supported on hollow ZSM-5 zeolite (MnFeOx/Hollow-ZSM-5) was synthesized (Figs. S39–S40). Although this catalyst exhibits comparable NH3-SCR activity (Fig. S41), its SO2 tolerance is markedly inferior (Fig. S42), establishing that the superior durability of MnFeOx@YS-ZSM-5 nanoreactor arises from the shielding effect of the intact zeolite shell rather than from the hollow structure alone. Moreover, MnFeOx@YS-ZSM-5 also shows outstanding alkali resistance. After K poisoning, MnFeOx@YS-ZSM-5 maintains high NOx conversion across the mid-temperature range with only minor losses at higher temperatures (Fig. 2c). In contrast, pure MnFeOx suffers severe deactivation (Fig. 2d), consistent with alkali-induced neutralization of acid sites. MnFeOx/ZSM-5 (Fig. S43) preserves moderate activity at medium-low temperatures but exhibits pronounced high-temperature deterioration, while MnFeOx/S-1 undergoes deactivation over nearly the entire temperature range (Fig. S44), reflecting its limited acidity. Under simultaneous SO2 and K poisoning, MnFeOx@YS-ZSM-5 nanoreactor sustains over 95% NOx conversion at 250 °C (Fig. S45), underscoring its superior multi-poison resistance.

Acidity-redox balance and structural stability

Given the crucial roles of surface acidity in NH3-SCR29–31, NH3 temperature-programmed desorption (NH3-TPD; Figs. 3a,  S46, and Table S6) was performed. The results demonstrate that pure MnFeOx possesses the weakest acidity, whereas all MnFeOx-zeolite composites feature enhanced total acidity. MnFeOx@YS-ZSM-5 nanoreactor delivers the highest acidity and the largest proportion of medium-strong acid sites. These medium-strong acid sites are critical for NH3 adsorption and the dual SCR pathways involving both Eley–Rideal and Langmuir–Hinshelwood mechanisms21, and correlate well with the superior activity and broad operating window of MnFeOx@YS-ZSM-5 as presented in Fig. 2a. Pyridine infrared (Py-IR; Fig. S47) experiments confirm abundant Brønsted and Lewis acid sites in zeolite-containing samples and negligible acidity in pure MnFeOx. After K poisoning (Figs. S47, S48 and Table S6), MnFeOx@YS-ZSM-5 still possesses abundant Brønsted acid sites and retains the highest total acidity as well as the largest proportion of surviving medium-strong acid sites, verifying that zeolite encapsulation buffers alkali attack. Hydrogen temperature-programmed reduction (H2-TPR; Fig. 3b) further reveals how zeolite confinement modulates redox property. For MnFeOx-zeolite composites, the low-temperature reduction peak (< 500  °C, α region) corresponds to Mn4+ reduction32, while the higher-temperature peak (β region) is attributed to Fe2O3 reduction33. MnFeOx@YS-ZSM-5 shows higher reduction temperatures in the α region, indicating weaker oxidative capacity. Pure MnFeOx possesses the strongest redox capability, consistent with its highest low-temperature activity but narrow active temperature window (Fig. 2a). Zeolite coupling therefore mitigates NH3 over-oxidation, and synergizes with enhanced acidity to improve high-temperature performance. The H2-TPR results of K-poisoned catalysts are shown in Fig. S49. Compared to the fresh catalysts, the reduction peak of K-MnFeOx shifts most significantly toward higher temperatures, whereas the reduction peak temperatures of other catalysts show negligible changes after K poisoning, indicating preserved oxidative capacity. K specie exerts a more severe poisoning effect on pure MnFeOx, impairing its oxidative capability. Combined with the significant loss of surface acid sites in K-MnFeOx observed in NH3-TPD results (Fig. S48 and Table S6), these two factors collectively caused the severe decline in catalytic activity and loss of the active temperature window for pure MnFeOx after K poisoning (Fig. 2d). The well-preserved Brønsted acid sites, together with the protected internal redox phase, maintain the acid-redox synergistic effect that dominates SCR activity, thus mitigating activity decline under K exposure over the yolk-shell structured MnFeOx@YS-ZSM-5 catalyst (Fig. 2c). XPS analysis (Figs. S50,  3c and Table S7) reveals that pure MnFeOx has the highest Mn4+ fraction (39.6%)22, which facilitates NO oxidation to NO2 and promotes low-temperature “fast SCR”34. Zeolite coupling decreases the Mn4+ proportion, with MnFeOx@YS-ZSM-5 nanoreactor showing the lowest (29.4%; Fig. 3c), consistent with its weakened oxidation property. XPS characterization was also performed on the catalysts after the SO2 poisoning resistance stability test at 246 °C (Fig. 2b). After SO2 exposure, the Mn4+ content of pure MnFeOx decreases sharply (Fig. S50), whereas MnFeOx@YS-ZSM-5 (Fig. 3c) shows only a slight reduction (29.4% → 28.3%). This decrease arises from electron transfer from SO2 to Mn4+35,36, leading to sulfate formation and active sites deactivation. Notably, MnFeOx@YS-ZSM-5 nanoreactor exhibits the lowest surface sulfur content (Table S7), further corroborated by SO2-TPD (Fig. S51). After K poisoning (Fig. S52), the Mn4+ fraction decreases markedly in pure MnFeOx but remains nearly unchanged in coupled catalysts. Accordingly, pure MnFeOx suffers substantial loss in low‑temperature activity (Fig. 2d), whereas coupled catalysts retain their low‑temperature performance (Figs. 2c and S43).

Fig. 3. Physicochemical characterization and analysis.

Fig. 3

Open in a new tab

a NH3-TPD curves and b H2-TPR profiles for MnFeOx@YS-ZSM-5 and related samples. c Mn 2p XPS spectra for fresh and corresponding SO2-poisoned samples. d, e Mn K-edge XANES spectra and f FTs of Mn K-edge EXAFS oscillations for fresh and corresponding SO2-poisoned samples. g–i WT plots of EXAFS spectra for MnFeOx, MnFeOx/ZSM-5, and MnFeOx@YS-ZSM-5.

Mn K-edge X-ray absorption near-edge structure (XANES) spectra (Fig. 3d) confirm that pure MnFeOx has the highest Mn oxidation state. Zeolite coupling shifts the absorption edge to lower energies, with MnFeOx/ZSM-5 exhibiting a higher Mn valence state than MnFeOx@YS-ZSM-5, consistent with XPS results. After SO2 poisoning (Fig. 2b), both MnFeOx/ZSM-5 and MnFeOx@YS-ZSM-5 exhibit decreased Mn valence (Fig. 3e). However, the magnitude of this shift is smaller for S-MnFeOx@YS-ZSM-5 than for S-MnFeOx/ZSM-5, indicating a less pronounced decrease in the average Mn valence state in S-MnFeOx@YS-ZSM-5 following SO2 exposure. The observations further confirm that the zeolite shell in the yolk-shell structured nanoreactor effectively shields Mn active sites from SO2 contact and reaction, thereby suppressing electron transfer from S to Mn and reducing manganese sulfate species formation. Fourier-transforms (FTs) of Mn K-edge extended X-ray absorption fine structure (EXAFS; Fig. 3f) show shorter Mn–O bond distances in the composite catalysts than in pure MnFeOx, indicative of stronger oxide-zeolite interactions. Wavelet transform (WT) analysis (Fig. 3g–i) identifies Mn–O and Mn–Mn/Mn–Fe coordination features37, confirming the structural integrity of the encapsulated MnFeOx phase.

Reaction pathway and dual-shielding mechanism

To elucidate the NH3-SCR reaction pathways, in situ diffuse reflectance infrared Fourier-transform spectroscopy (DRIFTS) was conducted at 200 °C. For MnFeOx@YS-ZSM-5 nanoreactor, NH3 adsorption (Fig. 4a, a1) reveals significantly stronger bands for both Lewis acid-bound NH3 (NH3-L; 1195 cm−1)38 and Brønsted acid-bound NH4+ species (NH4+-B; 1466 and 1735 cm−1)39,40 compared to pure MnFeOx (Fig. S53), confirming that zeolite encapsulation introduces abundant Lewis and Brønsted acid sites, thereby enhancing NH3 adsorption, consistent with the NH3-TPD results. Upon introducing NO + O2 to pre-adsorbed NH3, the NH3-L/NH4+-B bands on MnFeOx@YS-ZSM-5 decreased gradually without the formation of surface nitrate or nitrite intermediates over 40 min (Fig. 4b, b1). In contrast, pure MnFeOx (Fig. S54) showed rapid consumption of NH3-L species41 and immediate accumulation of monodentate nitrate42 and bridging nitrate43. Correspondingly, NO + O2 adsorption experiments on MnFeOx@YS-ZSM-5 (Fig. 4c) produced virtually no surface nitrogen species over 40 min, except for a weak transient NO2 band44 at 1633 cm−1 that disappears upon N2 purging, unlike pure MnFeOx, which formed substantial nitrates (Fig. S55). These results demonstrate that the NH3-SCR reaction over MnFeOx@YS-ZSM-5 predominantly follows the Eley–Rideal mechanism, involving gaseous NO reacting directly with adsorbed NH3 species. Conversely, pure MnFeOx primarily proceeds via a Langmuir–Hinshelwood pathway involving surface nitrate intermediates (Fig. S56), and this mechanistic divergence is further supported by NO-TPD observations (Fig. S57). The substantially stronger oxidative capability of pure MnFeOx, as evidenced by H2-TPR, XPS, and EXAFS analyses, facilitates NO oxidation to NO2 and enables the “fast SCR” pathway at low temperatures, yet simultaneously promotes nitrate accumulation and the formation of undesired N2O. In contrast, zeolite encapsulation weakens NO adsorption and oxidation, suppresses nitrate/nitrite formation, inhibits NH3 over-oxidation, and thus enhances N2 selectivity over a wide active temperature range. Density functional theory (DFT) calculations further support these findings. As shown in Figs. S58–S59 and  4d–g, NO adsorption on Mn and Fe sites of pure MnFeOx is strong, with adsorption energies (Eads) of −2.067 eV (Mn, Fig. 4d) and −4.100 eV (Fe, Fig. 4e), respectively. In contrast, NO adsorption on MnFeOx@YS-ZSM-5 is significantly weakened, with corresponding Eads of −1.686 eV (Mn, Fig. 4f) and −2.971 eV (Fe, Fig. 4g), consistent with the experimentally observed suppression of surface nitrate formation.

Fig. 4. Study on the mechanism of catalytic reaction.

Fig. 4

Open in a new tab

a, a1 In situ DRIFT spectra of NH3 adsorption on MnFeOx@YS-ZSM-5 at 200 °C. b, b1 In situ DRIFT spectra of pre-adsorbed NH3 species reacted with NO + O2 on MnFeOx@YS-ZSM-5 at 200 °C. c In situ DRIFT spectra of NO + O2 adsorption on MnFeOx@YS-ZSM-5 at 200 °C. Optimized adsorption configurations of NO molecules on d, e MnFeOx and f, g MnFeOx@YS-ZSM-5.

The adsorption behavior of SO2 was further examined by DFT calculations (Fig. 5a–d). Relative to pure MnFeOx, SO2 adsorption on MnFeOx@YS-ZSM-5 is substantially weakened, accompanied by elongated S–Mn and S–Fe bond lengths. Remarkably, SO2 adsorption on Fe sites becomes thermodynamically unfavorable (Eads = +0.383 eV; Fig. 5d), indicating that the zeolite shell effectively suppresses SO2 binding on active sites through confinement effect. Moreover, SO2 adsorption on Brønsted acid sites of ZSM-5 zeolite is also unfavorable (Eads =   + 0.496 eV; Fig. S60), suggesting electrostatic repulsion that further hinders sulfur deposition.

Fig. 5. Structure–activity relationship analysis.

Fig. 5

Open in a new tab

Optimized adsorption configurations of SO2 molecules on a, b MnFeOx and c, d MnFeOx@YS-ZSM-5. e In situ Mn K-edge XANES spectra for transient reaction over yolk-shell structured MnFeOx@YS-ZSM-5 nanoreactor after treatment for 0, 2, and 3 h at 246 °C in a 250 ppm SO2 + 5 vol.% O2 atmosphere. WT plots of EXAFS spectra for yolk-shell structured MnFeOx@YS-ZSM-5 nanoreactor after treatment at 246 °C in an atmosphere of 250 ppm SO2 + 5 vol.% O2 for f 0 h, g 2 h, and h 3 h (R and CN denote the average interatomic distances between Mn and the O atoms in the first coordination shell and the Mn–O coordination number, respectively). i Schematic diagram illustrating the enhanced co-poisoning resistance to SO2 and K of yolk-shell structured MnFeOx@YS-ZSM-5 nanoreactor via synergistic dual physical and chemical shielding effects.

From a mechanistic perspective, the SO2 poisoning of Mn-based oxide catalysts is intrinsically linked to their strong oxidative capability, which promotes SO2 oxidation to SO3 and subsequent sulfate formation at Mn4+ sites45. For unprotected MnFeOx, DFT results confirm that SO2 binds strongly to Mn and Fe sites, at which it is oxidatively activated to SO3 by surface Mn4+ species. This process induces electron transfer (Mn4+ + SO2 → Mn3+ + SO3) and leads to irreversible sulfate formation and progressive deactivation, consistent with the decreased Mn4+ content and accumulated surface sulfur species observed after SO2 exposure. Control experiments with MnFeOx supported on hollow ZSM-5 zeolite (MnFeOx/Hollow-ZSM-5) demonstrate that the hollow zeolite shell alone provides only limited protection, whereas the yolk-shell architecture with full encapsulation delivers dramatically superior resistance to SO2 chemical poisoning, confirming that physical isolation by the intact zeolite shell is the dominant protective factor. Although the zeolite micropores (~ 0.54 nm) are larger than SO2 molecules (~ 0.36 nm) and cannot completely block SO2 diffusion, the intact shell substantially reduces the concentration and residence time of SO2 near the core. Combined with the electrostatic repulsion originating from framework Brønsted acid sites validated by DFT, the synergistic physical and chemical shielding substantially alleviates sulfation of the inner MnFeOx active phase.

To directly probe the structural stability of active sites under SO2 exposure, in situ EXAFS measurements were performed during transient SO2 + O2 treatment at 246 °C (consistent with the temperature used in the SO2 resistance test shown in Fig. 2b). As shown in Fig. 5e, the Mn K-edge XANES profiles of yolk-shell structured MnFeOx@YS-ZSM-5 nanoreactor remain essentially unchanged after 2–3 h of SO2 + O2 exposure. Correspondingly, the Mn–O peak position in Fourier-transform EXAFS spectra (Fig. S61) and the Mn–O coordination number (~ 5.0), as well as the bond distance (~ 1.91 Å; Fig. 5f–h) show negligible variation, confirming that SO2 does not significantly interact with the encapsulated MnFeOx active sites. This mechanistic understanding reveals that the yolk-shell architecture moderates the intrinsic redox properties of the MnFeOx metal oxides through confinement, which suppresses excessive NH3 oxidation and avoids SO2 activation, while the zeolite shell provides physical and chemical shielding against SO2 chemical attack. Consequently, the catalyst partially sacrifices low-temperature redox activity and achieves substantially improved durability under SO2-containing flue gas conditions, representing a reasonable activity-stability trade-off.

Taken together, the above results establish a dual-shielding mechanism for simultaneous SO2 and alkali resistance of the yolk-shell structured MnFeOx@YS-ZSM-5 nanoreactor, as schematically illustrated in Fig. 5i. For SO2, the hollow ZSM-5 zeolite shell acts as a physical barrier to isolate the inner MnFeOx core from direct sulfur attack, while intrinsic framework Brønsted acid sites further suppress SO2 uptake via electrostatic repulsion. For alkali metal species such as K, consistent with the acidity characterization results, framework Brønsted acid sites on the zeolite shell serve as preferential trapping sites to neutralize alkali metal species before they diffuse inward. When SO2 and K coexist in the reaction system, this dual-shielding architecture further mediates spatially separated deactivation pathways for the two poisons. K is mainly trapped and neutralized by framework Brønsted acid sites on the shell with only minor impact on the encapsulated redox active phase, while SO2 permeates inward mildly under the combined restriction of physical diffusion barrier and electrostatic repulsion, causing limited sulfation of the inner oxides core. The two poisons act on distinct spatial regions without forming stable metal sulfate species directly on the active phase, thus avoiding synergistically aggravated deactivation and endowing the catalyst with stable performance under multi-component poisoning conditions. Such a synergistic dual-shielding effect preserves the structural integrity of active phases and maintains the dominant Eley–Rideal reaction pathway. It therefore underpins the outstanding NH3-SCR activity, high N2 selectivity, and long-term operational durability of the yolk-shell structured nanoreactor under harsh multi-poisoning flue gas environments.

Discussion

In summary, this work addresses the long-standing challenge of simultaneous SO2 and alkali metal poisoning in NH3-SCR by demonstrating a zeolite-confined nanoreactor design based on a yolk-shell MnFeOx@YS-ZSM-5 architecture. While the integration of metal oxides and zeolites has been widely investigated to improve catalyst poisoning tolerance, most conventional strategies relying on mixing, deposition or impregnation suffer from direct exposure of active oxides. Even emerging confinement strategies often struggle to balance high active phase content and intact zeolite framework integrity, failing to achieve both high intrinsic activity and robust poisoning resistance. We overcame this bottleneck via an impregnation–dissolution–recrystallization (IDR) strategy, which enables the encapsulation of high-content (> 10 wt.%) MnFeOx nanoparticles within a hollow ZSM-5 zeolite shell. The superior co-poisoning resistance of the catalyst originates from a synergistic dual-shielding mechanism. The ZSM-5 zeolite shell acts as a physical barrier, hindering SO2 from accessing and poisoning the inner active oxides core. Meanwhile, the intrinsic framework Brønsted acid sites suppress SO2 adsorption and capture alkali metal species, further preventing poisons from reaching the inner active phase. The combination of physical isolation and chemical trapping creates a synergistic protective effect, which is particularly effective against the combined deactivation induced by coexisting SO2 and alkali metal species. Benefiting from this dual-shielding mechanism, the optimal catalyst maintains over 95% NOx conversion at 250 °C under co-exposure to SO2 and potassium. The synthetic strategy further exhibits good generality and can be successfully extended to encapsulate other mixed metal oxides (e.g., MnCoOx, MnCeOx) within hollow zeolite shells without losing structural integrity. This work establishes the yolk-shell zeolite-confined nanoreactor as a versatile platform for designing durable NH3-SCR catalysts. The proposed dual-shielding design principle offers a viable solution to the challenge of multi-poison tolerance in industrial flue gas, and can be further extended to other high-temperature environmental catalytic processes requiring long-term operational stability.

Methods

Catalyst synthesis

The yolk-shell structured MnFeOx@YS-ZSM-5 catalyst was synthesized via an impregnation–dissolution–recrystallization (IDR) strategy. For comparison, MnFeOx metal oxides supported on the outer surface of ZSM-5 zeolite (MnFeOx/ZSM-5) was prepared by the conventional impregnation method. To reveal the formation mechanism of the yolk-shell structured MnFeOx@YS-ZSM-5 nanoreactor, the samples crystallized at each step were collected. Furthermore, to verify the universality of the IDR method, hollow ZSM-5 zeolite confining other metal oxides was also prepared. The detailed preparation procedures are discussed in the Supplementary Information.

Characterization method

Various testing methods—including X-ray diffraction (XRD), N2 adsorption–desorption analysis, X-ray photoelectron spectroscopy (XPS), pyridine infrared (Py-IR), NH3 temperature-programmed desorption (NH3-TPD), scanning electron microscopy (SEM), transmission electron microscopy (TEM), high resolution TEM (HRTEM), energy-dispersive X-ray spectroscopy (EDS) elemental mapping, aberration-corrected TEM (AC-TEM), AC-EDS elemental mapping, X-ray absorption near-edge structure (XANES), in situ extended X-ray absorption fine structure spectroscopy (in situ EXAFS), density functional theory (DFT) calculations, and the in situ diffuse reflectance infrared Fourier-transform spectroscopy (DRIFTS)—were adopted to characterize the physical and chemical properties and elucidate the plausible reaction mechanism over the yolk-shell structured MnFeOx@YS-ZSM-5 nanoreactor and related catalysts. The detailed characterization procedures are described in the Supplementary Information.

Activity and kinetic test

The detailed NH3-SCR activity and kinetic tests are also presented in the Supplementary Information.

DFT calculation

DFT calculations were performed using the Vienna Ab-initio Simulation Package (VASP 6.3.3). Computational parameters and slab model configurations are provided in the Supplementary Information.

Supplementary information

Supplementary Information (8MB, pdf)
Transparent Peer Review file (670.2KB, pdf)

Source data

Source data (9.9MB, xlsx)

Acknowledgements

The authors thank all supporting institutions and colleagues for their assistance with this research.

Author contributions

H.P.: conceptualization, methodology, project administration, funding acquisition, supervision, and writing—review and editing. Y.L.: conceptualization, methodology, visualization, data curation, writing-original draft, and writing—review and editing. G.L.: methodology, formal analysis, and DFT calculations. M.L.: methodology, investigation, and writing—review and editing. R.L., J.J., X.H., F.Y., H.L., and W.L.: methodology and investigation.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This work was supported by the National Natural Science Foundation of China (22476078, 22376222, 22276086, 22306086 and 22306117), National key R&D Program of China (2023YFA1508400 and 2024YFC3712104), the Natural Science Foundation of Jiangxi Province (20243BCE51169, 20232BCJ22003, 20232BAB213028, and 20242BAB25144), Science and Technology Innovation Program of Hunan Province (2023RC1012), Central South University Research Program of Advanced Interdisciplinary Studies (Grant 2023QYJC012) and the Natural Science Foundation of Chongqing (CSTB2023NSCQ-MSX0950 and CSTB2024NSCQ-MSX1295).

Data availability

All data presented in this study are included in the article and Supplementary Information. The data are available from the corresponding authors upon request. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

These authors contributed equally: Yonglong Li, Guobo Li.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-77159-y.

References

  • 1.Li, Z. et al. Unexpected redox role of WO3 in V2O5-WO3/TiO2 catalysts for selective reduction of NO by forming V–W dinuclear sites. Angew. Chem. Int. Ed.64, e202501957 (2025). [DOI] [PubMed] [Google Scholar]
  • 2.Shan, Y. et al. Strikingly distinctive NH3-SCR behavior over Cu-SSZ-13 in the presence of NO2. Nat. Commun.13, 4606 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Liu, G. et al. Atomic-level design of acid-base pairs in oxides for selective catalytic reduction of nitrogen oxides with ammonia. Angew. Chem. Int. Ed.64, e202509362 (2025). [DOI] [PubMed] [Google Scholar]
  • 4.Paolucci, C. et al. Dynamic multinuclear sites formed by mobilized copper ions in NOx selective catalytic reduction. Science357, 898–903 (2017). [DOI] [PubMed] [Google Scholar]
  • 5.Chen, M. et al. Unveiling secondary-ion-promoted catalytic properties of Cu-SSZ-13 zeolites for selective catalytic reduction of NOx. J. Am. Chem. Soc.144, 12816–12824 (2022). [DOI] [PubMed] [Google Scholar]
  • 6.Wu, Y. et al. Rate controlling in low-temperature standard NH3-SCR: implications from operando EPR spectroscopy and reaction kinetics. J. Am. Chem. Soc.144, 9734–9746 (2022). [DOI] [PubMed] [Google Scholar]
  • 7.Han, J. et al. Cu-OFF/ERI zeolite: intergrowth structure synergistically boosting selective catalytic reduction of NOx with NH3. J. Am. Chem. Soc.146, 7605–7615 (2024). [DOI] [PubMed] [Google Scholar]
  • 8.An, D. et al. Water-driven surface lattice oxygen activation in MnO2 for promoted low-temperature NH3–SCR. Environ. Sci. Technol. 58, 16974–16983 (2024). [DOI] [PubMed]
  • 9.Wang, R. et al. Arsenic impurity in flue gas significantly inhibiting N2O byproduct formation on the chromium-poisoned V2O5–WO3/TiO2 catalyst in the NH3–SCR reaction. Environ. Sci. Technol.59, 23602–23611 (2025). [DOI] [PubMed] [Google Scholar]
  • 10.Shen, Z. et al. Ultrahighly alkali-tolerant NOx reduction over self-adaptive CePO4/FePO4 catalysts. Environ. Sci. Technol.57, 14472–14481 (2023). [DOI] [PubMed] [Google Scholar]
  • 11.Han, L. et al. Environmental catalysis for NOx reduction by manipulating the dynamic coordination environment of active sites. Environ. Sci. Technol.59, 2306–2316 (2025). [DOI] [PubMed] [Google Scholar]
  • 12.Shen, Z. et al. Alkali and heavy metal copoisoning resistant catalytic reduction of NOx via liberating Lewis acid sites. Environ. Sci. Technol.56, 5141–5149 (2022). [DOI] [PubMed] [Google Scholar]
  • 13.Chen, Y. et al. Challenges and perspectives of environmental catalysis for NOx reduction. JACS Au. 4, 2767–2791 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Zhang, J. et al. Insight into the alkali resistance mechanism of FeMoTiOx catalysts for NH3 selective catalytic reduction of NO: self-defense effects of MoOx for alkali capture. Environ. Sci. Technol.58, 4145–4154 (2024). [DOI] [PubMed] [Google Scholar]
  • 15.Kim, H. S., Lee, H., Park, H., Song, I. & Kim, D. H. Revealing the two distinctive roles of HY zeolite in enhancing the activity and durability of manganese oxide-zeolite hybrid catalysts for low-temperature NH3-SCR. Appl. Catal. B: Environ. Energy355, 124199 (2024). [Google Scholar]
  • 16.Deng, J. et al. Catalyst deactivation in the abatement of atmospheric pollutants: origin, resistance, and regeneration. Chem. Rev.125, 11260–11357 (2025). [DOI] [PubMed] [Google Scholar]
  • 17.Zhang, Z. et al. NOx reconstruction triggered by zeolite reverses alkali metal poisoning in NO selective catalytic reduction. Environ. Sci. Technol.59, 8254–8262 (2025). [DOI] [PubMed] [Google Scholar]
  • 18.Song, I. et al. Simple physical mixing of zeolite prevents sulfur deactivation of vanadia catalysts for NOx removal. Nat. Commun.12, 901 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Deng, Z. et al. CeNbOx/SSZ-13 catalysts with zeolite-protected active sites for broad-temperature NH3-SCR and enhanced resistance to K-poisoning. Appl. Catal. B: Environ. Energy384, 126176 (2026). [Google Scholar]
  • 20.Li, Y. et al. Promoting metal oxides–zeolite electron interaction on MnCeOx/HY catalyst for boosting nitrogen oxides reduction. Appl. Catal. B: Environ. Energy360, 124535 (2025). [Google Scholar]
  • 21.Han, L. et al. Selective catalytic reduction of NOx with NH3 by using novel catalysts: state of the art and future prospects. Chem. Rev.119, 10916–10976 (2019). [DOI] [PubMed] [Google Scholar]
  • 22.An, D. et al. Facile H2O‑contributed O2 activation strategy over Mn-based SCR catalysts to counteract SO2 poisoning. Environ. Sci. Technol.57, 14737–14746 (2023). [DOI] [PubMed] [Google Scholar]
  • 23.Jixing Liu, L. et al. Insight into the potassium poisoning effect for selective catalytic reduction of NOx with NH3 over Fe/beta. ACS Catal.11, 14727–14739 (2021). [Google Scholar]
  • 24.Chen, Y. et al. Cu–Ce dual–atom sites embedded in zeolites boost resistance to impurity interference for environmental catalysis. Angew. Chem. Int. Ed. 137, e202517918 (2025). [DOI] [PMC free article] [PubMed]
  • 25.Wu, S. M., Yang, X. Y. & Janiak, C. Confinement effects in zeolite-confined noble metals. Angew. Chem. Int. Ed.58, 12340–12354 (2019). [DOI] [PubMed] [Google Scholar]
  • 26.Masoumifard, N., Guillet Nicolas, R. & Kleitz, F. Synthesis of engineered zeolitic materials: from classical zeolites to hierarchical core–shell materials. Adv. Mater.30, 1704439 (2018). [DOI] [PubMed] [Google Scholar]
  • 27.Xu, Y. et al. Yolk@Shell FeMn@Hollow HZSM-5 nanoreactor for directly converting syngas to aromatics. ACS Catal. 11, 4476–4485 (2021).
  • 28.Xu, Y., Wang, J., Ma, G., Zhang, J. & Ding, M. Hollow zeolite nanoparticles combined with Fe3O4@MnO2 tandem catalyst for converting syngas to aromatics-rich gasoline. ACS Appl. Nano Mater.3, 2857–2866 (2020). [Google Scholar]
  • 29.Cheng, S. et al. Modulating the activity and SO2 resistance of α-Fe2O3 catalysts for NH3–SCR of NOx via crystal facet engineering. Environ. Sci. Technol.58, 8955–8965 (2024). [DOI] [PubMed] [Google Scholar]
  • 30.Wu, Y. et al. Interplay between copper redox and transfer and support acidity and topology in low temperature NH3-SCR. Nat. Commun.14, 2633 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Qu, W. et al. Single-atom catalysts reveal the dinuclear characteristic of active sites in NO selective reduction with NH3. Nat. Commun.11, 1532–1537 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.He, X. et al. Constructing multi-active sites of LDH-derived MnCoFe layered mixed oxide catalysts for simultaneous removal of NO and toluene. Appl. Catal. B: Environ. Energy377, 125496 (2025). [Google Scholar]
  • 33.Soltan, W. B., Peng, J., Cao, Z., Fu, Z. & Liu, H. Bimetallic Fe-Mn loaded H-ZSM-5 zeolites for excellent VOCs catalytic oxidation at low-temperatures: synergistic effects and catalytic mechanisms. Chem. Eng. J.475, 146251 (2023). [Google Scholar]
  • 34.Yan, Q. et al. Insights into enhancement of NH3-SCR activity and N2 selectivity of LDHs-derived NiMnAlOx catalysts: combination of experiments and DFT calculations. Appl. Catal. B: Environ. Energy343, 123489 (2024). [Google Scholar]
  • 35.Zhao, T., Huang, X., Cui, R., Song, X. & Tang, Z. Design the confinement structure Mn-TNTs@Ce catalyst for selective catalytic reduction of nitrogen oxides with superior SO2 resistance over a wider operation window. Appl. Catal. B: Environ. Energy358, 124353 (2024). [Google Scholar]
  • 36.Wang, B. et al. Improved activity and SO2 resistance by Sm-modulated redox of MnCeSmTiOx mesoporous amorphous oxides for low-temperature NH3-SCR of NO. ACS Catal.10, 9034–9045 (2020). [Google Scholar]
  • 37.Liu, Z. et al. Fe-doped Mn3O4 spinel nanoparticles with highly exposed Feoct–O–Mntet sites for efficient selective catalytic reduction (SCR) of NO with ammonia at low temperatures. ACS Catal.10, 6803–6809 (2020). [Google Scholar]
  • 38.Li, L. et al. Highly efficient RuOx/NbOx-ZrOx catalysts for ammonia removal via tuning acidic active species. ACS Sustain. Chem. Eng.11, 17015–17030 (2023). [Google Scholar]
  • 39.Ma, S. et al. Effect of W on the acidity and redox performance of the Cu0.02Fe0.2WTiOx (a = 0.01, 0.02, 0.03) catalysts for NH3-SCR of NO. Appl. Catal. B: Environ. Energy248, 226–238 (2019). [Google Scholar]
  • 40.Wang, P. et al. Poisoning-resistant NOx reduction in the presence of alkaline and heavy metals over H-SAPO-34-supported Ce-promoted Cu-based catalysts. Environ. Sci. Technol.54, 6396–6405 (2020). [DOI] [PubMed] [Google Scholar]
  • 41.Wang, Z. et al. The superior performance of CoMnOx catalyst with ball-flowerlike structure for low-temperature selective catalytic reduction of NOx by NH3. Chem. Eng. J.381, 122753 (2020). [Google Scholar]
  • 42.Zhang, S. et al. Decoding the reactivity enhancement of Ln2Ce2O7 compounds (Ln = Yb, Y, Tb, and Gd) for soot combustion: the remarkable contribution of variable valence A-sites. Inorg. Chem.63, 6798–6812 (2024). [DOI] [PubMed] [Google Scholar]
  • 43.Yang, S. et al. Strikingly facile cleavage of N–H/N–O bonds induced by surface frustrated Lewis pair on CeO2(110) to boost NO reduction by NH3. Environ. Sci. Technol.58, 19027–19037 (2024). [DOI] [PubMed] [Google Scholar]
  • 44.Zhu, N. et al. Distinct NO2 effects on Cu-SSZ-13 and Cu-SSZ-39 in the selective catalytic reduction of NOx with NH3. Environ. Sci. Technol.54, 15499–15506 (2020). [DOI] [PubMed] [Google Scholar]
  • 45.Fang, X. et al. Atom pairing enhances sulfur resistance in low-temperature SCR via upshifting the lowest unoccupied states of cerium. Environ. Sci. Technol.58, 12272–12280 (2024). [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Supplementary Information (8MB, pdf)
Transparent Peer Review file (670.2KB, pdf)
Source data (9.9MB, xlsx)

Data Availability Statement

All data presented in this study are included in the article and Supplementary Information. The data are available from the corresponding authors upon request. Source data are provided with this paper.

Articles from Nature Communications are provided here courtesy of Nature Publishing Group

RESOURCES