Zero-emission NO₂ capture using divalent metal cation-exchanged zeolites for air purification

Abstract

The global adoption of hydrogen and biofuels, though central to decarbonization, risks aggravating NO₂ pollution as their combustion drives up NO_x emissions through hotter flames and excess oxygen. Effective ambient NO₂ removal is critical for protecting human health, yet existing adsorbents often release toxic NO, limiting practical deployment. Here, we overcome this grand challenge by developing divalent metal cation-exchanged zeolites that achieve complete NO₂ capture with “zero” NO emission. Exemplified by Ca²⁺- and Mn²⁺-exchanged zeolites, these materials operate under both dry and humid (70% RH) conditions, even at NO₂ concentrations up to 500 ppm. Mechanistic studies reveal that divalent cations, acting as surface active sites, can suppress NO formation via stabilizing its intermediate. We further demonstrate real-world applicability by integrating these zeolites into a wearable air-purifying respirator, enabling a “zero-NO_x shield” for personal protection. This work establishes NO₂ adsorption chemistry and offers a scalable pathway to clean air solutions.

Introduction

The global transition from conventional fossil fuels to clean alternatives, such as hydrogen (H₂) and biofuels, advances vital sustainability goals, including decarbonization and reductions in carbon monoxide, hydrocarbons, and particulate emissions^1,2. However, the combustion of these clean fuels can lead to increased nitrogen oxide (NO_x) emissions due to their higher flame temperatures and, in the case of biofuels, elevated intrinsic oxygen content^3,4. Since hard-to-abate combustion-based industries still account for approximately 30% of global energy consumption^5,6,7,8, adopting clean fuels in these sectors could exacerbate ambient NO_x pollution, predominantly in nitrogen dioxide (NO₂). NO₂ is a major environmental and public health hazard, contributing to respiratory diseases and driving the formation of fine particulate matter and ground-level ozone, which were linked to an estimated 38,000 premature deaths worldwide in 2015⁹. Consequently, combustion-related NO₂ emissions will remain a significant environmental challenge, underscoring the urgent need for effective mitigation strategies.

Existing NO_x abatement technologies, such as selective catalytic reduction (SCR), operate effectively only at elevated temperatures (above 250 °C)¹⁰, leaving ambient NO₂ largely unaddressed, precisely where human exposure is most acute. Adsorption-based capture using solid porous materials has emerged as a promising solution for ambient-temperature NO₂ removal^{11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27}. However, a critical challenge persists: all known adsorbents invariably generate and release toxic nitric oxide (NO) as a by-product during NO₂ adsorption, a phenomenon widely attributed to dissociative chemisorption^{11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27}. This “NO release”, often accounting for 20–30% of the NO₂ input^{11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27}, poses additional respiratory health risks and can be rapidly re-oxidized to NO₂, thereby undermining the goal of zero-emission NO₂ capture for air purification^28,29. This issue is especially problematic in occupational settings (e.g., steelmaking, cement production, power generation, welding, and mining) and enclosed or densely populated environments (e.g., enclosed parking facilities, transit hubs, ferry terminals, and residential spaces using biofuels)³⁰.

Attempts to mitigate NO release, such as co-adsorbing NO by designing NO-retaining adsorbents³¹, have been constrained by the inherently stronger binding affinity of NO₂ compared to NO, and the prevalence of surface active sites that promote NO formation. Most NO₂ adsorbents were reported to have surface active sites that spontaneously facilitate substantial NO formation^{15,16,17,18,19,20,21,22,23,24,25,26}. These active sites, although lacking definitive identification, are generally understood to be surface radical sites (e.g., -C*^{15,16,17,18,19}, -O*^20,21,22,23, and -N*²⁴ arising from surface defect) and basic groups^25,26, which trigger redox reactions during NO₂ adsorption, thereby forming NO. As a result, intrinsically suppressing NO formation through material-level design should be a way out. However, this approach remains largely underexplored and poorly realized.

Zeolites represent a compelling but overlooked platform in this context. Their crystalline, inorganic frameworks, composed of Si-O and Al-O tetrahedra, appear to lack the reactive surface sites commonly associated with NO formation. Preliminary experimental evidence suggests that zeolites, especially when modified, can achieve low levels of NO release^32,33,34,35. Yet, systematic efforts to engineer zeolites for true zero-emission NO₂ capture remain absent from the literature, largely due to the prevailing notion that NO formation during NO₂ adsorption is unavoidable.

Here, we overcome this long-standing challenge by incorporating divalent metal cations into LTA zeolites via ion exchange, yielding materials that achieve complete NO₂ capture with “zero” NO emission under both dry and humid ambient conditions. Through combined experimental and theoretical investigation, including dynamic adsorption/desorption experiments, temperature-programmed desorption (TPD), in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS), and density functional theory (DFT) calculations—we establish that divalent metal cations suppress NO formation by stabilizing a key intermediate (nitrous acid—HNO₂) and preventing its decomposition. We further demonstrate the generality of this strategy across multiple zeolite frameworks and highlight its practical relevance by integrating these materials into wearable air-purifying respirators, thereby establishing a “zero-NO_x shield” technology. A schematic overview of this study is presented in Fig. 1. This work not only redefines the fundamental chemistry of NO₂ adsorption in porous materials but also introduces a scalable pathway toward emission-free air purification technologies critical for public health.

Fig. 1: A schematic illustration depicting the background, design strategy, and application of divalent metal cation-exchanged zeolites for “zero-emission” NO₂ capture at ambient temperature.

NO_x emissions from fuel combustion primarily exist as NO₂ in the atmosphere, motivating targeted removal strategies. Existing adsorbents typically release NO during NO₂ capture, limiting their effectiveness. The adsorbents developed in this work achieve complete NO₂ capture with “zero” NO emission by incorporating divalent metal cations into the zeolite framework. This approach enables emission-free air purification and demonstrates potential for practical deployment.

Results

Synthesis and characterization of metal-exchanged LTA zeolites

LTA zeolites with silicon-to-aluminum (Si/Al) ratios of 3 and 6 were hydrothermally synthesized³⁶ and converted to their ammonium form (as an intermediate for further ion exchange) and proton (H⁺) form (as the benchmark). Ion exchange of metal cations featuring different charge-to-size ratios and/or electronic configurations (Na⁺, K⁺, Mg²⁺, Ca²⁺, Mn²⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺, Y³⁺, La³⁺, Ce³⁺, Eu³⁺, Tb³⁺, and Yb³⁺) was performed via aqueous salt solutions to tailor the adsorbents for optimal NO₂ removal with minimal NO release. The detailed procedures for sample preparation are provided in the Supplementary Information (Section 1.1). Powder X-ray diffraction (PXRD), energy dispersive spectroscopy (EDS), X-ray photoelectron spectroscopy (XPS), inductively coupled plasma optical emission spectroscopy (ICP‒OES), and N₂ adsorption‒desorption analyses at −196.15 °C confirmed high phase purity, successful cation exchange, and preserved porosity. Detailed explanations are provided in the Supplementary Information (Section 2.1, Supplementary Figs. 4–8 and Tables 1, 2).

Elimination of NO release during NO₂ adsorption

Column breakthrough dynamic adsorption tests at approximately 25 °C and 101.325 kPa using a 500 ppm NO₂/Air mixture revealed that our studied LTA zeolites exhibit exceptionally low NO release, ranging from 0.06% to 4.11% of the NO₂ input (Fig. 2a and Supplementary Table 3), which is vastly lower than the > 20% typically reported for conventional NO₂ adsorbents (Fig. 2b, Table 1, and Supplementary Table 4)^{9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40}. More importantly, unlike those materials that release NO at the outset of NO₂ adsorption^{9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40}, our di-/trivalent metal cation-exchanged LTA zeolites release NO only at or slightly before NO₂ breakthrough (Fig. 2c and Supplementary Fig. 9), enabling “zero-emission” NO₂ capture. Two prominent examples, Ca-LTA-3 and Mn-LTA-3 (Si/Al ratio = 3), maintained “zero” NO release until NO₂ breakthrough (Fig. 2a–c and Table 1), even at a reduced feed NO₂ concentration of 25 ppm (Fig. 2d and Supplementary Table 5), a condition that typically promotes NO release due to the prolonged adsorption duration. This performance positions di-/trivalent metal cation-exchanged LTA zeolites as uniquely capable of delivering “zero-emission” NO₂ capture under ambient conditions.

Fig. 2: NO₂ adsorption and accompanying NO release in metal cation-exchanged LTA zeolites (activated/regenerated at 300 °C) upon column breakthrough dynamic adsorption tests using low-concentration NO₂ at approximately 25 °C and 101.325 kPa.

a Amounts of NO₂ adsorbed and NO released, and percentages of released NO relative to NO₂ input in H⁺-, Na⁺-, Ca²⁺-, Mn²⁺-, Ni²⁺-, and La³⁺-exchanged LTA-3 zeolites in NO₂/Air (500 ppm NO₂ balanced with air). b Comparison of NO₂ uptake and the percentage of released NO relative to NO₂ input in Na⁺-, Ca²⁺-, and Mn²⁺-exchanged LTA-3 zeolites with benchmark adsorbents under similar conditions^{12,13,20,24,26,53,54}. c NO₂ breakthrough and NO release curves of H⁺-, Na⁺-, Ca²⁺-, Mn²⁺-, Ni²⁺-, and La³⁺-exchanged LTA-3 zeolites in NO₂/Air. d NO₂ breakthrough and NO release curves of Na⁺-, Ca²⁺-, and Mn²⁺-exchanged LTA-3 zeolites in 25 ppm NO₂/Air (25 ppm NO₂ balanced with air). e Cyclic adsorption performance of Na⁺-, Ca²⁺-, and Mn²⁺-exchanged LTA-3 zeolites in NO₂/Air. f Comparison of NO₂ adsorption regenerability and the percentage of released NO relative to NO₂ input in Na⁺-, Ca²⁺-, and Mn²⁺-exchanged LTA-3 zeolites with benchmark adsorbents under similar conditions^12,20,24.

Table 1 Comparison of Na⁺-, Ca²⁺-, and Mn²⁺-exchanged LTA-3 zeolites with benchmark NO₂ adsorbents under similar conditions

In addition to this breakthrough in eliminating NO release, our adsorbents also exhibit high NO₂ uptake—up to 3.9 mmol g^-1 (Figs. 2a–b, Table 1, and Supplementary Tables 3–4), rapid kinetics—as evidenced by achieving up to 86.43% of the saturating NO₂ uptake before NO₂ breakthrough (Fig. 2d and Supplementary Table 5), and excellent regenerability—approximately 100% (Figs. 2e–f and Table 1); further details are provided in the Supplementary Information (Section 2.2, Supplementary Figs. 9–11 and Tables 3–8).

Identified factors for eliminating NO release

The exceptionally low NO release observed on our LTA zeolites can stem from two potential factors: minimal NO formation and adsorptive NO retention^31,33. To identify the contributions of these factors, we compared representative Na⁺-, Ca²⁺-, Mn²⁺-, and Ni²⁺-exchanged LTA-3 zeolites (detailed in the Supplementary Information—Section 2.3) via binary adsorption tests: one with 500 ppm NO₂ alone (NO₂/N₂) and the other with 500 ppm NO₂ plus 50 ppm NO (NO₂/NO/N₂). The difference in outlet NO concentration (∆NO) served as a metric for NO retention.

For Na-LTA-3, ∆NO remained at approximately 50 ppm throughout (Fig. 3a), indicating negligible NO retention. This suggests that the low NO release observed (Fig. 2b, Table 1, and Supplementary Table 4) is due mainly to minimal NO formation. In contrast, Ca-LTA-3 had an initial ∆NO of 20 ppm, which increased to 76 ppm before NO₂ breakthrough (Fig. 3b). This indicates modest NO retention, constrained by competitive adsorption, with the virtually eliminated NO release (Fig. 2a–d, Table 1, and Supplementary Tables 3, 4) being attributed primarily to minimal NO formation, coupled with limited retention.

Fig. 3: NO₂ adsorption and accompanying NO release upon column breakthrough dynamic adsorption tests using NO₂/NO/N₂ (500 ppm NO₂ and 50 ppm NO, balanced with N₂) and NO₂/N₂ (500 ppm NO₂ balanced with N₂) at approximately 25 °C and 101.325 kPa, respectively.

NO₂ breakthrough and NO release/breakthrough curves of a Na-LTA-3, b Ca-LTA-3, c Mn-LTA-3, and d Ni-LTA-3 adsorbents activated at 300 °C.

On the other hand, Ni-LTA-3 and Mn-LTA-3 maintained ∆NO values below 50 ppm (i.e., ranging from 20 to 36 ppm for Mn-LTA-3 and 0 to 40 ppm for Ni-LTA-3) until NO₂ breakthrough (Fig. 3c and 3d), demonstrating superior NO retention. These findings align with the DFT-calculated NO binding energies and experimentally measured NO uptakes (Ni-LTA > Mn-LTA > Ca-LTA > Na-LTA, Supplementary Figs. 12, 13 and Table 9, with detailed explanations presented in the Supplementary Information—Section 2.3). Despite their superior NO retention capabilities, Ni-LTA-3 and Mn-LTA-3 exhibited significantly decreased NO₂ uptake in the presence of NO (i.e., the NO₂ uptake in NO₂/NO/N₂ was 1-1.5 mmol g^-1 lower than that in NO₂/N₂, Fig. 3c, d and Supplementary Table 10), underscoring the adverse effect of strong NO retention on NO₂ uptake. Therefore, we conclude that the virtually eliminated NO release observed for Mn-LTA-3 and Ni-LTA-3 (Figs. 2a–d, Table 1, and Supplementary Tables 3, 4) is attributed primarily to their minimal NO formation, which is further reinforced by their superior NO retention capabilities.

In summary, our data reveal that the exceptionally low NO release in our LTA zeolites primarily arises from their minimal NO formation. This result indicates that LTA zeolites inherently lack the active sites that trigger significant NO formation—a key distinction from conventional adsorbents—resulting in a unique profile that enables “zero-emission” NO₂ capture. Detailed mechanistic analyses of the NO formation in LTA zeolites and its suppression by divalent metal cations are provided in the following sections.

Mechanisms of NO formation during NO₂ adsorption

To understand the mechanism of NO formation in our LTA zeolites, we first identified the species formed during NO₂ adsorption via TPD (Supplementary Table 11 and Figs. 14, 15) and in situ DRIFTS (Fig. 4a–d and Supplementary Fig. 16). The results revealed the formation of nitric acid (HNO₃) and HNO₂ in addition to physically adsorbed NO₂, indicating that, alongside NO₂ physisorption, dissociative chemisorption of NO₂ occurs via NO₂‒H₂O reactions. The presence of residual H₂O within our LTA zeolites after activation at 300 °C was identified by the thermalgravimetric analysis (TGA, Supplementary Fig. 17) and differential thermogravimetric analysis (DTG, Fig. 4e) for Na⁺-, Ca²⁺-, Mn²⁺-, and Ni²⁺-exchanged LTA-3 zeolites³⁷.

Fig. 4: Mechanistic investigation of NO formation in LTA zeolites.

In situ DRIFTS results of a Na-LTA-3, b Ca-LTA-3, c Mn-LTA-3, and d Ni-LTA-3 activated at 300 °C upon NO₂ adsorption in NO₂/Air (500 ppm NO₂ balanced with Air) at approximately 25 °C and 101.325 kPa. e Differential thermogravimetric analysis (DTG) results of Na⁺-, Ca²⁺-, Mn²⁺-, and Ni²⁺-exchanged LTA-3 zeolites. f Comparative study of NO₂ adsorption and accompanying NO release in Na-LTA-3 with different amounts of retained H₂O. NO₂ breakthrough and NO release/breakthrough curves (activated at 300 and 450 °C, respectively) upon column breakthrough dynamic adsorption tests using NO₂/N₂ (500 ppm NO₂ balanced with N₂) and NO₂/NO/N₂ (500 ppm NO₂ and 50 ppm NO balanced with N₂), respectively, at approximately 25 °C and 101.325 kPa.

Further NO₂ adsorption experiments on samples with distinct H₂O content confirmed that lower H₂O content correlates with reduced NO formation. Na-LTA-3 was specifically selected for this study due to its negligible NO retention capability (Figs. 3a and 4f), which allows its NO release curve in NO₂/N₂ to accurately reflect NO formation. As demonstrated in Fig. 4f, Na-LTA-3 activated at 450 °C (with lower H₂O content than the 300 °C-activated sample, as evidenced by Fig. 4e) exhibited reduced NO release (3% of NO₂ input, Supplementary Table 12) compared to its 300 °C-activated counterpart (6.3%, Supplementary Table 12). This trend was consistently observed across Ca²⁺-, Mn²⁺-, and Ni²⁺-exchanged LTA-3 zeolites (Supplementary Table 12 and Fig. 18). This finding indicates that NO formation during NO₂ adsorption on LTA zeolites is primarily driven by the NO₂‒H₂O reaction, which is further supported by cyclic adsorption-desorption results (Supplementary Fig. 19).

Based on the identified dissociatively chemisorbed species (i.e., HNO₃, HNO₂, nitronium—NO₂⁺, and nitrosonium—NO⁺) and the established reaction pathways (i.e., NO₂–H₂O reactions), the mechanism of NO formation in LTA zeolites can be described as follows: NO₂ reacts with residual H₂O to form HNO₃ and the less stable HNO₂, with the latter subsequently decomposing to form NO, as presented in Eq. (1).

$$6N{O}_{2}+3{H}_{2}O\mathop{\longrightarrow }^{Reaction\,1}3HN{O}_{3}+3{{{\boldsymbol{HN}}}}{{{{\boldsymbol{O}}}}}_{2}\mathop{\longrightarrow }^{Reaction\,2}3HN{O}_{3}+{{{\boldsymbol{HN}}}}{{{{\boldsymbol{O}}}}}_{3}+2{{{\boldsymbol{NO}}}}+{{{{\boldsymbol{H}}}}}_{2}{{{\boldsymbol{O}}}}$$

(1)

Similarly, the reaction between dinitrogen tetroxide (N₂O₄, the dimer of NO₂, as shown in Eq. (2)) and H₂O also produces HNO₂ and NO, along with the formation of NO₂⁺, as depicted in Eq. (3).

$${2{NO}}_{2}\leftrightarrow {N}_{2}{O}_{4\,}$$

(2)

$$6{N}_{2}{O}_{4}+6{H}_{2}O\mathop{\longrightarrow }^{Reaction\,1} 6N{O}_{2}^{+}O{H}^{-}+6HN{O}_{2}\mathop{\longrightarrow }^{Reaction\,2}6N{O}_{2}^{+}O{H}^{-} \\ +2HN{O}_{3}+4NO+2{H}_{2}O$$

(3)

Additionally, the presence of O₂ in the feed gas can reduce NO formation by oxidizing the transient-formed NO into more favorable adsorbates such as NO₂ and NO⁺ within zeolite pores (Supplementary Figs. 20, 21 and Table 13), as illustrated by the mechanisms proposed in Eq. (4) and Eq. (5).

$$2{{{\rm{NO}}}}+{O}_{2}\to {2{NO}}_{2}$$

(4)

$$4{NO}+{O}_{2}+4\left({{{\rm{Al}}}}-{{{\rm{O}}}}-{H}^{+}\right)\to 2{H}_{2}O+4\left({{{\rm{T}}}}-{{{\rm{O}}}}-{{NO}}^{+}\right)$$

(5)

These experimental observations and analyses, together with the proposed mechanisms for NO formation, underpin our subsequent analysis of the suppressed NO formation in divalent cation-exchanged LTA zeolites.

Divalent metal cations as key suppressors of NO formation

Our comprehensive adsorption experiments and in situ DRIFTS analyses reveal that divalent metal cations (e.g., Ca²⁺, Mn²⁺, and Ni²⁺) are crucial for suppressing NO formation during NO₂ adsorption, whereas monovalent metal cations (e.g., Na⁺) lack this ability. Specifically, Na-LTA-3 exhibits a low-level NO release throughout NO₂ adsorption, while Ca-LTA-3, Mn-LTA-3, and Ni-LTA-3 eliminate NO release (Fig. 2c, d). Besides, increasing the degree of ion exchange of divalent metal cations (e.g., Ca²⁺, Mn²⁺, and Ni²⁺) within Na-LTA-3 constantly reduces NO release (Supplementary Figs. 22, 23 and Tables 14–16). Additionally, in situ DRIFTS show that Ca²⁺-, Mn²⁺-, and Ni²⁺-exchanged LTA-3 zeolites form greater amounts of HNO₃ and HNO₂ (the intermediate of NO formation) via NO₂–H₂O reactions than Na-LTA-3 does, as evidenced by the more prominent peaks observed in the range of 1100–1300 cm^-1 (Figs. 4a–d and Supplementary Fig. 16)^{38,39,40,41,42}. These results demonstrate that divalent metal cation-exchanged zeolites consistently exhibit suppressed decomposition of HNO₂ into NO. This contrasts with the monovalent metal cation-exchanged ones, which form less HNO₂ but exhibit more significant NO release, suggesting that monovalent metal cations lack the capability to suppress HNO₂ decomposition.

DFT calculations support this observation, revealing that HNO₂ binds significantly more strongly to divalent metal cations (i.e., Ni²⁺: 1.17 eV, Mn²⁺: 1.13 eV, and Ca²⁺: 0.71 eV) than to Na⁺ (0.42 eV), owing to their higher charge‒to‒size ratios and resultant stronger electrostatic interactions (Fig. 5b)^37,43,44. Although trivalent cation-exchanged zeolites could theoretically offer even stronger binding³⁷, their low degree of ion exchange (below 40%, with Na⁺ remaining the majority cation, Supplementary Table 1) limits their practical effectiveness (Supplementary Fig. 9 and Table 3). In summary, the strong binding of HNO₂ by divalent metal cations redirects the pathway of the NO₂–H₂O reaction away from NO formation (schematically illustrated in Fig. 5c), which is critical for eliminating NO release and achieving “zero-emission” NO₂ capture.

Fig. 5: Mechanistic investigation of suppressed NO formation in zeolites via ion‑exchanged divalent metal cations.

a Locations of cations in Na⁺ (cyan), Ca²⁺ (red), Mn²⁺ (green), and Ni²⁺ (purple)-exchanged LTA zeolites determined by synchrotron PXRD (Supplementary Fig. 24) analysis. b DFT-calculated binding energies of HNO₂ in LTA models containing a single cation of each metal, along with the corresponding optimized adsorption configurations. c Schematic illustration summarizing the mechanism of NO formation in zeolites and its suppression facilitated by ion‑exchanged divalent metal cations.

Moreover, the enhanced formation of NO⁺ in Mn-LTA-3 and Ni-LTA-3, as evidenced by more pronounced peaks at approximately 2100 cm^-1 (Fig. 4a–d and Supplementary Fig. 16), suggests that these cations may promote the self-disproportionation of NO₂ to yield NO₃^- and NO⁺ (Eq. (6)).

$${{{\rm{Al}}}}-{O}^{n-}-{M}^{n+}+{N}_{2}{O}_{4}\to {{{\rm{Al}}}}-{O}^{n-}-{{NO}}^{+}+{M}^{n+}{{NO}}_{3}^{-}$$

(6)

This pathway assists in diverting NO₂ away from the NO₂–H₂O reaction and facilitates effective electrostatic stabilization of produced NO₃^- and NO⁺ on positively-charged cations and negatively-charged zeolite frameworks, respectively, thereby mitigating HNO₂ and NO formation (schematically illustrated in Fig. 5c)^{40,42,45,46,47,48}. This desirable property of Mn²⁺ and Ni²⁺ may be attributed to their unfilled outer shell d-orbitals and greater Lewis acidity than Na⁺ and Ca²⁺, enabling them to act as stronger electron acceptors^47,49,50.

Broad applicability of divalent metal cations in zeolites to achieve “Zero” NO release

Our strategy of using divalent metal cations (e.g., Ca²⁺ and Mn²⁺) to suppress NO formation and thus eliminate NO release demonstrated broad applicability across various zeolite types. Besides its effectiveness on LTA zeolites, exchanging these cations into commercially available CHA, MOR, and FAU zeolites also results in reduced NO release during NO₂ adsorption, achieving “zero-emission” NO₂ capture (Fig. 6a, Supplementary Figs. 25–31 and Tables 17–25). These findings indicate that the NO formation/release behavior in zeolites is primarily governed by the type of the metal cation, with the influence of zeolite topology being comparatively limited. This broad applicability underscores the scalability and practical viability of our strategy for advanced NO_x removal technologies.

Fig. 6: Studying the effect of metal cations and zeolite topologies on NO releasing behaviors during NO₂ adsorption.

a NO₂ breakthrough and NO release curves for Na⁺-, Ca²⁺-, and Mn²⁺-exchanged LTA, CHA, MOR, and FAU zeolites (Si/Al ratio = 6, activated at 300 °C) in NO₂/Air (500 ppm NO₂ balanced with air) at approximately 25 °C and 101.325 kPa. Correlation between the quantity of retained water (after 300 °C activation) and released NO, along with differential thermogravimetric analysis (DTG) for studied b Na⁺-, c Ca²⁺-, and d Mn²⁺-exchanged LTA, CHA, MOR, and FAU zeolites.

Different zeolite topologies slightly influence the amount of NO released, primarily due to differences in retained H₂O content after identical activation processes. Specifically, for LTA, CHA, MOR, and FAU zeolites exchanged with the same metal cation, the amount of NO released positively correlates with the amount of water retained after activation at 300 °C (Fig. 6b–d). This finding reaffirms that the NO₂-H₂O reactions constitute the dominant pathway for NO formation.

Real-world application potential and demonstrations

Our adsorbents capable of “zero-emission” NO₂ capture were rigorously evaluated under conditions that closely mimic real-world challenges. First, we confirmed that our LTA zeolites operate effectively in humid environments—scenarios known to hinder conventional adsorbents by inducing substantial NO release (up to 45% of NO₂ input) via NO₂-H₂O reactions²⁷. Even in gas streams containing ~2 vol% H₂O (~70% relative humidity at ambient temperature, representative of tropical and subtropical regions) and various NO₂ concentrations—500 ppm (commonly used for evaluating NO₂ adsorbents) and 4 ppm (representative of typical indoor NO₂ pollution scenarios⁵¹)—our di-/tri-valent metal cation-exchanged LTA-3 zeolites consistently achieved “zero-emission” NO₂ capture, demonstrating their practical viability for ambient air purification (Fig. 7a–b and Supplementary Table 26). The representative Ca-LTA-3 exhibited excellent recyclability under humid conditions, retaining up to 97% of its NO₂ uptake and maintaining “zero-emission” NO₂ capture over 10 cycles (Fig. 7c–e). This performance surpasses the previously reported maximum retained NO₂ uptake (64%) in cyclic adsorption tests under humid conditions²⁴.

Fig. 7: “Zero-emission”, recyclable NO₂ adsorption on LTA zeolites under humid conditions and demonstration as “zero-NO_x shields” for occupational personal protection.

a NO₂ breakthrough and NO release curves for Na⁺-, Ca²⁺-, Mn²⁺-, and Ni²⁺-exchanged LTA-3 zeolites (activated at 450 ^oC) in humid NO₂/Air (500 ppm NO₂ balanced with air, containing ~2 vol% of H₂O, at ~70% relative humidity, 25 °C, and 101.325 kPa). b NO₂ breakthrough and NO release curves for Ca-LTA-3 (activated at 450 ^oC) in humid 4 ppm NO₂/Air (4 ppm NO₂ balanced with air, containing ~2 vol% of H₂O, at ~70% relative humidity, 25 °C, and 101.325 kPa). c NO₂ breakthrough and NO release curves and d corresponding NO₂ uptake and percentages of released NO relative to NO₂ input for Ca-LTA-3 (activated/regenerated at 450 ^oC) during cyclic NO₂ adsorption in humid NO₂/Air. e NO₂ breakthrough and NO release curves for Ca-LTA-3 zeolites (activated/regenerated at 450 ^oC) during cyclic NO₂ adsorption in humid 4 ppm NO₂/Air. f Schematic illustration of our designed filter for wearable air‑purifying respirators intended for occupational personal protection. g Photograph of the fabricated prototype of our designed filter. h Photograph of the filter integrated into a commercially available respirator (3 M 6200 Half Mask; this product is solely utilized for demonstrating the integration and does not imply endorsement, affiliation, or evaluation by the manufacturer). i NO₂ breakthrough and NO release curves, along with inhalation resistance of the filter during the standardized NO₂ service-life test, performed using humid 500 ppm NO₂ polluted air (containing ~2 vol% of H₂O, at ~70% relative humidity, 25 °C, and 101.325 kPa) at a flow rate of 32 L min^-1.

Afterward, we integrated our adsorbents into the filter of wearable air-purifying respirators, engineered as a “zero-NO_x shield” for occupational protection. This prototype (Fig. 7f–h) not only met but also exceeded the standardized criteria for NO₂ removal (outlined by the National Institute for Occupational Safety and Health (NIOSH), procedure codes: RCT-APR-STP-0062 and TEB-APR-STP-0007). Specifically, the filter effectively captured 500 ppm NO₂ from humid air with “zero” NO_x emissions for 132 minutes (Fig. 7i), which is significantly beyond the required 30-minute service life claimed in RCT-APR-STP-0062. Additionally, the inhalation resistance is 282.3 Pa (Fig. 7i), which falls well within the acceptable range (i.e., <400 Pa, claimed in RCT-APR-STP-0062 and TEB-APR-STP-0007). These results provide a compelling demonstration of how our fundamental discovery can translate into a robust, real-world protective device.

Furthermore, desktop estimations indicate that our adsorbents hold significant promise as a “zero-NO_x shield” for indoor air purification through integration into filters of flow-through air purifiers. Calculations based on NO₂ uptake suggest that only a small amount of material is required to clean confined spaces effectively. Specifically, 112.5 g of Ca-LTA-3 (corresponding to approximately 100 cm³ in volume) would be sufficient for complete NO₂ removal from a workshop of 60 m² area and 4 m height, contaminated with 25 ppm NO₂. These results further highlight the potential of our approach to address long-standing air quality challenges.

Moreover, leveraging the excellent regenerability of our LTA zeolites, the spent materials used for NO₂ capture can be regenerated via heating and gas purging, releasing the adsorbed NO₂ for catalytic reduction (e.g., by SCR). This approach enables a complete cycle for NO₂ adsorption and reduction.

Cost analysis of large-scale production

To assess the practical viability of our advanced NO₂ adsorbents, an industrial-scale cost analysis was conducted. The estimated production cost of our zeolite adsorbents is approximately USD 1.93 per kilogram (Supplementary Tables 28, 29) and could be further reduced through process scale‑up. Given that our zeolite adsorbents can be nearly fully regenerated following NO₂ adsorption under practical conditions (Fig. 7c–e), the spent materials can be collected, regenerated, and reused—paving another way for cost reduction. Compared with commercial porous adsorbents such as activated carbon, silica gel, and alumina, our zeolite-based adsorbents offer a highly competitive cost-performance profile (Detailed in Supplementary Table 30), underscoring their promise for widespread deployment in air purification technologies.

Discussion

The global transition to clean fuels (e.g., H₂ and biofuels) can drive increased NO_x emissions across indispensable combustion-based industries. Adsorptive removal of NO₂ at ambient temperature is a promising strategy for mitigating air pollution, but conventional porous materials are fundamentally limited by the formation and release of NO, a toxic by-product. In this work, we report a class of adsorbents capable of achieving complete NO₂ capture with “zero” NO release at ambient temperature. This realizes true “zero-emission” NO₂ capture under ambient conditions. This performance was enabled by precisely tuning the adsorption chemistry through ion-exchanging divalent metal cations (e.g., Ca²⁺ and Mn²⁺) into zeolites.

Mechanistic studies combining dynamic adsorption/desorption experiments, in situ spectroscopy, and DFT calculations reveal that these cations suppress NO formation by strongly stabilizing nitrous acid (HNO₂), a key intermediate, and, in some cases, promoting the self-disproportionation of NO₂ into NO₃^- and NO⁺ without NO release. The superior electronegativity, Lewis acidity, and ion exchange capacity of divalent cations are central to this effect, altering the adsorption landscape in a way that prevents redox pathways leading to NO formation.

The breakthrough not only advances the fundamental understanding of NO₂ adsorption chemistry in zeolitic frameworks but also opens a paradigm for ambient NO₂ abatement technologies. Our materials exhibit robust performance under both dry and humid conditions (up to 70% relative humidity) across a wide NO₂ concentration range (e.g., 4–500 ppm), confirming their practical relevance for real-world air purification.

To demonstrate applicability, we engineered these zeolites into functional devices, establishing a “zero-NO_x shield” for human protection. When incorporated into wearable air-purifying respirators, the adsorbents substantially reduce inhalation exposure to NO₂, enhancing occupational safety in high-risk environments such as firefighting, welding, traffic control, mining, agriculture, military operations, and healthcare. Additionally, integration into flow-through air purification systems could effectively mitigate NO₂ levels in urban microenvironments, including transit terminals, (semi-)enclosed parking facilities, tunnels, and roadside booths, as well as confined spaces such as workshops, indoor fireplaces, and vehicles. These deployment routes directly address the pressing public health need to reduce the NOₓ-related respiratory and cardiovascular risks.

Beyond air purification, this study reveals additional opportunities for zeolite-based systems in NO_x conversion, storage, and utilization. First, ambient adsorption of NO₂ in the presence of water yields HNO₃ within zeolite pores, presenting a low-energy pathway for nitric acid production, a sustainable alternative to the energy-intensive Ostwald process. Second, the stabilization of HNO₂ and NO⁺ within divalent metal cation-exchanged zeolites enables their function as solid-state NO reservoirs, with controlled NO release achievable via mild thermal or moisture triggers. This controlled release supports applications in biomedical NO delivery for wound healing and vasodilation, as well as in antimicrobial surface coatings. Third, the high NO₂ uptake and concurrent formation of reactive nitrogen species within confined zeolite environments suggest the feasibility of solid-phase nitration, offering a greener, solvent-free route for producing nitrated organics used in pharmaceuticals, dyes, agrochemicals, and energetic materials.

In summary, our work not only overcomes a critical limitation in ambient NO₂ adsorption—eliminating the release of toxic NO by-product—but also establishes a versatile platform for advanced air purification and broader chemical engineering applications involving nitrogen oxides and their derivatives.

Methods

Preparation of adsorbents

LTA and CHA zeolites with Si/Al ratios of 3 or 6 were synthesized hydrothermally following established procedures^36,52. MOR and FAU zeolites with the Si/Al ratio of 6 were purchased from Zeolyst Co. (CBV10A and CBV712, respectively). Ammonium (NH₄⁺) from zeolites and metal cation (Na⁺-, K⁺-, Mg²⁺-, Ca²⁺-, Mn²⁺-, Co²⁺-, Ni²⁺-, Cu²⁺-, Zn²⁺-, Y³⁺-, La³⁺-, Ce³⁺-, Eu³⁺-, Tb³⁺-, and Yb³⁺-)-exchanged zeolites were prepared via liquid-state ion exchange according to our previously reported methods^37,43. Proton (H⁺) from LTA zeolite was obtained by calcinating the NH₄⁺ form in the airflow. Detailed procedures for sample preparation are provided in the Supplementary Information (Section 1.1).

Characterization

The crystallinity and phase purity of our studied zeolites were evaluated via PXRD patterns in the 2θ range of 5–50° with a step size of 0.026°. PXRD patterns were recorded on a PANalytical X’Pert3 powder diffractometer equipped with a Cu Kα radiation source (λ = 0.15406 nm) operated at 40 V and 40 mA. The crystallinity, phase purity, and cation locations of our studied representative LTA zeolites were further confirmed by synchrotron PXRD (PD beamline, Australian Synchrotron, ANSTO) at wavelengths of 0.5903 Å and 0.6887 Å. The Si/Al ratios and ion exchange degrees of the studied zeolites were determined from their elemental compositions, which were obtained from EDS using an Oxford Aztec Energy X-MAX 50 system in conjunction with field emission gun-scanning electron microscopy (FEG-SEM, FEI Quanta 450 FEG), as well as from ICP‒OES using a Agilent 720ES. These parameters were calculated using Eqs. (7) and (8), respectively:

$${Si}/_{Al}{ratio}=\frac{{Molar\; ratio\; of\; Si}}{{Molar\; ratio\; of\; Al}}$$

(7)

$$ {Ion\; exchange\; degree}\left({also\; referred\; to\; as\; the\; degree\; of\; ion\; exchange}\right)=\\ \left(\frac{{Molar\; ratio\; of\; ion\; exchanged\; metal\; cation}\times {Charge\; of\; ion\; exchanged\; metal\; cation}}{{Molar\; ratio\; of\; Al}}\right) \\ \times \, 100\%$$

(8)

The valence states of the ion-exchanged metal cations were determined via XPS using Thermo Scientific ESCALAB Xi + . The porous properties of the studied zeolites were characterized by N₂ adsorption isotherms at −196.15 °C using a Micromeritics 3Flex gas adsorption analyzer. Prior to the measurements, the samples were activated at 300 °C under high vacuum (10^-6Pa) for 10 hours.

In situ DRIFTS during NO₂ adsorption were collected using a Thermo Nicolet i550 spectrometer. Prior to spectral acquisition, the samples were activated at 300 °C for 4 hours under continuous argon (Ar) flow at a rate of 40 mL min^-1. After activation, the background spectrum was recorded once the sample cooled to ambient temperature (~25 °C). Subsequently, in situ spectra were collected during NO₂ adsorption using a feed gas containing 500 ppm NO₂ (balance with either air or N₂—designated as NO₂/Air and NO₂/N₂, respectively) at a total flow rate of 50 mL min^-1. All spectra were recorded with 64 scans at a resolution of 4 cm⁻¹.

TGA was conducted using an SDT 650 thermal analyzer. The samples were first heated to 300 °C at a rate of 2 or 5 °C min⁻¹ and then held for 2 hours, followed by heating to 450 or 600 °C at 5 °C min⁻¹ and holding for an additional 1 hour. All measurements were performed under a continuous 200 mL min^-1 N₂ flow. DTG analysis was obtained by calculating the first derivative of the TGA result.

NO₂ adsorption experiments

NO₂ adsorption was measured at ambient temperature ( ~ 25 °C) and atmospheric pressure ( ~ 101.325 kPa) by column breakthrough dynamic adsorption experiments, with a description of the experimental setup provided in the Supplementary Information (Section 1.2.1). Prior to the adsorption test, adsorbents (0.2–0.3 g, particle size ~0.4 mm) were activated at 300 or 450 °C under a 40 mL min⁻¹ continuous Ar flow, with a heating rate of 2 °C min⁻¹ to the target temperature, followed by holding for 10 hours. During the adsorption tests, a low-concentration NO₂ feed (i.e., 500 ppm, 25 ppm, or 4 ppm NO₂ balanced with air or N₂, respectively) was introduced into the sample column at a flow rate of 200 mL min⁻¹. The outlet concentrations of NO₂ and NO were continuously monitored using a gas analyzer (MultiRAE Lite and/or MKS Multigas 6030). Data acquisition was terminated once the outlet NO₂ concentration reached 20 ppm.

The amount of NO₂ adsorbed (also referred to as NO₂ uptake, in mmol g^-1) was calculated using Eq. (9):

$${N}_{{{{{\rm{NO}}}}}_{2}}=\frac{Q}{{V}_{m}\cdot m}\left({\int }_{0}^{{T}_{s}}\left(C-{C}_{d}\right){dt}\right)$$

(9)

where N_NO2 denotes the NO₂ uptake in mmol g^-1; Q is the feed‑gas flow rate (200 mL min^-1); V_m is the molar volume of an ideal gas under ambient conditions; m is the mass of the activated adsorbent in g; C is the NO₂ concentration in the feed (500 ppm, 25 ppm, or 4 ppm); C_d is the detected outlet NO₂ concentration in ppm at time t; and T_s is the breakthrough time (the time at which C_d reaches 20 ppm) in minutes.

The amount of NO released (in mmol g^-1) during NO₂ adsorption was calculated via Eq. (10):

$${R}_{{{{\rm{NO}}}}}=\frac{Q}{{V}_{m}\cdot m}\left({\int }_{0}^{{T}_{s}}\left({C}_{d}\right){dt}\right)$$

(10)

where R_NO denotes the amount of NO released in mmol g^-1, and C_d is the detected outlet NO concentration in ppm at time t. All other parameters are defined as in Eq. (9). To facilitate comparison across different adsorbents, the amount of NO released was normalized as a percentage of the NO₂ input.

Binary adsorption experiments of NO₂ and NO were conducted under similar conditions using a feed gas of 500 ppm NO₂ and 50 ppm NO balanced with N₂. The amounts of NO₂ adsorbed and NO released were calculated using the methods described in Eqs. (9) and (10).

DFT calculations

DFT calculations were performed using the Vienna Ab initio Simulation Package (VASP 5.4.4) on the Australian Synchrotron Compute Infrastructure (ASCI). The generalized gradient approximation (GGA) and GGA + U methods were applied, employing the Perdew-Burke-Ernzerhof (PBE) exchange-correlation functional. Interactions between ionic cores and valence electrons were treated using ultrasoft pseudopotentials, and van der Waals (VdW) interactions were included via the zero-damping DFT-D3 dispersion correction method of Grimme. A plane-wave cutoff energy of 600 eV was used for all calculations, and a Monkhorst-Pack k-point mesh was employed to ensure total‑energy convergence within 1 meV per atom. The LTA zeolite framework used in the DFT calculations was obtained from the Rietveld refinement of the synchrotron PXRD data (Supplementary Fig. 24). The adsorption binding energy (E_ads) was calculated via Eq. (11):

$${E}_{{ads}}={E}_{{adsorbate}+{LTA\; framework}}-\left({E}_{{adsorbate}}+{E}_{{LTA\; framework}}\right)$$

(11)

Demonstration of zeolite adsorbents in wearable air-purifying respirators

We conducted both the standardized NO₂ service-life test and inhalation-resistance test in accordance with the National Institute for Occupational Safety and Health (NIOSH) procedure (code: RCT-APR-STP-0062 and TEB-APR-STP-0007). A customized tubing system was used to perform these experiments, as detailed in the Supplementary Information (Section 1.3). Prior to testing, the adsorbents (60 g of Ca-LTA-3 and 40 g of Ni-LTA-3) were pelleted into 0.7 mm particles to balance high NO₂ uptake, minimal NO release, and low breathing resistance (Supplementary Fig. 32 and Table 27) and subsequently packed into a customized simulated filter. The assembled device was activated by heating at 2 °C min^-1 to 450 °C under a 10 L min^-1 Ar flow and held at this temperature for 3 hours. During the NO₂ service-life test, a feed gas consisting of 500 ppm NO₂ balanced with air and containing ~2 vol% of H₂O at ~70% relative humidity, 25 °C, and 101.325 kPa was supplied at a flow rate of 32 L min^-1. The outlet concentrations of NO₂ and NO were continuously monitored using a flue gas analyzer (Betoo Instrument, CN), and data acquisition was terminated once the outlet NO₂ concentration reached 20 ppm. The inhalation resistance was measured simultaneously by recording the pressure drop across the device. A blank test (without adsorbent loading) was also conducted for background subtraction.

Evaluation of zeolite adsorbents for indoor air purification

The evaluation was performed through desktop calculations based on the NO₂ uptake obtained from single breakthrough dynamic adsorption results under dry conditions, assuming operation in effectively dehumidified environments achieved through a desiccant pre‑layer or a dehumidifier.

The specific air purification capacity V_s (in m³) was calculated via Eq. (12):

$${V}_{s}=\frac{{N}_{{{{{\rm{NO}}}}}_{2}}\times m\times {V}_{m}}{{C}_{{{{{\rm{NO}}}}}_{2}}}$$

(12)

where V_s denotes the specific air purification capacity in m³, C_NO2 is the initial NO₂ concentration in the atmosphere, and the other parameters are defined as in Eq. (9).