Cost-effective abatement of industrial sources of nitrous oxide with methane for urgent climate mitigation Authors: Yunshuo Wu, Rongrong Hong, Xuanhao Wu, Feng-Shou Xiao, Haiqiang Wang, Zhongbiao Wu Published: 2026 ABSTRACT Abatement of industrial emissions of nitrous oxide (N2O), a potent greenhouse gas and stratospheric ozone-depleting substance, offers a viable approach to urgent climate mitigation. However, existing N2O abatement processes require rigorous conditions ( > 400 °C) due to its inertia and high costs. Here we present a cost-effective technology enabling 99.99% N2O conversion at 150 °C and a high reaction rate of 9629.55 kgN2O kgPt−1 h−1 using CH4 as a reductant on Pt/HZSM-5. Mechanism studies reveal the key to this success lies in the reaction-induced Pt nanoparticles, which host monoatomic adsorbed oxygen species and achieve a rapid O* formation/consumption cycle. Techno-economic analysis shows that it offers the potential to reduce N2O abatement costs by 62.3% for industrial adipic acid emissions in China. The study can incentivize further technical development and improve the sustainability of N2O abatement, as well as low-temperature CH4 activation. Introduction The planetary boundaries of climate change and ozone depletion are under critical stress, posing profound risks to global sustainability and humanity’s future1,2. Among greenhouse gases, nitrous oxide (N2O) poses a dual threat: it is the third most significant contributor to global warming with a global warming potential 273 times that of CO23, and the leading anthropogenic ozone-depleting substance4. With a long lifetime of >110 years, N2O levels are steadily increasing due to anthropogenic emissions, necessitating urgent and effective mitigation strategies to gain climate and ozone co-benefits from reducing N2O emissions5. Agriculture emissions contribute about two-thirds of the global anthropogenic N2O and are challenging to control due to the non-point characteristic6. In contrast, industrial N2O emissions are unintended by-products from the production of adipic acid/caprolactam (mainly used in nylon) and nitric acid. Industrial emissions are characterized by high concentrations, point-source property, and low oxygen content, making them well-suited for immediate and impactful climate mitigation, especially in China7. Industrial sources have overtaken agricultural emissions to become the fastest-growing contributors to N2O release8, a challenge that will escalate in the coming decades. Prioritizing industrial N2O emissions thus offers a critical pathway to achieving the 1.5/2 °C climate target and global carbon neutrality target. Current catalytic direct decomposition technology (N2O → N2 + 1/2 O2) requires high temperatures (> 400 °C) since N2O is an inert molecule9,10, leading to significant energy consumption and high mitigation costs (denoted as N2O-DD-400). Inspired by NH3 selective catalytic reduction (SCR) technology for NOx abatement, the introduction of reductants such as CO, H2 or hydrocarbons has been proposed to lower reaction temperatures for N2O reduction11. Among these, methane (CH4) offers distinct advantages as a reductant: (i) CH4 poses significant environmental threats as the second potent GHG with global warming potentials around 28 times that of CO212,13. (ii) Its abundance and availability as low-quality methane (e.g., from landfills or coal-bed methane) further enhance its sustainability14. (iii) CH4’s high hydrogen-to-carbon ratio enables a CH4/N2O reaction ratio of 1/4. Despite these advantages, achieving complete abatement at low temperatures (for example, below 150 °C) remains a major challenge due to the inherent difficulty of CH4 activation associated with methane’s C–H bond cleavage (434 kJ mol−1) and low polarizability15,16. This process mostly employes iron zeolites17,18, as N2O is known to form a unique oxygen species (α-O) on iron zeolites19, which can activate methane20–22. However, using iron as the active site still suffers from high operating temperatures (> 350 °C) due to poor CH4 activation at lower temperatures, as well as deactivation from sintering or coke formation due to the accumulation of water in the reaction products or inadequate oxidation3. Also, effective mitigation necessitates catalysts with high reaction rates to efficiently convert large quantities of industrial N2O in a short time. The delicate balance between C–H dissociation and N2O activation severely restricts the choice of metal centers for efficient catalysis. As a result, achieving 100% N2O abatement below 150 °C remains a major challenge in the field. In this study, we present 99.99% N2O conversion and 100% N2 selectivity without any detected CH4 escape at temperatures below 150 °C by using CH4 as a reductant with a stoichiometric CH4/N2O ratio of 1/4 and an industrially-viable in-situ activated Pt/HZSM-5 (ZSM = Zeolite Socony Mobil) catalyst (denoted as CH4-SCR-150). It achieves a high reaction rate of 9629.55 kg N2O per kg of Pt per hour (kgN2O kgPt−1 h−1) and an estimated cost decrease of 62.3% for industrial adipic acid emissions in China. A combination of experimental characterizations and theoretical calculations reveals that the success lies in the rapid O–H consumption cycle on the surface of Pt nanoparticles (NPs). By addressing the urgent need for cost-effective technologies, this work establishes a strong foundation for the widespread adoption of industrial N2O mitigation efforts. Results Pt/HZSM-5 enables <150 °C N2O abatement by CH4 Pt/HZSM-5-F (F denoted as fresh) was prepared via wetness impregnation method using H2PtCl6 precursor and a commercial HZSM-5 support with 500 °C air calcination. Pt/HZSM-5-U (U denoted as used) and Pt/HZSM-5-R (R denoted as reduced) were obtained by pretreating Pt/HZSM-5-F in 2.5% CH4 + 10% N2O at 280 °C and 5% H2 at 500 °C, respectively. The actual Pt loadings were measured by inductively coupled plasma-optical emission spectroscopy (ICP-OES; Supplementary Table 1). CH4-SCR catalytic performance for the prepared samples was investigated with a gas hourly space velocity (GHSV) of 12,000 ml g−1 h−1. Industrial N2O is generated in significant quantities as a by-product during the production of adipic acid (40% N2O), caprolactam (10% N2O), and nitric acid (2% N2O). Activity data showed that the Pt/HZSM-5-U catalyst demonstrated unexpected low-temperature activity for these emissions (Fig. 1a). Notably, at N2O concentrations exceeding 10%, the catalyst enabled T100 at temperatures below 150 °C. When the concentration dropped to 2%, the catalyst still achieved 100% N2O conversion at 185 °C. At 40% N2O, since the reaction is highly exothermic under near-adiabatic conditions, localized hot spots in the catalyst bed may raise the actual catalyst temperature. Under such conditions, the reactor temperature (x-axis) should be interpreted as the ignition temperature rather than the actual catalyst temperature, due to temperature run-away associated with hot-spot formation. Therefore, the catalyst bed was diluted with crushed quartz at a high ratio (catalyst/diluent = 1/10 to 1/20), and the activity curves were re-measured after confirming the elimination of hot spots (Supplementary Fig. 1). The low-temperature activity was basically preserved, suggesting that it originates from the intrinsic catalytic activity. Both Pt/HZSM-5-U and Pt/HZSM-5-R exhibited 100% N2O/CH4 conversion below 150 °C, while Pt/HZSM-5-F failed to achieve it (Fig. 1b). We propose that the in situ pretreatment of Pt/HZSM-5-U led to the reaction-induced formation of active Pt species, which is subsequently confirmed by further characterization (vide infra). The products were also analyzed using Gasmet Dx4000, where no NO or NO2 was detected. Moreover, the reaction proceeded without forming any other nitrogen- or carbon-containing by-products, as confirmed by gas chromatography (GC; Supplementary Fig. 2) and temperature-programmed surface reaction with mass spectrometry (TPSR-MS; Supplementary Fig. 3), indicating that both CO2 selectivity and N2 selectivity reached 100%. We further calculated the thermodynamic equilibrium plots (Supplementary Fig. 4), indicating the feasibility of achieving this reaction below 150 °C. Then, we carried out a multi-cycle experiment, which involved raising the reaction temperature, measuring the conversion curve for 6 h, cooling down the system, re-measuring the conversion curve, and repeating this cycle. The conversion curves are almost the same, showing a reliable activity performance (Supplementary Fig. 5).Fig. 1Catalytic performance in N2O abatement.a Conversion curves for different N2O inlet concentrations on Pt/HZSM-5-U. b Conversion curves over different Pt/HZSM-5 catalysts. c N2O conversion and CH4 escape on Pt/HZSM-5-U in 10% N2O and different CH4 contents. d Conversion curves over Pt/HZSM-5 catalysts with different Si/Al ratios. e Conversion curves over Pt/HZSM-5 catalysts prepared by different methods. f Conversion curves over the Pt/HZSM-5-U with different O2 contents. Reaction conditions: unless otherwise specified, 2.5% CH4, 10% N2O, He balance, GHSV of 12,000 ml gcat−1 h−1. Catalytic performance in N2O abatement. a Conversion curves for different N2O inlet concentrations on Pt/HZSM-5-U. b Conversion curves over different Pt/HZSM-5 catalysts. c N2O conversion and CH4 escape on Pt/HZSM-5-U in 10% N2O and different CH4 contents. d Conversion curves over Pt/HZSM-5 catalysts with different Si/Al ratios. e Conversion curves over Pt/HZSM-5 catalysts prepared by different methods. f Conversion curves over the Pt/HZSM-5-U with different O2 contents. Reaction conditions: unless otherwise specified, 2.5% CH4, 10% N2O, He balance, GHSV of 12,000 ml gcat−1 h−1. The measured conversions of N2O and CH4 were equivalent within ±3% when the inlet gas maintained a CH4/N2O ratio of 1/4 (Supplementary Fig. 6), thus the y-axis was labeled as N2O/CH4 conversion for clarity. To investigate what would occur when deviating from the 1/4 ratio, we fixed the N2O concentration at 10% and varied CH4 supply from 2 to 3% to test N2O conversion rates and methane escape rates (Fig. 1c). It suggested that strict control of methane supply during operation, which could be achieved through automated feed control systems that allowed precise and real-time adjustment of gas compositions, was crucial to prevent insufficient N2O conversion or CH4 escape. We further lowered the CH4 concentration to 0.5% and measured the conversion curve (Supplementary Fig. 7). The result was consistent with those obtained under 0.5% CH4 + 2% N2O (Fig. 1a), indicating that the reaction was primarily constrained by the reactants concentration and continued to follow the 1/4 stoichiometric ratio. In addition, a series of HZSM-5-supported catalysts impregnated with other metals were tested in CH4-SCR but all failed to show a comparable activity with Pt-based HZSM-5 (Supplementary Fig. 8), suggesting that the superior activity of CH4-SCR lies in Pt species. Then, we contrasted the use of HZSM-5 as support for Pt with that of other supports (Supplementary Fig. 9), showing that HZSM-5 exhibited the highest activity, while CeO2 also achieved a T90 of 150 °C. Supports with different acidic properties were evaluated, including HZSM-5 with Si/Al ratios of 50 and 200, where a higher Si/Al ratio corresponds to fewer acidic sites. The activity order is consistent with the acidic sequence Pt/HZSM-5-U > Pt/HZSM-5-U-50 » Pt/HZSM-5-U-200 (Fig. 1d), showing that zeolite acidity has a pronounced effect on catalytic performance. To distinguish between Pt NPs located on external zeolite surfaces versus those within the matrix (Supplementary Fig. 10), we prepared Pt/HZSM-5-NP (NP denotes nanoparticle) by the sol immobilization method23,24, and Pt@HZSM-5 by the ion-exchange method using the Pt(NO3)2 precursor25,26, then evaluated the catalytic activity. The activity of Pt/HZSM-5-NP was comparable to that of Pt/HZSM-5-U, whereas Pt@HZSM-5 showed no activity below 200 °C (Fig. 1e). We therefore infer that the active species for CH4 + N2O reaction are Pt NPs. In contrast, Pt species located inside the zeolite matrix contribute negligibly to the low-temperature activity. Given that in typical SCR reactions, CH4 is reacted with NO/NO2, it is imperative to discuss whether similar interactions might occur and how they contribute to the overall catalytic performance. Therefore, we further evaluated the activity when replacing N2O with NO or NO2 (Supplementary Fig. 11). The results showed that neither NO nor NO2 could drive the reaction, indicating that N2O is indispensable. It is well-established that the presence of O2 can inhibit N2O conversion for most catalysts. Fortunately, O2 levels in industrial N2O exhaust gases are typically low, as high O2 impurities would severely compromise the quality of the target products (e.g., adipic acid or caprolactam, with O2 contents generally below 2%, though occasionally fluctuating to 2–12%). Besides, membrane-based O2 separation can be applied to lower O2 contents27. Therefore, it is necessary to evaluate the impact of O2 to establish comprehensive operation parameters (Fig. 1f), showing that CH4 conversion was strictly suppressed above 2.5% O2 for Pt/HZSM-5-U. For high-concentration N2O streams, a 0.5% O2 content was considered safe, and inhibition caused by higher O2 levels could be rapidly recovered once the O2 was removed (Supplementary Fig. 12). Furthermore, changing the oxidant from 10% N2O to 5% O2 (remaining the same O ratios), the activity showed a T10 (temperature at 10% CH4 conversion) of ~350 °C, suggesting the inability of O2 as the oxidant for CH4 combustion at low temperatures, likely due to the formation of an inactive PtOx surface layer, as verified and discussed later. Catalyst characterizations X-ray diffraction (XRD; Supplementary Fig. 13) and N2 physisorption results (Supplementary Fig. 14) were consistent with previous studies28, ensuring the reliability of the commercial HZSM-5. The catalyst preparation procedure is illustrated in Fig. 2a. Aberration-corrected scanning transmission electron microscopy (AC-STEM) images of pristine HZSM-5 (Fig. 2b) showed an obvious MFI topology29,30, with the size of ten-membered ring being 5.46 Å (Supplementary Fig. 15), consistent with the reported result21. Transmission electron microscopy (TEM; Supplementary Fig. 16) showed the size of Pt NPs supported on HZSM-5 for Pt/HZSM-5-F (4.08 ± 1.48 nm), Pt/HZSM-5-U (4.69 ± 0.98 nm), and Pt/HZSM-5-R (5.04 ± 1.95 nm), suggesting the mild growth of Pt NPs in Pt/HZSM-5-F during treatment. The mobility of Pt species was evident from the TEM observations (Supplementary Fig. 17a−l). Moreover, AC-STEM images of the Pt/HZSM-5-U and Pt/HZSM-5-R catalysts were presented in Fig. 2c, d, respectively. Elemental mapping clearly demonstrated the Pt NPs on HZSM-5 surface (Supplementary Fig. 18). XRD showed MFI topology remained after the impregnation of Pt species and a diffraction peak corresponding to Pt(111) showed at 39.8° (Fig. 2e), characterized by a lattice spacing of 0.23 nm (Supplementary Fig. 19). And the signal at 39.8° for Pt/HZSM-5-U and Pt/HZSM-5-R was higher than that for Pt/HZSM-5-F, suggesting that both pretreatments facilitated the growth of Pt(111), consistent with TEM results.Fig. 2Catalyst characterizations.a Schematic illustration of different synthesis methods for Pt/HZSM-5 catalysts. b–d AC-STEM images of HZSM-5 (b), Pt/HZSM-5-U (c), and Pt/HZSM-5-R (d). The standard data of MFI is from the Databases of Zeolite Structures from the International Zeolite Association website (IZA, http://www.iza-structure.org). e XRD patterns. Inset: the enlarge of selected area. f The normalized XANES spectra. g Fitted curves of k2-weighted Pt L3-edge EXAFS spectra for Pt/HZSM-5-U, Pt/HZSM-5-R, Pt/HZSM-5-F, and Pt foil. h EXAFS WT plots for Pt/HZSM-5-U, Pt/HZSM-5-R, Pt/HZSM-5-F, and Pt foil. Catalyst characterizations. a Schematic illustration of different synthesis methods for Pt/HZSM-5 catalysts. b–d AC-STEM images of HZSM-5 (b), Pt/HZSM-5-U (c), and Pt/HZSM-5-R (d). The standard data of MFI is from the Databases of Zeolite Structures from the International Zeolite Association website (IZA, http://www.iza-structure.org). e XRD patterns. Inset: the enlarge of selected area. f The normalized XANES spectra. g Fitted curves of k2-weighted Pt L3-edge EXAFS spectra for Pt/HZSM-5-U, Pt/HZSM-5-R, Pt/HZSM-5-F, and Pt foil. h EXAFS WT plots for Pt/HZSM-5-U, Pt/HZSM-5-R, Pt/HZSM-5-F, and Pt foil. Pt L3-edge X-ray absorption near edge structure (XANES) was performed to show the oxidation state of Pt species31, and the oxidation states of these three Pt/HZSM-5 samples were close to that of Pt foil (Fig. 2f). The white-line peak originates from the transfer of the Pt 2p core-electron to the 5d states, and the gradual decrease in white-line intensity indicates a lower oxidation state of Pt32. Therefore, the white-line intensity gradually decreased in the order of Pt/HZSM-5-F, Pt/HZSM-5-U, Pt/HZSM-5-R, and Pt foil, indicating a more metallic character of the Pt species after the CH4 + N2O reaction or H2 reduction. The Pt coordination was further revealed by extended X-ray absorption fine structure spectroscopy (EXAFS, Fig. 2g and Supplementary Table 2)32,33. The EXAFS fitting of the Pt/HZSM-5-U sample yielded a Pt–Pt coordination number (CN) of 8.01. For Pt/HZSM-5-R, the fitted Pt–Pt CN increased to 8.91. In contrast, Pt/HZSM-5-F showed a lower Pt–Pt CN of 6.90, together with an additional Pt–O contribution (CN = 0.43), indicating the presence of partially oxidized Pt species. Moreover, the wavelet transform (WT) of the k2-weighted EXAFS spectra was shown in Fig. 2h. Pt foil and Pt/HZSM-5 samples exhibited the WT signal at 2.7 Å in R space and 9.8 Å−1 in k space, ascribing to Pt–Pt bond in the first coordination shell, with Pt/HZSM-5-R being the highest intensity among these three samples. These results suggested that both CH4 + N2O pretreatment and H2 reduction promote the growth of Pt nanoparticles and the removal of oxidic Pt species. The CH4 + N2O pretreatment, however, appears to be milder and potentially self-limiting: the Pt particle size gradually approaches a saturation point as the system evolves toward a comparable chemical potential for Pt species34,35. The apparent activation energies for CH4 conversion with N2O over Pt/HZSM-5-U and Pt/HZSM-5-F were determined from Arrhenius analysis to be 65.2 and 94.6 kJ mol–1, respectively (Supplementary Fig. 20), suggesting that the size and oxidation state of Pt NPs played roles in lowering the energy barrier for CH4 activation, thereby affecting the overall catalytic activity. Moreover, it is worth noting that the use of H2PtCl6 as the precursor could raise concerns, since residual chloride is known to affect Pt dispersion, reducibility, and catalytic behaviour36. EXAFS fitting revealed no Pt–Cl scattering path for Pt/HZSM-5-F (Supplementary Table 2)33, indicating the complete removal of chloride after calcination, consistent with ion chromatography results (Supplementary Fig. 21). To elucidate the evolution of chloride during catalyst preparation, samples were collected at various stages of the synthesis process and subjected to EXAFS analysis (Supplementary Fig. 22 and Table 3). The results showed a gradual loss of chloride, concomitant with the formation of Pt–Pt and Pt–O bond. Pt nanoparticles serve as the active sites For Pt/HZSM-5-U, the precursor [PtCl6]2− complex has an estimated size of about 8.28 Å (Supplementary Fig. 23), larger than the ten-membered ring windows of HZSM-5, as observed by the optimum bright-field TEM (OBF-TEM; Supplementary Fig. 24). As a result, most Pt species were expected to form on the external surface rather than migrate into the micropores. It indicated that Pt NPs primarily resided on the outer surface of the zeolite. Nevertheless, we cannot entirely exclude the presence of a minor fraction of Pt species confined within the channels. Such internally dispersed Pt, although limited, could also affect N2O/CH4 activation. Distinguishing the contributions from external and internal Pt locations would provide deeper insight into the structure–activity relationship. Therefore, Pt/HZSM-5-NP and Pt@HZSM-5 were prepared. Theoretical illustrations of Pt locations for Pt/HZSM-5-U, Pt/HZSM-5-NP, and Pt@HZSM-5 are presented in Fig. 3a–c, respectively. In Pt/HZSM-5-U, Pt NPs exhibited a broad size distribution (Fig. 3d). For Pt/HZSM-5-NP, only Pt NPs larger than 5 nm were observed (Fig. 3e). In contrast, Pt species confined within the zeolite matrix were shown in Pt@HZSM-5 (Fig. 3f). EXAFS fitting was conducted to determine the Pt coordination environment (Fig. 3g, h and Supplementary Table 4). Pt/HZSM-5-NP displayed a higher Pt–Pt CN of 9.06 compared to that of Pt/HZSM-5-U (8.01), whereas Pt@HZSM-5 exhibited a lower Pt–Pt CN of 5.15, accompanied by a Pt–O path with a CN of 1.26, likely arising from slight oxidation of Pt clusters or the presence of isolated Pt sites anchored to framework oxygen atoms.Fig. 3Microscopic and spectroscopic identification of active sites on Pt/HZSM-5.a–c Schematic illustration for Pt/HZSM-5-U (a), Pt/HZSM-5-NP (b), and Pt@HZSM-5 (c). d–f AC-STEM images of Pt/HZSM-5-U (d), Pt/HZSM-5-NP (e), and Pt@HZSM-5 (f). g–i Fitted curves of k2-weighted Pt L3-edge EXAFS spectra for Pt/HZSM-5-U (g), Pt/HZSM-5-NP (h), and Pt@HZSM-5 (i). Microscopic and spectroscopic identification of active sites on Pt/HZSM-5. a–c Schematic illustration for Pt/HZSM-5-U (a), Pt/HZSM-5-NP (b), and Pt@HZSM-5 (c). d–f AC-STEM images of Pt/HZSM-5-U (d), Pt/HZSM-5-NP (e), and Pt@HZSM-5 (f). g–i Fitted curves of k2-weighted Pt L3-edge EXAFS spectra for Pt/HZSM-5-U (g), Pt/HZSM-5-NP (h), and Pt@HZSM-5 (i). Further, propane dehydrogenation (PDH) was employed as a probe reaction to distinguish Pt species located on the external surfaces of the zeolite from those confined within the matrix in Pt/HZSM-5, as PDH has been widely recognized as a structure-sensitive reaction to Pt species37,38. As an initial evaluation, we tested a blank quartz-sand-filled reactor and a commercial UOP (Universal Oil Products) catalyst (after pellet crushing) to confirm the reliability of our setup. The blank reactor with only quartz-sand filler showed neglectable activity (Supplementary Fig. 25). Severe coke formation was observed during PDH for the UOP catalyst, leading to a rapid decline in conversion with time on stream (Supplementary Fig. 26), consistent with previous report39. Thus, the catalytic activity after 15 min of PDH reaction was used for comparison (Supplementary Fig. 27). Pt@HZSM-5 exhibited high catalytic performance in PDH, with a C3H8 conversion of 39.8% and a C3H6 selectivity of 89.7%, characteristic of Pt species confined within the zeolite matrix. In contrast, Pt/HZSM-5-U showed inferior activity compared to Pt@HZSM-5, but displayed performance comparable to Pt/HZSM-5-NP, consistent with the AC-STEM and EXAFS results. The catalytic performance was shown in Fig. 1e. The Pt/HZSM-5-NP catalyst, prepared directly from pre-synthesized metallic Pt nanoparticles, exhibited activity comparable to Pt/HZSM-5-U. In contrast, using a smaller ionic size precursor, Pt(NO3)2, led to internal incorporation of Pt species, and these confined Pt sites showed no measurable activity below 200 °C. Moreover, Pt/CeO2-U, which exhibited a T90 of 150 °C (Supplementary Fig. 9), also showed well-defined Pt NPs dispersed on the CeO2 support (Supplementary Fig. 28). These results confirmed that metallic Pt NPs were the active sites for low-temperature CH4-SCR of N2O, whereas Pt species located inside the zeolite matrix remained as spectator sites. Although Pt/HZSM-5-U and Pt/HZSM-5-NP exhibited comparable activity, Pt/HZSM-5-U was considered as the priority for industrial applications. All raw materials are commercially available, and the wet impregnation method is widely used in industry; thus, no major barriers are expected for large-scale catalyst manufacturing26. Moreover, the observed activity trend of Pt/HZSM-5-U > Pt/HZSM-5-U-50 » Pt/HZSM-5-U-200 (Fig. 1d) motivated a detailed investigation into the role of zeolite acidity. Acidic HZSM-5 facilitates hydrogen spillover (proton mobility) during the reaction, which can promote CH4 activation and dehydrogenation processes28,40. To directly verify the acidity, NH3-TPD and pyridine-IR were conducted. In NH3-TPD profiles (Supplementary Fig. 29a), the desorption peak below 350 °C corresponds to weak acid sites, while the high-temperature peak above 350 °C corresponds to strong acid sites. In pyridine-IR spectra (Supplementary Fig. 29b−d), the bands at around 1450 and 1540 cm−1 are assigned to Lewis and Brønsted acid sites, respectively41. The desorption at 200 °C represents total acidity, while that at 350 °C reflects strong acid sites. The NH3-TPD and pyridine-IR quantitative results (Supplementary Table 5) showed that the acidity decreased progressively with increasing Si/Al ratio, and Pt/HZSM-5-U-200 exhibited the lowest acidity and correspondingly the lowest activity, whereas Pt/HZSM-5-U-50 and Pt/HZSM-5-U-27 displayed comparable acidity levels. In addition, it has been widely reported that acidic HZSM-5 zeolites facilitate the formation of metallic Pt species42–44. Pt L3-edge XANES measurements were performed on Pt/HZSM-5-U samples with different Si/Al ratios (Supplementary Fig. 30a). The spectra of all samples closely resembled that of Pt foil, indicating that Pt predominantly exists in the metallic state. However, the white-line intensity of the sample with a Si/Al ratio of 200 was slightly higher, suggesting a relatively higher oxidation state of Pt species. Further EXAFS fitting analysis revealed that decreasing acidity led to changes in the Pt coordination environment (Supplementary Fig. 30b and Table 6). CO diffuse reflectance infrared Fourier-transform spectroscopy (DRIFTS) results (Supplementary Fig. 31a) showed a characteristic band at 2090 cm−1, which is assigned to linearly adsorbed CO on metallic Pt NPs45. In contrast, CO adsorbed on electron-deficient oxidized Ptδ+ species typically appeared at higher wavenumbers due to weakened back-donation from Pt to the CO antibonding orbital. Compared with Pt/HZSM-5-U, the CO adsorption band for Pt/HZSM-5-U-200 shifted by 5 cm−1 to higher wavenumbers, indicating a little oxidized character of Pt species, consistent with the XANES results. Furthermore, HR-TEM of Pt/HZSM-5-U-200 (Supplementary Fig. 31b) showed that the average Pt particle size decreased, while the overall particle size distribution became broader, with more pronounced coexistence of small and large particles. It suggested that stronger acidity promoted a more uniform redistribution of Pt nanoparticles during the CH4 + N2O reaction-induced pretreatment. Smaller particles were more prone to be oxidized, which provided a structural basis for the XANES and CO-DRIFTS results. To further decouple the effect of acidity from Pt species, Na-exchanged ZSM-5 was prepared via repeated ion exchange, and Pt/NaZSM-5-NP was synthesized using the same method as Pt/HZSM-5-NP. Under identical reaction conditions (2.5% CH4, 10% N2O, 12000 ml gcat−1 h−1), Pt/NaZSM-5-NP exhibits slightly lower activity than Pt/HZSM-5-NP (Supplementary Fig. 32), but significantly higher activity than Pt/HZSM-5-U-200. Acid characterization (Supplementary Fig. 33 and Table 5) showed that the total Brønsted acidity of Pt/NaZSM-5-NP was only 39.3% of Pt/HZSM-5-U-200. Although a positive correlation between catalytic activity and zeolite acidity was observed among the HZSM-5-supported catalysts (Fig. 1d), the experimental evidence suggests that Brønsted acidity alone cannot account for the observed activity differences. In particular, the comparison between Pt/HZSM-5-NP and Pt/NaZSM-5-NP indicates that the nature of Pt species is the key factor, while acidity plays an assisting role. However, the substantially larger activity differences observed among catalysts containing distinct Pt species (i.e., nanoparticles located outside the ZSM-5 channels and Pt species confined within the channels) indicate that the nature of Pt species is the primary factor governing catalytic performance. Brønsted acid sites mainly function as promoters by facilitating hydrogen spillover, thereby enhancing the consumption of surface oxygen species and regenerating active sites for the subsequent reaction cycle. Therefore, the superior activity of Pt/HZSM-5-U arises from the synergistic contributions of Pt nanoparticle morphology/electronic state and Brønsted acid site-promoted hydrogen spillover. Mechanistic insights into evolution of Pt nanoparticles CH4 is an inert molecule with a high C–H bond energy and low polarizability, resulting in a significant challenge for its efficient activation below 150 °C over most heterogeneous catalysts16,46,47. Interestingly, N2O is known as a milder oxidant compared to O248, but why N2O was more reactive in CH4 activation on metallic Pt species than O2 and why it can achieve such a superior low-temperature activity? To figure it out, we carried out ex-situ, quasi-in situ, and in-situ characterizations to identify the differences between CH4 + N2O (2.5% CH4, 10% N2O, He balance) and CH4 + O2 (2.5% CH4, 5% O2, He balance) with the same CH4/O ratio at 150 °C. Quasi-in situ AC-STEM was applied to compare the states of Pt NPs on Pt/HZSM-5-U after CH4 + N2O and CH4 + O2 reactions. After CH4 + N2O, the Pt NP remained unchanged, exhibiting a lattice spacing of 0.23 nm corresponding to the Pt(111) plane (Fig. 4a). This was further supported by quantitative analyses, including FFT (fast Fourier transform) patterns and profile intensity measurements. In contrast, after CH4 + O2, although Pt NP still mainly displayed the Pt(111) structure, the gas–solid interface underwent a transformation, giving rise to a monolayer oxide with an interspacing of 0.26 nm (Fig. 4b), indicative of a PtOx layer formed on the Pt(111) surface49–52. Moreover, XRD of the Pt/HZSM-5-U sample after CH4 + O2 reaction showed a comparable intensity of band at 39.8°, but a little shift towards lower degree and a decreased intensity were observed (Supplementary Fig. 34), indicating the lattice expansion due to the surface oxidation. This was further supported by H2 temperature-programmed reduction (TPR) in Fig. 4c, which showed an increased peak at 337 °C (H2 consumption by surface PtOx species) after CH4 + O2 reaction53.Fig. 4Evolution of Pt nanoparticles.a, b Quasi-in situ AC-STEM images, corresponding FFT pattern and profile intensity of Pt NPs supported on Pt/HZSM-5-U after CH4 + N2O (a) and after CH4 + O2 (b). c H2 temperature-programmed reduction of Pt/HZSM-5-U after CH4 + N2O and CH4 + O2 reaction. d Fitted curves of k2-weighted Pt L3-edge quasi-in situ EXAFS spectra for Pt/HZSM-5-U after CH4 + O2. e CO-DRIFTS spectra of Pt/HZSM-5-U after in situ CH4 + N2O and CH4 + O2 reaction. f Operando NAP-XPS Pt 4 d spectra of Pt/HZSM-5-U in CH4 + N2O and CH4 + O2 reaction. “a.u.” denotes arbitrary units throughout this study. Evolution of Pt nanoparticles. a, b Quasi-in situ AC-STEM images, corresponding FFT pattern and profile intensity of Pt NPs supported on Pt/HZSM-5-U after CH4 + N2O (a) and after CH4 + O2 (b). c H2 temperature-programmed reduction of Pt/HZSM-5-U after CH4 + N2O and CH4 + O2 reaction. d Fitted curves of k2-weighted Pt L3-edge quasi-in situ EXAFS spectra for Pt/HZSM-5-U after CH4 + O2. e CO-DRIFTS spectra of Pt/HZSM-5-U after in situ CH4 + N2O and CH4 + O2 reaction. f Operando NAP-XPS Pt 4 d spectra of Pt/HZSM-5-U in CH4 + N2O and CH4 + O2 reaction. “a.u.” denotes arbitrary units throughout this study. To further confirm the formation of a PtOx layer upon O2 exposure, in situ aberration-corrected environmental scanning transmission electron microscopy (ESTEM, Hitachi HF5000) was employed54,55. Initially, the temperature was increased to 200 °C, and the Pt/HZSM-5-U sample was exposed for 60 min to completely remove residual ethanol from sample preparation. Oxygen was introduced to a pressure of 4.5 Pa. Under such condition for 5 min, the Pt NP supported on HZSM-5 exhibited the Pt(111) feature with a lattice spacing of 0.23 nm (Supplementary Fig. 35a). Further exposure to the oxidizing environment for 90 min induced surface variations (partially highlighted by the red box in Supplementary Fig. 35b), whereas other regions of the Pt NP remained metallic. These surface changes were characterized by lattice expansion associated with oxygen incorporation, leading to an interplanar spacing increase to 0.26 nm and indicating the formation of a PtOx layer. The coordination environment of Pt NPs in CH4 + O2 was examined by quasi-in situ EXAFS (Fig. 4d and Supplementary Table 7). In this procedure, the catalyst was rapidly collected immediately after reaction, evacuated, stored under liquid-nitrogen conditions to preserve its chemical state56, and then measured. After the CH4 + O2 reaction, the Pt–Pt CN decreased to 6.53 and a Pt–O path emerged with a CN of 1.15. For a Pt NP with a diameter of 5 nm and a face-centered cubic (fcc) structure, approximately 4388 Pt atoms are present, of which 992 are surface atoms (22.6%; Supplementary Fig. 36). The observed decrease in Pt–Pt CN together with the appearance of Pt–O coordination is therefore consistent with oxidation occurring predominantly at the outer surface of the Pt NPs. CO-DRIFTS measurements were further performed in an in-situ IR cell after reaction to distinguish changes in Pt species (Fig. 4e)45. After the in situ CH4 + N2O reaction, only the CO adsorption band on metallic Pt NPs was observed. In contrast, bands attributable to Ptδ+(CO) and Ptδ+(CO)n emerged after the in situ CH4 + O2 reaction, indicating partial oxidation of the surface of Pt NPs. Operando near-ambient-pressure X-ray photoelectron spectroscopy (NAP-XPS) of the Pt 4d region was carried out to determine the Pt valence state after the in situ reactions. Because Pt 4f overlaps with the Al 2p signal in HZSM-5 (an Al-containing zeolite)57, which makes peak deconvolution subjective, the Pt 4d region was selected for analysis instead (Fig. 4f). After the CH4 + N2O reaction, the Pt 4d signal remained essentially identical to that of the metallic Pt in Pt/HZSM-5-U. In contrast, after the CH4 + O2 reaction, the Pt 4d peak shifted by 1.1 eV toward higher binding energy, consistent with partial oxidation of the Pt surface. Reaction intermediates and computational studies Operando NAP-XPS of the O 1s region was performed to probe the evolution of oxygen species after the CH4 + N2O and CH4 + O2 reactions. For clarity, the binding energy peaks and their assignments are summarized in Supplementary Table 8. For HZSM-5, an oxygen-rich support, the O 1s peak has been widely studied and typically appears at around 532 eV, corresponding to Si–O–Si and Si–O–Al linkages (Supplementary Fig. 37). The higher-binding-energy component (Ohydro, above 532.6 eV) is mainly associated with Si–OH species58. Upon cofeeding CH4/N2O (1/4) to the Pt/HZSM-5-U catalyst, the O 1s peak shifted toward higher binding energy, accompanied by an increased intensity above 532.6 eV (Fig. 5a), indicating an increase in Ohydro species, likely arising from Pt–OH structures. In contrast, cofeeding 2.5% CH4 and 5% O2 resulted in an O 1s peak with an increased full width at half maximum (FWHM), allowing for a more pronounced Oα component at around 530 eV (Fig. 5b). The peak below 531 eV is attributed to lattice oxygen coordinated to transition metals59, which could also support the formation of a surface PtOx layer. Identical NAP-XPS experiments on bare HZSM-5 (without Pt) under CH4 + N2O and CH4 + O2 conditions were conducted (Supplementary Fig. 38), and neither showed observable spectral changes compared to pristine HZSM-5, indicating that the variation observed for Pt/HZSM-5-U is not primarily due to changes in the hydration state of the support, but is associated with the presence of Pt species.Fig. 5Reaction intermediates and computational studies.a, b Operando NAP-XPS O 1s spectra of Pt/HZSM-5-U in CH4 + N2O reaction (a) and CH4 + O2 reaction (b). c EPR spectra of spin-trapping evidence for radicals captured by DMPO in different reactions. d, e In situ DRIFTS hydroxy-region spectra of Pt/HZSM-5-U in CH4 + N2O reaction (d) and CH4 + O2 reaction (e). f In situ DRIFTS spectra of Pt/HZSM-5-U in CH4 + N2O reaction. g Potential energy diagram for the CH4 activation over different Pt-based surfaces. TS, transition state. h, i Schematic illustration of reaction mechanisms in low-temperature CH4 activation with N2O (h) and O2 (i) on Pt/HZSM-5-U catalyst. Reaction intermediates and computational studies. a, b Operando NAP-XPS O 1s spectra of Pt/HZSM-5-U in CH4 + N2O reaction (a) and CH4 + O2 reaction (b). c EPR spectra of spin-trapping evidence for radicals captured by DMPO in different reactions. d, e In situ DRIFTS hydroxy-region spectra of Pt/HZSM-5-U in CH4 + N2O reaction (d) and CH4 + O2 reaction (e). f In situ DRIFTS spectra of Pt/HZSM-5-U in CH4 + N2O reaction. g Potential energy diagram for the CH4 activation over different Pt-based surfaces. TS, transition state. h, i Schematic illustration of reaction mechanisms in low-temperature CH4 activation with N2O (h) and O2 (i) on Pt/HZSM-5-U catalyst. Moreover, spin-trapping reaction of DMPO for electron paramagnetic resonance (EPR) experiments were carried out to further validate the formation of −OH structure during the CH4 + N2O reaction (Fig. 5c)60. Introduction of CH4 (10 mbar) and N2O (40 mbar) with the spin-trapping agent DMPO at 150 °C on the Pt/HZSM-5-U catalyst led to the formation of a DMPO/∙OH adduct, characteristic of a four-line spectrum with relative heights of 1:2:2:1. It might origin from the desorption of hydroxy group on the Pt−OH structure. In contrast, when substituting the oxidant from N2O to O2, no signal was observed. To further investigate the difference of N2O and O2 for CH4 conversion on Pt NPs, in situ DRIFTS experiments were carried out to probe the surface species on the Pt/HZSM-5-U catalyst at 150 °C. For the CH4 + N2O reaction in Fig. 5d, in addition to the band at 3737 cm–1 assigned to Si−OH, peaks at 3643 and 3597 cm−1 appeared, corresponding to monomer H2O and the hydroxyl group (Pt−OH), respectively61. The hydrogen atoms in H2O and Pt−OH likely originated from CH4 dissociation on Pt NPs. In contrast, for the CH4 + O2 reaction, only the Si−OH band was observed (Fig. 5e), and the Pt−OH structure was absent, suggesting no hydrogen was supplied and indicating that CH4 activation on the surface PtOx layer was inhibited. This was further identified by density functional theory (DFT) calculations. For N2O activation, such surface-adsorbed monatomic O* species were formed on metallic Pt sites, exhibiting a small energy barrier of 0.39 eV (Supplementary Fig. 39). But it did not mean Pt sites served as a competitive candidate for N2O direct decomposition62, due to the challenges posed by O* recombination (1.76 eV) and O2 desorption (0.86 eV) in Supplementary Fig. 40. Meanwhile, the Hdiss* atoms were widely present on the surface, serving as reactants to form OH* upon combining with an adsorbed O* species, leading to the consumption of OH* and the formation of H2O (Supplementary Fig. 41). The Hdiss* species readily reacted with an O* site to form OH* with an energy barrier of 0.43 eV. Then, Hdiss* species effectively reacted with OH* to form H2O* with a low energy barrier of 0.33 eV. The desorption of water from the Pt(111) surface occurred easily with a low adsorption energy of 0.57 eV, refreshing the active sites for the next reaction cycle. As a result, the coverage of O* may vary over the course of reaction (Supplementary Fig. 42), given that its formation and consumption occur simultaneously. When CH4 and N2O were introduced into the IR chamber, both molecules were spontaneously adsorbed onto the surface of Pt/HZSM-5-U. The DRIFTS vibrational bands and their assignments are provided in Supplementary Table 9. At 100 °C, only the characteristic features of CH4 (3014 and 1304 cm−1) and N2O (2237 and 2204 cm−1)63 were observed immediately after introducing the reactants (Supplementary Fig. 43). The peaks at 3014 and 1304 cm–1 correspond to the antisymmetric stretching νas(C–H) and bending δ(C–H) modes of adsorbed CH4, which rapidly disappeared once the CH4 feed was cut off. Upon increasing the temperature to 150 °C (Fig. 5f), bands assigned to the C = O stretching of adsorbed CO2 at 2360 and 2337 cm−1 were also detected immediately, indicating that the reaction proceeds rapidly. Meanwhile, the intensities of other intermediate-related features became stronger, including bands at 1039, 1200, 1304 (δ(C–H)), 1342, and 1616 cm–1. The bands at 1000–1200 cm–1 are attributed to ν(C–O) stretching vibrations of surface methoxy (CH3O*) species64. A broad feature centered at 1342 cm–1 emerged, which is assigned to the C–H stretching vibration of adsorbed CH2O* or HCO* species. In addition, the peak at 1616 cm–1 corresponds to δ(OH) bending signals. It suggested that these surface species served as key reaction intermediates. In contrast, when CH4 and O2 were introduced at 150 °C, only CH4 bands at 3014 and 1304 cm–1 were detected (Supplementary Fig. 44). Upon further heating to 400 °C, CO2 was formed, accompanied by the emergence of intermediate-related bands in the 1000–1700 cm−1 region. This indicated that CH4 oxidation proceeded through a similar reaction pathway when O2 was used as the oxidant, analogous to the case with N2O. EXAFS of the Pt/HZSM-5-U sample after CH4 + O2 reaction at 400 °C revealed the oxidation of Pt NPs (Supplementary Fig. 45 and Table 7), demonstrating that the active Pt species shifted from Pt0 NPs to Ptδ+ (PtOx) when the oxidant changes from N2O to O2. Consequently, a higher activation barrier must be overcome, consistent with the requirement of higher temperatures for CH4 oxidation when O2 serves as the oxidant. It has been widely reported that the rate-determining step (RDS) of CH4 oxidation over most heterogeneous catalysts is the initial C–H bond activation, whereas the subsequent oxidation steps proceed rapidly and are difficult to control65. Moreover, the activation of methanol has been shown to occur readily on Pt-based catalysts at low temperatures (below 150 °C)66. We therefore infer that the initial activation of CH4 to a surface-bound methoxy species represents the RDS; once this step is overcome, subsequent oxidation steps proceed easily, which makes the overall reaction feasible below 150 °C. To verify that methanol indeed served as a key intermediate in the CH4 + N2O reaction, we adopted a direct strategy by introducing CH3OH into the reaction system as a reductant. The conversion of CH3OH by N2O was observed at 150 °C (Supplementary Fig. 46), supporting its role as a reaction intermediate. Therefore, the key to the low-temperature activity of the CH4 + N2O reaction on Pt/HZSM-5-U lies in the fact that the special oxygen species generated on Pt NPs upon N2O activation can effectively promote the first C–H bond cleavage of CH4. We attribute the oxygen species to monoatomic adsorbed oxygen species (O*, where * denotes an adsorbed state) upon N2O activation, where the N ≡ N–O molecule undergoes N–O bond scission to generate adsorbed O* and desorbed N2. Because N2O is a “softer” oxidant than O2, the O* generated from N2O is not sufficiently strong to oxidize the Pt surface to form PtOx. Furthermore, the surface reactivity of the monoatomic O* induced by N2O (mono-oxygen adsorbate) and the PtOx layer formed by O2 was compared by DFT calculations. The energy diagram for CH4 activation involving a hydrogen atom abstraction reaction (Fig. 5g and Supplementary Fig. 47) revealed that after the C–H bond cleavage, a hydrogen atom from CH4* was transferred to a nearby O atom, forming dissociated H* (denoted as Hdiss*) and CH3* on the monoatomic O* and PtOx layer. The released H atom spilled over to the neighboring O atom to form OH* species67, here, the Pt−OH structure. The results showed that the energy barriers for C–H activation were 0.51 eV and 0.99 eV on the monoatomic O* and PtOx layer model, respectively. This suggested that CH4 was more readily activated on the mono-oxygen model. Moreover, the geometric structures indicated that during methane activation, CH3OH* was formed and dehydrogenated to CH3O*, which aligned with in situ DRIFTS. To simulate the Pt sites after initial N2O activation, we constructed a Pt–O bond model with one adsorbed O* species on the Pt(111) surface. The C–H bond elongated toward the O atom of the Pt–O bond, forming a CH3OH intermediate (Ea = 0.53 eV), which was similar to the value for the monoatomic O* model. In contrast, the direct CH4 dissociation barrier for CH4 → H* + CH3* on bare Pt(111) was 0.58 eV (Supplementary Fig. 48), consistent with previously reported results68, which was higher than that on the monoatomic O*. This suggested that neither bare Pt NPs nor PtOx layer can effectively activate CH4, whereas Pt NPs with adsorbed surface oxygen can. The charge density calculations revealed an electron-rich surface for the monoatomic O* model, whereas the PtOx surface exhibited electron deficiency (Supplementary Fig. 49). Electrons from the Pt phase were trapped by the inner oxygen layer, acting as an electron isolation layer. The different electronic densities might be a key factor for the different activity. Next, we calculated the complete reaction pathway for CH4-SCR on the Pt monoatomic O* (Supplementary Fig. 50). The essence of the monoatomic O* model lies in the presence of sufficient mono-oxygen species to facilitate the reaction, and thus, for simplicity, we focused on the Pt–O bond in this simulation. After CH4 activation and the migration of a hydrogen atom to a nearby O* species, CH3O* intermediate and OH* were formed with a barrier energy of 0.34 eV. Following subsequent hydrogen atom abstractions, the HCO* species was formed. Then, HCO* reacted with a neighboring OH*, forming a *HCOOH intermediate (Ea = 0.05 eV). From an electron structure perspective, the disparity in electron density between the C in *HCO and the O in *OH allowed them to readily bind to each other (Supplementary Fig. 51), forming the *HCOOH intermediate. The subsequent reaction HCOOH* + O* → HCOO* + OH* has a low energy barrier of 0.17 eV. The subsequent reaction HCOO* + O* → CO2* + OH* also readily occurred, given a low energy barrier of 0.44 eV. Finally, CO2* desorbed (Eads = −0.43 eV), completing the entire reaction. We summarized the reaction mechanisms. For CH4 + N2O (Fig. 5h), N2O is a relatively mild oxidizing agent, as revealed by projected crystal orbital Hamiltonian population (pCOHP; Supplementary Text 5 and Supplementary Fig. 52), making it challenging to form PtOx layer. Thus, Pt sites existed as monoatomic oxygen species then monoatomic O* upon N–O dissociation, enabling efficient methane activation below 150 °C. CH3OH intermediates were formed and further dehydronated to CH3O*. The consumption and formation of the Pt−OH structure reached a dynamic equilibrium in the combined reduction and oxidation atmosphere, achieving a rapid O–H consumption cycle. In contrast, for CH4 + O2 (Fig. 5i), the strong reactivity of O2 induced the surface phase reconstruction into a PtOx layer, which was unfavorable for low-temperature CH4 activation, resulting in a low activity. Application-potential assessments of CH4-SCR-150 After elucidating the mechanism of CH4-SCR-150, it is crucial to transition from fundamental studies to practical evaluations to evaluate its industrial viability due to the urgent abatement of industrial sources of N2O (Supplementary Text 6). Under simulated practical conditions (40% N2O, 10% CH4, 5% H2O vapor, 4% CO2, 0.1% O2, with N2 balance, 60 °C), the Pt/HZSM-5-U catalyst achieved 100% N2O conversion at 125 °C with a gas velocity of 12,000 ml g–1 h–1. Space velocity tests showed that even at 120,000 ml g–1 h–1, the catalyst maintained 100% N2O conversion below 150 °C (Fig. 6a). At 150 °C and 120,000 ml g–1 h–1, a Pt loading of 0.8% was sufficient to achieve a maximum reaction rate of 9629.55 kgN2O kgPt−1 h−1 (Fig. 6b). More importantly, no NO, NO2, CO, or hydrocarbons were detected, achieving 100% selectivity for N2 and CO2 without any measurable CH4 escape. Moreover, higher water vapor contents were introduced to evaluate the water tolerance of the catalyst and to identify suitable operating humidity conditions that broaden its practical applicability. Even at 10% H2O, the catalyst maintained stable performance, indicating good moisture resistance; however, 20% H2O caused rapid deactivation (Supplementary Fig. 53). In practical applications, the moisture stability of Pt/HZSM-5-U is generally sufficient. Specifically, in adipic acid plants, the water vapor content in the tail gas is typically below 5%69. For exhaust streams involving an NH3-washing step (Supplementary Fig. 54), such as those from caprolactam and nitric acid production, the moisture content is often close to saturation; however, condensation and heat-exchange operations can effectively reduce the temperature and, according to the Antoine equation, lower the water content. Besides, low water contents could greatly reduce the requirements for equipment maintenance. Furthermore, stability tests using simulated tail gas revealed no activity loss over 168 h (Fig. 6c), reinforcing the reliability of the catalyst for long-term operation.Fig. 6Comprehensive evaluation of industrial N2O abatement: large-scale application potential and techno-economic assessment of CH4-SCR-150 technology.a–c Conversion curves under practical conditions for varying gas hourly space velocity on 1.0 wt.% Pt/HZSM-5-U (a); different Pt loadings on Pt/HZSM-5-U at GHSV of 120,000 ml gcat–1 h–1 (b); long-term stability evaluation of 0.8 wt.% Pt/HZSM-5 at 150 °C and 120,000 ml gcat–1 h–1 (c). Practical reaction conditions: unless otherwise specified, 40% N2O, 10% CH4, 5% H2O vapor, 4% CO2, 0.1% O2, N2 balance, GHSV of 120,000 ml gcat–1 h–1. d Operating fee comparison of 1 kg N2O abatement considering the electricity and CH4 price, where negative values mean cost and positive values mean profit. e Overall cost comparison of 1 kg N2O abatement between N2O-DD-400 and CH4-SCR-150. f Carbon profit of 1 kg N2O abatement considering future carbon price. Comprehensive evaluation of industrial N2O abatement: large-scale application potential and techno-economic assessment of CH4-SCR-150 technology. a–c Conversion curves under practical conditions for varying gas hourly space velocity on 1.0 wt.% Pt/HZSM-5-U (a); different Pt loadings on Pt/HZSM-5-U at GHSV of 120,000 ml gcat–1 h–1 (b); long-term stability evaluation of 0.8 wt.% Pt/HZSM-5 at 150 °C and 120,000 ml gcat–1 h–1 (c). Practical reaction conditions: unless otherwise specified, 40% N2O, 10% CH4, 5% H2O vapor, 4% CO2, 0.1% O2, N2 balance, GHSV of 120,000 ml gcat–1 h–1. d Operating fee comparison of 1 kg N2O abatement considering the electricity and CH4 price, where negative values mean cost and positive values mean profit. e Overall cost comparison of 1 kg N2O abatement between N2O-DD-400 and CH4-SCR-150. f Carbon profit of 1 kg N2O abatement considering future carbon price. To further assess real-world applicability under industrially relevant conditions, shaped catalysts were prepared by calcining extrudates produced from catalyst powder using clay as a binder (Supplementary Fig. 55). The shaped catalyst displayed activity comparable to the powder counterpart under identical conditions. Under fluctuating N2O/CH4 ratios, elevated oxygen and water vapor contents, and partial-conversion conditions, the shaped catalyst showed reliable long-term stability (Supplementary Fig. 56), indicating that the proposed technology is industrially relevant. Before practical application, the variations in gas composition and the presence of additional components such as NOx, CO, and volatile organic compounds in specific emission scenarios should be considered. In addition to the technology’s low-temperature activity and stability, its real-world implementation should be economically and environmentally viable, particularly in terms of catalyst scale-up and the use of CH4 as a reductant. To evaluate the feasibility of real-world implementation, technical–economic analysis (TEA) was performed based on adipic acid tail-gas streams. The application potential of CH4-SCR-150 over the Pt/HZSM-5-U catalyst was benchmarked against the currently employed N2O-DD-400 (industrial pilot-scale data), for urgent N2O abatement from adipic acid production (Supplementary Text 7 and 8). The price of Pt/HZSM-5-U (Supplementary Fig. 57) and Fe/Beta (Supplementary Fig. 58) was calculated using the CatCost software70, showing that raw materials are the primary driver of catalyst cost. Considering the catalyst amount and lifetime, the catalyst costs are $0.109 and $0.227 per kg N2O for CH4-SCR-150 (3 years) and N2O-DD-400 (2 years), respectively (Supplementary Table 10). We then calculated direct operating cost/benefit for abating 1 kg of N2O, considering three key components: catalyst cost, energy/electricity consumption & heat recovery, and methane cost (Supplementary Text 7 and Table 11). Operating costs were evaluated based on electricity and methane prices. Sensitivity analyses revealed that CH4-SCR-150 consistently displayed lower operating costs than N2O-DD-400 under reasonable electricity and methane price scenarios (Fig. 6d). The analysis indicated that higher electricity prices and lower CH4 prices made CH4-SCR-150 more economical. Thus, the economic advantage of CH4-SCR-150 was especially notable in regions with highly competitive methane prices, such as the United States. O2 removal unit was also considered in the TEA for CH4-SCR-150 (Supplementary Text 8). Finally, we calculated the overall cost (Supplementary Tables 12 and 13), consisting of fixed costs, operating costs, capital costs, and other expenses71. As a result, CH4-SCR-150 reduced overall costs by 62.3% compared to existing N2O-DD-400 for industrial adipic acid emissions in China (Fig. 6e). This substantial cost reduction not only alleviated policy-driven pressures for N2O abatement but also provided strong economic incentives for industries to adopt widespread sustainable practices. Incorporating carbon price significantly boosts future economic appeal of CH4-SCR-150 (Fig. 6f and Supplementary Table 14). The high global warming potential of N2O results in substantial carbon credit revenues, corresponding to a 99.9% CO2e (e denotes equivalent) reduction in environmental benefits for this technology (Supplementary Text 9). Even so, given the reaction CH4 + 4N2O = 4N2 + CO2 + 2H2O, a 100% conversion of N2O at a rate of 9629.55 kgN2O kgPt−1h−1 would correspond to the formation of approximately 2407.39 kgCO2kgPt−1h−1, which is non-negligible but can be well mitigated72. For each kg of N2O mitigated, this process generates an estimated carbon price profit of 0.273 t CO2e for N2O-DD-400 and 0.2728 t CO2e for CH4-SCR-150. Utilizing methane from landfill gas or coal-bed methane exhaust further enhances the revenue potential through additional carbon reduction benefits to 0.2753 t CO2e. Sensitivity analysis showed that the technology’s profitability remained largely unaffected by methane transportation costs (Supplementary Fig. 59), confirming its viability across various operational scenarios. As a result, a profit of 0.2753 t CO2e is earned, providing a strong economic incentive for promoting this technology even in the absence of regulations or incentives. Discussion In summary, we reported a breakthrough technology enabling 99.99% N2O conversion and 100% N2 selectivity at temperatures below 150 °C. This is accomplished using CH4 as a reductant at a CH4/N2O ratio of 1/4 with a Pt/HZSM-5 catalyst, delivering an exceptional reaction rate of 9629.55 kgN2O kgPt−1 h−1 with no detectable CH4 escape. Mechanistic studies reveal that Pt nanoparticles are the active component, while a rapid O* formation/consumption cycle is proposed. Techno-economic analysis highlights the potential of this approach to reduce industrial N2O abatement costs by at least 62.3% for industrial adipic acid emissions in China, supporting sustainable climate mitigation solutions. Beyond its immediate application, our findings have important implications for the low-temperature activation of methane, providing insights into how methane can be effectively utilized in N2O abatement. Additionally, this work offers valuable perspectives on various industrial sources, including adipic acid, caprolactam, and nitric acid production, and other potential synergistic removal strategies in future scenarios where CH4 and N2O co-exist, such as in agriculture and wastewater treatment. Moving forward, this technology could also inspire applications utilizing CH4 from natural gas or shale gas as an alternative fuel for methane-powered vehicles or electricity generation, enabling ultra-low environmental impacts and paving the way for sustainable energy solutions. Methods Chemical and materials All chemicals were purchased from Aladdin Chemical Reagent Company in analytical reagent (AR) grade. HZSM-5 zeolites (MFI topology) were obtained from Nankai University Catalyst Company and Zeolyst International40. All gases utilized in this study were sourced from Hangzhou Jingong Special Gases Co., Ltd. Experimental solutions were prepared using deionized (DI) water. Note on safe operation of experiments Any reaction involving catalytic oxidation must take care to work under conditions outside of the explosive mixture composition of the reagents. This is the case for all our experiments. Synthesis of Pt/HZSM-5-F Commercial HZSM-5 was directly used as a support without further treatment. The commercial HZSM-5 zeolite supplied from Nankai University Catalyst Company and Zeolyst International has been widely used as a support40. The 1.0 wt.% Pt/HZSM-5 catalysts were all synthesized by wet impregnation method using commercial zeolites. Briefly, 1 g of commercial HZSM-5 zeolite was dispersed in 100 ml of an aqueous H2PtCl6·6H2O solution (0.10 mgPt ml−1). The suspension was sonicated for 30 min and stirred at 40 °C for 6 h. The resulting mixture was dried at 110 °C overnight and subsequently calcined in static air at 500 °C for 4 h with a heating rate of 5 °C min−1. The obtained catalyst was denoted as Pt/HZSM-5-F, where “F” indicates the fresh catalyst. Note: if not specified, the Si/Al ratio of HZSM-5 is 27. Synthesis of Pt/HZSM-5-U The Pt/HZSM-5-U catalyst was obtained by pretreating the Pt/HZSM-5-F catalyst in 2.5% CH4 + 10% N2O at 280 °C for 4 h with a heating rate of 5 °C min−1. Other catalysts marked by -U were prepared following the same procedure, except changing the metal precursor or support. Synthesis of Pt/HZSM-5-R The Pt/HZSM-5-R catalyst was obtained by pretreating the Pt/HZSM-5-F catalyst in 5% H2/Ar at 500 °C for 4 h with a heating rate of 5 °C min−1. Other catalysts synthesis is presented in Supplementary Text 1. Catalytic performance evaluation The catalytic activity was performed in a continuous-flow fixed-bed reactor, with thermocouples placed at both the catalyst bed center and the furnace wall. In a typical experiment, 500 mg of sieved sample (40–60 mesh) was loaded with gas hourly space velocity (GHSV) of 12,000 ml·gcat−1·h−1. The catalytic performances of various catalysts for the CH4 + N2O reaction were investigated with a stoichiometric ratio of CH4/N2O = 1/4. O2 was introduced in the reaction system when used. The products of the reaction were analysed by an online gas chromatography (GC, Agilent 7890 A) equipped with a thermal conductivity detector (TCD) using He as the carrier gas with Porapak Q and TDX-01 packed columns and a flame ionization detector (FID) with a HP-PLOTQ capillary column. Methane was quantified using GC-FID, where signals with peak areas below 5 were excluded from quantification due to potential baseline fluctuations. Based on the calibration curve for CH4, with a factor of 2058.8 (corresponding to 1% CH4), this threshold corresponds to a detection limit of approximately 24 ppm. The quantitative N2O analysis of the feed and effluent gas was measured by FTIR spectrometers, G210 (0.1–100%) and G200 (0–1000 ppm), purchased from Geotechnical Instruments (UK) Ltd, with a detection limit of 1 ppm. NO and NO2 were monitored using Gasmet Dx4000 (Dekati Ltd, Finland) with a detection limit of 0.1 ppm. Each single data point in conversion curves were tested for at least 3 times under the same conditions, and every single uncertainty of the data collected from a parallel experiment was typically below 5%. Other information about catalytic performance evaluation is described in Supplementary Text 2. Characterizations The identification of active sites employed XRD, N2 physisorption, XAFS, TEM/AC-STEM, H2-TPR, CO-DRIFTS, quasi-in situ XAFS/AC-STEM, operando NAP-XPS, in situ ESTEM. The intermediates were characterized by TPSR, in situ DRIFTS and spin-trapping EPR. The details are illustrated in Supplementary Text 3. Density functional theory calculations DFT calculations were performed using the Vienna ab initio simulation package (VASP) code (version 5.4.4)73,74. Detailed descriptions are shown in Supplementary Text 4. Techno-economic analysis TEA was performed for ascertaining the economic potential of scalable synthesis of Pt/HZSM-5 and CH4-SCR-150 application in actual industrial N2O abatement. The catalyst price was estimated from CatCost (https://catcost.chemcatbio.org)70. The catalyst production process was designed to meet the targeted annual production of catalyst with a 95% capacity factor. The design production rate was set at 1.00 tonne per hour using a wet impregnation method, based on the manufacturing capabilities of one of our cooperating companies located in Sichuan Province, China. To account for losses during the production process, a waste/spoilage factor of 2% was applied. The production process is assumed to operate for 10 years, with an estimated 80% recovery rate for the Pt metal content during the active phase of the catalyst. Supplementary information Supplementary Information Transparent Peer Review file Supplementary Information Transparent Peer Review file Source data Source Data Source Data