Shielded bifunctional nanoreactor enabled tandem catalysis for plasma methane coupling

Abstract

The direct conversion of methane into valuable unsaturated C₂ hydrocarbons (C₂H₂ and C₂H₄) attracts growing attention. Non-thermal plasma offers a promising approach for this process under mild conditions. However, the competing formation of C₂H₆ and excessive dehydrogenation limit the selectivity toward C₂H₂ and C₂H₄. Herein, we develop a promising shielded bifunctional nanoreactor with a hollow structure and mesoporous channels (Na₂WO₄-Mn₃O₄/m-SiO₂) that effectively limits CH₄ overactivation and promotes selective coupling to form C₂H₂ and C₂H₄ under plasma activation, achieving 39% CH₄ conversion with 42.3% C₂H₂ and C₂H₄ fraction. This nanoreactor features isolated Na₂WO₄ embedded within the channels and Mn₃O₄ confined in the cavity of the SiO₂ hollow nanospheres, enabling internal tandem catalysis at co-located active sites. Na₂WO₄ induces the conversion of diffused CH₄ and CH₃ into reactive intermediates (^*CH and ^*CH₂), which subsequently couple on the Mn₃O₄ surface to form C₂H₂ and C₂H₄. Furthermore, the mesoporous channels inhibit the plasma discharge within the nanoreactor, preventing deep dehydrogenation of CH_x species to solid carbon. This nanoreactor demonstrates a highly selective route for the nonoxidative conversion of methane to valuable C₂ hydrocarbons, offering a new paradigm for the rational design of catalysts for plasma-driven chemical processes.

Introduction

Unsaturated C₂ hydrocarbons, crucial building blocks in the chemical industry, are primarily produced from the energy-intensive processing of crude oil. The development of sustainable alternatives via gas conversion has attracted increasing interest^1,2. While conventional gas conversion methods rely on indirect syngas routes, direct methane conversion to unsaturated C₂ hydrocarbons offers a more attractive pathway^3,4. Direct conversion of CH₄ to value-added fuels and chemicals can be achieved through either nonoxidative or oxidative catalytic routes^5,6,7. The nonoxidative pathway offers high C₂ selectivity and atom utilization efficiency but demands endothermic conditions (up to 1100 °C), incurring substantial energy costs and CO₂ emissions¹. In contrast, oxidative coupling of methane (OCM) can significantly lower the reaction temperature to ~600 °C in the presence of O₂^8,9. However, this method suffers from overoxidation of CH₄ to thermodynamically favored CO and CO₂, limiting the yield of unsaturated C₂ hydrocarbons^8,10. These challenges highlight the urgent need for innovative strategies to enhance methane activation and coupling efficiency.

Non-thermal plasma (NTP) offers a promising solution for the homolytic activation of CH₄ into radicals, enabling nonoxidative coupling of methane (NOCM) under mild conditions^11,12. Among NTP techniques, dielectric barrier discharge (DBD) has been extensively explored for methane coupling. However, the main challenge is the limited control over product distribution, particularly in achieving selective production of unsaturated C₂ hydrocarbons, such as acetylene (C₂H₂) and ethylene (C₂H₄). Instead, ethane (C₂H₆) is predominantly produced in NOCM using DBD reactors^{11,13,14,15,16}. This selectivity issue originates from the significantly longer lifetime of CH₃ radicals (>1 ms) compared to CH₂ (<30 ns) and CH (<5 ns) radicals, which favors undesired reaction pathways¹⁷. Integrating tailored catalysts into DBD reactors has great potential to enhance selectivity toward targeted products^14,18,19. Ideally, such catalysts would regulate the dehydrogenation process and selectively stabilize CH and CH₂ intermediates, directing the reaction toward the formation of C₂H₂ and C₂H₄.

The plasma-catalytic NOCM process involves heterogeneous surface reactions, highlighting the importance of the rational design of catalysts with tailored active sites and morphologies to maximize performance. In conventional OCM, Na₂WO₄ and Mn₂O₃ have proven effective in CH₄ activation and C-C coupling²⁰. Specifically, Na₂WO₄ generates reactive oxygen species to activate CH₄, while Mn₂O₃-supported O species promote the coupling of ^*CH₃ intermediates^21,22,23. Given their critical roles in methane coupling, integrating Na₂WO₄ and Mn₂O₃ into plasma-catalytic NOCM offers a promising avenue to enhance the production of unsaturated C₂ hydrocarbons through their synergistic effects – a strategy that remains underexplored. Notably, Na₂WO₄ and Mn₂O₃ function at distinct stages of the methane coupling pathway. Plasma pre-activates CH₄, facilitating Na₂WO₄-mediated enhancement of CH_x dehydrogenation into ^*CH and ^*CH₂ surface species, while Mn₂O₃ promotes subsequent coupling of these intermediates to form C₂H₂ and C₂H₄. Inspired by this synergistic interplay, we propose that a tandem plasma-catalysis system using spatially structured Na₂WO₄-MnO_x catalysts—designed to sequentially optimize dehydrogenation and coupling steps—could potentially achieve enhanced efficiency in C₂ hydrocarbon formation compared to plasma-catalysis systems using randomly structured catalysts.

To address the challenges of overactivation and selectivity control in plasma-assisted methane coupling, we propose a promising assembled nanoreactor (Na₂WO₄-Mn₃O₄/m-SiO₂, denoted as WMO/m-SiO₂) designed for integration into a DBD plasma reactor (Supplementary Fig. 1). This nanoreactor features a hollow structure with mesoporous channels, accessibility to reactants both internally and externally. This design provides a shielding effect for CH₄ molecules within the channels, preventing excessive dehydrogenation by plasma-generated reactive species. Furthermore, the WMO/m-SiO₂ reactor demonstrates tandem catalytic functionality, significantly enhancing selectivity toward target C₂ products. Specifically, Na₂WO₄ located within the m-SiO₂ channels initiates CH₄ activation, generating ^*CH₂ and ^*CH surface intermediates. Concurrently, Mn₃O₄ species confined within the cavity of the SiO₂ spheres promote facile coupling of these CH and CH₂ intermediates to form C₂H₂ and C₂H₄. Our study demonstrates significantly enhanced C₂H₂ and C₂H₄ yields compared to conventional catalysts under analogous DBD plasma conditions. This work offers a strategy to overcome critical limitations in plasma-catalysis, advancing efficient methane conversion to high-value C₂ hydrocarbons.

Results

Structural characterization of the catalysts

The synthesized mesoporous SiO₂ (m-SiO₂) featured interconnected nanospheres with diameters ranging from 95 to 135 nm (Fig. 1a and Supplementary Fig. 2a). Each nanosphere exhibited a hollow structure with a cavity diameter of ~65 nm and a shell thickness of ~20 nm (Supplementary Fig. 2b), along with a relatively high specific surface area of 240 m² g⁻¹ (Supplementary Table 1). Loading manganese and tungsten species onto m-SiO₂ did not affect the hollow structure, and the resulting WMO/m-SiO₂ nanoreactor retained surface porosity with particles localized both within the cavity and on the external surface (Fig. 1b, d and Supplementary Fig. 3). X-ray diffraction (XRD) confirmed the presence of Mn₃O₄ and amorphous SiO₂ in the WMO/m-SiO₂ composite (Supplementary Fig. 4). High-resolution transmission electron microscopy (HRTEM) revealed that the dispersed particles on the shell and cavity walls of m-SiO₂ were Na₂WO₄ and Mn₃O₄, respectively (Fig. 1d–f and Supplementary Fig. 3). The measured lattice spacings of 0.38 and 0.25 nm corresponded to Na₂WO₄ (111) and Mn₃O₄ (211), respectively (Fig. 1e). Notably, although XRD confirmed the presence of Mn₃O₄ (Supplementary Fig. 4), scanning electron microscopy (SEM) elemental distribution mapping (Supplementary Figs. 5, 6) of WMO/m-SiO₂ revealed negligible Mn and W signals compared to WMO/SiO₂ and WMO/ZSM-5 (Zeolite Socony Mobil-5). Unless otherwise specified, WMO/m-SiO₂, WMO/SiO₂, and WMO/ZSM-5 refer to samples with 1% Na₂WO₄-5% Mn loaded on the support material. These findings indicate that Na₂WO₄ and Mn₃O₄ particles are predominantly encapsulated within the m-SiO₂ spheres. Additionally, energy-dispersive X-ray spectroscopy (EDS) scans in Fig. 1f further confirmed the presence of Mn species inside the m-SiO₂ cavity.

Fig. 1: Characterization and catalytic performance.

a TEM image and particle size distribution of m-SiO₂. b SEM image of WMO/m-SiO₂. c Production rate and molar fraction of C₂H₄ and C₂H₂ within C₂-C₃ hydrocarbons (Conditions: 1 bar, specific energy input (SEI) 5.1 kJ L⁻¹, where SEI is defined as plasma discharge power divided by the gas flow rate; Feed gas 5 vol% CH₄/Ar, total flow rate 200 mL min⁻¹, discharge power 17 W, experiment duration 60 min). Error bars (standard deviation) in the figure were obtained from three sampling runs. d, e TEM and HRTEM images of WMO/m-SiO₂. f EDS line scans (from Point 1 to Point 2) from Fig. 1d. g Pore size distributions of m-SiO₂, WMO/m-SiO₂, WMO/ZSM-5, and WMO/SiO₂. h Production rate of C₂-C₃ hydrocarbons and methane conversion for catalysis-only, plasma-only and plasma-catalysis systems.

Plasma-catalytic NOCM reaction

We evaluated CH₄ conversion under three different conditions: plasma-only (no catalyst, no external heating), catalysis-only (external heating at 250 °C, no plasma), and plasma-catalysis (coupled plasma and catalysts, no external heating). Under plasma-only and plasma-catalysis conditions, the measured temperature was ~250 °C. In plasma-only mode, C₂-C₃ hydrocarbons dominated, with a maximum production of 31.2 μmol min⁻¹ at a CH₄ conversion of 33% (Supplementary Figs. 7, 8). Hydrocarbons with four or more carbon atoms were excluded due to negligible concentrations (<5% relative selectivity compared to C₂ products). Figure 1c shows the distributions of C₂H₄ and C₂H₂ within the C₂-C₃ range. A synchronized increase in C₂H₂ and C₂H₄ proportions was observed, which can be attributed to the closely aligned energetic thresholds of electron-induced CH₄ conversion into CH₂ and CH species¹³. WMO/m-SiO₂ exhibited no catalytic activity for methane conversion at 250 °C in the absence of plasma. However, under NTP conditions with WMO/m-SiO₂, the production of C₂H₄ and C₂H₂ significantly increased to 30.3 μmol g⁻¹ min⁻¹ (12.8 μmol g⁻¹ min⁻¹ for C₂H₄ and 17.5 μmol g⁻¹ min⁻¹ for C₂H₂), surpassing WMO/SiO₂ and WMO/ZSM-5 by factors of ~5 and 3.4, respectively (Fig. 1c). With the WMO/m-SiO₂ nanoreactor, the proportion of unsaturated hydrocarbons in the C₂–C₃ range increased significantly from 17.7% (plasma-only) to 42.3% (Fig. 1c and Supplementary Fig. 9). Simultaneously, the total yield of C₂H₄ increased markedly from 2.6 to 6.4 μmol min⁻¹, while C₂H₂ increased from 3.1 to 8.8 μmol min⁻¹. Notably, all three catalysts selectively promoted C₂H₄ and C₂H₂ production via deep dehydrogenation and coupling of CH₄, rather than promoting overall methane conversion (Fig. 1h).

The pore sizes of m-SiO₂ and WMO/m-SiO₂ ranged from 20 to 40 nm, distinct from those of WMO/SiO₂ and WMO/ZSM-5 (Fig. 1g and Supplementary Fig. 10). As shown in Supplementary Table 1, WMO/ZSM-5 exhibited the highest surface area of 226 m² g⁻¹, significantly exceeding that of WMO/m-SiO₂ (67 m² g⁻¹). However, plasma-only and plasma-catalysis conditions (using WMO/ZSM-5, WMO/SiO₂ and WMO/m-SiO₂) exhibited similar discharge properties (Supplementary Figs. 11–13). Additionally, packing the discharge zone with silica supports yielded consistent C₂H₂ and C₂H₄ levels (Supplementary Fig. 14). Furthermore, despite its lower surface area, WMO/m-SiO₂ demonstrated better CH₄ coupling efficiency to C₂H₂ and C₂H₄, highlighting the critical role of its unique hollow mesoporous structure in enhancing the synergistic interaction between NTP and catalysis.

Evaluation of the effect of catalyst position on performance

TEM and scanning transmission electron microscopy (STEM) images depict the morphology of the catalysts and the location of metal oxide particles, respectively (Fig. 2a and Supplementary Fig. 15). The metal oxide particles were selectively deposited in three configurations: (1) exclusively inside m-SiO₂ (In-m-SiO₂), (2) partially distributed within m-SiO₂ (Both-m-SiO₂), and (3) predominantly deposited on the exterior of m-SiO₂ (Out-m-SiO₂). For both “In-m-SiO₂” and “Both-m-SiO₂”, the Mn₃O₄ sizes ranged from 5 to 35 nm (Supplementary Fig. 15e, f). TEM analysis of Both-m-SiO₂ revealed a significant reduction in the Si signal around Particle 1 (50–90 nm), while no similar decrease was observed near Particle 2 (150–200 nm) (Fig. 2a). Moreover, the Mn signal intensified in both particle types, suggesting that Particle 1 and Particle 2 are located on the exterior and interior surfaces of the m-SiO₂ nanosphere, respectively (Fig. 2a). XRD patterns further confirmed the spatial distribution of Na₂WO₄ and Mn₃O₄ particles. Samples with higher diffraction intensities and more pronounced peaks corresponded to increased exposure of Mn₃O₄ species on the external surface of m-SiO₂ (Supplementary Fig. 16).

Fig. 2: Effect of catalyst location m-SiO₂ on NTP-CM performance.

a TEM image and EDS spectra of spent catalysts In-m-SiO₂, Out-m-SiO₂, and Both-m-SiO₂ (after 60 min of reaction). b Schematic illustration of the role of m-SiO₂ and the effect of catalyst position on methane coupling. c Mn³⁺/(Mn³⁺+Mn²⁺) ratio for fresh and spent catalysts. d Yields of C₂H₂ and C₂H₄ (defined as the product of CH₄ conversion and the selectivity of C₂H₂ and C₂H₄) and equivalent carbon deposition rate (ECR, defined as the solid carbon selectivity on the catalyst divided by the methane conversion) (Conditions: 1 bar, SEI 5.1 kJ L⁻¹, total flow rate 200 mL min⁻¹, discharge power 17 W, experiment duration 60 min).

EDS analysis of Particle 1 (internal) revealed weaker carbon and stronger oxygen signals compared to Particle 2 (external) (Fig. 2a and Supplementary Fig. 17). This suggests that m-SiO₂ protects Mn₃O₄ particles from direct exposure to CH₄ plasma. This shielding effect arises from the “Debye shielding” mechanism, where plasma discharge cannot penetrate the mesopores of m-SiO₂ due to their pore diameters being smaller than the Debye length (typically hundreds of nanometers)^19,24,25. Thus, the shielded internal cavity prevents the reduction of Mn₃O₄ particles and mitigates direct carbon deposition. This hypothesis is further supported by the observed decrease in the Mn³⁺/(Mn³⁺+Mn²⁺) ratio as more manganese oxide particles are located outside the m-SiO₂ layer (Fig. 2c and Supplementary Fig. 18). Previous studies have shown that low-valent metal catalysts (oxides, carbides, or metals) are highly effective for the deep dehydrogenation of CH₄^26,27. Consistent with these findings, Supplementary Fig. 19 shows that Mn₃O₄ particles exhibited a stronger carbon signal than the m-SiO₂ surface. Furthermore, as Na₂WO₄ and Mn₃O₄ particles are encapsulated within m-SiO₂ nanospheres, the yield of C₂H₂ and C₂H₄ decreased from 7.0% (WMO/m-SiO₂) to 4.5% (Na₂WO₄-Mn₃O₄, denoted as WMO). This trend implies that converting an equivalent amount of methane leads to more carbon deposition when fewer encapsulated Na₂WO₄ and Mn₃O₄ particles are present (Fig. 2d and Supplementary Fig. 20).

The role of Na₂WO₄ and Mn₃O₄ in plasma-catalytic NOCM reaction

The 1% Na₂WO₄-5% Mn₃O₄/m-SiO₂ catalyst achieved a combined C₂H₄ and C₂H₂ selectivity of 18.1%, surpassing that of Na₂WO₄/m-SiO₂ (14.5%), 5% Mn₃O₄/m-SiO₂ (13.2%) and plasma-only conditions (7.9%) (Fig. 3a and Supplementary Fig. 21). A comparison of CH₄ conversion and C₂-C₃ hydrocarbon distribution in the DBD reactor with previous studies is provided in Supplementary Table 2. The 1% Na₂WO₄-5% Mn₃O₄/m-SiO₂ (WMO/m-SiO₂) catalyst demonstrated the high selectivity for C₂H₂ and C₂H₄, while maintaining competitive methane conversion. Among reported studies, this work achieved a lower energy cost (EC) for CH₄ conversion (6.8 MJ/mol), demonstrating the effectiveness of the catalyst in plasma-catalytic NOCM reaction. Notably, the catalyst exhibited stable performance for over 25 h (Supplementary Figs. 22, 23). X-ray photoelectron spectroscopy (XPS) and XRD analyses confirmed that the dominant oxidation states of tungsten and manganese species in WMO/m-SiO₂, as well as in 5% Mn/m-SiO₂ and 1% Na₂WO₄/m-SiO₂, remained unchanged after the reaction (Supplementary Figs. 24–26). Supplementary Fig. 21 shows that increasing the Na₂WO₄ loading from 0.5 to 5% increased carbon deposition from 16.4 to 24.6%. In contrast, higher Mn₃O₄ content reduced carbon deposition. These findings suggest that Na₂WO₄/m-SiO₂ more effectively promotes the further dehydrogenation of methane compared to Mn₃O₄/m-SiO₂.

Fig. 3: Performance of Mn and W species.

a Selectivity of products (carbon deposited on the catalyst, C₂H₄ and C₂H₂) for plasma-only and plasma-catalysis systems (Conditions: 1 bar, SEI 5.1 kJ L⁻¹, total flow rate 200 mL min⁻¹, discharge power 17 W, experiment duration 60 min). Error bars (standard deviation) in the figure were obtained from three sampling runs. b Selectivity of carbon deposited on the catalyst and selectivity of C₂H₄ and C₂H₂ for reduced WMO/m-SiO₂ (reduced at 450 °C with H₂). c Mn 2p XPS spectra of reduced WMO/m-SiO₂ (pretreated with H₂ for 5, 10, 20, and 25 min). d IR spectra of WMO/m-SiO₂ (black), Mn₃O₄/m-SiO₂ (blue), and Na₂WO₄/m-SiO₂ (red) under plasma-catalysis conditions. e Quasi-in situ DRIFT spectra of catalysts after plasma-catalytic NOCM reaction. f CH₂D₂ and C₂HD₃ species generated under plasma-only and plasma-catalysis (with WMO/m-SiO₂) conditions (feed gas 2.5 vol% CH₄-2.5 vol% CD₄/Ar, SEI 5.1 kJ L⁻¹, total flow rate 200 mL min⁻¹; experiment duration 60 min).

The involvement of catalyst-bound oxygen species is well established in methane dehydrogenation^5,9 and the coupling of intermediate species^28,29,30, highlighting their crucial roles in these reactions. To investigate this effect, we pretreated WMO/m-SiO₂ with H₂ at 450 °C for different durations, generating a series of WMO/m-SiO₂ samples with varying oxygen contents. H₂ pretreatment at 450 °C primarily reduced Mn₃O₄ but not tungsten species, as evidenced by H₂ temperature-programmed reduction (H₂-TPR) and XPS (Fig. 3c and Supplementary Figs. 27, 28). H₂ treatment resulted in increased carbon deposition and decreased selectivity for C₂H₄ and C₂H₂ when the WMO/m-SiO₂ nanoreactor was reduced for 5–20 min, accompanied by a decline in Mn³⁺ content and oxygen loading within the catalyst (Fig. 3b, c). These observations align with the trend that increasing the Mn content (β) from 2 to 10% in 1% Na₂WO₄-β Mn₃O₄/m-SiO₂ slightly enhanced the production of C₂-C₃ hydrocarbons (Supplementary Fig. 29). These findings suggest that MnO_x acts as an oxygen carrier, promoting the coupling of active species (CH_x and H) and reducing methane cracking. This highlights the importance of optimizing oxygen content in catalysts to maximize performance.

CH_x species absorbed on the catalyst and in the gas phase

In situ plasma-coupled Fourier transform infrared (FTIR) spectroscopy was used to investigate plasma-assisted surface reactions on WMO/m-SiO₂, 5% Mn₃O₄/m-SiO₂ and 1% Na₂WO₄/m-SiO₂. As shown in Supplementary Fig. 30, the intensities of the IR peaks at 628 cm⁻¹ (≡C-H) and 950 cm⁻¹ (=C-H) decreased as the reaction progressed. These absorbed ≡C-H and =C-H bands are associated with key intermediates involved in the formation of C₂H₂ and C₂H₄ during the catalytic process³¹. Additional IR bands at 1465, 1585, 1708, 3265, and 3320 cm⁻¹ correspond to C = C stretching vibrations on the catalyst surface (Fig. 3d)^31,32,33. Notably, WMO/m-SiO₂ exhibited the highest intensities for peaks associated with ≡C-H, =C-H, and C=C compared to catalysts containing only Mn or W. This finding suggests that the synergistic interaction between Mn₃O₄ and Na₂WO₄ sites effectively promotes the formation of surface-adsorbed ^*CH and ^*CH₂ groups, ultimately enhancing the yield of C₂H₄ and C₂H₂.

Quasi-in situ DRIFTS characterization provided further insights into the types of adsorbed species remaining on the catalyst surface after the reaction (Supplementary Figs. 31, 32). As shown in Fig. 3e, several key peaks were observed, including C–H stretching from ^*CH₃ (2960 and 2872 cm⁻¹)³⁴, C–H stretching from ^*CH₂ (2930 cm⁻¹)³⁴, C=C bonds (1585 and 1465 cm⁻¹)^32,33 and C-H bond bending/deformation modes (1384 cm⁻¹)³⁵. In this study, the angles (∠1 = 12°, ∠2 = 0°, and ∠3 = 3°) between the standard slope and tangents (peak 2960 to peak 2930 cm⁻¹) were used to measure the relative proportions of absorbed ^*CH₃ and ^*CH₂ species on the catalyst surface. These species serve as precursors for the formation of ethane and ethylene, respectively³⁶. Among the catalysts, 1% Na₂WO₄-5% Mn₃O₄/m-SiO₂ exhibited the highest ^*CH₂ intensity, followed by 1% Na₂WO₄/m-SiO₂ and 5% Mn/m-SiO₂. Notably, compared to 1% Na₂WO₄/m-SiO₂, 5% Mn/m-SiO₂ promoted the formation of surface C=C (1585 and 1465 cm⁻¹), indicating that ^*CH₂ formation predominantly occurred at W sites, while Mn sites facilitated the coupling of CH₂ to form C=C bonds. Although the internal Mn₃O₄ sites are not directly exposed to plasma (Supplementary Fig. 3), varying the Mn₃O₄ loading within the m-SiO₂ spheres led to observable changes in product distribution and the relative intensity of adsorbed ^*CH₃ and ^*CH₂ species (Supplementary Figs. 21, 31). This suggests that CH_x radicals can diffuse or migrate at least 20 nm to reach Mn₃O₄ sites within their lifetime, enabling them to access the interior of the catalyst for subsequent reactions.

Methane isotopic labeling experiments

Methane isotopic labeling experiments, conducted using a parallel flow of CH₄ and CD₄, revealed an increase in CH₃ and CH₂ radicals during the plasma-catalyzed reaction over WMO/m-SiO₂ compared to the plasma-only condition (Fig. 3f). Detailed experimental procedures are provided in the Methods section, and the corresponding conversions of CD₄ and CH₄, along with product distributions, are shown in Supplementary Fig. 33. The experiment was designed to probe the complex dynamics within the discharge field, where multiple collisions and coupling reactions occur, leading to the reversible activation of C-H bonds, facilitating the reformation of nascent methane from activated C_xH_y intermediates³⁶. This behavior contrasts with the essentially irreversible C-H activation in traditional OCM²⁹. The detection of mixed CH₂D₂ and CHD₃ isotopes during plasma-catalyzed NOCM on WMO/m-SiO₂ confirmed this mechanism. Compared to the plasma-only system, the plasma-catalytic system with WMO/m-SiO₂ demonstrated significantly higher production of CH₂D₂ and CHD₃ (Fig. 3f). This result suggests that WMO/m-SiO₂ enhances the activation of CH₄ and the generation of CH₃/CD₃ and CH₂/CD₂ radicals. The enriched pool of CH₂ species ultimately promotes dimerization into C₂H₂ and C₂H₄¹¹. In summary, the synergistic interaction between Mn₃O₄ and Na₂WO₄ sites in the WMO/m-SiO₂ nanoreactor effectively promotes the formation of surface-adsorbed ^*CH and ^*CH₂ species, as well as gas-phase CH₂ radicals, leading to enhanced yields of C₂H₂ and C₂H₄.

Density functional theory (DFT) calculations

Synchrotron radiation-based X-ray absorption spectroscopy (XAS) was employed to elucidate the chemical state and local structure of the WMO/m-SiO₂ catalysts. The X-ray absorption near-edge structure (XANES) spectra at the Mn K-edge (Mn₃O₄/m-SiO₂) and W K-edge (Na₂WO₄/m-SiO₂) in WMO/m-SiO₂ closely resembled those of the individual Mn₃O₄/m-SiO₂ and Na₂WO₄/m-SiO₂ reference catalysts (Supplementary Fig. 34). This observation, in substantial contrast to the combined Na₂WO₄-Mn₃O₄ spectrum, strongly suggests the presence of isolated Mn₃O₄ and Na₂WO₄ active sites dispersed on the m-SiO₂ support. Further validation was provided by Fourier transform (FT) extended X-ray absorption fine structure spectroscopy (EXAFS) results (Fig. 4a, b). The peaks at 1.4 Å (Fig. 4a) and 1.2 Å (Fig. 4b) mainly represent the single scattering of Mn−O and W–O bonds, respectively, confirming the existence of isolated Mn and W species and the absence of direct Mn-W binding in the WMO/m-SiO₂ catalyst. The bond lengths for Mn–O and W–O in WMO/m-SiO₂ were determined to be 1.86 and 1.69 Å, respectively (Supplementary Table 3). The proposed Mn-O model exhibited excellent agreement with the experimental spectra, as evidenced by its superior fit in the XANES analysis and negligible deviations from DFT calculations (Supplementary Figs. 34–36).

Fig. 4: The structures of WMO/m-SiO₂ and DFT calculations.

a Fourier transform (FT) of the EXAFS spectrum of Mn and the DFT-optimized structure of Mn₃O₄ on WMO/m-SiO₂. b FT-EXAFS spectra of W and DFT-optimized structure of Na₂WO₄ on WMO/m-SiO₂. c DFT-optimized geometries of ^*CH₄ dehydrogenation to ^*CH₃, ^*CH₂, and ^*CH on Mn₃O₄ (211) and Na₂WO₄ (111) surfaces. d DFT-optimized geometries of ^*CH₃ to C₂H₄ on Mn₃O₄ (211) and Na₂WO₄ (111) surfaces. e DFT-optimized geometries of ^*CH₂ coupled to C₂H₂ on Mn₃O₄ (211) and Na₂WO₄ (111) surfaces. f DFT-optimized geometries of intermediates in the plasma-catalytic conversion of CH₄ to radicals, C₂H₂, C₂H₄, and C₂H₆. g Wavelet transform plots of the Mn K-edge and W K-edge. h Schematic illustration of C₂H₂ and C₂H₄ formation pathways on WMO/m-SiO₂.

To elucidate the fundamental mechanism of C₂H₂ and C₂H₄ production within WMO/m-SiO₂, spin-polarized periodic DFT calculations were employed to unravel the distinct roles of its components during plasma-catalytic NOCM reaction. Based on experimental results (Fig. 1e and Supplementary Fig. 3), we selected Mn₃O₄ (211) and Na₂WO₄ (111) surfaces as model systems (Supplementary Fig. 35). The calculated binding energies of reaction intermediates involved in CH₄ dehydrogenation and subsequent coupling to C₂H₂ and C₂H₄ are presented in Supplementary Tables 4, 5. Notably, the binding strength of the C_xH_y intermediates followed the order of Mn₃O₄ (211) > Na₂WO₄ (111), with intermediates binding via C or C-C on both surfaces (Supplementary Figs. 37, 38 and Supplementary Tables S4, S5).

The calculated Gibbs free energy change profiles for CH₄ dehydrogenation and intermediate active species coupling on the Mn₃O₄ (211) and Na₂WO₄ (111) catalysts are shown in Fig. 4c–e. For CH₄ dehydrogenation (Fig. 4c), the Na₂WO₄ (111) surface can facilitate CH₄ dehydrogenation to produce ^*CH₃, ^*CH₂, and ^*CH species. The most challenging step was ^*CH₃ dehydrogenation, with an uphill energy of 0.22 eV, followed by ^*CH₂ dehydrogenation, with an uphill energy of 0.19 eV on the Na₂WO₄ (111) surface. In contrast, CH₄ dehydrogenation on the Mn₃O₄ (211) surface required significantly higher uphill energy (0.73 and 0.75 eV) for producing ^*CH₂ and ^*CH species, respectively. In addition, the reaction energy barriers for ^*CH₃ → ^*CH₂ + ^*H and ^*CH₂ → ^*CH + ^*H were calculated (Supplementary Table 8). These two elementary reactions were identified as the rate-determining steps in CH₄ dehydrogenation to generate ^*CH₃, ^*CH₂, and ^*CH species on both surfaces (Fig. 4c), with energy barriers of 0.97 and 0.69 eV on Na₂WO₄ (111), significantly lower than those on Mn₃O₄ (211) (1.93 and 2.02 eV). This indicates that the Na₂WO₄(111) surface is more effective in facilitating CH₄ dehydrogenation, consistent with experimental results (Supplementary Fig. 21). The Mn₃O₄ (211) surface was found to be more favorable for coupling ^*CH₂ and ^*CH species to generate C₂H₄ and C₂H₂ (Fig. 4d, e). The most challenging steps were the desorption of ^*C₂H₄ (Fig. 4d) and ^*CH₂ dehydrogenation (Fig. 4e), with uphill energies of 0.49 and 0.19 eV, respectively. In contrast, on the Na₂WO₄ (111) surface, the most difficult steps were ^*CH₃ dehydrogenation (Fig. 4d) and desorption of ^*C₂H₂ (Fig. 4e), with significantly higher uphill energies of 0.73 and 2.52 eV, respectively. In addition, we investigated CH₄ dehydrogenation and radical coupling reactions under plasma-only conditions (Fig. 4f). The results suggest that plasma-driven NOCM without a catalyst favors the production of C₂H₆, in stack contrast to plasma-catalyzed reactions.

To elucidate the role of SiO₂ in the WMO/m-SiO₂ nanoreactor, SiO₂/Mn₃O₄ (211) and SiO₂/Na₂WO₄ (111) models were built (Supplementary Figs. 36, 39, 40). The calculated binding energies of reaction intermediates involved in CH₄ dehydrogenation and subsequent coupling to form C₂H₂ and C₂H₄ are presented in Supplementary Tables 6 and 7, with the most stable adsorption configurations illustrated in Supplementary Figs. 39, 40. Notably, SiO₂/Na₂WO₄ (111) exhibited significantly weaker binding for ^*C₂H₂ compared to the isolated Na₂WO₄ (111) surface, with adsorption energies of −0.54 versus −0.95 eV, respectively (Supplementary Fig. 41 and Supplementary Tables 6, 7). In contrast, SiO₂/Mn₃O₄ (211) favored the desorption of ^*C₂H₄ compared to pure Mn₃O₄ (211), with an adsorption energy of −0.62 eV (compared to −1.58 eV on pure Na₂WO₄ (111)). The binding energies of ^*CH/^*CH₂ on SiO₂/Mn₃O₄ (211) and SiO₂/Na₂WO₄ (111) were −6.75 and −5.77 eV and −2.61 and −5.43 eV, respectively, indicating that SiO₂/Mn₃O₄ has a stronger adsorption capacity for ^*CH₂ and ^*CH than SiO₂/Na₂WO₄. In summary, DFT calculations reveal that Na₂WO₄ facilitates CH₄ dehydrogenation to ^*CH or ^*CH₂ intermediates, while Mn₃O₄ promotes the coupling of ^*CH and ^*CH₂ to form C₂H₂ and C₂H₄, respectively. Furthermore, the presence of SiO₂ in combination with Na₂WO₄ and Mn₃O₄ enhances the desorption of the generated ^*C₂H₂ and ^*C₂H₄ species. These findings are consistent with experimental observations.

Reaction mechanisms

The enhanced catalytic performance of Na₂WO₄-Mn₃O₄/m-SiO₂, compared to Na₂WO₄/m-SiO₂ and Mn₃O₄/m-SiO₂, highlights the synergistic effects between Na₂WO₄ and Mn₃O₄ sites (Supplementary Fig. 21). Quasi-in situ DRIFTS characterization (Fig. 3e and Supplementary Figs. 31, 32) indicates that W sites promote ^*CH₂ generation, while Mn sites facilitate the coupling of ^*CH₂ to form C=C bonds.

If radicals were fully converted at Na₂WO₄ before reaching Mn₃O₄ sites within m-SiO₂, significant carbon deposition would be expected on Na₂WO₄. However, compared to Na₂WO₄/m-SiO₂, the 1% Na₂WO₄-5% Mn₃O₄/m-SiO₂ catalyst exhibits lower carbon accumulation (Supplementary Fig. 21) and higher ^*CH₂ and C=C intensities (Fig. 3e). These findings suggest that radicals initially generated on Na₂WO₄ sites undergo further transformation into ^*CH and ^*CH₂, which subsequently migrate to Mn₃O₄ sites for coupling reactions to produce C₂H₂ and C₂H₄.

Wavelet transforms of the EXAFS spectra at the Mn K-edge and W K-edge for WMO/m-SiO₂, 5% Mn₃O₄/m-SiO₂, and 1% Na₂WO₄/m-SiO₂ reveal similar Mn−O and W−O scattering peaks at (4.1, 1.4 Å) and (3.9, 1.2 Å), respectively (Fig. 4g). These findings indicate that Na₂WO₄ and Mn₃O₄ are independently distributed on the m-SiO₂ support, consistent with TEM observations (Supplementary Fig. 3). Notably, TEM analysis also demonstrates the close spatial proximity of Na₂WO₄ and Mn₃O₄, providing potential pathways for intermediate species migration.

DFT calculations further reveal that the adsorption energies of ^*CH and ^*CH₂ on Mn₃O₄ are −5.21 and −3.96 eV, respectively, significantly stronger than those on Na₂WO₄ (−3.27 and −3.08 eV) (Supplementary Tables 6, 7). Notably, when supported on m-SiO₂, the adsorption energy gap increases, suggesting that radicals preferentially stabilize on Mn₃O₄ rather than remaining on Na₂WO₄ (Supplementary Fig. 41). This thermodynamic preference, combined with the close proximity of two active sites, indicates that ^*CH and ^*CH₂ species likely undergo surface diffusion or desorption-reabsorption migration, facilitating C-C coupling on Mn₃O₄. Similar bifunctional catalysis mechanisms have been reported in thermal catalysis systems, where intermediate spillover between distinct active sites enhances reaction efficiency^37,38.

The distribution and synergy between Na₂WO₄ and Mn₃O₄ sites further facilitate the sequential activation of C-H bonds and C-C coupling within the WMO/m-SiO₂ nanoreactor. Herein, a tandem reaction mechanism is proposed for CH₄ conversion to C₂H₂ and C₂H₄ (Fig. 3h): (I) CH₄ dissociation and CH₃ diffusion: Energetic electrons from the plasma induce CH₄ dissociation into CH_x fragments, which subsequently diffuse toward the interior of the silica sphere due to the concentration gradient across m-SiO₂. (II) Dehydrogenation at Na₂WO₄ sites (support on the channel): ^*CH₃ species undergo dehydrogenation at W sites, forming surface-adsorbed ^*CH and ^*CH₂ species. (III) Surface species coupling on Mn sites: ^*CH and ^*CH₂ species migrate from Na₂WO₄ to Mn₃O₄ sites, leading to C-C coupling for C₂H₂ and C₂H₄ production.

Notably, the mesoporous channels of m-SiO₂ restrict plasma penetration into the interior of m-SiO₂, thereby mitigating excessive CH₄ activation (e.g., methane cracking) and suppressing carbon deposition. This structural confinement, combined with the synergistic tandem catalysis of Na₂WO₄ and Mn₃O₄, significantly enhances the yield of C₂ products. Moreover, the shielding effect of the nanoreactor reduces plasma-induced product decomposition and recombination, further enhancing selectivity.

Discussion

This work presents a promising nanoreactor catalyst design strategy that significantly improves the yield and selectivity of C₂H₂ and C₂H₄ through plasma-catalytic NOCM under mild conditions. The nanoreactor features a hollow nanosphere structure with Na₂WO₄ nanoparticles anchored on the interconnected channels and monodispersed Mn₃O₄ nanocrystals hosted within the internal cavity. By positioning the nanoreactors in the discharge area, methane conversion reached 34%, with a selectivity of 42.3% toward C₂H₂ and C₂H₄. This represents a nearly 4.5-fold increase in yield and a fourfold increase in selectivity for unsaturated C₂ hydrocarbons compared to the plasma-only system. Importantly, no deactivation was observed during the 25-h catalyst stability test. Mechanistic investigations revealed that Na₂WO₄ promotes the dehydrogenation of diffused CH₄ and CH₃, leading to the formation of ^*CH and ^*CH₂ intermediates. These species subsequently undergo C-C coupling on the Mn₃O₄ surface to form C₂H₂ and C₂H₄. The excellent catalytic performance, supported by in situ plasma-coupled FTIR characterization, is further corroborated by DFT calculations. These calculations demonstrate that a tandem catalytic effect is achieved through the isolated Na₂WO₄ and Mn₃O₄ active sites, which are responsible for the enhanced selectivity. Furthermore, the mesoporous nanoreactor design prevents the reduction of internal Mn₃O₄ by CH₄ plasma through the Debye shielding effect. This reduces carbon deposition on the catalyst and protects the generated ^*CH and ^*CH₂ intermediates from further decomposition due to the absence of plasma discharge within the mesopores. This catalyst design strategy offers significant potential for advancing plasma-catalysis. It enables highly selective and directional conversion, improving the energy efficiency of plasma-catalytic systems and paving the way for a high-value route to transform methane into unsaturated light olefins under mild conditions.

Methods

Synthesis of SiO₂ nanospheres

Mesoporous SiO₂ (m-SiO₂) was synthesized using a double template method in an ethanol solution. First, ethanol and polyacrylic acid were added to a vial and stirred for 30 min. Then, diluted ammonia water was added to the above solution, followed by the introduction of polyether, and the mixture was stirred for another 30 min. After that, ethyl orthosilicate was added dropwise to the vial, resulting in a suspension after 4 h of stirring. Subsequently, the suspension was centrifuged and washed three times with ethanol. Finally, the precursors were evaporated overnight and calcination at 550 °C.

Synthesis of α Na₂WO₄-β Mn₃O₄/m-SiO₂

α Na₂WO₄/m-SiO₂, β Mn₃O₄/m-SiO₂, and α Na₂WO₄-β Mn₃O₄/m-SiO₂ (where α and β denote the respective weight percentages of each metal oxide) were synthesized using the incipient wetness method. The Mn loading on the m-SiO₂ support varied from 2 to 10 wt% (2, 5, 7.5, and 10 wt%), while the Na₂WO₄ loading ranged from 0.5 to 5% (0.5, 1, 2, and 5 wt%). For the preparation of Na₂WO₄/m-SiO₂ and Mn₃O₄/m-SiO₂, aqueous solutions of manganese nitrate tetrahydrate (Mn(NO₃)₂·4H₂O) or sodium tungstate dihydrate (Na₂WO₄·2H₂O) were mixed with m-SiO₂ in a water bath. The mixture was stirred until it reached a paste-like consistency and then dried overnight at 70 °C. Subsequently, the dried samples were calcined in air at 550 °C. A two-step impregnation method was used for the synthesis of β Na₂WO₄-α Mn₃O₄/m-SiO₂ (denoted as WMO/m-SiO₂). First, Mn₃O₄/m-SiO₂ was prepared as described above. Then, Na₂WO₄ was loaded onto the Mn₃O₄/m-SiO₂ using the same drying and calcination procedures. This method was also used to prepare 1% Na₂WO₄-5% Mn₃O₄/SiO₂ and 1% Na₂WO₄-5% Mn₃O₄/ZSM-5. Before testing, all catalysts were crushed and sieved to obtain particles with sizes between 30 and 60 mesh.

Synthesis of Mn₃O₄-deposited m-SiO₂ catalysts

Three types of catalysts with varying Mn₃O₄ locations were prepared: In-m-SiO₂, where Mn₃O₄ was deposited exclusively within the mesopores of m-SiO₂; Both-m-SiO₂, where Mn₃O₄ was partially distributed within the mesopores and on the exterior surface of m-SiO₂; Out-m-SiO₂, where Mn₃O₄ was primarily located on the exterior surface of m-SiO₂. In-m-SiO₂ was synthesized as described above. Both-m-SiO₂ was prepared using a rapid heating and drying method, where Na₂WO₄ was first loaded onto m-SiO₂, followed by the deposition of Mn₃O₄. The precursors were rapidly heated from room temperature to 300 °C at a heating rate of 20 °C min⁻¹ for 2 h and then calcined at 550 °C. Out-m-SiO₂ was synthesized by mechanically mixing pre-synthesized Mn-O and Na-W-O precursors. Specifically, solutions of Mn or W were dissolved in deionized water and stirred at 50 °C for 3 h. After drying overnight, the precursors were mechanically mixed with m-SiO₂ and calcined at 550 °C.

Catalyst activity test

The performance of the catalyst in the thermal catalytic NOCM reaction (denoted as catalyst only) was evaluated in a DBD reactor (plasma off) equipped with heating tape and operated at atmospheric pressure. The reactor was loaded with 1% Na₂WO₄-5% Mn₃O₄/m-SiO₂ and secured with quartz wool on both ends. Prior to testing, the catalyst was pretreated with argon (200 mL min⁻¹) at 100 °C for 20 min to remove impurities. The temperature was then increased to 250 °C and monitored using a K-type thermocouple placed within the catalyst bed. A diluted methane feed (5 vol% CH₄ in Ar) at a total flow rate of 200 mL min⁻¹ was used to minimize mass transfer limitations and accurately evaluate the intrinsic catalytic activity. The feed gas was preheated to 30 °C, and the experiments were conducted for 60 min.

For plasma-catalysis and plasma-only conditions, the DBD reactor was operated without external heating. The experimental procedure involved the following steps: (1) The catalyst was pretreated with argon (100 mL min⁻¹) at 100 °C for 20 min, followed by cooling to room temperature. (2) A 5 vol% CH₄/Ar mixture (200 mL min⁻¹) flowed through the DBD reactor for 10 min before switching on the plasma. (3) The plasma was ignited at an SEI of 5.1 kJ L⁻¹ (calculated as the discharge power divided by the gas flow rate) and maintained at a discharge power of 17 W with a flow rate of 200 mL min⁻¹ for 60 min. (4) After switching off the plasma, the DBD reactor was purged with argon (100 mL min⁻¹) for 10 min. (5) The spent catalyst was removed, and a 10 vol% O₂/Ar mixture (100 mL min⁻¹) was introduced to oxidize solid carbon deposited in the DBD reactor. This step was carried out at an SEI of 6.14 kJ L⁻¹ until no CO or CO₂ was detected in the exhaust gas. (6) The spent catalyst was reintroduced into the cleaned reactor, and a 10 vol% O₂/Ar mixture (100 mL min⁻¹) was used to oxidize carbon deposited on the catalyst. This step was continued, and no CO or CO₂ were detected in the exhaust gas. The amount of carbon deposited was calculated using the equation N_C = V × (C_CO2 + C_CO) × 22.4, where V is the total volume of exhaust gas during the oxidation process, and C_CO2 and C_CO are the concentrations of CO₂ and CO in the exhaust gas, respectively. To avoid interfering with the plasma field, the reactor temperature was measured using an infrared thermometer.

Isotopic labeling experiments

To investigate the relative contributions of CH₂ and CH₃ radicals to CH₄ decomposition, isotopic labeling experiments were conducted. The catalyst was first pretreated in a 5 vol% CH₄/Ar mixture (discharge power 17 W, flow rate 200 mL min⁻¹) for 60 min and then cooled to room temperature. Next, argon (200 mL min⁻¹) was flowed through the reactor for 10 min to purge residual gases. A gas mixture of 2.5 vol% CD₄ and 2.5 vol% CH₄ in argon (200 mL min⁻¹) was subsequently introduced into the DBD reactor for 10 min. The plasma was then switched on and sustained for 60 min, during which mass spectrometric signals at m/z = 15, 18, and 19 were monitored. The concentrations of CH₄ and CD₄ in the mixture were determined using mass spectrometry. From these measurements, the conversion of CH₄ and CD₄, as well as the selectivity of the product, was determined. Under plasma discharge conditions, CH₄ and CD₄ dissociate into radicals including CH₃, CH₂, CH, CD₃, CD₂, CD, H, and D. These radicals recombine to form isotopically labeled methane species, such as CD₃H (m/z = 19), and CD₂H₂ (m/z = 18). The formation of CD₃H and CD₂H₂ results from the recombination reactions of CD₃ with H and CD₂ with two H atoms, respectively. The signals at m/z = 15 and m/z = 19 reflect the concentrations of CH₄ and CD₃H, respectively. When analyzing the peak intensity of CD₂H₂, it is critical to consider the contribution of CD₄ fragmentation (m/z = 20), which generates CD₃ (m/z = 18). Specifically, the signal at m/z = 18 reflects contributions from both CD₂H₂ and the CD₃ fragment derived from CD₄ fragmentation.

In situ plasma-coupled FTIR characterization

To elucidate plasma-induced surface reactions during the plasma-catalytic NOCM process, we employed in situ plasma-coupled FTIR spectroscopy using a custom-designed plasma gas cell¹⁹. The experimental procedure is described in detail as follows: (I) Prior to analysis, each sample was pretreated with argon plasma (99.999% purity) at a flow rate of 100 mL min⁻¹ in a DBD reactor at 100 °C for 20 min to remove residual surface species. (II) A 5% CH₄/Ar mixture (40 mL min⁻¹) was introduced to purge the cell for 30 min. During this step, the temperature was decreased from 100 to 35 °C, after which the IR background spectrum was collected. (III) The plasma was switched on, and the plasma-catalytic NOCM was conducted for 15 min. (IV) After switching off the gas flow, IR spectra were collected every 3 min for a total duration of 18 min.

Quasi-in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFT) analysis was conducted using an FTIR spectrometer (IS50, Thermo Fisher Co. Ltd.) equipped with a liquid nitrogen N₂-cooled mercury-cadmium-telluride (MCT) detector. The background spectrum was obtained by pretreating the catalyst with Ar plasma in a DBD reactor under conditions identical to those used in the plasma-catalytic NOCM experiments (SEI = 5.1 kJ L⁻¹, flow rate = 200 mL min⁻¹). Following pretreatment, the catalyst was cooled to room temperature before conducting the quasi-in situ DRIFTS measurements. To avoid air exposure, the pretreated catalysts were transferred from the DBD reactor to the DRIFT cell within a glovebox. For the plasma-catalytic NOCM reaction, the plasma was switched off after the experiment, and the inlet and outlet of the DBD reactor were sealed. Subsequently, the spent catalysts were then transferred from the DBD reactor to the DRIFT cell within a glovebox, and IR spectra were collected at room temperature.

Computational details

Spin-polarized density functional theory (DFT)^39,40 calculations were performed using the Vienna ab initio simulation package (VASP) code⁴¹. The exchange-correlation interactions between electrons were described using the Perdew–Burke–Ernzerhof (PBE) functional within the generalized gradient approximation (GGA)^42,43. Following convergence tests, plane-wave pseudopotentials with kinetic cutoff energy of 420 eV⁴⁴ for Mn₃O₄ and 500 eV⁴⁵ for Na₂WO₄ method were employed within the projector augmented wave (PAW) method. The Mn₃O₄ (211) and Na₂WO₄ (111) surfaces were selected as the computational models based on the experimental results (Fig. 1e and Supplementary Fig. 3). To minimize interactions between the slab and its periodic images, a vacuum layer of ~15 Å was added above the slab. During geometry optimization, the bottom two atomic layers were fixed, while all other atoms and adsorbates were allowed to relax until the force on each atom was less than 0.01 eV Å⁻¹. A convergence criterion of 1 × 10⁻⁵ eV/atom was used for structural optimization. Brillouin zone integration was performed using a 2 × 2 × 1 Monkhorst-Pack grid with a Methfessel-Paxton smearing width (σ) of 0.2 eV. Due to the presence of localized 3d states on Mn, the electronic structure of Mn was treated within the DFT + U formalism with a U-J parameter of 4.00 eV⁴⁶. In addition, to account for weak interactions within the catalyst, van der Waals corrections were incorporated using the DFT-PBE-D3 method⁴⁷. Further details regarding the DFT calculation methods are provided in the Supplemental Information.