With the gradual depletion of crude oil, growing demand for chemicals, and cumulative greenhouse gas emissions, the catalytic conversion of CH4 and N2O to value-added chemicals is an urgent and significant issue1,2,3,4. Traditionally, the CH4 resource utilization is first through steam methane reforming (SMR) process to produce syngas5,6, then Fischer-Tropsch (FT) synthesis to form methanol7,8,9, and followed by the methanol-to-olefin (MTO) conversion to generate olefins10,11. However, the long and energy-intensive reaction process needs to be simplified. In addition, the cleaning of N2O (N2O → N2 + O2) relies on high temperature or catalytic thermal decomposition12. Direct catalytic conversion of CH4 and N2O to methanol or other value-added chemicals is possible to alleviate the above-mentioned problems13,14. Actually, the direct conversion of syngas to hydrocarbons over metal-oxide/zeolite bifunctional catalysts via a tandem reaction coupling approach is the focus of recent research and has achieved breakthrough progress15. The strategy was designed to implement the short process of merging two steps into one4. In addition, direct conversion of methane to valuable chemicals is able to be realized via oxidative coupling of methane (OCM)16,17, methane dehydroaromatization (MDA)18,19,20, and the newly emerging strategy of methane conversion to olefins, aromatics, and hydrogen (MTOAH)21,22. However, the three strategies require a quite high temperature (å 700 oC) and the products distribution is limited in ethylene, ethane, and benzene21. In the recent work of Sushkevich et al. oxidative methane C − C coupling over copper-exchanged zeolites was reported23. However, the coupling products are light paraffins, not light olefins.

To our knowledge, the strategy of direct conversion of methane to hydrocarbons (DMTH) over metal-containing zeolite bifunctional catalysts via a tandem reaction coupling approach has not been sufficiently studied. In the past decades, the focus of direct oxidation of methane to methanol was on the yield and selectivity of methanol13,24. In order to improve methanol selectivity, lower temperatures (200–300 oC) were regularly used to prevent methanol to hydrocarbons (MTH) reaction and the over-oxidation of methanol to CO and CO213,14,24,25,26. Therefore, hydrocarbons as the secondary reaction products of methane oxidation have received little attention. Recent work of direct oxidation of methane to methanol reported by Xu et al. has detected dimethyl ether (DME) and hydrocarbons at 320 oC, however, the coke selectivity was up to 80% and the selectivity of hydrocarbon has not been fully researched due to the use of large pore Cu/BEA zeolite27. Very recently, we performed direct oxidation of methane to methanol (DMTM) reaction at 350 oC on the small pore Cu/AEI zeolite and observed hydrocarbons being converted from methanol on acid sites of Cu/AEI zeolite, where AEI is the abbreviation of Aluminophosphate-eighteen and the aluminosilicate AEI-type zeolite is well-known as SSZ-3928,29,30. Simultaneously, we noticed that some Cu/AEI zeolites contained a considerable amount of Brønsted Acid Sites (BAS), however, no hydrocarbons were observed. Thus, other factors affecting the generation of hydrocarbons should be considered.

Inspired by the achievements in conversion of syngas to olefins via methanol/DME intermediates on metal-oxide/zeolite catalysts31, the effect of spatially separated Cu2+ and BAS in the Cu-exchanged AEI zeolite on the reaction performance of direct oxidation of methane was studied in this work. Noteworthy, the Ångstrom and nanoscale intimacy are quite difficult to describe and investigate. A study by Martens and coworkers uncovered that variation in the distance between Pt nanoparticles and BAS in the association of Pt-Al2O3/Y zeolite and Pt-Y zeolite/Al2O3 achieved different product selectivity in n-Decane hydrocracking reactions32. Only Pt was observable in the HAADF-STEM images, however, visualization of Pt and acid sites over spatial distance remained elusive. Similar research has been reported several times by de Jong and coworkers33,34. In addition, Deng and coworkers adopted the 1H–67Zn double-resonance solid-state NMR technique to detect the spatial proximity/interaction between Zn2+ and H+ from BAS in Ångstrom scale35. However, it was still difficult to specifically quantify the spatially separated metal ions and BAS in a large-scale range.

Herein, we presented the simple strategy to control the spatially separated Cu and BAS by ion exchange and physical mixture and thus to investigate its influence on the reaction performance of direct oxidation of methane. When AEI zeolite was ion-exchanged with different concentrations of Cu(NO3)2 solution, the density of Cu and BAS were calculated based on the chemical composition and cell parameters. When Cu/AEI was physically mixed with acidic zeolite in different integration manners, the spatially separated Cu and BAS were modeled. The reaction performance and spatial distance were well associated. These results would guide reasonable design catalysts and reaction conditions according to the target products.

Results and discussion

Effect of the spatially separated Cu and acid sites in Cu/AEI zeolite

The basic characterization results of xCu/AEI zeolites have been presented (Supplementary Figs. 1-4). Since no peak at 35.6 and 38.8o belonging to aggregated CuO nanoparticles was observed, the highly dispersed Cu species were obtained (Supplementary Fig. 1)36. The UV-vis spectra, HRTEM images, HADDF-STEM images, and EDS mapping images further confirmed the highly dispersed feature of Cu species in xCu/AEI samples (Fig. 1a–d) and Supplementary Figs. 58)37.

Fig. 1: Effects of Cu density and acid amount on reaction performance of direct oxidation of methane.
figure 1

EDS elemental mapping images of overlay Al and Cu for (a) 1Cu/AEI, (b) 5Cu/AEI, (c) 50Cu/AEI, (d) 500Cu/AEI zeolite catalysts. Reaction performance of (e) 1Cu/AEI, (f) 5Cu/AEI, (g) 50Cu/AEI, (h) 500Cu/AEI at 350 oC. Reaction conditions: 100 mg catalyst, CH4/N2O/H2O/Ar = 10/10/2/3 ml·min−1, WHSV = 15000 ml·g−1·h−1. rHydrocarbons = 2(rC2H4 + rC2H6) + 3(rC3H6 + rC3H8)+4(rC4H8 + rC4H10)+5rC4+. i The relation of Cu density with the strong acid amount; (j) product selectivity as a function of Cu density (TOS = 1.16 h); (k) product selectivity as a function of the strong acid amount (TOS = 1.16 h).

The xCu/AEI zeolite catalysts with varying Cu content (Supplementary Figs. 812 and Supplementary Tables 13) were used to perform the intimacy effect between Cu and BAS on the reaction performance of DMTM (Supplementary Fig. 13). Although Al and Cu can be visualized through EDS mapping images (Fig. 1a–d, Supplementary Figs. 68), it is still full of challenges to quantify the variation in spatially separated BAS and Cu. The reaction performance is shown in Fig. 1e–h, MeOH was produced at the initial stage (TOS = 0.16 h) for all the samples and a close MeOH formation rate was obtained (ca. 18-25 μmol.g−1.min−1). However, as the reaction proceeded, MeOH was massively and rapidly converted to hydrocarbons on the catalysts with low Cu content within 5 h (Fig. 1e, f). The massive formation of hydrocarbons on 1Cu/AEI zeolite led to a severe decrease in the formation rate and selectivity of MeOH (from 25 to 1.3 μmol·g−1·min−1 and from 25 to 2%, respectively) (Fig. 1e). After a few hours of reaction, the coke was deposited on the acid sites, resulting in the inactivation of acid sites, which were confirmed by comparison of the results of NH3-TPD and TG-DTA for the fresh and spent zeolites (Supplementary Figs. 14, 15). In addition, another reason that caused the gradually declined hydrocarbon formation rate was dealumination in the process of activation and reaction, which has been proven by comparison of the 27Al MAS NMR curves of the fresh and spent zeolites (Supplementary Fig. 16). It is noteworthy that the activity of Cu sites was still maintained, thus the productivity of MeOH was able to recover (Supplementary Figs. 14c, d). As for hydrocarbons, C2= was formed first with the highest selectivity in hydrocarbons. With the Cu content increasing and the corresponding acid amount decreasing (5Cu/AEI), the formation time of hydrocarbons was prolonged in comparison with 1Cu/AEI, indicating that the transformation of MeOH from the Cu sites to the acid sites required a longer time possibly due to the farther separation (Fig. 1f). On the contrary, almost no hydrocarbons were observed on the catalysts with more Cu content and less acid sites (Fig. 1g, h). We considered that the significant difference in the formation of hydrocarbons was due to the reduced amount of BAS (Supplementary Table 2), or the farther and more unattainable spatially separated Cu and the acid sites (Supplementary Fig. 17). 50 Cu/AEI and 500 Cu/AEI showed different CO2 selectivity compared to 1Cu/AEI and 5Cu/AEI, signifying that the excessive Cu amount led to over-oxidation23.

Since it was agreed that the peak centered ca. 475 oC in the NH3-TPD curve was due to NH3 strongly adsorbed on the BAS, this peak was used to inspect the impact of BAS content or strong acid amount on the MTH reaction38. The impact of Cu density (Supplementary Table 1) and strong acid amount (Supplementary Table 2) on the product selectivity of direct oxidation of methane at TOS = 1.16 h was visualized (Fig. 1i–k). Cu density was inversely proportional to strong acid amount (Fig. 1i), due to the Cu ions replacing H of Si(OH)Al. CO2 selectivity proportionally grew to Cu density, while the opposite trend was observed for hydrocarbon selectivity. In addition, the selectivity of (MeOH+DME) displayed the hill shape along with Cu density (Fig. 1j). Therefore, the hydrocarbon selectivity increased along with the strong acid amount (Fig. 1k), the selectivity of CO and CO2 showed opposite trends, and that of MeOH and DME peaked at the medium amount of strong acid sites.

Herein, we considered schematic diagrams of the spatially separated Cu and BAS for 1Cu/AEI and 500Cu/AEI (Supplementary Figs. 17a, b). Based on the cell parameters and the chemical composition, and most importantly, assuming that all elements were uniformly distributed and ignoring Al zoning, a unit cell (a = 13.7 Å, b = 12.6 Å, c = 18.5 Å) of 1Cu/AEI and 500Cu/AEI contained 0.24 and 1.44 Cu atoms (Supplementary Table 1), respectively. Namely, 4 unit cells shared ca. 1 Cu atom in 1Cu/AEI and ca. 6 Cu atoms in 500Cu/AEI. Simply assume that a single Cu corresponds to a single Al, and Cu pairs correspond to Al pairs. When 4 unit cells shared ca. 1 Cu atom, 24 Al atoms, and thus 23 H atoms from BAS, the spatially separated Cu and acid sites would be relatively close in Ångstrom scale. Consequently, it was quite possible to transport the intermediate MeOH from Cu sites to acid sites and convert it to hydrocarbons. However, when 4 unit cells shared ca. 6 Cu atoms, 24 Al atoms, and thus 18 H atoms from bridging hydroxyl. The possibility that MeOH produced on Cu sites was transported to acid sites was reduced. Therefore, the maximum selectivity of hydrocarbons declined from 95% for 1Cu/AEI to 0.3% for 500Cu/AEI. Furthermore, the increased Cu density resulted in the closely located Cu and Cu sites, which was visible in the representative and comparison HAADF-STEM images of 1Cu/AEI and 500Cu/AEI (Supplementary Fig. 7), and thus the methanol intermediate was over-oxidized to CO and CO223.

To verify our hypothesis, the diffusion velocity was adjusted by varying the flow rate and catalyst amount of 5Cu/AEI and 500Cu/AEI zeolites (Supplementary Figs. 18, 19). When the total flow rate was doubled by raising Ar flow rate from 3 to 28 ml·min−1, which meant that the weight hourly space velocity (WHSV) augmented from 15,000 to 30,000 ml·g−1·h−1, the peak selectivity of hydrocarbons for 5Cu/AEI was enlarged from ca. 45 to 65% and the massive formation time of hydrocarbons was advanced from 4.16 to 2.16 h (Supplementary Fig. 18a, b). While for 500Cu/AEI, growing WHSV improved the selectivity of CO2 and DME (Supplementary Fig. 18c, d). These results suggested that intensification of space velocity was able to improve the mass transportation between Cu and acid sites as well as between Cu and Cu sites.

Moreover, similar results were obtained by regulating the catalyst amount from 25 to 200 mg. For 5Cu/AEI, the peak selectivity of hydrocarbons was amplified from ca. 3 to 75%, and the massive formation time of hydrocarbons was advanced to 1.16 h (Supplementary Figs. 19a, b). Because MeOH is first formed on the Cu sites of Cu/AEI zeolite in the front bed, and then, the produced MeOH goes through the acid sites of Cu/AEI in the back bed and is converted to hydrocarbons30. In the case of 500Cu/AEI, more catalyst amount resulted in higher CO2 selectivity due to the dominant role of Cu sites of 500Cu/AEI in the back bed (Supplementary Fig. 19c, d).

Subsequently, the acid density was adjusted by calcination at 550–850 oC after Cu exchange to guarantee a similar total Cu and Al content39. Aggregated Cu clusters were not observed even after calcination at 850 oC (Supplementary Figs. 2026). The strong acid amount decreased with the calcination temperature growing from 550 to 850 oC (Supplementary Fig. 26 and Supplementary Table 2). As for the reaction performance (Fig. 2a–c), the selectivity of total hydrocarbons peaked ca. 40% at 4.16 h on 5Cu/AEI-550 (i.e. 5Cu/AEI zeolite). Continuing to diminish the strong acid amount, the selectivity of total hydrocarbons peaked at ca. 20 % at 6.16 h on 5Cu/AEI-750 and was less than 2% within 10.16 h on 5Cu/AEI-850. One should note that the methanol formation rate of 5Cu/AEI-850 amplified and stabled at ca. 33 μmol·g−1·min−1 (1980 μmol·g−1·h−1) due to the low selectivity of total hydrocarbons, which was the source of carbon deposit. The relationship between the strong acid amount and product formation rate was depicted in Fig. 2d40. Only the formation rate of hydrocarbons was proportional to the strong acid amount, both the formation rate of total products and (MeOH+2*DME) were inversely proportional to the strong acid amount. Considering the mechanism of acid amount reduction by calcination (Fig. 2e), the schematic diagram of the relationship between Cu and BAS was depicted in Fig. 2f, g and Supplementary Fig. 27. It is worth pointing out that cations such as Cu2+ and Na+ on the extra framework of zeolite can stabilize Al in the framework41. Thus, dealumination via high-temperature calcination mainly occurred on Al of Si(OH)Al (Fig. 2f, g). When the sample was calcined at a low temperature (550 oC), most of Al atoms in the framework were reserved. The content of BAS was sufficient and the spatially separated Cu and acid sites was relatively close (Fig. 2f). Therefore, the MeOH and DME intermediates were transported to the acid sites and converted to hydrocarbons (Fig. 2f). When 5Cu/AEI was calcined at higher temperatures (750 and 850 oC), the vacancies and reduced acid content happened because of dealumination (Fig. 2g). Correspondingly, the intimacy between Cu and acid sites weakened, and thus delivery of MeOH from Cu sites to acid sites took longer time or even was difficult to deliver from Cu sites to the farther acid sites (Fig. 2g)42. Simultaneously, because the Cu content and Cu states offered negligible changes (Supplementary Figs. 24, 25), few MeOH and DME were over-oxidized to CO2 (Fig. 2c). The carbon deposit amount of 5Cu/AEI-t decayed from 8.0 to 5.2% with t increasing from 550 to 850 oC (Supplementary Fig. 28).

Fig. 2: Effects of acid amount on reaction performance of direct oxidation methane.
figure 2

Stability test of (a) 5Cu/AEI-550, (b) 5Cu/AEI-750, (c) 5Cu/AEI-850 in DMTM reaction at 350 oC. Reaction conditions: 100 mg catalyst, CH4/N2O/H2O/Ar = 10/10/2/3 ml·min−1, WHSV = 15000 ml·g−1·h−1. d Product formation rate as a function of strong acid amount at TOS = 4.16 h, rHydrocarbons = 2(rC2H4 + rC2H6) + 3(rC3H6 + rC3H8)+4(rC4H8 + rC4H10)+5rC4+, rTotal = rCO + rCO2 + rMeOH + 2(rDME + rC2H4 + rC2H6) + 3(rC3H6 + rC3H8)+ 4(rC4H8 + rC4H10) + 5rC4+. e FTIR spectra in the OH stretching region of 5Cu/AEI-t samples collected at −120 oC after activation at 500 oC in vacuum for 1 h. Schematic depictions of (f) closely (5Cu/AEI-550) and (g) distantly (5Cu/AEI-850) separated Cu and acid sites on the reaction performance of direct oxidation of methane.

Effect of the spatially separated Cu/AEI and H-AEI

Since it was difficult to measure the spatial distance between Cu and acid sites in Cu/AEI zeolites, the physical mixture of Cu/AEI zeolite and acidic zeolite was adopted to explore the effect of spatial separation on reaction performance. Here, 50Cu/AEI zeolite catalyst was adopted as the Cu component because hydrocarbons were not generated and the CO2 selectivity was moderate (Fig. 1g). Meanwhile, H-AEI zeolite was used as the acidic component (Supplementary Table 2). A series of integration manners between 50Cu/AEI and H-AEI zeolite were used to regulate the spatially separated Cu and acid sites from the millimeter to nanometer. Based on the literature, the spatially separated two functional granules was roughly considered as the sum of two radii for the granule stacking and powder-grinding manner30. In this study, the same size of pellets (400–600 μm) and particles (200 nm* 600 nm* 600 nm) of Cu/AEI and H-AEI zeolite catalysts were used, and the spatially separated Cu and acid sites was approximately equal to the size of pellets (400–600 μm) and particles (200–600 nm) of AEI zeolite.

The spatially separated Cu species in 50Cu/AEI and acid sites in H-AEI was regulated from millimeter (Fig. 3a), micrometer (Fig. 3b) to nanometer (Fig. 3c) by attempting the dual-bed, granule stacking, and powder-grinding modes. The separation of 50Cu/AEI and H-AEI by quartz wool in one reactor, i.e., the dual-bed model, led to no methanol being observed during the initial 6 h (Fig. 3a), suggesting that all the intermediate MeOH was transferred to hydrocarbons on the acid sites of the second-bed H-AEI zeolite. Meanwhile, the total selectivity of hydrocarbons reduced from 42 to 13% within 6 h. In this model, the formation rate of total hydrocarbons peaked at the beginning (TOS = 0.16 h), signifying that the intermediate MeOH was transported to the acid sites quickly and easily. In addition, the acid sites in this model were easy to inactivate, possibly due to the acidic zeolite located in the second bed. When we stacked granules of 50Cu/AEI and H-AEI in the size of 400-600 μm, the selectivity of total hydrocarbons was improved from 25 to 75% as the reaction proceeded to 6 h (Fig. 3b). The gradually increased selectivity of hydrocarbons implied that the MeOH produced on 50Cu/AEI was converted at the acid sites on H-AEI step-by-step. Furthermore, when the powder-grinding 50% 50Cu/AEI and 50% H-AEI was applied in the oxidation of methane reaction, the maximum selectivity of total hydrocarbons was up to 92% at TOS = 1.16 h (Fig. 3c). Compared to the results of Fig. 3, the selectivity of CO2 was decreased to 24–7% from 58–50% with the spatially separated Cu and acid sites declining. Hence, it has been intuitively proved that the distantly separated Cu and acid sites was beneficial to the production of CO2, while the close separation was favorable to the formation of hydrocarbons.

Fig. 3: Effects of mixing methods of 50Cu/AEI and H-AEI on reaction performance of direct oxidation of methane.
figure 3

a Dual bed (50 mg 50Cu/AEI + 50 mg H-AEI); (b) granule-stacking of 50 mg 50Cu/AEI and 50 mg H-AEI with size of 400–600 μm; (c) grinding-mixing 50% 50Cu/AEI and 50% H-AEI and then compressing to granules with size of 400–600 μm. Reaction conditions: total catalyst mass: 100 mg, 350 oC, CH4/N2O/H2O/Ar=10/10/2/3 ml·min−1, WHSV = 15,000 ml·g−1·h−1. rHydrocarbons = 2(rC2H4 + rC2H6) + 3(rC3H6 + rC3H8)+4(rC4H8 + rC4H10)+5rC4+.

It was worth pointing out that the product distribution, especially CO and CO2, was highly dependent on the Cu component. The different Cu species on 50Cu/AEI, 500 Cu/AEI, and 5Cu/AEI-850 zeolite catalysts were used as the Cu sites (Supplementary Figs. 9c, d, 25c) since almost no hydrocarbons were obtained (Figs. 1g, h, 2c). The reaction performance of stacking granules (50Cu/AEI and H-AEI), (500Cu/AEI and H-AEI), and (5Cu/AEI-850 and H-AEI) was compared (Supplementary Fig. 29). In the premise of a similar spatially separated Cu and BAS, ca. 27, 48, and 6% CO2 selectivity were obtained on (50Cu/AEI and H-AEI), (500Cu/AEI and H-AEI), and (5Cu/AEI-850 and H-AEI), respectively, due to the different Cu component.

Additionally, AEI and BEA zeolites were adopted as the small and large pore zeolites, respectively, to research the effect of Cu component and spatially separated Cu and BAS (Supplementary Fig. 30a, b). 5Cu/BEA zeolite was prepared using the same method as 5Cu/AEI. Analogous to 5Cu/AEI, the XRD pattern, UV-vis spectrum, HAADF-STEM images and EDS mapping images of 5Cu/BEA indicated that aggregated CuO nanoparticles were not observed (Supplementary Fig. 30c–e). The broad adsorption band at 650-850 nm of 5Cu/BEA was assigned to the d → d transition43. Combine with the NO adsorption FTIR spectra of 5Cu/BEA, the Cu2+(NO) bands at higher wavenumbers of 1959, 1948, and 1936 cm−1 possibly indicated the Cu clusters in 5Cu/BEA zeolite (Supplementary Fig. 30f). The NH3-TPD curves specified that the acid amount of 5Cu/BEA was less than that of 5Cu/AEI due to the higher Si/Al ratio and larger pore size of Beta zeolite (Supplementary Fig. 30g and Supplementary Table 2). The Cu clusters, less acid amount, weaker acid strength, and the distantly separated Cu and BAS on 5Cu/BEA zeolite resulted in ca. 75 % CO2 selectivity and no hydrocarbons, while the main product of 5Cu/AEI was MeOH (48 %) (Supplementary Fig. 30h, i). Additionally, the compared characterization and reaction results of 1Cu/AEI and 1Cu/BEA zeolites further confirmed the statement (Supplementary Fig. 31). Therefore, the result reasonably validated our hypothesis that the distantly separated Cu and acid sites, and the closely separated Cu and Cu sites were convenient for the formation of CO2.

Effect of the acidic zeolite

The reaction performance of different topology structure zeolites as the acidic zeolites, including H-AEI (BASF SE), H-MFI (Japan Reference Catalyst, JRC-Z5-30NH4), and H-BEA (Zeolyst International, CP814E*) (Supplementary Table 3 and Supplementary Fig. 32), and 5Cu/AEI-850 as the Cu component, in the granule-stacking model with the size of 400–600 μm were compared in Fig. 4. First of all, the product distribution was different. For the small pore zeolite AEI, the selectivity of hydrocarbons was up to 90%, and neither MeOH nor DME was remained as the intermediates (Fig. 4d). While for the medium (MFI) and large (BEA) pore zeolites, the selectivity of hydrocarbons dropped to 30 and 14 % (Fig. 4e, f), respectively. Meanwhile, ca. 60 and 75% selectivity of (MeOH+DME) were left through MFI and BEA, respectively, indicating that both topology structure and acidity of acidic zeolite played a vital role. The remaining MeOH and DME were the apparent indication of the insufficient acid content, which was confirmed by the NH3-TPD results (Fig. 4b, c and Supplementary Table 2), and an indication of inaccessible acid sites. Moreover, another possible reason was the spatially separated Cu and acid sites. As previously mentioned, medium and large pore-size zeolites meant a distantly separated acid site. When medium and large pore-size zeolites were used as the acid sites, the spatially separated Cu sites in 5Cu/AEI-850 zeolite and acid sites in H-MFI or H-BEA zeolites became farther, resulting in the lower selectivity of hydrocarbons. Meanwhile, MeOH and DME that have not been promptly converted were detected by GC-FID as the final products.

Fig. 4: Effects of the acidic zeolite structure on reaction performance of direct oxidation of methane.
figure 4

Deconvolution of NH3-TPD curves of (a) H-AEI, (b) H-MFI, (c) H-BEA zeolites. Reaction performance of 5Cu/AEI-850 granular stacking with (d) H-AEI, (e) H-MFI, (f) H-BEA in the direct oxidation of methane. Reaction conditions: 50 mg 5Cu/AEI-850 and 50 mg H-zeolite with the granule size of 400–600 μm, 350 oC, CH4/N2O/H2O/Ar=10/10/2/3 ml·min−1, WHSV = 15,000 ml·g−1·h−1. rHydrocarbons = 2(rC2H4 + rC2H6) + 3(rC3H6 + rC3H8)+4(rC4H8 + rC4H10)+5rC4+.

In our previous work, CHA zeolite was used in DMTM reaction and ethylene (C2=) was observed as the main hydrocarbon component, and we ascribed to the “hydrocarbon pool (HCP)” mechanism44, which was proposed and accepted in the MTO or MTH process10. There was no dispute under that circumstance since the C2= selectivity of CHA zeolites was reported higher than other olefins in the MTH reaction due to the cage effect45,46,47. Similarly, in the publication of Su et al. the mechanism of syngas to light olefins on ZnCrOx/AlPO-18 catalyst was also ascribed to the “HCP” mechanism14. However, in this study, we found that C2= was always the main contributor among the hydrocarbons regardless of the acidic properties and the spatially separated Cu and acid sites (Supplementary Figs. 3335). It was different from the MTH reaction result taking H-AEI zeolite as an example (Supplementary Fig. 36) and also different from the reports that propylene (C3=) was the main hydrocarbon product39,40,45,48. To shed light on the doubt, we proposed the possible reaction route in Fig. 5, where MeOH is produced on the Cu sites as the initial product, and two MeOH molecules would form DME on the acid sites through the dehydration reaction49. In this study, the production of hydrocarbons was considerably dependent on the transportation of MeOH to the acid sites. Thus, both the closely located Cu and acid sites (Figs. 13) and the accelerated mass transfer (Supplementary Fig. 18) were able to realize the high selectivity and earlier formation of hydrocarbons. When MeOH is promptly transported to the acid sites and CnH(2n+1) intermediates are not released from the ((CnH(2n+1))···O)- sites, it is possible to add another methyl group and then yield to CmH2m, where m is equal to n + 1 (n ≥ 1). The hydrocarbons distribution of 5Cu/AEI (Supplementary Fig. 37) exposed that various alkenes and alkanes were gradually produced then disappeared. In addition, the hydrocarbons selectivity was increased with the flow rate growing (Supplementary Fig. 18b). Both the two phenomena were able to verify the conjecture. In fact, the methoxymethyl cation mechanism and methane–formaldehyde mechanism44,50 were proposed as competitive mechanisms and important complements to the “HCP” mechanism in MTH reaction51. In this work, formaldehyde (HCOH) was not observed due to the limitation of the high reaction temperature (350 oC) and analytical instrument. Thus, there is not enough evidence for methane–formaldehyde mechanism. On the other hand, the formative sequence of hydrocarbons in this study was consistent with the methoxymethyl cation mechanism, which has been reported as the initial C-C bond formation mechanism in MTH reaction50,52,53 and syngas to olefins reaction54,55. Furthermore, when the spatially separated Cu and Cu sites is close enough, methanol that is released from the Cu sites is possibly over-oxidized to CO and CO2 on the adjacent Cu sites (Fig. 1c, d) and Supplementary Figs. 18c, d, 19c, d, 29a, b).

Fig. 5: Possible reaction pathway of direct oxidation of methane to hydrocarbons via MeOH/DME intermediates.
figure 5

a Methane is directly converted to methanol (DMTM) on active Cu sites and (b) methanol is converted to hydrocarbons (MTH) on acid sites.

We have successfully revealed the effect of spatially separated Cu and BAS on the reaction performance of direct oxidation of methane by ion exchange of different concentrations of Cu(NO3)2 solution with AEI zeolite, and by physical mixture of Cu/AEI with acidic zeolite in different integration manners. It has been exposed that the low Cu density in Cu/AEI zeolite implied the closely separated Cu and BAS, while the high Cu density in Cu/AEI zeolite meant the closely separated Cu and Cu sites and thus the distantly separated Cu and BAS. Closely located Cu and BAS were beneficial for forming hydrocarbons due to the tandem conversion of methanol/DME on the nearby acid sites. Closely located Cu and Cu sites were prone to over-oxidize methanol to CO2 on the adjacent Cu sites. The steady and efficient production of methanol from methane was based on the trade-off of spatially separated Cu and BAS. In addition, the physical mixture Cu/AEI with acidic zeolite in different integration manners also exhibited that the closely separated Cu and BAS was advantageous for the generation of hydrocarbons, and the product distribution was highly dependent on the acidic properties and structure of acidic zeolites. These findings will guide to design of efficient catalysts to control side reactions and increase methanol yield in direct oxidation of methane to methanol reaction, as well as open up an avenue for direct oxidation of methane to hydrocarbons at low temperatures.


Catalyst preparation


Aluminosilicate AEI-type zeolite (NH4+ form, Si/Al=6.9) provided by BASF SE was used as the parent zeolite56. The Cu density was regulated by using the parent zeolite to ion exchange with 1, 5, 50, 500 mmol/L Cu(NO3)2 (FUJIFILM, Wako Special Grade) solutions, respectively, at 80 oC stirring for 24 h (Solid/Liquid ratio was 1 g/100 ml). The solid products were washed, dried at 100 oC overnight, and calcined at 550 °C for 5 h in air. The obtained samples were denoted as xCu/AEI, where x means the Cu(NO3)2 concentration.


To adjust the acid density, the dried 5Cu/AEI was calcined at 550, 750, and 850 °C, respectively, for 5 h in air. The obtained samples were denoted as 5Cu/AEI-t, where t means the calcination temperature. Note that, 5Cu/AEI and 5Cu/AEI-550 are the same sample.

Acidic zeolites

The H-type aluminosilicate AEI (H-AEI), MFI (H-MF), BEA (H-BEA) zeolites were prepared by calcination AEI (BASF SE, NH4+ form), MFI (Japan Reference Catalyst, JRC-Z5-30NH4, NH4+ form), Beta (Zeolyst International, CP814E*, NH4+ form) zeolites at 550 °C for 5 h in air.

Characterization of catalysts

XRD pattern was collected on a Rint-Ultima III (Rigaku) using a Cu Kα X-ray source (40 kV, 40 mA).

Elemental analysis was performed on an inductively coupled plasma-atomic emission spectrometer (ICP-AES, Shimadzu ICPE-9000).

Field-emission scanning electron microscopic (FE-SEM) images of the powder samples were obtained on SU9000 (Hitachi) microscope operating at 1 kV.

High-angle annular dark field scanning transmission electron microscopy (HAADF-STEM) imaging and energy dispersive spectrometry (EDS) mapping were performed on a FEI Themis Z microscope equipped with an XFEG field electron source and double aberration corrector operated on 300 keV. The HAADF-STEM images were acquired with camera length as 115 mm while the beam convergence was 25.1 mrads. The pixle size is 37 pm and the dwell time is 2 us/pixle. The collection angle of HAADF detector was set to 48–200 mrads. STEM-EDS elemental maps were acquired with 4 us/pixel acquisition time using Velox software

Nitrogen adsorption and desorption measurements to obtain information of the micro- and meso-porosities were conducted at −196 oC on a Belsorp-mini II (MicrotracBEL).

The amount of coke in the spent samples was determined by the weight loss from 250 to 800 oC in a thermogravimetric (TG) profile, which was performed on a thermogravimetric-differential thermal analyzer (TG-DTA, RigakuThermo plus EVO II).

Solid-state 29Si and 27Al MAS NMR spectra were measured on a JEOL ECA-600 spectrometer at a resonance frequency of 156.4 MHz using a 4 mm sample rotor with a spinning rate of 15.0 kHz. The 29Si and 27Al chemical shifts were referenced to −34.12 and −0.54 ppm, polydimethylsiloxane (PDMS) and AlNH4(SO4)2·12H2O, respectively.

Temperature-programmed ammonia desorption (NH3-TPD) profiles were recorded on Multitrack TPD equipment (Japan BEL). Typically, 25 mg of catalyst was pretreated at 600 oC in He (50 mL·min−1) (Taiyo Nippon Sanso Co., Ltd., 99.99995%) for 1 h and then cooled to 100 oC. Prior to the adsorption of NH3 (Toho Chemical Industry Co., Ltd., 28.8% NH3 in He), the sample was evacuated at 100 oC for 1 h. Approximately 2500 Pa of NH3 was allowed to contact with the sample at 100 oC for 10 min. Subsequently, the sample was evacuated to remove weakly adsorbed NH3 at the same temperature for 30 min. Finally, the sample was cooled to 100 oC and heated from 100 to 600 oC at a ramping rate of 10 oC min−1 in a He flow (50 mL·min−1). A thermal conductivity detector (TCD) and BELMass (MicrotacBEL Corp.) were used to monitor desorbed NH3. The amount of acid sites was determined by the fitting peak area of the profiles.

UV-vis spectra were collected in the range of 190–900 nm on a V-650DS spectrometer (JASCO). The diffuse reflectance spectra were converted into the absorption spectra using the Kubelka-Munk function.

Fourier Transform Infrared (FTIR) spectra were obtained by using a JASCO 4100 FTIR spectrometer equipped with a triglycine sulfate (TGS) detector. IR spectra of the clean disk were recorded in vacuo at 25 oC to obtain background spectra. The sample was pressed into a self-supporting disk (20 mm diameter, 30 − 60 mg) and placed in an IR cell attached to a closed-gas circulation system. The sample was pretreated by evacuation at 500 oC for 1 h, followed by measuring the hydroxyl vibration at −120 oC and then adsorption of 5 − 120 Pa NO (Taiyo Nippon Sanso Co., Ltd., 99%) or 5−1000 Pa CO (Taiyo Nippon Sanso Co., Ltd., 99.95%) at −120 oC. The IR spectra resulting from the subtraction of the spectra before adsorption from those with NO or CO adsorption are shown unless otherwise noted.

Catalyst tests

The continuous oxidation of methane reaction was performed in a fixed-bed flow reactor. The online-reaction-analysis system was equipped with two six-port inlet valves. In a typical test, 100 mg of catalyst in a granular form (400 − 600 μm) was charged into a quartz tube (inner diameter 4 mm), which was placed in an electric tube furnace. The catalyst was pretreated at 500 °C for 1 h in the Argon flow. The reaction was conducted at 350 °C in a flowing gas mixture of CH4 (Taiyo Nippon Sanso Co., Ltd., Ltd., 99.999%), N2O (Koike Precision Instruments, 99.99%), H2O (Homemade deionized water), and Ar (Toho Chemical Industry Co., Ltd., General industrial) with flow rates of 10, 10, 2, and 3 mL·min−1. The outlet gas, containing products, unreacted CH4 and N2O, was analyzed by two on-line gas chromatographs (GC; GC-2014, Shimadzu). One of the GCs was used with a Shin carbon ST 50/80 packed column (Agilent Technologies, inner diameter 3 mm, length 6 m) and a TCD detector. Specifically, GC-TCD was used to detect H2, N2O, CO, CO2 and CH4. The other GC was equipped with a Porapak Q 80/100 packed column (Agilent Technologies, inner diameter 3 mm, length 6 m), a flame ionization detector (FID), and a methanizer. The GC-FID was used to investigate CH4, and the produced methanol (MeOH), dimethyl ether (DME), alkanes, and alkenes. The yield of each carbon-containing product was calculated by considering the number of carbon atoms. The methane conversion in this study was defined as the total obtained products, and calculated as:

$${{{{{{\rm{C}}}}}}}_{{{{{{\rm{CH}}}}}}4}=\frac{\sum (i*{Ci})}{\sum (i*{Ci})+{CH}4}$$

where CCH4 is the CH4 conversion, i is the number of carbon atoms in product Ci, Σ(i*Ci) is the total amount of carbon of all the products, and CH4 is the amount of CH4 detected at the same time.

The N2O conversion was calculated as:


where CN2O is the N2O conversion, ni is the initial N2O molar weight, na is the N2O molar weight after reaction.

Note that, CH4 and N2O conversion were calculated according to different methods, thus they were not equal in value.

The product selectivity was calculated as:

$${{{{{{\rm{S}}}}}}}_{{{{{{\rm{Ci}}}}}}}=\frac{i*Ci}{\sum (i*Ci)}$$

where SCi is the selectivity of the product Ci, Σ(i*Ci) is the total amount of carbon of all the products.

The product yield was calculated as:

$${{{{{{\rm{Y}}}}}}}_{{{{{{\rm{Ci}}}}}}}=\frac{i*Ci}{\sum (i*Ci)+CH4}$$

where YCi is the yield of the product Ci, Σ(i*Ci) is the total amount of carbon of all the products, and CH4 is the amount of CH4 detected at the same time.

The product formation rate was calculated as:


where RCi is the formation rate of product Ci, FCH4 is the initial flow rate of CH4, mcat is the mass of catalyst. The total formation rate rCH4 = rCO + rCO2 + rCH3OH + 2*rDME + 2*(rC2H4 + rC2H6)+3*(rC3H6 + rC3H8) + 4*(rC4H8 + rC4H10) + 5*rC5+.

Statistics and reproducibility

We repeated the main catalysts for the catalyst test. All the experimental results were able to be reproduced within a small margin of error. No statistical method was used to predetermine the sample size. No data were excluded from the analyses. The experiments were not randomized. The investigators were not blinded to allocation during experiments and outcome assessment.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.