Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.
Nature Communications volume 16, Article number: 1622 (2025 ) Cite this article adsorption carbon capture
Crystalline metal-organic frameworks (MOFs) exhibit enormous potential application in gas separation, thanks to their highly porous structures and precise pore size distributions. Nevertheless, the inherent limitations in mechanical stability of crystalline MOFs cause challenges in processing MOF powders into bulky structures, particularly for membrane filtrations. Melt-quenched MOF glasses boast excellent processability due to liquid-like properties. However, the melting process diminishes the inherent porosity, leading to reduced gas adsorption capacities and lower gas diffusion coefficients. In this work, we demonstrated that enhancing the porosity of MOF glasses is achievable through topological engineering on the crystalline precursors. Crystalline zeolitic imidazolate frameworks (ZIFs) with large 12-membered rings pores, including AFI and CAN topology, were synthesized by using both structure-directing agents and mixed organic ligands. The large pores are partially preserved in the melt-quenched glass as evidenced by high-pressure CO2 absorption at 3000 kPa. The agAFI-[Zn(Im)1.68(bIm)0.32] glass was then fabricated into self-supported membranes, which shows high gas separation performance, for example, CO2 permeance of 3.7 × 104 GPU with a CO2/N2 selectivity of 14.8.
Excessive energy consumption, particularly from the continuous burning of fossil fuels, has significantly worsened climate change1,2,3,4. Extensive research has focused on clean energy production and carbon capture to reduce the release of greenhouse gases, mainly CO2 emissions, into the atmosphere during the process of extracting, producing, and using energy2,3,4. MOFs are cage-like structures containing metal nodes and organic ligands connected in infinite arrays5. The highly porous structures and precise pore size distributions of crystalline MOFs give them potential applications in various energy and environment-related gas separations, such as natural gas purification and CO2 capture6,7,8,9,10,11. Crystalline MOFs typically require processing, such as pelletization, coating, or sintering, to form bulky structures for device applications, especially for membrane modules12,13. However, their practical implementation is still hampered by inherent limitations in mechanical stability of microcrystalline powders14,15.
The discovery of MOF glasses, recognised as a fourth category of glass materials, given their dissimilarity to existing inorganic (non-metallic), organic and metallic families, opens up possible opportunities to process MOFs into macroscale, grain-boundary free morphologies16. ZIFs are a class of MOFs containing tetrahedral metal ions linked by imidazolate ions. A few types of ZIFs have been discovered with the capability to form glasses via melt-quenching17. These melt-quenched MOF glasses possess similar structures to amorphous silica and display viscoelastic states at temperatures above their glass transition temperatures (Tg). Several forming methods such as hot-pressing and remelting have been explored to fabricate bulk samples of both MOF glasses and MOF crystal glass composites (CGCs)18,19,20. However, upon glass formation, the inherent porosity in their crystalline precursors is reduced, which would be unfavourable for possible applications in gas separation compared with other pure polycrystalline MOFs10,21,22,23,24,25. Hence, a substantial enhancement in porosity is imperative for MOF glasses to emerge as practical solutions for gas separation challenges. The introduction of pore-inducing agents can enhance the porosity of MOF glass through the decomposition of chemicals added during glass formation, such as polymers21,26. Nevertheless, achieving precise manipulation of porosity in MOF glass at the Ångstrom scale remains a significant challenge.
Chemical composition of the crystalline MOFs, i.e., the type of metal nodes and organic ligands, is a decisive factor in determining their melting point and decomposition temperature27,28. Thus, to make both meltable and highly porous crystalline MOFs, it is logical to engineer a more porous topology by using similar metal nodes and organic ligands to those present in reported meltable MOFs, e.g., ZIF-4, [Zn(Im)2, Im=imidazolate], and ZIF-62, [Zn(Im)1.75(bIm)0.25, bIm=benzimidazolate]. The commonly employed ligand-directed strategy, which utilizes large-sized ligands to enhance porosity in crystalline MOFs, is not readily applicable to meltable MOFs. This is due to the introduction of extra ligands which may potentially lower the decomposition temperature of MOFs21.
Large solvent molecules have previously been used as the structure-directing agents (SDAs) to synthesize crystalline MOFs with highly porous topology without introducing extra ligands. For example, Shi et al. synthesized two large pore ZIF compounds with Im organic ligands, by using dibutylformamide (DBF) and dipropylformamide (DPF) solvents as SDAs, respectively29. The obtained ZIFs possess the zeolite CAN and AFI topology with 12-membered rings (MR), and are termed CAN-[Zn(Im)2] and AFI-[Zn(Im)2], respectively. AFI-[Zn(Im)2] displays pore apertures as large as 15.6 Å, which is larger than that of ZIF-4 (2.1 Å, 6 MR) and ZIF-62 (1.4 Å, 6 MR)30. However, CAN-[Zn(Im)2] and AFI-[Zn(Im)2] will recrystallize to the dense ZIF-zni phase with 4 MR before melting, similar with ZIF-424,31. This leads to a reduction in the porosity of the resultant melt-quenched glass, agZIF-zni32.
In this work, meltable MOFs that have highly porous topologies of 12 MR have been synthesised by using both SDAs and mixed ligands of imidazole (Im) and benzimidazole (bIm) as shown in Fig. 1. Bulky amide solvents were used as SDAs to create highly porous topologies, without introducing extra ligands. The mixed ligands of Im and bIm were used to suppress the recrystallization to dense ZIF-zni phase during the melting process and provide spatial support for porosity preservation. The obtained crystalline frameworks, termed AFI-[Zn(Im)1.68(bIm)0.32] and CAN-[Zn(Im)1.73(bIm)0.27], possess similar chemical compositions to the prototypical glass former ZIF-62, though they have far greater porosity. The enhanced porosity carries over to the glass upon melt quenching, resulting in high H2, CO2, and CH4 adsorption capacities of agAFI-[Zn(Im)1.68(bIm)0.32], while decomposition was found during the melt quenching process of CAN-[Zn(Im)1.73(bIm)0.27]. Moreover, a free-standing membrane was fabricated with agAFI-[Zn(Im)1.68(bIm)0.32], featuring record-breaking CH4/N2 and CO2/N2 separation performance. This porous MOF glass shows promising applications in energy gas purification and CO2 capture.
SDAs create highly porous topologies without introducing extra ligands. The mixed ligands of Im and bIm suppress the recrystallization to dense ZIF-zni phase during the melting process, and provide spatial support for porosity preservation. a Synthesis of large-pore AFI-[Zn(Im)2-x(bIm)x] and agAFI-[Zn(Im)2-x(bIm)x] by using both SDAs and mixed ligands. b Synthesis of small-pore ZIF-62 (CCDC code SIWJAM) and agZIF-62 by using only mixed ligands. c Synthesis of AFI-[Zn(Im)2] (CCDC code PAJROK), thermal recrystallization ZIF-zni (CCDC code IMIDZB), and agAFI-[Zn(Im)2] by using only SDAs.
AFI-[Zn(Im)2] and CAN-[Zn(Im)2] with 12MR pore openings were fabricated by solvothermal synthesis29. The typical ZIF topologies were listed in Supplementary Fig. 1 and Supplementary Table 1. The porous crystalline structures of AFI-[Zn(Im)2] and CAN-[Zn(Im)2] were demonstrated by XRD and N2 adsorption isotherms (Supplementary Figs. 2 and 3, and Supplementary Table 2). Both AFI-[Zn(Im)2] and CAN-[Zn(Im)2] exhibited glass transition temperatures at ca. 590–600 K, and recrystallize to the dense ZIF-zni phase at ca. 800 K, before subsequently melting at 853 K, through partial decomposition of the organic ligand was observed during melting in each case (Supplementary Figs. 4 and 5). Although the chemical composition Zn(Im)2 is meltable according to previous results, existing Zn(Im)2 polymorphs all recrystallize to dense phase ZIF-zni before melting24,31. Partial substitution of Im by bIm however has been shown to result in crystalline frameworks of identical topology to ZIF-4, though which melt directly, rather than collapse to a dense framework first16. The effects of bIm content on the performance of ZIF-6227 and ZIF-832 have been reported. Henke et al.27 found that the bIm linker partially occupies two of the four available Im salt positions in the ZIF-62 structure. By reducing the ratio of Im and bIm, the Tm of ZIF-62 can be reduced from 703 K to 643 K. The pore size distribution indicates that during the melting and formation of agZIF-62, small cavities (with a diameter of approximately 3.5 Å) completely disappear, but medium-sized cavities (with a diameter of 5–6 Å) are significantly preserved, and even new larger cavities (with a diameter of 8 Å) are formed. In previous studies32,33, bIm was used to replace one of the three methylimidazole (mIm) joints in ZIF-8. As the content of bIm increased, the pore size significantly decreased, the pore size distribution narrowed, and the thermal stability also decreased. To obtain the highly porous structure of MOF glasses, the mixed-linker MOFs were synthesised by controlling ligand ratios of Im and bIm. The NMR results (Fig. 2a, and Supplementary Fig. 8) confirmed the existence of bIm in both frameworks, and the compositions were AFI-[Zn(Im)1.68(bIm)0.32] and CAN-[Zn(Im)1.73(bIm)0.27]. The powder XRD patterns of AFI-[Zn(Im)1.68(bIm)0.32] and CAN-[Zn(Im)1.73(bIm)0.27] were similar to the simulated patterns of AFI-[Zn(Im)2] and CAN-[Zn(Im)2] (Fig. 2b). Pawley refinement (Supplementary Fig. S6 and Table S4) indicated the unit cell dimensions and the space group of AFI-[Zn(Im)1.68(bIm)0.32] and CAN-[Zn(Im)1.73(bIm)0.27] which have little change comparing with AFI-[Zn(Im)2] and CAN-[Zn(Im)2]. Although single crystal structures were not obtained, the Pawley refinement results implied that the topology structures of AFI-[Zn(Im)1.68(bIm)0.32] and CAN-[Zn(Im)1.73(bIm)0.27] were similar with that of AFI-[Zn(Im)2] and CAN-[Zn(Im)2], respectively. The N2 adsorption amounts decrease after adding bulky bIm ligand in the framework, for example, the BET decreased from 1155 m2/g to 1073 m2/g for AFI-[Zn(Im)2] and agAFI-[Zn(Im)1.68(bIm)0.32], from 1033 m2/g to 603 m2/g for CAN-[Zn(Im)2] and agCAN-[Zn(Im)1.55(bIm)0.45]. The lower BET surface area of CAN-[Zn(Im)1.73(bIm)0.27] than CAN-[Zn(Im)2] (Supplementary Fig. 3) is attributed to that the synthesized CAN-[Zn(Im)1.73(bIm)0.27] has low crystallinity, resulting in low porosity of the crystals.
a1H NMR results of AFI-[Zn(Im)1.68(bIm)0.32] and CAN-[Zn(Im)1.73(bIm)0.27]. b XRD patterns of AFI-[Zn(Im)1.68(bIm)0.32], CAN-[Zn(Im)1.73(bIm)0.27], and simulations from the published crystallographic information files (cifs)29. c, d XRD patterns of amorphous MOFs: (c) aTAFI-[Zn(Im)2-x(bIm)x] and agAFI-[Zn(Im)1.68(bIm)0.32], and (d) aTCAN-[Zn(Im)2-x(bIm)x] and agCAN-[Zn(Im)1.55(bIm)0.45]. Enthalpic responses and weight loss of (e) AFI-[Zn(Im)1.68(bIm)0.32] and (f) CAN-[Zn(Im)1.73(bIm)0.27] measured by DSC and TGA separately, with a heating rate of 10 K/min. Thermal amorphization (TA), melting temperature (Tm), and glass transition temperature (Tg) are indicated. The light yellow and blue shaded region highlights the integrated area for the determination of ΔHm.
AFI-[Zn(Im)1.68(bIm)0.32] becomes thermally amorphous at 633 K, denoted as aT, and then melts at 705 K (Fig. 2c, e). The glass obtained from quenching at 735 K possesses an identical composition of agAFI-[Zn(Im)1.68(bIm)0.32] confirmed by 1H NMR (Supplementary Fig. 9a and Fig. 10a-b), and a glass transition temperature of 607 K (Fig. 2e). CAN-[Zn(Im)1.73(bIm)0.27] melts at 670 K, i.e. lower than AFI-[Zn(Im)1.68(bIm)0.32] (Fig. 2d), which was ascribed to the lower bIm content in CAN-[Zn(Im)1.73(bIm)0.27]27. NMR confirms that the agCAN glass had composition Zn(Im)1.55(bIm)0.45 (Supplementary Fig. 9b and Fig. 10c-d), which indicated some decomposition of the organic linkers upon melt quenching. The agCAN-[Zn(Im)1.55(bIm)0.45] possessed a Tg of 591 K that was also lower than agAFI-[Zn(Im)1.68(bIm)0.32]. The synthesized CAN-[Zn(Im)1.73(bIm)0.27] has low crystallinity and parasitic non-porous phase, resulting in low porosity of the crystals. The melting points of both AFI-[Zn(Im)1.68(bIm)0.32] and CAN-[Zn(Im)1.73(bIm)0.27] were similar to that of the prototypical ZIF-62 (Tm = 710 K). In addition, the SEM images displayed aggregate needle-like morphologies for both AFI-[Zn(Im)1.68(bIm)0.32] and CAN-[Zn(Im)1.73(bIm)0.27], and bulk boundary-free morphology for corresponding MOF glasses (Supplementary Fig. 11). Here, we observed solid-state thermal amorphization, prior to melting for both AFI-[Zn(Im)1.68(bIm)0.32] and CAN-[Zn(Im)1.73(bIm)0.27]. Importantly, we did not observe recrystallization to ZIF-zni phase for either AFI-[Zn(Im)1.68(bIm)0.32] or CAN-[Zn(Im)1.73(bIm)0.27] (Fig. 2c, d). Moreover, DSC testing of AFI-[Zn(Im)1.68(bIm)0.32] and CAN-[Zn(Im)1.73(bIm)0.32] were evaluated at varying heating rates (10 K, 20 K, and 30 K), the TA, Tm, and Tg were marked in Supplementary Figs. 12, 13 and Supplementary Table 5. The DSC results demonstrated that both AFI-[Zn(Im)1.68(bIm)0.32] and CAN-[Zn(Im)1.73(bIm)0.32] exhibited melting characteristics.
The amorphous nature of the melt quenched agAFI-[Zn(Im)1.68(bIm)0.32] and agCAN-[Zn(Im)1.55(bIm)0.45] glasses was confirmed by the structure factors S(Q) obtained through room-temperature X-ray total scattering (Supplementary Fig. 15). Subsequent Fourier transform yielded the pair distribution functions (PDFs) D(r) (Fig. 3 and Supplementary Fig. 16). The PDFs of both agAFI-[Zn(Im)1.68(bIm)0.32] and agCAN-[Zn(Im)1.55(bIm)0.45] were essentially featureless above 8 Å, indicating long-range disorder. Below 8 Å, PDF traces for both crystals and glasses are virtually identical, with only small differences in peak intensities.
Pair distribution functions, D(r), of (a) AFI-[Zn(Im)1.68(bIm)0.32] and ag AFI-[Zn(Im)1.68(bIm)0.32], and (b) CAN-[Zn(Im)1.73(bIm)0.27] and agCAN-[Zn(Im)1.55(bIm)0.45], and the inserts assign PDF peaks to the main short atom-atom distances in the structure. Zn, blue; C, dark green; N dark yellow; H, omitted.
The processability of MOF glasses is an important advantage compared with MOF crystals, allowing customization for various applications. Assessing the material’s mechanical properties is crucial for determining its ability to withstand industrial conditions20. The mechanical properties of agAFI-[Zn(Im)1.68(bIm)0.32] and agCAN-[Zn(Im)1.55(bIm)0.45] were measured by using nano-indentation (Supplementary Fig. 17). The Young’s modulus (E) was 6.58 ± 0.10 GPa and 6.63 ± 0.05 GPa, and the hardness (H) was 0.670 ± 0.013 GPa and 0.656 ± 0.013 GPa for agAFI-[Zn(Im)1.68(bIm)0.32] and agCAN- [Zn(Im)1.55(bIm)0.45], respectively, which are similar to those of agZIF-62 (E = 6.58 ± 0.02 GPa, H = 0.656 ± 0.005 GPa).
Due to the relatively low BET surface area of CAN-[Zn(Im)1.73(bIm)0.27] and the slight decomposition during melting, the investigation of gas separation performance was focused on AFI-[Zn(Im)1.68(bIm)0.32] and agAFI-[Zn(Im)1.68(bIm)0.32]. The gas isotherms of AFI-[Zn(Im)1.68(bIm)0.32] and agAFI-[Zn(Im)1.68(bIm)0.32] were tested by H2, CO2, N2, and CH4 (Fig. 4a–d). As expected, crystalline AFI-[Zn(Im)1.68(bIm)0.32] with open nano-pores was porous to a range of small gases, and hysteresis in the desorption branches was observed for H2 at 77 K and CH4 at 273 K. Glassy agAFI-[Zn(Im)1.68(bIm)0.32] showed lower gas uptake than the crystal for all tested gases. The partial collapse of open pores in crystalline MOFs during melt-quenching results in a decrease in porosity in MOF glasses21,34. No hysteresis was observed for N₂, indicating that N₂ molecules desorb rapidly due to their weaker interactions with agAFI-[Zn(Im)1.68(bIm)0.32], resulting in the absence of a hysteresis effect during N₂ desorption, similar case has been reported for agZIF-8-mIm0.15Im0.74bIm0.1135. Moreover, the first step of pore collapse also occurred at the point of thermal amorphization, as significant decrease in N2 uptake was observed for aTAFI-[Zn(Im)1.68(bIm)0.32] (Supplementary Fig. 18). The porosity and pore size distribution of agAFI-[Zn(Im)1.68(bIm)0.32] and agCAN-[Zn(Im)1.55(bIm)0.45] were measured by CO2 gas adsorption at 195 K (Fig. 4e and Supplementary Fig. 19), and the comparison was made with other MOF glasses in Table 134,35. As shown in Fig. 4f, the high-pressure CO2 gas adsorption at 273 K and 3000 kPa demonstrated the enhanced CO2 adsorption capacity of agAFI-[Zn(Im)1.68(bIm)0.32]. As shown in Supplementary Fig. 20, following the melting and quenching process, both pre-pressing and post-pressing agAFI-[Zn(Im)1.68(bIm)0.32] exhibited an almost identical CO2 adsorption capacity. Although mechanical pressure slightly damages the porous structure of AFI-[Zn(Im)1.68(bIm)0.32] crystals, it has almost no effects on the adsorption capacity of the melt-quenching glass.
Results were tested by (a) H2 at 77 K, (b) CO2 at 273 K, (c) N2 at 77 K, (d) CH4 at 273 K, (e) CO2 at 195 K, and (f) CO2 at 273 K up to 3000 kPa. The filled and hollow circles represent adsorption and desorption, respectively.
Furthermore, as shown in Supplementary Fig. S21, the agAFI-[Zn(Im)1.68(bIm)0.32] was heated above the glass transition temperature (Tg = 673 K) for 2 h, resulting in the formation of the agAFI-[Zn(Im)1.68(bIm)0.32]T. The slight reduction in CO₂ adsorption capacity is indicative of the sufficient collapse of the pores in the agAFI-[Zn(Im)1.68(bIm)0.32] following the initial adequate heat treatment.
Remarkably, agAFI-[Zn(Im)1.68(bIm)0.32] displayed significantly higher gas adsorption quantity for H2, CO2, and CH4 than other MOF glasses and CGCs such as agZIF-62 and (MIL-118)0.5(agZIF-62)0.536 (Supplementary Table 6). For example, agAFI-[Zn(Im)1.68(bIm)0.32] absorbs 1.56 mmol g−1 H2 at 77 K and a pressure of 100 kPa, and the adsorption quantity of agZIF-62 and ag[(ZIF-8)0.2(ZIF-62)0.8] is 0.42 and 1.24 mmol g−1, respectively. The low density of agAFI-[Zn(Im)1.68(bIm)0.32] when compared with agZIF-62 is consistent with the higher porosity. Moreover, the low density of agAFI-[Zn(Im)1.68(bIm)0.32] possibly indicates the presence of a multitude of blocked or inaccessible pores. During the transformation from crystal to glass, the collapse of the pore structure results in the formation of varying degrees of pore channels. Furthermore, it is unavoidable that some pores will become obstructed, preventing He from entering and resulting in a reduction in the He pycnometric density. The agAFI-[Zn(Im)1.68(bIm)0.32] demonstrated a similar H2 adsorption quantity as (MIL-118)0.5(agZIF-62)0.5 and (UL-MOF-1)0.5(agZIF-62)0.5, both which contain 50 wt% porous MOF crystals37. Although agAFI-[Zn(Im)1.68(bIm)0.32] and agZIF-62 showed high similarity in chemical composition, glass transition temperature, PDFs, and mechanical properties, these two MOF glasses showed significant differences in density and gas uptake, which is believed to be attributed to the difference in topology of the crystalline precursors. It was the avoidance of recrystallization (to ZIF-zni) that we believe is responsible for imparting a higher porosity to the resultant glasses.
Unlike the conventional inorganic glasses whose microporosity is generally produced by leaching, microporosity is an intrinsic property of MOF glasses21. The molecular linkers of MOFs, especially the bulkier linkers, play an important role in suppressing the formation of densely packed structures upon melting32,38. Moreover, the large solvent molecules used as SDAs in the synthesis can create highly porous topologies in crystalline MOFs without introducing extra ligands that possibly decrease the decomposition temperature or increase the melting point of the resultant MOF crystals. Thus, the porosity of MOF glasses can be significantly improved by increasing the porosity of crystalline precursors through using both SDAs and mixed ligands of Im and bIm in the synthesis process.
The increased material processability, combined with pore structure, makes the glassy MOFs prepared in this work attractive for applications such as separation membranes, adsorbents, and beyond. To illustrate this, the membranes were successfully prepared by easily processing the agAFI-[Zn(Im)1.68(bIm)0.32] into mechanically durable pallets with a diameter of 2 cm (Fig. 5a) through their molten phase. The obtained membrane is transparent and self-supported (Fig. 5b). The cross-section (Fig. 5c) and surface (Fig. 5e–g) of the membrane were examined by SEM, showing a smooth and dense membrane structure without visible voids or cracks. Overall, the membrane had a high degree of integrity and uniformity. The thickness of the membranes, ranged from ~100 µm, ~200 µm, and ~300 µm (Supplementary Fig. 22) and could be controlled by varying the mass of AFI-[Zn(Im)1.68(bIm)0.32] used in the pressing before melting. Although the membrane displayed a dense morphology in the SEM images, the microstructure nature of agAFI-[Zn(Im)1.68(bIm)0.32] could be demonstrated from FIB lamella under the HR-TEM dark field mode (Fig. 5d). The microstructure of agAFI-[Zn(Im)1.68(bIm)0.32] was observed to exhibit slight looseness than that of agZIF-62 (Supplementary Fig. 23), and the packing density was decreased when analysed using HR-TEM.
a Fabrication process. b Photograph. c Cross-section SEM image. d HR-TEM dark field image. e–g Surface SEM images at different magnification.
a, b Gas permeance and gas selectivity for pure gas and mixed gas. Comparison of (c) CO2/N2 and (d) CH4/N2 separation performance of the agAFI-[Zn(Im)1.68(bIm)0.32] membrane with other reported high-performance membranes. Error bars represent standard deviations.
For comparison, a self-supported agZIF-62 membrane (~200 µm) was fabricated and tested by using mixed gases. The agAFI-[Zn(Im)1.68(bIm)0.32] membrane (~200 µm) possessed much higher gas permeance and selectivity than the agZIF-62 membrane for both CO2/N2 and CH4/N2 gas pairs (Fig. 6a, b). The higher gas selectivity of agAFI-[Zn(Im)1.68(bIm)0.32] membrane than that of the agZIF-62 membrane (14.8 vs 8.3 for CO2/N2 and 5.2 vs 1.2 for CH4/N2) was reasonably attributed to the higher porosity of agAFI-[Zn(Im)1.68(bIm)0.32] that leads to higher gas uptake as well as higher CO2/N2 and CH4/N2 solubility selectivity (Supplementary Tables 8, 9). The difference in gas permeance between agAFI-[Zn(Im)1.68(bIm)0.32] and agZIF-62 membranes (3.7 × 104 vs 305 GPU for CO2/N2 and 1.3 × 104 vs 45 GPU for CH4/N2) seemed not to be simply attributable to the different porosities of the glasses, considering the lesser difference in CH4 uptake between agAFI-[Zn(Im)1.68(bIm)0.32] and agZIF-62 powders (0.45 vs 0.17 mmol g−1) (Supplementary Table 1). The possible reason is that, besides of the higher porosity, the agAFI-[Zn(Im)1.68(bIm)0.32] membrane has a larger pore size and better pore connectivity, which results in higher diffusivity coefficient of gas molecules.
The gas separation performance of agAFI-[Zn(Im)1.68(bIm)0.32] membrane was compared with the reported high-performance micro-porous membranes including crystalline MOFs, agZIF-62, 2D materials, porous organic cages, and mixed matrix membranes (MMMs) as shown in Fig. 6c, d and Supplementary Tables 10, 11. The agAFI-[Zn(Im)1.68(bIm)0.32] membrane has the highest CO2 permeance and the next-highest CH4 permeance and the gas selectivity is comparable with the reported membranes. The high gas permeance of the agAFI-[Zn(Im)1.68(bIm)0.32] membrane is favourable to reduce the required membrane area of the separation processes, which could largely decrease the operating cost of gas separation processes39,40,41. The agAFI-[Zn(Im)1.68(bIm)0.32] membrane is suitable for assembling into disc type membrane modules. Although the packing density of disc type membrane modules is typically lower than that of spiral and hollow fibre modules, they can effectively inhibit concentration polarization of permeation gas, which is conducive to high-permeance membrane assembly. Disc type membrane modules can be designed to effectively handle gases in confined spaces, such as small pollution sources or compact waste gas treatment devices.
Comparing with the reported three types of agZIF-62 membranes22,26, the agAFI-[Zn(Im)1.68(bIm)0.32] membrane shows significant advances in gas separation performance. Wang22 et al. prepared an agZIF-62 thin-film composite membrane by using porous anodic aluminium oxide (AAO) as the support, which is a practical strategy to reduce the thickness of the separation layer and increase the gas permeance. However, the agZIF-62/AAO composite membrane displays only a CO2 permeance of 37.6 GPU with a CO2/N2 selectivity of 34, due to the penetration of agZIF-62 into the pores of AAO. The support layer on agZIF-62/AAO surface can penetrate and increase the actual thickness of the composite membrane, thereby enhancing the CO₂/N₂ separation performance. Consequently, the CO₂/N₂ selectivity of the agZIF-62/AAO membrane as in the literature is higher than that of the self-supported agZIF-62 in this work. Knebel et al.42 produced agZIF-62 with varying degrees of pore collapse by modifying the process parameters of melt-quenching. This strategy prevents the pore collapse during melt-quenching, and results in a CO2 diffusivity of 3.53 × 10-8 cm2/s measured by an infrared microscope that cannot be compared directly to the results measured by gas permeation used in this work. In our previous work26, a ZIF-62 glass foam, i.e., agfZIF-62, was developed by melting ZIF-62 with polyethyleneamine as the pore-forming agent, and the porosity of the formed glass was largely improved through the decomposition of polyethyleneamine. The agfZIF-62 membrane displays a higher CH4 permeance of 4.85 × 104 GPU and a lower CH4/N2 selectivity of 4.08 than those of agAFI-[Zn(Im)1.68(bIm)0.32] in this work. The higher gas selectivity of agAFI-[Zn(Im)1.68(bIm)0.32] than that of agfZIF-62 demonstrates that the strategy of designing meltable crystalline MOFs with large-pore topology in this work has advantages in achieving precise manipulation of porosity in MOF glass at the Ångstrom scale.
Meltable crystalline MOFs AFI-[Zn(Im)1.68(bIm)0.32] and CAN-[Zn(Im)1.73(bIm)0.27] were synthesised by using bulky amide solvents as SDAs and mixed ligands of Im and bIm. AFI-[Zn(Im)1.68(bIm)0.32] and CAN-[Zn(Im)1.73(bIm)0.27] exhibit large pore topologies, and no recrystallization to ZIF-zni was observed before melting, which partially preserves the highly porous topology in the resultant melt-quenched glasses. The gas adsorption capacity of agAFI-[Zn(Im)1.68(bIm)0.32] is higher than reported MOF glasses, for example, H2 and CO2 adsorption capacities of 1.56 mmol g−1 and 1.12 mmol g−1, respectively. Self-supported agAFI-[Zn(Im)1.68(bIm)0.32] membranes were easily fabricated due to its high processability and they show record-breaking gas separation performance tested by mixed gases, for example, CO2 permeance of 3.7 × 104 GPU with CO2/N2 selectivity of 14.8, and CH4 permeance of 1.3 × 104 GPU with CH4/N2 selectivity of 5.2. The porosity and gas separation performance of MOF glasses have been significantly improved by designing highly porous topologies into the crystalline precursors, which may lay a foundation for the application of MOF glass in low-cost gas separation process.
AFI-[Zn(Im)1.68(bIm)0.32] and CAN-[Zn(Im)1.73(bIm)0.27] were synthesized according to the previous report with some modification29. Specifically, AFI-[Zn(Im)1.68(bIm)0.32]: Zinc acetate (286 mg, 1.56 mmol) was dissolved in N,N-dipropylformamide (DPF, 5 mL) in a 20 mL glass vial. Imidazole (309 mg, 4.55 mmol) and benzimidazole (76.6 mg, 0.65 mmol) were dissolved in DPF (10 mL) in a separate glass vial. The solutions were mixed and placed in an oven at 328 K for 7 days. After cooling to room temperature, the product was replaced with dry acetone three times a day for a total of three days. CAN-[Zn(Im)1.73(bIm)0.27] was synthesized by replacing DPF with dibutylformamide (DBF) as the solvent, and other conditions were as same as that for AFI-[Zn(Im)1.68(bIm)0.32]. According to the reported literature, ZIF-62 was synthesized by solvothermal method20. Zinc nitrate hexahydrate (Zn(NO3)2·6H2O, 1.65 g, 5.54 mmol), Im (8.91 g, 131 mmol), and bIm (1.55 g, 131 mmol) were dissolved in 75 mL N,N-dimethylformamide (DMF) in a glass jar. The mixture was statically reacted in an oven at 403 K for 48 h. The product was then cooled to room temperature, and washed several times with dichloromethane dried in an oven at 313 K.
Powder X-ray diffraction (XRD) analysis (2θ = 5° to 40°) was collected at room temperature with a Bruker-AXS D8 diffractometer with Cu Kα (λ = 1.540598 Å) radiation detector. The Pawley refinement was performed using TOPAS (Version 3) program. X-ray total scattering data were collected on the I15-1 beamline at the Diamond Light Source (UK) with an X-ray wavelength of 0.161669 Å (76.7 keV). The BL14B1 beamline at the Shanghai Synchrotron Radiation Facility (SSRF, China) was used by an X-ray wavelength of 0.6889 Å (18 keV). A series of PDFs pre-experiments were carried out using SSRF in the initial and validation stages of the investigation, with the main purpose of verifying the feasibility of the experimental design and confirming the structure of the samples, while the data shown in this manuscript were obtained from the I15-1 beamline at the Diamond Light Source. Samples were loaded into borosilicate glass capillaries of 1.17 mm (inner) diameter. Data from the samples, empty instrument and empty capillary were collected and processed using the GudrunX program. The surface morphologies of samples were observed by using a scanning electron microscope (SEM), FEI Nova Nano 450. The density of the glasses was measured by pycnometry on a Micromeritics AccuPyc 1340 instrument with He gas. Gas sorption isotherms were conducted on a Micromeritics ASAP 2020 instrument with around 200 mg samples. High-pressure CO₂ adsorption tests were conducted on ZIF glasses using a high-pressure gas adsorption analyser (BSD-PH, Bei Shi De) at 273 K, with a pressure range of 0–3000 kPa. Prior to the adsorption/desorption testing, all samples were degassed under vacuum at 353 K for 6 h. The saturation pressure of 191 kPa was used for specific surface area calculation from CO2 sorption isotherms recorded at 195 K34,35. The thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) analysis were conducted using TA Q600 instrument in an argon atmosphere at a heating rate of 10 K/min, 20 K/min, and 30 K/min. The solution 1H nuclear magnetic resonance (1H NMR) spectra were recorded on a Bruker Avance III 500 MHz spectrometer at 293 K. The digested samples were prepared by mixing a mixture of DCl (35%)/D2O (0.1 mL) and DMSO-d6 (0.5 mL) with MOFs (about 6 mg). The chemical shift refers to the DMSO-d6 solvent signal and the spectrum is processed using the MestreNova Suite. The Young’s modulus (E) and hardness (H) of MOF glasses were studied via depth-sensing indentation testing with a nanoindenter (G200, Agilent Inc.), equipped with a three-sided Berkovich diamond tip (Synton-MDP Inc.). The MOF glasses were first mounted in a thermosetting epoxy resin (Araldite CY212, Agar Scientific Ltd.) and polished to optical finish using a water-based diamond suspension. The values of E were estimated for a theoretical Poisson ratio, v, of 0.20. Prior to the first experiments, the area function of the Berkovich indenter tip as well as the instruments frame compliance were calibrated on a fused silica reference glass sample of known elastic properties (Corning Code 7980, Corning Inc.).
The glassy MOF membranes were fabricated by pressing and melt-quenching. Crystalline AFI-[Zn(Im)1.68(bIm)0.32] powders were pressed at 10 MPa pressure to form a pellet at first. The crystalline pellet was then put into a tube furnace under Ar atmosphere, heated to 705 K with a heating rate of 10 K/min, and kept at 705 K for 30 min. After cooling down to the room temperature, the agAFI-[Zn(Im)1.68(bIm)0.32] membrane was obtained. The agZIF-62 membranes were prepared by using a maximum temperature of 713 K, and other conditions were the same as that for the agAFI-[Zn(Im)1.68(bIm)0.32] membranes.
The surface and cross-section morphology of membranes were observed by using scanning electron microscope-dual beam focused ion beam (SEM-FIB, CrossBeam 550, Zeiss). The membrane microstructure was observed by high resolution transmission electron microscope (HR-TEM, JEM-F200, Jeol) with a Gatan CMOS CCD camera. The instrument was operated at 200 kV. The TEM samples were prepared by ultrathin-sectioning by the FIB equipped with a ZEISS CrossBeam 550 instrument. The samples were sectioned at 30 kV ion gun voltage, rough sectioning at 30 nA current, and polished at 1 nA.
The mixed gas separation performance of membranes was evaluated by constant pressure-variable volume method. The gas selectivity (ɑ) was the ratio between two components Pi and Pj. For the mixed gas testing, a gas chromatograph (Agilent 8890) was used to measure the concentration of mixed gases on the permeation side driven by the sweep gas.
The gas permeance of glass membranes were measured by using constant pressure-variable volume method with Eq. (1):
where Pi is the gas permeance of i-component (GPU, 1 GPU = 10−6 cm3 (STP) cm−2·s−1·cm·Hg−1), Qi is the molar flow rate of i-component (cm3 (STP)/s), S is the membrane area (cm2), and ΔPi is the pressure difference across the membrane (Pa).
The solubility coefficient (S) and diffusion coefficient (D) of membranes were measured at 50 kPa and 298 K via the constant volume-variable pressure method with pure gas (time-lag method) on the Labthink instrument, G2/110-A. The initial step involved the application of a vacuum treatment to both the feed side and the permeation side. Subsequently, the cavity on the permeation side was sealed, the feed gas side was filled with test gas at a specified pressure, and it was ensured that a pressure differential of 50 kPa was established on both sides of the glass membrane. Subsequently, the gas will permeate from the feed side to the permeation side under the pressure gradient. The permeance, solubility coefficient, and diffusion coefficient of the membrane can be calculated by processing the pressure data from the permeable side using the software.
The diffusion coefficient (D, cm2/s) was calculated by time-lag method (2):
where l was the membrane thickness and θ was the time lag (s) before the steady straight-line relationship was calculated by the downstream pressure versus time plot.
The solubility coefficient S (cm3(STP)·cm−3·cmHg−1) was given by Eq. (3):
The authors declare that all data supporting the findings of this study are available within the paper and Supplementary Information files. Source data are provided with this paper.
Davis, S. J. et al. Net-zero emissions energy systems. Science 360, eaas9793 (2018).
Dai, Y. et al. Manipulating local coordination of copper single atom catalyst enables efficient CO(2)-to-CH(4) conversion. Nat. Commun. 14, 3382 (2023).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Brethomé, F. M., Williams, N. J., Seipp, C. A., Kidder, M. K. & Custelcean, R. Direct air capture of CO2 via aqueous-phase absorption and crystalline-phase release using concentrated solar power. Nat. Energy 3, 553–559 (2018).
Pang, S. et al. A phenazine-based high-capacity and high-stability electrochemical CO2 capture cell with coupled electricity storage. Nat. Energy 8, 1126–1136 (2023).
Article ADS CAS MATH Google Scholar
Cai, G., Yan, P., Zhang, L., Zhou, H. C. & Jiang, H. L. Metal-Organic Framework-Based Hierarchically Porous Materials: Synthesis and Applications. Chem. Rev. 121, 12278–12326 (2021).
Article CAS PubMed MATH Google Scholar
Moghadam, P. Z., Chung, Y. G. & Snurr, R. Q. Progress toward the computational discovery of new metal–organic framework adsorbents for energy applications. Nat. Energy 9, 121–133 (2024).
Belmabkhout, Y. et al. Natural gas upgrading using a fluorinated MOF with tuned H2S and CO2 adsorption selectivity. Nat. Energy 3, 1059–1066 (2018).
Article ADS CAS MATH Google Scholar
Liang, Y. et al. A cage-on-MOF strategy to coordinatively functionalize mesoporous MOFs for manipulating selectivity in adsorption and catalysis. Nat. Commun. 14, 5223 (2023).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Ghalei, B. et al. Enhanced selectivity in mixed matrix membranes for CO2 capture through efficient dispersion of amine-functionalized MOF nanoparticles. Nat. Energy 2, 17086 (2017).
Qian, Q. et al. MOF-Based Membranes for Gas Separations. Chem. Rev. 120, 8161–8266 (2020).
Article CAS PubMed MATH Google Scholar
Jo, Y. M. et al. MOF-Based Chemiresistive Gas Sensors: Toward New Functionalities. Adv. Mater. 35, e2206842 (2023).
Article PubMed MATH Google Scholar
Zhou, S. et al. Electrochemical synthesis of continuous metal–organic framework membranes for separation of hydrocarbons. Nat. Energy 6, 882–891 (2021).
Article ADS CAS MATH Google Scholar
Ma, X. et al. Zeolitic imidazolate framework membranes made by ligand-induced permselectivation. Science 361, 1008–1011 (2018).
Article ADS CAS PubMed MATH Google Scholar
Faust, T., Farha, O. & Hernandez, B. MOFs move to market. Nat. Chem. 8, 990–991 (2016).
Article CAS PubMed MATH Google Scholar
Su, Z., Miao, Y.-R., Zhang, G., Miller, J. T. & Suslick, K. S. Bond breakage under pressure in a metal organic framework. Chem. Sci. 8, 8004–8011 (2017).
Article CAS PubMed PubMed Central Google Scholar
Bennett, T. D. et al. Looking into the future of hybrid glasses. Nat. Chem. 16, 1755–1766 (2024).
Gaillac, R. et al. Liquid metal-organic frameworks. Nat. Mater. 16, 1149–1154 (2017).
Article ADS CAS PubMed MATH Google Scholar
Li, S. et al. A new route to porous metal-organic framework crystal-glass composites. Chem. Sci. 11, 9910–9918 (2020).
Article CAS MATH Google Scholar
Hou, J. et al. Metal-organic framework crystal-glass composites. Nat. Commun. 10, 2580 (2019).
Article ADS PubMed PubMed Central MATH Google Scholar
Li, S. et al. Mechanical Properties and Processing Techniques of Bulk Metal-Organic Framework Glasses. J. Am. Chem. Soc. 141, 1027–1034 (2019).
Article CAS PubMed MATH Google Scholar
McHugh, L. N. & Bennett, T. D. Introducing porosity into metal–organic framework glasses. J Mater. Chem. A 10, 19552–19559 (2022).
Wang, Y. et al. A MOF Glass Membrane for Gas Separation. Angew. Chem. Int. Edit. 59, 4365–4369 (2020).
Article CAS MATH Google Scholar
Xia, H. et al. A long-lasting TIF-4 MOF glass membrane for selective CO2 separation. J. Membr. Sci. 655, 120611 (2022).
Tan, J.-C., Civalleri, B., Erba, A. & Albanese, E. Quantum mechanical predictions to elucidate the anisotropic elastic properties of zeolitic imidazolate frameworks: ZIF-4 vs. ZIF-Zni. CrystEngComm 17, 375–382 (2015).
Ao, D. et al. Metal-Organic Framework Crystal-Glass Composite Membranes with Preferential Permeation of Ethane. Angew. Chem. Int. Edit. 135, e202304535 (2023).
Yang, Z. et al. ZIF-62 glass foam self-supported membranes to address CH4/N2 separations. Nat. Mater. 22, 888–894 (2023).
Article ADS CAS PubMed MATH Google Scholar
Frentzel-Beyme, L., Kloss, M., Kolodzeiski, P., Pallach, R. & Henke, S. Meltable Mixed-Linker Zeolitic Imidazolate Frameworks and Their Microporous Glasses: From Melting Point Engineering to Selective Hydrocarbon Sorption. J Am. Chem. Soc. 141, 12362–12371 (2019).
Article CAS PubMed Google Scholar
Nozari, V. et al. Ionic liquid facilitated melting of the metal-organic framework ZIF-8. Nat. Commun. 12, 5703 (2021).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Shi, Q. et al. Zeolite CAN and AFI-Type Zeolitic Imidazolate Frameworks with Large 12-Membered Ring Pore Openings Synthesized Using Bulky Amides as Structure-Directing Agents. J. Am. Chem. Soc. 138, 16232–16235 (2016).
Article CAS PubMed Google Scholar
Gandara-Loe, J., Bueno-Perez, R., Missyul, A., Fairen-Jimenez, D. & Silvestre-Albero, J. Molecular Sieving Properties of Nanoporous Mixed-Linker ZIF-62: Associated Structural Changes upon Gas Adsorption Application. ACS Appl. Nano. Mater. 4, 3519–3528 (2021).
Tao, H., Bennett, T. D. & Yue, Y. Melt-Quenched Hybrid Glasses from Metal-Organic Frameworks. Adv. Mater. 29, 1601705 (2017).
Krokidas, P. et al. On the efficient separation of gas mixtures with the mixed-linker zeolitic-imidazolate framework-7-8. ACS Appl. Mater. Interfaces 10, 39631–39644 (2018).
Article CAS PubMed Google Scholar
Thompson, J. A. et al. Hybrid zeolitic imidazolate frameworks: controlling framework porosity and functionality by mixed-linker synthesis. Chem. Mater. 24, 1930–1936 (2012).
Article CAS MATH Google Scholar
Frentzel-Beyme, L. et al. Quantification of gas-accessible microporosity in metal-organic framework glasses. Nat. Commun. 13, 7750 (2022).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Xue, W. L. et al. Highly porous metal–organic framework liquids and glasses via a solvent-assisted linker exchange strategy of ZIF–8. Nat. Commun. 15, 4420 (2024).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Longley, L. et al. Flux melting of metal–organic frameworks. Chem. Sci. 10, 3592–3601 (2019).
Article CAS PubMed PubMed Central MATH Google Scholar
Ashling, C. W. et al. Guest size limitation in metal–organic framework crystal–glass composites. J Mater. Chem. A 9, 8386–8393 (2021).
Zhou, C. et al. Metal-organic framework glasses with permanent accessible porosity. Nat. Commun. 9, 1–9 (2018).
Article ADS MATH Google Scholar
Sholl, D. S. & Lively, R. P. Seven chemical separations to change the world. Nature 532, 435–437 (2016).
Article ADS PubMed MATH Google Scholar
Koros, W. J. & Zhang, C. Materials for next-generation molecularly selective synthetic membranes. Nat. Mater. 16, 289–297 (2017).
Article ADS CAS PubMed MATH Google Scholar
Zhou, S. et al. Asymmetric pore windows in MOF membranes for natural gas valorization. Nature 606, 706–712 (2022).
Article ADS CAS PubMed MATH Google Scholar
Smirnova, O. et al. Precise control over gas-transporting channels in zeolitic imidazolate framework glasses. Nat. Mater. 23, 262–270 (2024).
S.L. acknowledges Natural Science Foundation of China (21606212) and China Scholarship Council (CSC). TDB would like to thank the Royal Society for a University Research Fellowship and Research Grant (URF\R\211013 and RGS\R2\212221). Z.Q. acknowledges Natural Science Foundation of China (22122810). J.H. acknowledges the financial support from the Australian Research Council (FT210100589 and DP230101901), the University of Queensland, ARC Centre of Excellence for Green Electrochemical Transformation of Carbon Dioxide (CE230100017) funded by the Australian Government. S.Y. acknowledges the Natural Science Research Project of Education Department of Anhui Province (2022AH030135), Ph.D. research funding of Suzhou University (2021BSK041) and China Scholarship Council (CSC). The authors acknowledge the I15-1 beamline at the Diamond Light Source, UK, and the BL14B1 beamline at the Shanghai synchrotron radiation facility (SSRF), China.
These authors contributed equally: Shichun Li, Chao Ma.
Institute of Chemical Materials, China Academy of Engineering Physics, Mianyang, China
Shichun Li & Yu Liu
State Key Laboratory of Separation Membranes and Membrane Processes, Tiangong University, Tianjin, China
Chao Ma, Zhihua Qiao & Chongli Zhong
School of Chemical Engineering, The University of Queensland, St Lucia, QLD, Australia
Jingwei Hou & Shuwen Yu
ARC Centre of Excellence for Green Electrochemical Transformation of Carbon Dioxide, Brisbane, Australia
Jingwei Hou & Shuwen Yu
School of Chemistry and Chemical Engineering, Suzhou University, Suzhou, China
College of Chemical and Pharmaceutical Engineering, Hebei University of Science and Technology, Shijiazhuang, China
Aibing Chen & Juan Du
Diamond Light Source Ltd, Diamond House, Harwell Science & Innovation Campus, Didcot, Oxfordshire, UK
Philip A. Chater & Dean S. Keeble
ISIS Facility, Rutherford Appleton Laboratory, Harwell Campus, Didcot, Oxon, UK
Department of Materials Science and Metallurgy, University of Cambridge, Cambridge, UK
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
S.L., Z.Q. and T.D.B. designed the project. MOFs were synthesized by S.L. and J.H. Membrane preparation and characterization were carried out by C.M., A.C., J.D., Z.Q. and C.Z. PDF measurements were carried out by S.L., P.A.C., D.S.K. and D.A.K. Gas isotherm measurements were carried out by S.L. S.Y. and Y.L. S.L. and Z.Q. wrote the manuscript with T.D.B, with input from all authors. All authors performed certain experiments and discussed and revised the paper.
Correspondence to Shichun Li, Zhihua Qiao or Thomas D. Bennett.
The authors declare no competing interests.
Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Li, S., Ma, C., Hou, J. et al. Highly porous metal-organic framework glass design and application for gas separation membranes. Nat Commun 16, 1622 (2025). https://doi.org/10.1038/s41467-025-56295-x
DOI: https://doi.org/10.1038/s41467-025-56295-x
Anyone you share the following link with will be able to read this content:
Sorry, a shareable link is not currently available for this article.
Provided by the Springer Nature SharedIt content-sharing initiative
Nature Communications (Nat Commun) ISSN 2041-1723 (online)
heat and air conditioning unit Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.