RESEARCH ARTICLE www.advancedscience.com Ultralarge Free-Standing Imine-Based Covalent Organic Framework Membranes Fabricated via Compression Jesús Á. Martín-Illán, José Antonio Suárez, Julio Gómez-Herrero, Pablo Ares, Daniel Gallego-Fuente, Youdong Cheng, Dan Zhao, Daniel Maspoch, and Félix Zamora* membrane development.[2] While these Demand continues for processing methods to shape covalent organic materials do not show intrinsic porosity, frameworks (COFs) into macroscopic objects that are needed for their the tortuous transport pathways formed practical applications. Herein, a simple compression method to prepare between these nanosheets allow the size- large-scale, free-standing homogeneous and porous imine-based specific permeation of molecules.[3] Alter- natively, other porous materials such as COF-membranes with dimensions in the centimeter range and excellent classical inorganic-based materials (e.g., ze- mechanical properties is reported. This method entails the compression of olites, silicas, or carbon)[4] and more novel imine-based COF-aerogels, which undergo a morphological change from an organic-based materials (e.g., conjugated elastic to plastic material. The COF-membranes fabricated upon compression microporous polymers (CMPs),[5] metal– show good performances for the separation of gas mixtures of industrial organic frameworks (MOFs),[6] and poly- mers of intrinsic microporosity (PIMs)[7] ) interest, N2 /CO2 and CH4 /CO2 . It is believed that the new procedure paves have been incorporated in the fabrication the way to a broader range of COF-membranes. of membranes because of their advantages (i.e., simultaneous high permeability and high selectivity) over traditional organic 1. Introduction polymeric membranes.[8] Covalent organic frameworks (COFs) comprise an emerging The fabrication of efficient membranes is crucial for the sep- class of porous materials that integrates organic subunits into aration processes found in myriad applications. In this con- periodic 2D or 3D crystalline structures held together by strong text, membrane technology is evolving fast to offer straightfor- covalent bonds.[9] COFs are thermally and chemically stable in ward and more ecofriendly solutions with lower energy con- harsh environments, such as extreme humidity, strong acids, and sumption than conventional separation processes.[1] In the last organic solvents.[10] Moreover, in contrast to classical 2D materi- years, the development of new materials has played a key role als in which their permeable pathways rely only on interlamellar in advancing membrane separation technology. Among the dif- transport and/or deliberate introduction of in-plane defects, 2D- ferent materials, graphene, graphene oxide, and other alterna- COFs also display ordered in-plane pores. In addition, COFs al- tive 2D materials have certainly played a significant role in low accurate and predictable control over composition, topology, J. Á. Martín-Illán, J. A. Suárez, F. Zamora J. Gómez-Herrero, P. Ares, D. Gallego-Fuente Departamento de Química Inorgánica Departamento de Física de la Materia Condensada Universidad Autónoma de Madrid Universidad Autónoma de Madrid Madrid 28049, Spain Madrid 28049, Spain E-mail: felix.zamora@uam.es J. Gómez-Herrero, P. Ares J. A. Suárez, D. Maspoch Condensed Matter Physics Center (IFIMAC) Catalan Institute of Nanoscience and Nanotechnology (ICN2) Universidad Autónoma de Madrid CSIC and BIST Madrid 28049, Spain Campus UAB Bellaterra Y. Cheng, D. Zhao Barcelona 08193, Spain Department of Chemical and Biomolecular Engineering National University of Singapore 4 Engineering Drive 4, Singapore 117585, Singapore D. Maspoch ICREA Pg. Lluís Companys 23, Barcelona 08010, Spain The ORCID identification number(s) for the author(s) of this article F. Zamora can be found under https://doi.org/10.1002/advs.202104643 Instituto Madrileño de Estudios Avanzados en Nanociencia (IMDEA-Nanociencia) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH. Cantoblanco, Madrid 28049, Spain This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in F. Zamora any medium, provided the original work is properly cited. Institute for Advanced Research in Chemical Sciences (IAdChem) Universidad Autónoma de Madrid DOI: 10.1002/advs.202104643 Madrid 28049, Spain Adv. Sci. 2022, 9, 2104643 2104643 (1 of 9) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com and porosity. In this context, imine-based COFs are an excellent The crystalline structure of COF-membranes was confirmed choice for many applications[11] since they can also be processed by powder X-ray diffraction (PXRD) in reflection (parallel) and into different macroscopic morphologies, including foams, aero- transmission (perpendicular) mode and deduced by theoretical gels, and membranes.[12] simulation.[22] Indeed, the different membranes exhibited good Before this work, several methods have been developed to crystallinity, comparable with the PXRD patterns of their COF- fabricate COFs and COF-based membranes, which have al- aerogels counterparts and matching the simulated patterns of ready demonstrated remarkable applications in catalysis and the AA-eclipsed stacking models. Thus, in parallel mode, TAPB- separation processes.[11c] Some of the most common fabrica- BTCA-MCOF showed an intense peak at 5.7° corresponding to tion methods include COF blending into polymer matrices to the (100) plane, along with the peaks at 9.9°, 11.5°, and 25.3° form mixed-matrix membranes (MMMs)[13] and the in situ attributed to the (110), (200), and (001) reflections (Figure 1A). growth method.[14] Recently, new ways have been reported to pro- As the pore size increases, the peak corresponding to the (100) duce continuous COF-membranes, either on surfaces or free- plane gets shifted to lower 2𝜃 values of 4.8° for PPDA-BTCA- standing.[15] The first approach used the conventional solvother- MCOF (Figure 1C) and 2.9° for TAPB-PDA-MCOF (Figure 1E). mal synthesis of COFs on the surface of solid substrates.[16] Thus, the corresponding (001) reflection peak displayed a high- For the fabrication of free-standing COF-membranes, a widely intensity increase. In contrast, the (100) reflection peak showed explored method is interfacial polymerization. This fabrica- a significant attenuation. On the other hand, in perpendicular tion method occurs at the interface of liquid/liquid[17] or mode (turning the pellet 90°), the corresponding (hk0) peaks liquid/gas.[18] However, it is limited to laboratory scale because displayed a slight increase of intensity (Figures 1B,D,F). We at- it is a tedious procedure that requires slow monomer diffusion tributed this behavior to a certain degree of preferred orientation between two phases, leading to crystalline polymerization. Addi- of the COFs as showed by a higher intensity of the (100) reflec- tionally, the so-formed membrane must be transferred before its tion peak in the perpendicular mode due to the pressure applied use, making its real application hampered for large-scale prepa- during the formation of the COF-membranes. Nevertheless, this ration. Alternatively, an approach based on solution casting and preferred crystallographic orientation is considerably lower than baking was reported to transform the molecular precursors into those observed for other imine-based COFs showing stacking.[23] COF-membranes. This process allows the fabrication of COF- Here, it is also important to mention that we tested the influ- membranes with dimensions in the hundreds of microns. How- ence of pressure on thickness and crystallinity. We observed that ever, it is so far limited to beta-ketoenamine COFs.[19] More re- higher pressures (>360 MPa) induce a decrease of crystallinity cently, Jiang et al. developed a solution-processing method,[20] in in the COF-membranes (Figure S1, Supporting Information). which large-area COF-membranes are fabricated by exposing an Moreover, we found that the initially reported density values of amorphous polymeric membrane to a monomer exchange pro- the COF-aerogels (≈0.02 g cm−3 ) changed to values ≈1 g cm−3 , cess under solvothermal conditions. which are in the range of those expected for monocrystalline COF Herein, we show that large-scale free-standing COF- structures. membranes can be prepared via a simple compression method. Having demonstrated the possibility to fabricate COF- Our method to produce COF-membranes started from ob- membranes using this simple compression method, we then servation during a mechanical study of our recently reported scaled it up to produce membranes with diameters ranging imine-based COF-aerogels.[21] These COF-aerogels showed from 1 to 5 cm (Figure 2 and Figure S2, Supporting Informa- to behave elastically below 25–35% strain. However, further tion). Field-emission scanning electron microscopy (SEM) im- compression to a maximum deformation of 90% gave rise to a ages showed the formation of COF-membranes with thicknesses plastic behavior without exhibiting failure, with partial recovery, in the range of 50–60 μm (Figure 2), which are thinner than those even at maximum deformation. We, therefore, postulated that previously obtained from their COF-powder counterparts.[24] control of this compression process in COF-aerogels could Here, we hypothesize that the lower density of COF-aerogels become a straightforward method to produce free-standing, compared to their COF-powder counterparts is crucial for form- homogeneous COF-membranes with sizes in the centimeter ing thinner COF-membranes. Moreover, the use of COF-aerogels range and excellent mechanical properties. We anticipate that revealed the formation of continuous, homogeneous, and com- these membranes show good performance for gas separation of pact membranes (Figures 2B,F,J). The surface roughness of mixtures of industrial interest, e.g., N2 /CO2 and CH4 /CO2 . these membranes was measured by atomic force microscope (AFM) (Figures 2D,H,L), using WSxM software for acquisition 2. Results and Discussion and analysis.[25] We determined an average surface roughness of 31 ± 8, 25 ± 4, and 70 ± 7 nm for TAPB-BTCA-MCOF, PPDA- The fabrication of COF-membranes started with the prepara- BTCA-MCOF, and TAPB-PDA-MCOF, respectively. These im- tion of three COF-aerogels, TAPB-BTCA-AGCOF, PPDA-BTCA- ages agreed with the expected material densification that occurs AGCOF, and TAPB-PDA-AGCOF, following our previously re- upon pressure, allowing the production of free-standing COF- ported procedure.[21] Then, these COF-aerogels were gently bro- membranes. ken into smaller pieces (size: ≈1–3 mm) in the presence of AcOH We also characterized the COF-membranes by Fourier trans- (5 μL of AcOH for 10 mg of COF). Finally, a fixed amount of form infrared (FT-IR) spectroscopy, 13 C CP-MAS solid-state the resulting pieces (0.0037 g cm−2 ) was pressed under 120 MPa NMR, and thermogravimetric analysis (TGA) under N2 atmo- for 5 min to produce free-standing membranes of the corre- sphere. The spectroscopic features of the COF-membranes were sponding COFs; hereafter named as TAPB-BTCA-MCOF, PPDA- almost identical to those of their corresponding COF-aerogels. BTCA-MCOF, and TAPB-PDA-MCOF. The main difference was ascribed to the presence of AcOH Adv. Sci. 2022, 9, 2104643 2104643 (2 of 9) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 1. Top: Schematic representation of the synthesis (sol–gel processes) of a COF-aerogel and its processing to a free-standing membrane. This method comprises i) mixture of the molecular building blocks in AcOH; ii) formation of the COF gels; iii) solvent exchange and supercritical CO2 acti- vation; iv) breaking process; and v) pressing of the COF-aerogel pieces. Photograph corresponds to a TAPB-BTCA-MCOF-membrane. Scale bar: 0.3 cm. Bottom left: Representation of the structures of the 2D imine-based COFs used in this work. Bottom right: A,C,E) Reflection and B,D,F) transmission PXRD patterns of the corresponding membranes (red), aerogels (green), and simulated patterns (blue) for A,B) TAPB-BTCA-COF, C,D) PPDA-BTCA-COF, and E,F) TAPB-PDA-AGCOF. Adv. Sci. 2022, 9, 2104643 2104643 (3 of 9) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 2. A,E,I) Photographs, B,C,F,G,J,K) SEM micrographs, and D,H,L) atomic force microscope topographies of A–D) TAPB-BTCA-MCOF, E–H) PPDA-BTCA-MCOF, and I–L) TAPB-PDA-MCOF. guest molecules trapped in the porous structure of the mem- Next, the chemical stability of the COF-membranes was evalu- branes. The appearance of the typical stretching band at ≈1620 ated upon their immersion in acidic, neutral, and basic (pH = 1– cm−1 confirmed no change of the imine bond (C═N) (Figures 14) aqueous solutions and common organic solvents. Analysis of S3–S5, Supporting Information). Furthermore, the characteris- the treated membranes showed no significant changes by PXRD, tic vibrations of the amino and carbonyl groups of the initial confirming their high chemical stability (Figure S12, Supporting monomers were detected, suggesting the presence of unreacted Information). groups likely placed at the defective edges of nanolayers. As we We also studied their mechanical properties by performing in- have previously reported,[24] the unreacted amine and/or alde- dentation experiments using atomic force microscopy (AFM). We hyde groups can connect COF-polymine nanolayers mediated by acquired a force versus distance curve in each pixel of the im- AcOH molecules by H-bonding formation with the amine and age and determined Young’s moduli of the films by fitting the aldehyde groups. The AcOH stretching bands were also observed corresponding indentation curves to the Hertz model.[26] We ob- at 1698 cm−1 (Figures S3–S5, Supporting Information). Addi- tained Young’s modulus of 0.8 ± 0.3, 0.5 ± 0.2, and 0.4 ± 0.2 GPa tionally, 13 C CP-MAS solid-state NMR spectra confirmed the for- for TAPB-BTCA-MCOF, PPDA-BTCA-MCOF, and TAPB-PDA- mation of imine bonds (≈156 ppm) (Figures S6–S8, Supporting MCOF, respectively (Figures S13 and S14, Supporting Informa- Information). Finally, TGA under N2 atmosphere showed that tion). A similar tendency was already observed in the mechanical the COF-membranes are thermally stable up to 500 °C (Figures properties of imine-based COF-aerogels, in which the enhance- S9–S11, Supporting Information). From this analysis, one guest ment of the pore size correlated to the fragility of the framework molecule of AcOH per formula unit was estimated, as also cor- due to the decrease of 𝜋–𝜋 stacking force.[27] roborated by elemental analysis (Table S1, Supporting Informa- N2 gas adsorption experiments at 77 K were performed to tion). evaluate the permanent porosity of all COF-membranes. The Adv. Sci. 2022, 9, 2104643 2104643 (4 of 9) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com isotherms of the membranes were very similar to those obtained from their aerogel counterparts (Figures S17–S19, Supporting Information).[21] From the aforementioned isotherms, we cal- culated Brunauer−Emmett−Teller surface areas (SABET ) of 247 m2 g−1 for TAPB-BTCA-MCOF, 186 m2 g−1 for PPDA-BTCA- MCOF, and 170 m2 g−1 for TAPB-PDA-MCOF. Despite being still porous, these values confirmed a loss of porosity when COF- aerogels were compressed to produce the corresponding COF- membranes. This behavior is somehow expected after the mem- brane formation, where a “gate-closing” effect can occur, as re- ported by Zhong et al.[28] However, the pore size distribution ob- tained using the DFT method (Figures S20–S22, Supporting In- formation) revealed that COF-membranes had crystallographic pores of the same size as their COF-aerogels counterparts. To further confirm the porous character of our COF- membranes, we also measured the adsorption capacity of CH4 and CO2 at several temperatures for both COF-aerogels and COF- membranes (Figure 3). Notably, COF-aerogels showed high ad- sorption of both CO2 and CH4 : CO2 adsorptions (e.g., 22.6 mmol g−1 at 200 K for TAPB-PDA-AGCOF) are in the range of other COFs with great CO2 adsorption capacity (Figure S29, Support- ing Information); and CH4 adsorptions (e.g., 3.8 mmol g−1 for TAPB-BTCA-AGCOF) are among the highest described for 2D- COFs at low-pressures (1 bar) (Figure S30, Supporting Infor- mation). Compared to their corresponding aerogels, the max- imum adsorption capacities of the membranes for CO2 and CH4 (200 K, 100 kPa) decreased in the range of 40–50% for TAPB-BTCA-MCOF, 40–44% for PPDA-BTCA-MCOF, and 63– 76% for TAPB-PDA-MCOF. Thus, the maximum CO2 uptake for each membrane was 8.1 mmol g−1 at 200 K (1.2 mmol g−1 at 298 K), 6.6 mmol g−1 at 200 K (1.0 mmol g−1 at 298 K), and 5.4 mmol g−1 at 200 K (0.6 mmol g−1 at 298 K) for TAPB- BTCA-MCOF, PPDA-BTCA-MCOF, and TAPB-PDA-MCOF, re- spectively. For CH4 , the maximum uptakes were 1.8 mmol g−1 at 200 K (0.3 mmol g−1 at 298 K) for TAPB-BTCA-MCOF; 1.7 mmol g−1 at 200 K (0.3 mmol g−1 at 298 K) for PPDA-BTCA-MCOF; and 1.0 mmol g−1 at 200 K (0.1 mmol g−1 at 298 K) for TAPB- PDA-MCOF. Together with PXRD data, these results confirm that, although the compression of COF-aerogels reduces their gas adsorption capacity, the resulting COF-membranes are still porous. Once demonstrated that imine-based COF-membranes syn- thesized via compression are porous and show good mechani- cal properties, we evaluated their gas permeation properties by single and mixed gas measurements at 298 K under a trans- membrane pressure of 1 bar (Figure 4). A Knudsen diffusion phenomenon was observed in TAPB-PDA-MCOF during the sin- Figure 3. A,B) CO2 and C,D) CH4 sorption isotherms of aerogels (black) and membranes (red) of TAPB-BTCA-COF (sphere-symbol), PPDA-BTCA- gle gas tests (Figure 4A). This behavior agrees with the fact that COF (square-symbol), and TAPB-PDA-COF (triangle-symbol). Uptakes TAPB-PDA-MCOF is a mesoporous membrane with a pore size were measured at A–C) 200 K and B–D) 298 K at 1 bar. of 3.2 nm, which is much larger than the diameters of gases. For instance, this membrane exhibited an ideal H2 /CO2 selec- tivity of 3.8, close to the H2 /CO2 selectivity of 4.7 calculated water that can enhance the CO2 /CH4 separation factor, as previ- from the Knudsen diffusion theory. In the case of TAPB-BTCA- ously reported for related COFs.[29] MCOF and PPDA-BTCA-MCOF, these factors were higher than These observations agree with the presence of imine groups the corresponding ideal selectivity for separating CO2 /CH4 and into the COF cavities since it is known that the presence of some CO2 /N2 . For example, while the ideal CO2 /CH4 selectivity is 8.2, heteroatoms may enhance adsorption capacities of COFs.[30] A the CO2 /CH4 separation factor for TAPB-BTCA-MCOF reached separation mechanism based on an adsorption-diffusion process 16.8. These results could be rationalized by the coadsorption pro- is expected for imine-based COFs, in which imine groups located cess in the mixed gas separation process or by the adsorption of at the COFs pores favor the CO2 adsorption due to weak interac- Adv. Sci. 2022, 9, 2104643 2104643 (5 of 9) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 4. A) Gas permeation properties of TAPB-BTCA-MCOF, PPDA-BTCA-MCOF, and TAPB-PDA-MCOF obtained by single gas measurements at 298 K under a transmembrane pressure of 1 bar. B) Ideal gas pair selectivity (solid) and mixed gas pair selectivity (or separation factor, empty) of TAPB- BTCA-MCOF, PPDA-BTCA-MCOF, and TAPB-PDA-MCOF. C) CO2 /CH4 and (D) CO2 /N2 upper bound plots for all the three COF-membranes. tions between the nitrogen of imine bond (Lewis bases) and CO2 compared them with other membranes on the Robeson upper (Lewis acid). In contrast, these interactions are missing with the bound plots (Figure 4C). Therefore, these new COF-membranes CH4 molecules. Thus, when CO2 is adsorbed into the COF pore show higher CO2 permeability than commercial membranes induces a coadsorption of additional CO2 molecules from the gas (such as CA, PSf, PC, and Matrimid) and separation performance mixture giving rise to a pore size reduction that enables the sepa- close to the 2008 upper bound limits and better than reported ration of CO2 versus CH4 . Therefore, the sorption and diffusion for some MOFs (ZIF-8 and Bio-MOF-1) and Zeolites (MFI and of CH4 molecules into the COF-membranes are retarded.[29] SAPO-34) (Tables S8 and S9, Supporting Information). With this Additionally, it has been already reported that the presence of data, we tested the working performance of the best membrane: moisture in the COF pores can enhance the CO2 /CH4 separa- TAPB-BTCA-MCOF. To this end, we studied the pressure influ- tion factor due to water adsorption,[29] in which the presence of ence (transmembrane pressure from 1 to 6 bar) on the separa- water in the membranes reduces gas permeabilities due to lower tion performance of TAPB-BTCA-MCOF in mixed gas tests (Fig- gas diffusivities and solubilities (i.e., higher CO2 sorption in wa- ure S31A,B, Supporting Information). With increasing the trans- ter overwhelms the reduced CO2 diffusivity due to the strong membrane pressure, the gas permeability rapidly increased, and affinity between CO2 molecules and water molecules in terms the gas pair selectivity quickly decreased, which indicated that the of permeability.[31] defective sites in the membrane at higher pressures increased. It is worth noting that permeability is the product of solu- These defects may arise from intercrystalline boundaries present bility and diffusivity. Therefore, the CH4 permeability decreases in TAPB-BTCA-MCOF. from 55 Barrer in the single gas test to 21 Barrer in the mixed Next, we evaluated the separation performance of TAPB- gas test. We then plotted the separation performance of TAPB- BTCA-MCOF under different temperatures for two different BTCA-MCOF, PPDA-BTCA-MCOF, and TAPB-PDA-MCOF and mixed gas tests (Figure S32C,D, Supporting Information). We Adv. Sci. 2022, 9, 2104643 2104643 (6 of 9) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com found that the gas permeability increased by increasing the tem- (80 μL) acetic acid (AcOH). Then, a pressure of 30 MPa was applied, using perature due to the fast diffusion rates of gases at higher tem- a uniaxial hydraulic press, for 60 min. peratures. From data collected in Figure S32C,D (Supporting In- Chemical Stability Test: Chemical stability tests were carried out by im- mersing the COF-membranes in (1 mL) aqueous solutions of hydrochloric formation), we calculated activation energies for TAPB-BTCA- acid (HCl) (12 m), sodium hydroxide (NaOH) (14 m), and common sol- MCOF of 9.39 (CO2 ) and 13.88 (CH4 ) kJ mol−1 for the separa- vents (toluene, dimethylformamide, and hexane) under stirring for 72 h. tion of CO2 /CH4 , and 9.31 (CO2 ) and 14.13 (N2 ) kJ mol−1 for the Then, the samples were washed with water (only for the incubations with separation of CO2 /N2 confirming a stronger interaction of TAPB- HCl and NaOH), tetrahydrofuran (THF), and ethanol (EtOH) and dried BTCA-MCOF with CO2 than with CH4 and N2 . under vacuum at 333 K for 24 h. In the case of the acidic and basic treat- Finally, we evaluated the long-term stability of TAPB-BTCA- ments, 93% and 98% of the solid material were recovered. 100% of the ma- terial was recovered after chemical stability tests with common solvents. MCOF for the separation of CO2 /CH4 mixtures (Figure S33, Sup- In all of these experiments, there was no significant loss of crystallinity. porting Information). We introduced moisture (relative humid- Characterizations: PXRD patterns were collected with a Bruker D8 Ad- ity: 85%) during the test, and the stabilized membrane showed a vance X-ray diffractometer (Cu K𝛼 radiation; 𝜆 = 1.5418 Å) equipped with rather steady separation performance. The reduction of gas per- a Lynxeye detector. Samples were mounted on a flat sample plate. Patterns meability is due to the partial block of pores by water under hu- were collected in the 3.5° < 2𝜃 < 35° range with a step size of 0.025° and ex- mid conditions.[32] posure time of 1.3 s step−1 . Elemental analyses were obtained using LECO CHNS-932 elemental analyzer. FT-IR spectra were recorded in a Perkin Elmer Spectrum 100 with a PIKE Technologies MIRacle Single Reflec- tion Horizontal ATR (attenuated total reflection) accessory with a spectral 3. Conclusion range of 4000–650 cm−1 , signals are given in wavenumbers (cm−1 ). N2 ad- sorption isotherms were measured using a Micromeritics ASAP2020 vol- In summary, we report a simple compression method for the fab- umetric instrument under static adsorption conditions. Before measure- rication of centimeter-scale imine-based COF-membranes from ment, samples were heated at 323 K overnight and outgassed to 10−6 Torr. COF-aerogel. Applying pressure on COF-aerogels allows their Brunauer–Emmet–Teller (BET) and Langmuir analyses were carried out to determine the total specific surface areas from the N2 isotherms at 77 transformation into free-standing COF-membranes that preserve K. The pore size distribution (PSD) was determined from N2 adsorption their original crystallinity and are porous. Moreover, the fabri- isotherms at 77 K using nonlocal DFT for a model with cylindrical pores cated free-standing COF-membranes have shown an enhance- present in software MicroActive Version 4.06 of Micromeritics. Solid-State ment in their mechanical properties, i.e., Young modulus and 13 C NMR spectra were carried out on a Bruker AV 400 WB Spectrometer. flexibility, enabling them to be used for gas separation pro- Carbon chemical shifts are expressed in parts per million (𝛿 scale). SEM cesses under pressure conditions. The separation performances studies were performed on a JSM-7600 equipped with an OXFORD X-Max XEDS, operating at an accelerating voltage of 10 kV. Samples were previ- of TAPB-BTCA-MCOF for the gas mixture CO2 /CH4 or CO2 /N2 ously coated with gold in a sputter Quorum Q150T-S. The samples were are close to the upper bound limits. Therefore, we believe that this prepared by gluing, in perpendicular, a piece of COF-membrane in carbon novel processing method, once optimized, opens new avenues tape. TGA were run on a Thermobalance TGA Q-500 thermal gravimetric to produce large-scale COF-membranes with controlled size and analyzer with samples held in an aluminum pan under a nitrogen atmo- thickness for different applications, going from nanofiltration to sphere. The samples were heated at 10 K min−1 within a temperature range proton membranes. of 25–1000 °C. Mechanical Properties and Topography Characterization by AFM Imaging: AFM measurements were carried out using a Cervantes Full mode AFM from Nanotec Electronica SL. WSxM software (www.wsxmsolutions.com) 4. Experimental Section was employed both for data acquisition and image processing.[25] Topog- raphy images in amplitude modulation mode using a hemispherical cone- Materials: Most chemicals and solvents were obtained from Aldrich shaped tip with a radius of 120 nm from Team Nanotec, with a resonance Chemical Co. and used without further purification. Ethanol (EtOH) and frequency of 348 kHz and a spring constant of 46 N m−1 , calibrated using tetrahydrofuran (THF) were dried using standard methods and stored Sader’s method were acquired.[34] The same tip was used to carry out the under an inert gas atmosphere.[33] TAPB was synthesized as previously nanoindentation experiments. This consisted in acquiring a force versus reported.[22a] BTCA, PDA, and PPDA were commercially available and were distance curve in each pixel of the image to determine Young’s modulus used as purchased with no further treatment. of the membranes by fitting the corresponding indentation curves to the Synthesis of COF-Aerogels: The synthesis and characterization of COF- Hertz model.[26] Further measurements were repeated using a different aerogels were made according to the published procedures described pre- tip with a nominal radius of 250 nm, obtaining similar results, proving the viously by us.[21] robustness of the measurements. After acquiring a force versus distance Membrane Fabrication: Membranes of TAPB-BTCA-AGCOF, PPDA- curve in each pixel, the indentation curve is obtained by using the spring BTCA-AGCOF, and TAPB-PDA-AGCOF were prepared by breaking (10 mg) constant of the cantilever to subtract the effect of the cantilever deflection of the COF-aerogel material into small pieces (size: ≈1–3 mm) in the pres- from the aforementioned curve, leaving just the indentation length of the ence of (5 μL) acetic acid (AcOH). Then, a pressure of 120 MPa was ap- tip into the sample. With these indentation curves it is possible to obtain plied, using a uniaxial hydraulic press, for 5 min. Under these conditions, the Young’s Modulus by fitting the model of Hertzian Contact for a spheri- COF-membranes showed around 13 mm of diameter and a thickness of cal indenter: F = 4/3 × E × R1/2 × d3/2 , where F is the normal force, R is the 50–60 μm (Figure S1A, Supporting Information). For the 20 mm in diame- tip radius, d is the indentation length, and E is Young’s Modulus obtained ter COF-membranes (Figure S1B, Supporting Information), they were pre- from the fit. pared by breaking (25 mg) of the COF-aerogel material into small pieces Membrane Separation Performance: It was determined by single gases (size: ≈1–3 mm) in the presence of (15 μL) acetic acid (AcOH). Then, or gas mixtures (50:50 vol%) in a Wicke–Kallenbach gas permeation setup. a pressure of 120 MPa was applied, using a uniaxial hydraulic press, for The volumetric flow rates of the feed gas and sweep gas (argon) were 20 min. maintained at 50 and 50 mL min−1 , respectively. The gas compositions at In the case of 50 mm in diameter COF-membranes (Figure S1C, Sup- the permeate side were analyzed using a gas chromatograph (Shimadzu porting Information), they were prepared by breaking (150 mg) of the COF- GC-2014). All membranes were prepared at least three times to verify their aerogel material into small pieces (size: ≈1–3 mm) in the presence of reproducibility. The gas permeability (Pi , barrer, 1 barrer = 10−10 cm3 (STP) Adv. Sci. 2022, 9, 2104643 2104643 (7 of 9) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com cm−2 s−1 cm Hg−1 ) is defined as Pi = (Qi l)/(AΔPi ), where Qi is the vol- [7] X. Zou, G. Zhu, Adv. Mater. 2018, 30, 1700750. ume flow rate of gases (cm3 s−1 ), l is the membrane thickness (μm), A [8] Z. Kang, H. Guo, L. Fan, G. Yang, Y. Feng, D. Sun, S. Mintova, Chem. is the effective area of the membrane (cm2 ), and ΔPi is the partial pres- Soc. Rev. 2021, 50, 1913. sure difference across the membrane (cm Hg). The mixed gas selectivity [9] A. P. Cote, A. I. Benin, N. W. Ockwig, M. O’Keeffe, A. J. Matzger, O. (𝛼 ij ) is defined as: 𝛼 ij = (yi /yj )/(xi /xj ), where xi /xj and yi /yj represent the M. Yaghi, Science 2005, 310, 1166. molar fraction of i/j in the feed and permeate side, respectively. When the [10] H. Xu, J. Gao, D. Jiang, Nat. Chem. 2015, 7, 905. membranes were tested under moisture, the feed gas was humidified by [11] a) V. Singh, H. R. Byon, Mater. Adv. 2021, 2, 3188; b) F. Haase, B. V. passing through a water bottle at room temperature (relative humidity: Lotsch, Chem. Soc. Rev. 2020, 49, 8469; c) Z. Wang, S. Zhang, Y. Chen, ≈85%). Z. Zhang, S. Ma, Chem. Soc. Rev. 2020, 49, 708; d) S. Cao, B. Li, R. Zhu, H. Pang, Chem. Eng. J. 2019, 355, 602. [12] a) D. Rodriguez-San-Miguel, F. Zamora, Chem. Soc. Rev. 2019, 48, Supporting Information 4375; b) G. He, R. Zhang, Z. Jiang, Acc. Mater. Res. 2021, 2, 630; c) D. Zhu, Y. Zhu, Q. Yan, M. Barnes, F. Liu, P. Yu, C.-P. Tseng, N. Supporting Information is available from the Wiley Online Library or from Tjahjono, P.-C. Huang, M. M. Rahman, E. Egap, P. M. Ajayan, R. Ver- the author. duzco, Chem. Mater. 2021, 33, 4216. [13] a) H. Yang, H. Wu, Z. Yao, B. Shi, Z. Xu, X. Cheng, F. Pan, G. Liu, Z. Jiang, X. Cao, J. Mater. Chem. A 2018, 6, 583; b) Y. Peng, G. Xu, Z. Acknowledgements Hu, Y. Cheng, C. Chi, D. Yuan, H. Cheng, D. Zhao, ACS Appl. Mater. Interfaces 2016, 8, 18505. This work was supported by the Spanish MINECO (PID2019-106268GB- [14] a) S. Kandambeth, B. P. Biswal, H. D. Chaudhari, K. C. Rout, H. S. C31, PID2019-106268GB-C32, PCI2018-093081, and RTI2018-095622-B- Kunjattu, S. Mitra, S. Karak, A. Das, R. Mukherjee, U. K. Kharul, R. I00), the Catalan AGAUR (project 2017 SGR 238), and through the “María Banerjee, Adv. Mater. 2017, 29, 1603945; b) H. Fan, A. Mundstock, A. de Maeztu” Programme for Units of Excellence in R&D (CEX2018-000805- Feldhoff, A. Knebel, J. Gu, H. Meng, J. Caro, J. Am. Chem. Soc. 2018, M). It was also funded by the CERCA Program/Generalitat de Catalunya. 140, 10094. ICN2 was supported by the Severo Ochoa program from the Spanish [15] S. Yuan, X. Li, J. Zhu, G. Zhang, P. Van Puyvelde, B. Van der Bruggen, MINECO (Grant No. SEV-2017-0706). Y.C. and D.Z. acknowledge the Chem. Soc. Rev. 2019, 48, 2665. National Research Foundation Singapore (NRF2018-NRF-ANR007 POCE- MON) and the Ministry of Education – Singapore (MOE2018-T2-2-148, [16] J. W. Colson, A. R. Woll, A. Mukherjee, M. P. Levendorf, E. L. Spitler, MOE2019-T2-1-093) for their financial support of this project. V. B. Shields, M. G. Spencer, J. Park, W. R. Dichtel, Science 2011, 332, 228. [17] a) D. Yadav, A. Kumar, J. Y. Kim, N.-J. Park, J.-O. Baeg, J. Mater. Chem. A 2021, 9, 9573; b) Q. Hao, C. Zhao, B. Sun, C. Lu, J. Liu, M. Liu, L. J. Conflict of Interest Wan, D. Wang, J. Am. Chem. Soc. 2018, 140, 12152. [18] a) L. Cao, H. Wu, Y. Cao, C. Fan, R. Zhao, X. He, P. Yang, B. Shi, X. The authors declare no conflict of interest. You, Z. Jiang, Adv. Mater. 2020, 32, 2005565; b) N. A. Khan, R. Zhang, H. Wu, J. Shen, J. Yuan, C. Fan, L. Cao, M. A. Olson, Z. Jiang, J. Am. Chem. Soc. 2020, 142, 13450. Data Availability Statement [19] S. Karak, S. Kandambeth, B. P. Biswal, H. S. Sasmal, S. Ku- mar, P. Pachfule, R. Banerjee, J. Am. Chem. Soc. 2017, 139, The data that support the findings of this study are available from the cor- 1856. responding author upon reasonable request. [20] C. Fan, H. Wu, J. Guan, X. You, C. Yang, X. Wang, L. Cao, B. Shi, Q. Peng, Y. Kong, Y. Wu, N. A. Khan, Z. Jiang, Angew. Chem., Int. Ed. Engl. 2021, 60, 18051. Keywords [21] J. A. Martin-Illan, D. Rodriguez-San-Miguel, O. Castillo, G. Beobide, J. Perez-Carvajal, I. Imaz, D. Maspoch, F. Zamora, Angew. Chem., Int. aerogels, COF-membranes, covalent organic frameworks, gas mixtures Ed. 2021, 60, 13969. separation, porous materials [22] a) A.de la Pena Ruigomez, D. Rodriguez-San-Miguel, K. C. Stylianou, M. Cavallini, D. Gentili, F. Liscio, S. Milita, O. M. Roscioni, M. Received: October 18, 2021 L. Ruiz-Gonzalez, C. Carbonell, D. Maspoch, R. Mas-Balleste, J. Revised: November 19, 2021 L. Segura, F. Zamora, Chem. – Eur. J. 2015, 21, 10666; b) S. Published online: January 17, 2022 Y. Ding, J. Gao, Q. Wang, Y. Zhang, W. G. Song, C. Y. Su, W. Wang, J. Am. Chem. Soc. 2011, 133, 19816; c) B. J. Smith, A. C. Overholts, N. Hwang, W. R. Dichtel, Chem. Commun. 2016, 52, 3690. [1] S. Wang, L. Yang, G. He, B. Shi, Y. Li, H. Wu, R. Zhang, S. Nunes, Z. [23] D. A. Vazquez-Molina, G. S. Mohammad-Pour, C. Lee, M. W. Logan, Jiang, Chem. Soc. Rev. 2020, 49, 1071. X. Duan, J. K. Harper, F. J. Uribe-Romo, J. Am. Chem. Soc. 2016, 138, [2] a) S. Kim, H. Wang, Y. M. Lee, Angew. Chem., Int. Ed. 2019, 58, 17512; 9767. b) J. Shi, W. Wu, Y. Xia, Z. Li, W. Li, Desalination 2018, 441, 77. [24] C. Montoro, D. Rodriguez-San-Miguel, E. Polo, R. Escudero-Cid, M. [3] C. Chi, X. Wang, Y. Peng, Y. Qian, Z. Hu, J. Dong, D. Zhao, Chem. L. Ruiz-Gonzalez, J. A. R. Navarro, P. Ocon, F. Zamora, J. Am. Chem. Mater. 2016, 28, 2921. Soc. 2017, 139, 10079. [4] Y. S. Lin, Ind. Eng. Chem. Res. 2018, 58, 5787. [25] a) A. Gimeno, P. Ares, I. Horcas, A. Gil, J. M. Gomez-Rodriguez, J. [5] J. M. Lee, A. I. Cooper, Chem. Rev. 2020, 120, 2171. Colchero, J. Gomez-Herrero, Bioinformatics 2015, 31, 2918; b) I. Hor- [6] a) Q. Qian, P. A. Asinger, M. J. Lee, G. Han, K. Mizrahi Rodriguez, S. cas, R. Fernandez, J. M. Gomez-Rodriguez, J. Colchero, J. Gomez- Lin, F. M. Benedetti, A. X. Wu, W. S. Chi, Z. P. Smith, Chem. Rev. 2020, Herrero, A. M. Baro, Rev. Sci. Instrum. 2007, 78, 013705. 120, 8161; b) H. Wang, S. Zhao, Y. Liu, R. Yao, X. Wang, Y. Cao, D. Ma, [26] K. L. Johnson, Contact Mechanics, Cambridge University Press, Cam- M. Zou, A. Cao, X. Feng, B. Wang, Nat. Commun. 2019, 10, 4204. bridge 1987. Adv. Sci. 2022, 9, 2104643 2104643 (8 of 9) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com [27] C. H. Feriante, S. Jhulki, A. M. Evans, R. R. Dasari, K. Slicker, W. R. [32] Y. Ying, M. Tong, S. Ning, S. K. Ravi, S. B. Peh, S. C. Tan, S. J. Penny- Dichtel, S. R. Marder, Adv. Mater. 2020, 32, 1905776. cook, D. Zhao, J. Am. Chem. Soc. 2020, 142, 4472. [28] M. Tong, Q. Yang, Q. Ma, D. Liu, C. Zhong, J. Mater. Chem. A 2016, [33] W. L. F. Armarego, C. L. L. Chai, Purification of Laboratory Chemicals, 4, 124. Butterworth-Heinemann, Oxford 2009. [29] L. Yang, H. Yang, H. Wu, L. Zhang, H. Ma, Y. Liu, Y. Wu, Y. Ren, X. Wu, [34] a) J. E. Sader, R. Borgani, C. T. Gibson, D. B. Haviland, M. J. Hig- Z. Jiang, J. Mater. Chem. A 2021, 9, 12636. gins, J. I. Kilpatrick, J. Lu, P. Mulvaney, C. J. Shearer, A. D. Slattery, P. [30] G. Lim, K. B. Lee, H. C. Ham, J. Phys. Chem. C 2016, 120, 8087. A. Thoren, J. Tran, H. Zhang, H. Zhang, T. Zheng, Rev. Sci. Instrum. [31] H. W. Kim, H. W. Yoon, B. M. Yoo, J. S. Park, K. L. Gleason, B. D. 2016, 87, 093711; b) J. E. Sader, J. W. M. Chon, P. Mulvaney, Rev. Sci. Freeman, H. B. Park, Chem. Commun. 2014, 50, 13563. Instrum. 1999, 70, 3967. Adv. Sci. 2022, 9, 2104643 2104643 (9 of 9) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH