Research ArticleENGINEERING

Vapor-phase linker exchange of metal-organic frameworks

See allHide authors and affiliations

Science Advances  01 May 2020:
Vol. 6, no. 18, eaax7270
DOI: 10.1126/sciadv.aax7270


Metal-organic frameworks (MOFs) have been attracting intensive attention because of their commendable potential in many applications. Postsynthetic modification for redesigning chemical characteristics and pore structures can greatly improve performance and expand functionality of MOF materials. Here, we develop a versatile vapor-phase linker exchange (VPLE) methodology for MOF modification. Through solvent-free and environment-friendly VPLE processing, various linker analogs with functional groups but not for straightforward MOF crystallization are inserted into frameworks as daughter building blocks. Besides single exchange for preparing MOFs with dual linkers, VPLE can further be performed by multistage operations to obtain MOF materials with multiple linkers and functional groups. The halogen-incorporated ZIFs exhibit good porosity, tunable molecular affinity, and impressive CO2/N2 and CH4/N2 adsorption selectivities up to 31.1 and 10.8, respectively, which are two to six times higher than those of conventional adsorbents. Moreover, VPLE can substantially enhance the compatibility of MOFs and polymers.


Metal-organic frameworks (MOFs) are crystalline materials coordinated by metal ions/centers and organic linkers (1). Because of ultrahigh specific surface areas, well-defined apertures, variable topological and chemical characteristics, and tunable affinities for different molecules, MOF materials have been attracting substantial attention for gas storage/separation (24), catalysis (5, 6), sensing (7, 8), and optics (9, 10). The functions and applications of MOFs are intimately related to their intrinsic chemical and crystalline features. If the linkers with appropriate geometries and specific functional groups can be synthesized and used for crystallization, MOFs can be functionalized purposefully. Up to now, numerous MOFs have been designed and reported, thanks to the diversity of metal salts and linkers (11). Unfortunately, hydro/solvothermal methods, which are most widely workable for MOF preparation, are failing to obtain certain desired MOF crystals because of the issues of solubility and stability and the negative influences of functional groups toward metal-linker coordination reactions (12). Therefore, postsynthetic modification is proposed for changing the physicochemical properties of frameworks or designing new MOFs.

Postsynthetic modification refers to incorporation of functional groups, ions, and molecules in MOFs and replacement of building blocks by analogous metal ions and linkers (1226). The former requires the availability of active sites in the frameworks for reacting with introduced reagents (19, 20), thereby limiting the types of MOFs for modification. The latter, linker/cation exchanges, can regulate pore structures (21, 22), sorption properties (23, 24), reactivities (25), and crystalline topologies (3) and even construct new MOFs that are unattainable via conventional synthesis (26). Relative to cation substitution (12, 14), linker exchange has been investigated more intensively owing to the diversity of linkers. Linker exchange is generally carried out by immersing MOF crystals in linker substitute solutions accompanied by heat treatment, named solvent-assisted linker exchange (SALE). The liquid-solid reaction of SALE is strongly affected by the dynamic process of reactant transport and the nature of solvents.

Solvent-free processing includes mechanochemical synthesis (2729), hot pressing (30, 31), vapor deposition (3236), etc. Vapor deposition realized by gas-solid reaction shows many merits, e.g., environmental friendliness, low cost, easy scale up, and high precursor utilization, and has been demonstrated with robust availability for MOF formation (3236). Stassen et al. (32) fabricated zeolitic imidazolate framework-8 (ZIF-8) films through exposing zinc oxide layers in 2-methylimidazole (MeIM) vapor. In our previous study, vapor deposition was developed to obtain ultrathin MOF membranes through transforming Zn-based gels composed of zinc acetate/ethanolamine complexes (35). Ma et al. (36) further prepared MOF membranes by all vapor–phase processing. These studies seem to involve the coordination reaction between linkers and metallic oxides or gels to form MOFs but can also be regarded as the replacement of nonmetallic parts of oxygen in zinc oxide or organic chains in Zn-based gel by linkers. Consequently, we envisage that solvent-free vapor deposition has the capability to accomplish linker exchange of MOFs.

Here, we report a solvent-free strategy for postsynthetic modification of MOFs, which we refer to as vapor-phase linker exchange (VPLE) (Fig. 1A). Through environment-friendly VPLE processing, single and multiple linker analogs, which have specific functional groups and are unable to synthesize MOFs by hydro/solvothermal methods, are inserted in parent MOFs via single- and multistage operations. The porosity and adsorption properties of the exchanged MOFs are tunable. Moreover, the incorporation of halogens distinctly enhances the compatibility between MOFs and polymers.

Fig. 1 VPLE of MOFs.

(A) Schematic of VPLE strategy. Yellow balls and colored sticks represent metal centers and linkers, respectively. (B) Possible hexagonal window structures of ZIF-8 after first and second linker exchanges. The window structures are surrounded by zinc centers and organic linkers. (C) Molecular structures of parent and daughter linkers. IIM, 4-iodoimidazole; BrIM, 4-bromoimidazole; ClBIM, 2-chloromethylbenzimidazole.



To examine the feasibility of the VPLE strategy, ZIF-8 (37, 38), one of the most studied MOF with zinc centers and MeIM linkers, was synthesized by the solvothermal method and used as a probe (Fig. 1B). 4-Iodoimidazole (IIM) with a melting point of 137°C was applied as the daughter linker (Fig. 1C). The relatively lower molecular weight and polarizability of imidazolates than benzene-carboxylic linkers mean higher volatility, saturated vapor pressure, and linker concentration. Unlike SALE that usually needs aggressive solvents such as methanol and N,N-dimethylformamide (26), VPLE was executed by exposing ZIF-8 powders in IIM vapor produced by heat treatment at 150°C. In the postsynthetic process, the IIM vapor surrounded the ZIF-8 crystals and substituted MeIM to form the exchanged MOF nanoparticles with mixed linkers, labeled as ZIF-8/I. To remove any unreacted IIM on surfaces and in frameworks, ZIF-8/I was washed with heated methanol several times and dried at 120°C in vacuum for 12 hours. Fourier transform infrared (FTIR) spectra suggest that the characteristic peaks of N─Zn at 420 cm−1, C─H at 760 cm−1, C─N at 995/1146 cm−1, and C═N at 1582 cm−1 for ZIF-8 are retained after IIM exposure (Fig. 2A), implying the analogous chemical composition and zinc-imidazolate coordination state of ZIF-8/I and ZIF-8. The emerging C─I peak at 1088 cm−1 demonstrates the integration of IIM linkers in ZIF-8/I. There are two possible combination modes between IIM and ZIF-8, IIM replacement of MeIM and IIM encapsulation in cavities of the frameworks. If IIM is encapsulated in the cavities of ZIF-8, then the proton of the daughter linker will be retained, whereas IIM is deprotonated if linker substitution reaction occurs. No N─H peak of IIM at 2500 to 3000 cm−1 and 1600 to 1850 cm−1 in the FTIR spectrum of ZIF-8/I proves that the daughter linker is incorporated in the framework by linker exchange rather than cavity occupation (fig. S1). The chemical bonding states were examined by x-ray photoelectron spectroscopy (XPS). The high-resolution C 1s XPS spectrum of ZIF-8/I presents a new peak for C─I with a binding energy of 284.4 eV, besides intrinsic C─C/C═C/C─H and C─N/C═N bonds (Fig. 2B and fig. S2). Because IIM has no C─C bond, the content of C─C/C═C/C─H bonds decreases from 53.4 to 47.3% after linker exchange, as indicated by the reduction of peak areas. I 3d peaks in the XPS spectrum of ZIF-8/I confirm the presence of IIM in the framework again (Fig. 2C). High-resolution N 1s XPS spectra of both ZIF-8 and ZIF-8/I mainly consist of C─N─Zn bond (fig. S2). Theoretically, skewing of electron cloud causes binding energy to change. The almost identical location and shape of the N 1s peak for two ZIFs imply no unreacted IIM with N─H bond in ZIF-8/I because the proton can affect the distribution of electron cloud. This finding also suggests the linker exchange rather than occupation. These chemical characterizations illustrate the successful linker exchange of MOFs by the solvent-free VPLE strategy.

Fig. 2 Characterizations of ZIF-8 and ZIF-8/I.

(A) FTIR spectra of ZIF-8 and ZIF-8/I. The characteristic peaks of IIM are marked by asterisks. (B) High-resolution C 1s XPS spectrum of ZIF-8/I. (C) High-resolution I 3d XPS spectra of ZIF-8 and ZIF-8/I. (D) 1H NMR spectra of ZIF-8 and ZIF-8/I. The peaks of IIM and MeIM in ZIF-8/I are marked by triangles and asterisks, respectively. ppm, parts per million. (E) X-ray diffraction (XRD) patterns of ZIF-8 and ZIF-8/I. XRD pattern of simulated ZIF-8 is presented for comparison. (F) Nitrogen adsorption-desorption isotherms and pore size distributions of ZIF-8 and ZIF-8/I. The pore distribution is calculated on the basis of the slit-pore model. (G) Transmission electron microscopy (TEM), high-resolution TEM (HRTEM), and energy dispersion spectroscopy (EDS) mapping images of ZIF-8. (H) TEM, HRTEM, and EDS mapping images of ZIF-8/I.

The effect of exposure time on linker exchange was studied (figs. S3 to S5). As expected, more MeIM linkers are exchanged by IIM as time passes, which can be identified by the gradually increased intensity of the C─I peak in FTIR spectra (fig. S3). The stronger C─I/I 3d peaks and weaker C─C/C═C/C─H peak in XPS spectra of ZIF-8/I after exposure for 24 hours prove that there are more IIM linkers in crystals as well (fig. S4). The ZIF-8/I composition was measured by proton nuclear magnetic resonance (1H NMR) (15, 39), through digestion of crystalline powders by dimethyl sulfoxide–d6 and deuterium chloride. With the extension of exposure time, the linker exchange ratio increases at the beginning and then reaches a plateau (Fig. 2D and fig. S6). The exchange ratios of ZIF-8/I with exposure times of 12 and 72 hours are 16 and 26%, respectively. As demonstrated by previous studies (12, 21, 40, 41), the linker in parent MOFs can be completely replaced by the daughter one for parent and daughter linkers with substantively different physicochemical properties (e.g., length and acidity coefficient). If two linkers have similar features, the linker exchange can be regarded as equilibrium between linkers to solid crystals. We exposed ZIF-8/I in MeIM vapor at 150°C for reverse exchange. FTIR spectrum displays reduced intensity of C─I peak (fig. S7), suggesting the substitution of IIM by MeIM and the reversibility of VPLE. Therefore, the reaction process of VPLE can be described in five steps in turn (fig. S8): IIM vapor is generated, IIM replaces MeIM, MeIM escapes as vapor, MeIM partial pressure increases, and reaction approaches equilibrium. This may be the reason for the exchange ratio plateau as time increases because the concentration and partial pressure of exchanged-out MeIM increase with the progress of replacement reaction. With the larger partial pressure from the fresh IIM vapor, the exchange ratio of ZIF-8/I-2nd after the second sequential VPLE treatment further increases, confirming the proposed reaction mechanism (fig. S9).

X-ray diffraction (XRD) patterns suggest that ZIF-8/I has a pure sodalite crystalline structure like ZIF-8 (Fig. 2E). All ZIF-8/I particles with different processing times exhibit well-maintained crystalline structures (fig. S10). The porosity of ZIF-8 and ZIF-8/I was measured by nitrogen adsorption-desorption experiments. The isotherms reveal that the Brunauer-Emmett-Teller (BET) specific surface area of ZIF-8/I is slightly smaller than that (1667 ± 40 m2 g−1) of ZIF-8 but is maintained at 1522 ± 40 m2 g−1 (Fig. 2F, fig. S11, and table S1), which is in the normal range of 1000 to 1700 m2 g−1 of ZIF-8 materials from previous studies (5, 24, 25, 35, 37, 38). The typical gate opening phenomenon of ZIF-8 can be obviously observed from the nitrogen adsorption-desorption isotherms (fig. S12) (24). The main half pore width of MOFs derived from the sodalite cage is approximately 5.4 Å [Fig. 2F (inset) and fig. S12], consistent with that of typical ZIF-8 materials (37, 38). The slight variation in specific surface area is interpreted by the notion that 2-methyl with a van der Waals radius of 2.0 Å and C─C bond with a length of 1.5 Å for MeIM are partially replaced by 2.2-Å 4-iodine and 2.1-Å C─I bond for IIM. Transmission electron microscopy (TEM) images of the MOF nanoparticles show that both ZIF-8 and ZIF-8/I crystals have rhombic dodecahedra structures with exposed {011} facet (Fig. 2, G and H). Although the electron beam with high energy of high-resolution TEM (HRTEM) may destroy MOF crystalline structures (42), ZIF-8 and ZIF-8/I nanoparticles still display the uniform dark spots and white speckles from zinc centers and framework cavities, respectively (43). Energy-dispersive spectroscopy (EDS) mapping reveals homogeneous carbon, nitrogen, and zinc distributions of ZIF-8 and ZIF-8/I and the existence of iodine in ZIF-8/I (Fig. 2, G and H, and fig. S13). Note that linker exchange may occur during the methanol washing step after loading IIM on ZIF-8 during vapor exposure. The existence of an intensive C─I peak but no/weak N─H peaks in the FTIR spectrum of the as-synthesized ZIF-8/I (ZIF-8/I-AS) without washing indicates almost no/very few nondeprotonated IIM on crystals (fig. S14A). NMR reveals that ZIF-8/I-AS has a similar IIM content to ZIF-8/I after washing (fig. S14B). Nitrogen adsorption-desorption shows the good porosity of ZIF-8/I-AS (fig. S14, C and D). These results suggest the pure chemical structures of ZIF-8/I after VPLE even without washing in methanol. The control SALE experiments of washing ZIF-8 by IIM solutions under the same conditions as ZIF-8/I were also performed (fig. S15). The weak C─I peak in FTIR spectra and low IIM ratios imply that linker exchange can be triggered by SALE, but the replacement is much slower than VPLE. The above characterizations and experiments confirm that the linker exchange mainly happens during the vapor exposure step and VPLE has a faster exchange rate than conventional SALE. We attempted to fabricate ZIF-8/I by the solvothermal method like ZIF-8 synthesis, through thermal treatment of IIM/zinc salt-contained solution with different metal salts, solvents, salt/linker ratios, and temperatures. All resulting powders have amorphous or poor crystalline structures and a weak N─Zn coordination bond (figs. S16 and S17), indicating that ZIF-8/I is unattainable through solvothermal synthesis, at least under various conditions applied here. These experiments demonstrate that VPLE can incorporate the linker analogs that cannot be directly used for crystallization in MOFs without destruction of crystalline and porous features.

Versatility of VPLE

To verify the versatility of the VPLE strategy, two additional imidazolate molecules, 4-bromoimidazole (BrIM) and 2-chloromethylbenzimidazole (ClBIM) (Fig. 1C), were used for linker exchange. The temperature for vapor generation of BrIM and ClBIM with melting points of 135° and 148°C was 150° and 160°C, respectively. Both ZIF-8/Br and ZIF-8/Cl have intact sodalite crystalline frameworks (Fig. 3A). There is an obvious C─Br peak in the FTIR spectrum of ZIF-8/Br (Fig. 3B), suggesting successful linker exchange after vapor exposure. The probable explanation for the C─Cl peak in the FTIR spectrum of ZIF-8/Cl at 700 to 750 cm−1 is the C─H overlap at a similar position. The exchange ratios of ZIF-8/Br and ZIF-8/Cl are 16 and 14%, respectively (table S1). Because of the analogous molecular geometry, acidity coefficient, and melting point of IIM and BrIM, ZIF-8/I and ZIF-8/Br exhibit similar exchange ratios. Although the processing temperature for ClBIM exchange is higher than that for IIM and BrIM, ZIF-8/Cl has a lower content of daughter linker than ZIF-8/I and ZIF-8/Br, originating from the lower acidity coefficient and larger molecular size (12). The morphology of ZIF-8/Cl is clean-cut granular (Fig. 3C and fig. S18), while ZIF-8/Br shows certain deformation owing to the thermal treatment in BrIM vapor (Fig. 3C and fig. S19). HRTEM images of ZIF-8/Cl and ZIF-8/Br crystals are consistent with those of ZIF-8 and ZIF-8/I (figs. S18 and S19). The existence of halogens in ZIF-8/Br and ZIF-8/Cl nanoparticles validates the insertion of BrIM and ClBIM in frameworks (Fig. 3C). The BET specific surface areas of ZIF-8/Cl and ZIF-8/Br are 1599 ± 40 and 621 ± 15 m2 g−1, respectively (figs. S20 and S21). We tried to synthesize ZIF-8/Cl and ZIF-8/Br in precursor methanol solutions by the solvothermal method as well. Although the produced powders synthesized in BrIM-contained solution display some peaks in XRD pattern and the peak of the N─Zn coordination bond in FTIR spectrum, the low intensity of XRD peaks and massively nondeprotonated BrIM shown in FTIR spectrum demonstrate the inferior crystallinity (fig. S22). For the solvothermal synthesis with ClBIM-contained solution, there is even no solid formation (fig. S23). These suggest the robust implementability of VPLE for linker exchange of unattainable MOFs. In addition to fabrication of MOFs with dual linkers, VPLE can be further performed by multistage operations. The ZIF-8/I powders were exposed in ClBIM vapor for a second linker exchange. The obtained ZIF-8/I/Cl has pure crystalline structure, intact morphology, and good porosity, as shown in XRD pattern, TEM images, and nitrogen adsorption-desorption isotherms (Fig. 3, A and C, and figs. S24 and S25). FTIR and EDS characterizations of ZIF-8/I/Cl testify the introduction of IIM and ClBIM (Fig. 3, B and C, and fig. S24). The above results illuminate that the VPLE strategy is commendably versatile for linker exchange and available for multistep linker exchange to obtain MOF materials with multiple linkers and functional groups.

Fig. 3 Versatility of VPLE.

(A) XRD patterns of ZIF-8, ZIF-8/Cl, ZIF-8/Br, and ZIF-8/I/Cl. The simulated ZIF-8 XRD pattern is presented for comparison. (B) FTIR spectra of ZIF-8, ZIF-8/Cl, ZIF-8/Br, and ZIF-8/I/Cl. (C) TEM and EDS mapping images of ZIF-8/Cl, ZIF-8/Br, and ZIF-8/I/Cl.

Adsorption behaviors of exchanged ZIFs

We tested the gas adsorption properties of the prepared MOFs for CO2, CH4, and N2 at 25°C. All ZIFs exhibit gas uptakes with the order of CO2, CH4, and N2. This result is ascribed to the polarizabilities and quadrupole moments of CO2 (26.3 × 10−25 cm−3 and 14.3 × 10−40 C m2), CH4 (26.0 × 10−25 cm−3 and 0), and N2 (17.6 × 10−25 cm−3 and 4.7 × 10−40 C m2) (Fig. 4, A to E, and table S2) (4446). The CO2 uptake of ZIF-8 is 18.0 ± 0.5 ml g−1 at 100 kPa. After linker exchange, exchanged MOFs may display decreased or increased CO2 adsorption capacity. For example, the CO2 uptake of ZIF-8/Cl increases to 18.5 ± 0.5 ml g−1, while that of ZIF-8/I reduces to 16.6 ± 0.4 ml g−1. The CO2 uptake is related to the specific surface area of ZIFs. Meanwhile, the halogen incorporation affects adsorption capability as well, because the strengthened polarizability reinforces the interaction of frameworks and gas molecules (47, 48). For CH4 adsorption, the uptake mainly ranges in the order of the specific surface area of MOFs, except that ZIF-8/Br and ZIF-8/Cl have better adsorption than ZIF-8/I/Cl and ZIF-8, respectively. These may be attributed to the strengthened polarizability, steric-hindrance effect, and some amorphous phase of the exchanged MOFs (49, 50). With respect to N2, ZIF-8/I and ZIF-8/Cl exhibit higher uptake than ZIF-8, while ZIF-8/Br and ZIF-8/I/Cl have a smaller one than ZIF-8. These results confirm that VPLE can tune the sorption properties of MOFs.

Fig. 4 Gas adsorption properties of ZIFs.

(A) Adsorption isotherms of ZIF-8 for CO2, CH4, and N2. STP, standard temperature and pressure. (B) Gas adsorption isotherms of ZIF-8/I. (C) Gas adsorption isotherms of ZIF-8/Cl. (D) Gas adsorption isotherms of ZIF-8/Br. (E) Gas adsorption isotherms of ZIF-8/I/Cl. (F) CO2/N2 and CH4/N2 selectivities of ZIF-8 and linker-exchanged ZIFs.

Capture of CO2 from N2 has notable environmental implications. We calculated the sorption selectivity of the MOFs using Henry’s law (fig. S26). ZIF-8 has an expectable CO2/N2 selectivity of 15.0, which is in accordance with that reported in a previous study (Fig. 4F) (51). The actual separation factor of gas mixture may be lower than the calculated one owing to the competitive adsorption. Because of the enhanced polarizability of frameworks, the relatively larger increment in N2 adsorption results in smaller CO2/N2 selectivity for ZIF-8/I and ZIF-8/Cl. Fortunately, since the N2 adsorption capacity has the largest reduction in proportion that originated from the relatively lower surface area and some amorphous phase in frameworks, ZIF-8/Br achieves a markedly improved CO2/N2 selectivity of 31.1, which is approximately double that for ZIF-8 materials, despite the degeneration of gas uptakes to some extent. Besides carbon capture, separation of CH4 and N2, such as natural gas/biogas purification and coal bed methane enrichment, is of great importance for chemical and energy-related industries (52, 53). However, CH4 and N2 are extremely difficult to separate, as they share highly similar kinetic diameters and polarizabilities. The variation in CH4/N2 selectivity of the MOFs is analogous to that for CO2/N2. The CH4/N2 selectivity of ZIF-8 is 4.9. After VPLE processing, ZIF-8/I, ZIF-8/Cl, and ZIF-8/I/Cl display some reduction in selectivity, whereas ZIF-8/Br exhibits an outstanding CH4/N2 selectivity of 10.8, which is two to six times larger than that of commonly investigated porous carbons (54), zeolites (55), and MOFs (56) (table S3). The impressive selectivities suggest that the MOFs exchanged by VPLE have great potential for gas separation.

Compatibility of hybrids

Because of poor processability, MOF materials are often hybridized with polymers for better application (57). Nevertheless, particle agglomeration and inferior compatibility are bottlenecks. Surface modification of MOFs can enhance the dispersibility of particles and strengthen the interaction between polymers and fillers (57, 58). We dispersed ZIF-8, ZIF-8/I, and ZIF-8/Br powders in chloroform by ultrasonic treatment and then added polysulfone (PSF) to obtain MOF/PSF composites. The ZIF-8/I and ZIF-8/Br fillers are homogeneously distributed in polymer, yet the ZIF-8/PSF composite shows serious filler agglomeration and large gaps at phase interfaces (fig. S27), verifying the admirably positive influence of VPLE in filler dispersity and hybrid compatibility. Agglomeration of MOFs usually occurs in the drying process (59); thus, it is difficult to prepare the suspended solutions with commendably monodisperse fillers. Benefitting from halogen incorporation, modified ZIFs can be easily redispersed in halogenated hydrocarbon-chloroform (fig. S28). For the reason of superior compatibility, the enhancement in polarity of MOFs improves the adhesion between fillers and polymers.


We have developed a solvent-free VPLE method for postsynthetic modification of MOF materials. Through VPLE, the daughter linkers, which have functional groups but are unable to be used for fabricating MOFs by hydro/solvothermal treatments, can be inserted into the parent MOFs by linker exchange. The VPLE process is environment friendly and versatile for various linkers and also has the practicability of multistep linker exchange to incorporate multiple linkers and functional groups in one MOF material. The linker-exchanged MOF materials exhibit high porosity and tunable gas sorption property. Relative to the parent ZIF-8, the CO2/N2 and CH4/N2 selectivities of ZIF-8/Br are doubled to 31.1 and 10.8, respectively. Moreover, VPLE can obviously improve the compatibility between MOFs and polymers. Overall, the postsynthetic modification based on gas-solid reaction described herein offers an alternative route to adjust pore structures, sorption properties, and other physicochemical features of porous materials for achieving desired functions.


Preparation of ZIF-8

Zinc nitrate hexahydrate (0.30 g) and MeIM (0.66 g) were dissolved in methanol (14.0 ml), respectively. These two solutions were completely mixed by stirring and transferred into a Teflon-lined stainless steel autoclave. For crystallization, the autoclave was heat-treated at 150°C for 12 hours. After thermal treatment, the autoclave was naturally cooled to room temperature. The produced powder was isolated by centrifugation at 5000 rpm for 6 min. The obtained ZIF-8 was washed with heated methanol at 60°C three times and dried at 120°C in vacuum for 12 hours.

Linker exchange of ZIF-8

IIM powder (0.10 g) was placed in a Teflon-lined stainless steel autoclave. The ZIF-8 powder (0.05 g) was placed on a porous dish and then sealed in the Teflon lining. There was no direct contact between ZIF-8 and IIM. For linker exchange, the autoclave was heat-treated at 150°C for a certain period of time (1, 3, 6, 12, 24, 48, and 72 hours). After natural cooling to room temperature, the exchanged ZIF-8/I was obtained. To remove the unreacted IIM, the obtained ZIF-8/I was washed three times with heated methanol at 60°C and dried at 120°C in vacuum for 12 hours. ZIF-8/Cl and ZIF-8/Br were prepared using the same procedures as those for ZIF-8/I, except that IIM was replaced by ClBIM and BrIM and temperature was changed from 150° to 160°C for ClBIM. ZIF-8/I/Cl was fabricated by exposing ZIF-8/I in ClBIM vapor as the synthesis of other linker-exchanged ZIFs.

Preparation of ZIF/PSF composites

ZIF powder (0.05 g) was dispersed in chloroform (0.56 ml) by stirring and ultrasonic treatment for 30 min. PSF (0.95 g) was added into the suspension and stirred overnight at room temperature. The casting solution was defoamed by slow stirring for 5 min and ultrasonic treatment for 5 min, repeating the procedure three times, and then letting the solution stand for 1 hour. The hybrid composite was prepared by pouring the solution on a flat glass plate. After evaporation for 1 hour, the film was torn off and thermally treated at 80°C in vacuum for 12 hours.


An FTIR spectrophotometer (IRTracer-100, Shimadzu Co.) was used to investigate the chemical structure of the ZIFs. XPS experiment was performed by using an RBD upgraded PHI-5000C ESCA system (PerkinElmer) with an incident radiation of monochromatic Mg Kα x-rays (hν = 1253.6 eV). XRD characterization was carried out by using an x-ray diffractometer (D2 PHASER and D8 ADVANCE, Bruker Co.). 1H NMR characterization was carried out by using an NMR apparatus (Ascend 400 MHz, Bruker Co.). Before characterization, the sample was digested by dimethyl sulfoxide–d6 and deuterium chloride. Nitrogen adsorption-desorption isotherms and gas adsorption data were collected by using a physisorption analyzer (Autosorb iQ Station 1, Quantachrome Instruments). Before measurement, the sample was thermally treated in vacuum at 150°C for 12 hours. The multipoint BET method and the nonlocal density functional theory method were applied to calculate the specific surface area and half pore width, respectively. Nitrogen adsorption-desorption isotherms for obtaining BET surface area and pore size distribution were recorded at 77.35 K held using a liquid nitrogen bath. N2, CO2, and CH4 uptake data were collected at a temperature of 25°C and a pressure of 0 to 100 kPa. TEM images were captured by using an electron microscope (JEM-2100, JEOL Ltd.) with an accelerating voltage of 200 kV. The element distribution of the samples was analyzed by the attached x-ray energy dispersion spectroscope.


Supplementary material for this article is available at

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.


Acknowledgments: Funding: This work was financially supported by the National Natural Science Foundation of China (grant no. 51708252), the Fundamental Research Funds for the Central Universities (grant no. 21617322), and Jinan University (grant no. 88016674). Author contributions: W.L. conceived the research idea and formulated the project. W.W., J.S., and M.J. performed experiments including fabrication and modification of MOFs. W.W., Z.L., G.L., and W.L. carried out sample characterizations including XRD, FTIR, gas adsorption, XPS, NMR, and TEM. W.W. analyzed and processed experimental data. W.L. wrote the paper. All authors contributed to revising the paper. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.
View Abstract

Stay Connected to Science Advances

Navigate This Article