Application of ZIF-67 as a crosslinker to prepare sulfonated polysulfone mixed-matrix membranes for enhanced water permeability and separation properties

High-performance sulfonated polysulfone (SPSf) mixed-matrix membranes (MMMs) were fabricated via a nonsolvent-induced phase separation (NIPS) method using zeolitic imidazolate frameworks-67 (ZIF-67) as a crosslinker. Acid-base crosslinking occurred between the sulfonic acid groups of SPSf and the tertiary amine groups of the embedded ZIF-67, which improved the dispersion of ZIF-67 and simultaneously improved the membrane strzcture and permselectivity. The dispersion of ZIF-67 in the MMMs and the acid-base crosslinking reaction were verified by energy-dispersive X-ray spectroscopy (EDX), X-ray diffractometry (XRD), Fourier-transform infrared spectroscopy (FTIR), and X-ray photoelectron spectroscopy (XPS). The pore structure analysis of MMMs indicated that filling ZIF-67 into SPSf enhanced the average surface pore sizes, surface porosities and more micropore in cross-sections. The crossflow filtrations showed the MMMs have higher pure water fluxes (57 to 111 L m 2 h ) than the SPSf membrane (55 L m 2 h ) but also higher bovine serum albumin (BSA) rejection rate of 93.9–95.8%, a model protein foulant. The MMMs showed a higher water contact angle than the SPSf membrane due to the addition of hydrophobic ZIF-67 and acid-base crosslinking, and also maintained high thermal stability evidenced by the thermogravimetric analysis (TGA) results. At the optimal ZIF-67 concentration of 0.3 wt%, the water flux of the SPSf-Z67-0.3 membrane was 82 L m 2 h 1 with a high BSA rejection rate of 95.3% at 0.1 MPa and better antifouling performance (FRR1⁄4 70%).


INTRODUCTION
In recent years, sulfonated polymers have been widely used for membrane fabrication due to their hydrophilicity, high proton conduction capacity, selective ion transport, and chlorine tolerance (Luo et al. 2017;Liu et al. 2019).Sulfonated polysulfone (SPSf), sulfonated polyethersulfone (SPES), sulfonated polyetheretherketone (SPEEK), and sulfonated polyphenylenesulfone (SPPSU) are sulfonated aromatic polymers that are widely used in proton exchange membranes (PEMs) (Martos et al. 2015;Zhang et al. 2019a), pervaporation (Chen et al. 2009), solvent filtration (Abdulhamid et al. 2020), and membrane filtration for water purification (Widjojo et al. 2013;Yuan et al. 2017;Zhao et al. 2019;Zhou et al. 2019).For filtration purification, the applications mainly involve ultrafiltration, nanofiltration, reverse osmosis, and forward osmosis.In most studies, sulfonated polymers are mainly applied in the following three ways: (1) sulfonated polymers are used as additives to improve the hydrophilicity of the hydrophobic polymer matrix of ultrafiltration membranes by blending modification (Li et al. 2016;Gumbi et al. 2018;Hu et al. 2019); (2) a sulfonated polymer solution is poured onto the polymer matrix membrane to form a dense film on the surface of the membrane through interfacial polymerization and surface coating to enhance the salt retention performance of nanofiltration and reverse osmosis membranes (Lv et al. 2016;Zhao et al. 2019;Ormanci-Acar et al. 2020); and (3) a sulfonated polymer is directly used as the matrix material, and its hydrophilicity is used to improve the performance of the membrane (Widjojo et al. 2013;Zhou et al. 2019).
However, the existence of highly-polar sulfonic acid groups in the sulfonated polymer backbone decreases the aggregation of the polymer, expands the polymer backbone, and increases the polymer chain mobility, making the material more flexible (Genova-Dimitrova et al. 2001;Widjojo et al. 2013).When a highly-sulfonated polymer is used as the membrane matrix, the hydrophilic sulfonic acid groups destabilize the membrane at high pressures because they allow water to plasticize and swell the membranes, which compromises the mechanical properties (Peeva et al. 2011;Ma et al. 2013).Therefore, few studies using sulfonated polymers to prepare ultrafiltration membranes via a phase inversion process.Chung and coworkers (Luo et al. 2017) chose SPPSU with a sulfonation degree of 2.5 mol% as the matrix material to fabricate flat-sheet ultrafiltration (UF) membranes by blending with a basic polyethyleneimine polymer.Due to the intermolecular interactions between SPPSU and polyethyleneimine molecular chains, ionic crosslinking, chain entanglement, and polymer aggregation significantly improved the mechanical stability of the membrane by increasing its resistance to water-induced swelling.In our previous research (Yin et al. 2021), hydrophilic SPSf with a sulfonation degree of 5% was chosen as the membrane matrix to prepare a UF membrane.The mechanical strength of the matrix membrane was enhanced by introducing Tröger's base polymer, whose tertiary amine underwent acid-base crosslinking with the sulfonic acid group of the matrix.Through acid-base crosslinking, the mechanical properties and permeation flux of the sulfonated polymer matrix membrane were improved, but the selectivity-permeability trade-off was still observed.
Metal-organic frameworks (MOFs) are promising materials that can be used to overcome this trade-off.MOFs are porous crystalline materials that are formed by the self-assembly of organic and inorganic building units linked through coordination bonds (Guo et al. 2020).Due to their highly-tunable structure and chemical functionalization, they have been widely investigated for catalysis (Ren et al. 2019;Zhang et al. 2020), adsorption (Guo et al. 2020;Wu et al. 2020), and separation applications (Liu et al. 2015;Fu et al. 2020;Wei et al. 2020).Zeolitic imidazolate frameworks (ZIFs) are a subclass of MOFs that are constructed by joining a central bivalent metal ion (e.g., Zn 2þ or Cu 2þ ) with four imidazole ligands.They have the advantages of MOFs, including a tunable pore size, adjustable functional groups, better water stability (Karimi et al. 2019;Feng et al. 2020), and sustainable synthesis (Hardian et al. 2020).Recently, ZIFs have been used as additives to fabricate mixed-matrix membranes (MMMs).Their small pore-aperture size (,3.8Å) allows water to pass through while rejecting ions and water-soluble organic molecules to overcome the trade-off constraints of conventional polymer membranes (Yang et al. 2017;Liu et al. 2018).Organic ligands improve the compatibility of pure inorganic nanoparticles with the polymeric matrix (Nordin et al. 2015;Wang et al. 2016;Zeng et al. 2020) to prevent the formation of nonselective voids during the phase inversion process due to incompatibility between pure inorganic nanoparticles and the polymer matrix.Meshkat et al. (2020) added different concentrations of ZIF-67 and ZIF-8 to Pebax MH-1657 to prepare MMMs for separating CO 2 from CH 4 and N 2 .The Pebax/ZIF-67 system overcame the Robeson upper bound for CO 2 /N 2 separation.The polymer substrate-supported ultra-thin ZIF membranes with no defect were fabricated by an in situ interface self-repair strategy.The prepared polyvinyl alcohol (PVA)-Co 2þ interfacial layer as anchor sites for the nucleation of ZIF-67 membrane assisted by the interface-enhanced deprotonation of 2-methylimidazole.The resulting ZIF-67 membranes were 80-280 nm thick and were demonstrated unprecedented CO 2 permeance and good CO 2 /N 2 selectivity (Yu et al. 2021).
In our research, sulfonated polysulfone with a sulfonation degree of 5% was used as the matrix, and self-made ZIF-67 was incorporated to mixed-matrix membranes at a low filler loading from 0.1wt% to 0.5wt%.As illustrated in Figure 1, the weak acid-base reaction between the imidazole ligands and the sulfonic acid groups contributed to the good dispersion of ZIF-67, which would produce membranes surface with larger average surface pore sizes and higher porosities.ZIF-67 also produced a small pore size distribution, which simultaneously improved the membrane's permeability-selectivity.To the best of our knowledge, this is the first report on the fabrication of SPSf-ZIF-67 mixed-matrix membranes for the separation of organic contaminants.

Preparation of ZIF-67
ZIF-67 was prepared using a co-precipitation method.Briefly, 3 mmol of Co(NO 3 ) 2 •6H 2 O and 12 mmol of 2-MeIm were fully dissolved in a solution containing 20 mL of methanol and 20 mL of ethanol, respectively.Then, the above two solutions were quickly mixed under stirring at 550 rpm with a magnetic bar.Following that, the mixture was aged for 24 h at room temperature.Finally, a purple product was obtained by centrifugation, washed with 25 mL ethanol per gram of ZIF-67 per washing cycle four times, and dried at 60 °C.The photograph of the ZIF-67 was given in Figure 2.

Membrane preparation
The SPSf ultrafiltration membrane and SPSf-ZIF-67 nanocomposite membranes were prepared via non-solvent induced phase separation (NIPS) at room temperature.Various amounts of ZIF-67 nanocrystals (0, 0.1, 0.3, and 0.5 wt% based on 10 g of polymer and solvent) were ultrasonically dispersed in NMP solvent for 30 min.Then, the 2 wt% PVP-360 as a pore-forming agent and 15 wt% SPSf were added to the above solution.The casting solutions were stirred at 60 °C for  12 h and vacuumed for another 3 h without stirring to ensure the complete release of bubbles.The resultant casting solution was cast on a clean glass plate to form a uniform 150 μm-thick film using a steel knife with a 1.5 m/min casting speed.After exposure to air for 30 s, the film was immersed in a coagulation bath of deionized water and an UF membrane was obtained, and the corresponding prepared membranes are marked as SPSf-Z67-X, where X represents the mass fraction of ZIF-67, which varied from 0.1 to 0.5 wt%.The membranes were kept in deionized water for 48 h, and water was changed frequently to ensure the removal of NMP.

Characterization of the ZIF-67 and membranes
The structures of synthesized ZIF-67 crystals and all membranes were measured by XRD (Smartlab 9 kW, Cu Ka radiation) in the 2θ scanning range of 5-60°at a scan rate of 10°/min for ZIF-67 and 2°/min for the membrane under a voltage of 40 kV and current of 40 mA.The chemical composition of the materials was detected by Fourier-transform infrared spectroscopy (FTIR, Thermo Scientific Nicolet iS5) in the wavelength range of 400-4,000 cm À1 .The thermal stability of the ZIF-67 and all membranes was investigated by thermogravimetric analysis (TGA, DTG-60 Shimadzu, Japan) under nitrogen at a heating rate of 10 °C min À1 from 40 °C to 800 °C.A Brunauer-Emmett-Teller (BET)-ASAP2460 analyzer was used to examine the specific surface area and pore size distribution of the synthesized ZIF-67 samples.
Atomic force microscopy (AFM, Bruker Dimension Edge) images were used to analyze the morphology and roughness of the membrane surfaces.The surface roughness was characterized in terms of the mean roughness (Ra) and the root-meansquare roughness (Rq).The surface chemistry of the membranes was studied by X-ray photoelectron spectroscopy (XPS, Thermo Fisher Nexsa).XPS was used to analyze the S 2p binding energy on the top surface of membranes using a monochromatic X-ray of 12 kV and 6 mA.The hydrophilicity of the membranes was characterized using contact angle goniometry (OCA60) at room temperature by measuring the contact angles five times for each sample to obtain an average.The top surface and cross-sectional morphology of the prepared membranes and ZIF-67 images were observed using a scanning electron microscope (SEM, S-4800, Hitachi, Japan).The membranes were first cryogenically fractured in liquid nitrogen and then coated with gold prior to SEM analysis.The SEM images were analyzed using ImageJ v1.48 software to quantitatively calculate the average surface pore size and surface porosity of the membranes.For each sample, at least three pieces of the membrane were observed to avoid chance and reduce errors, and surface images were taken from a random location on each membrane (Masselin et al. 2001).The energy-dispersive X-ray spectroscopy (EDX) mappings were obtained to show the ZIF-67 distribution on the top of the surface and cross-section of the hybrid membranes.

Ultrafiltration experiments
The permeation fluxes, BSA rejection, and dynamic antifouling performance of the membranes were tested using a cross-flow filtration apparatus with an effective membrane area of 35 cm 2 .The apparatus was run at a room temperature and a constant pressure of 1 bar.The pressure required in the system is supplied by pump and the schematic diagram of the apparatus is shown in Figure 3.The membrane was pre-pressurized at 0.15 MPa for 30 min to stabilize the pure water flux of the membrane.The initial pure water flux was obtained by filtering deionized water through the membrane at 0.1 MPa for 30 min, and then the BSA solution (0.2 g/L, in phosphate buffered saline (PBS) solution, pH ¼ 7.4) was filtrated for 1 h.Next, the membranes were rinsed; the water consumption was 5 Â 10 3 L m À2 per washing cycle, and the water flux of the cleaned membranes was measured for another 30 min.The BSA solution filtration and water washing processes were repeated three times.The permeate solution was weighed every 2 min by an electronic balance connected to a computer.The flux of the membranes J (L m À2 h À1 ) was calculated using Equation (1): where m (kg) is the mass of permeated water, A (m 2 ) is the effective membrane area, ρ refers to the density of the permeate (1.0 kg/L), and Δt (h) is the permeation time of deionized water.
The BSA rejection rate (R) of the membranes was calculated from Equation (2): where C f is the BSA concentration in the feed, and C p refers to the BSA concentration in the permeate, which were determined by measuring the absorption at 278 nm using UV-Vis spectrophotometry.
The dynamic protein antifouling performance of ultrafiltration membranes was expressed in terms of the flux recovery ratio (FRR), which was calculated using Equations (3): where J wv is the initial steady pure water flux and J wc is the recovered water flux after the filtration of BSA solution.

RESULTS AND DISCUSSION
Characterization of the ZIF-67 and membranes The morphology and crystal structure of the prepared ZIF-67 was analyzed using SEM and XRD.The surface area and pore size of ZIF-67 were characterized by the Brunauer-Emmett-Teller (BET) method.Figure 2 shows that ZIF-67 had a rhombic dodecahedral shape with a smooth surface, which is a typical morphology for ZIF-67, with an average particle size of around 750-1,300 nm (Ma et al. 2013).The XRD patterns of ZIF-67 have characteristic diffraction peaks with strong intensities that match well with reported data (Guo et al. 2020).This indicates that ZIF-67 was successfully synthesized with a large particle size.Figure 4 shows that the BET surface area of particles was 1,328.8m 2 /g, and the internal pore size distribution ranged from 0.3-0.9nm (mode ¼ 0.36 nm; median ¼ 0.389 nm).The specific area and suitable pore size distribution (0.36 nm) in MMMs can provide passageways for water molecules while obstructing target contaminants (Dai et al. 2019(Dai et al. , 2020;;Yang et al. 2019a).
In order to further investigate the existence of ZIF-67 nanocrystals in MMMs and determine their effect on the polymer matrix, XRD, FTIR, XPS, AFM, CA, and TGA measurements were performed.As shown in Figure 5, no obvious peaks  were observed for the pure SPSf membrane, confirming its amorphous structure.The characteristic peaks of the (0 1 1), (1 1 2), and (2 2 2) crystal planes of ZIF-67 were also detected in the XRD pattern of the MMM.Due to the peak overlap and lower ZIF-67 loading, the characteristic peak at 424 cm À1 of the Zn-N stretching of ZIF-67 was only detected in the FTIR spectrum of the SPSf-Z67-0.5 membrane (Zeng et al. 2020).Moreover, according to Figure 6, the XPS spectra of the pristine SPSf membrane and the hybrid membrane both contain peaks for C 1s, N 1s, O 1s, and S 2p, but the hybrid membrane shows extra peaks for Co 2p derived from the ZIF-67 additive.These results confirm the existence of ZIF-67 on the membrane.As for the S 2p narrow-scan XPS spectrum of the membrane surfaces (Figure 6), the binding energies of a sulfonic acid group (-SO 3 H) and O ¼ S ¼ O showed slight shifts.The peaks for O ¼ S ¼ O at 167.72 eV (Lindberg et al. 1970) and -SO 3 H at 169.02 eV (Dizon & Venault 2018) in SPSf shifted to 167.90 eV and 169.18 eV in SPSf-Z67-0.5 membrane, respectively.These changes may be attributed to the weak interactions between ZIF-67 and sulfonic acid groups .The thermal stability of ZIF-67, SPSf, and all MMMs were investigated by TGA under a N 2 atmosphere from 40 to 800 °C (Figure 7).The TGA curve of ZIF-67 shows a sharp decomposition stage above 400 °C, indicating high thermal stability .The SPSf polymer and membranes with 0.1% and 0.3% ZIF-67 displayed similar TGA curves with a typical two-step degradation.The first weight loss occurred from about 200 to 400 °C due to the thermal degradation of sulfonic acid groups (Martos et al. 2015;Lei et al. 2017).The second weight loss above 450 °C is attributed to main chain decomposition (Xing et al. 2004;Barjola et al. 2018); however, when the ZIF-67 content increased to 0.5%, the first step of SPSf-Z67-0.5 decomposition disappeared, and the TGA curve only showed a single decomposition stage above 450 °C due to main chain decomposition.This is mainly because of the consumption of -SO 3 H groups during acid-base crosslinking with tertiary amine groups in ZIF-67, which decreased the decomposition of -SO 3 H groups and improved the thermal stability of the blended membranes.This result reconfirmed the interactions between SPSf and ZIF-67.
AFM was used to investigate the membrane surface roughness.Figure 8 illustrates the 2D and 3D surface morphologies of the membranes.The 3D surface morphologies exhibit a typical ridge-valley morphology on the membrane surface.The Ra and Rq parameters are used to indicate the roughness of the membranes.As shown in Figure 8, compared with the pristine SPSf membrane, Ra and Rq increased as the ZIF-67 content increased.Particularly, the Ra and Rq of the SPSf-Z67-0.5 membrane were 9 times higher than those of the pristine SPSf membrane.This means that the surface roughness of the SPSf-ZIF-67 MMMs significantly increased after incorporating ZIF-67 in the membrane matrix.This may have been caused by two factors.First, the increased surface roughness was accompanied by an increased surface porosity, as shown in Table 1.Secondly, this may be related to an increase in the distribution of ZIF-67 on the membrane surface due to the increased amount of ZIF-67 in the membrane matrix.
Hydrophilicity is another important membrane surface property that significantly affects the antifouling performance of a membrane.The effect of ZIF-67 filler nanoparticles on the surface hydrophilicity of the MMMs was studied by water contact angle measurements.The results presented in Figure 9 show that the water contact angle slightly increased from 75°(SPSf) to 86°(SPSf-Z67-0.5)upon increasing the hydrophobic ZIF-67 content, but all water contact angles were still lower than 90°,  indicating that the membranes were hydrophilic.This resulting change in the antifouling performance of the membrane will be discussed in the following section.

Membrane morphology
The morphology of the pristine SPSf and SPSf MMMs containing various concentrations of ZIF-67 nanoparticles was characterized by SEM (Figure 10), and the average surface pore size and surface porosity are presented in Table 1.All MMMs showed larger average surface pore sizes and higher porosities than the SPSf UF membrane.The SPSf-Z67 MMMs had surface porosities in the range of 3.38-25.7%,which were much higher than that of the pristine SPSf membrane, and the porosity of SPSf-Z67-0.5 was 7.6 times higher than that of the pristine SPSf membrane.The average surface pore size of the MMMs  ranged from 8.4-23.5 nm, and the average surface pore size of SPSf-Z67-0.5 was 2.5 times higher than the pristine SPSf membrane (9.5 nm).Therefore, filling the SPSf with ZIF-67 nanoparticles promoted the formation of membrane surface pores during the NIPS process.Moreover, all prepared membranes exhibited a dense and homogeneous cross-section morphology, indicating that ZIF-67 presents excellent dispersibility and compatibility with the SPSf matrix (Liu et al. 2018;Liu et al. 2019).
The EDX mapping was also performed to confirm the dispersion of the ZIF-67 nanoparticles in the surface and cross-section of the SPSf matrix.As shown in Figure 11, ZIF-67 was homogeneously distributed in the membrane matrix at lower concentrations from 0.1% to 0.5%.ZIF-67 nanoparticles were mainly distributed on the cross-section of the membranes rather than the surface during the membrane formation process.When enlarged, it worth noting that the morphology of the cross-section sublayer with more micropores after increasing the ZIF-67 concentration.
The larger average surface pore sizes, higher porosities, and micropore cross-sections of the MMMs compared with pristine SPSf can be explained by the presence of ZIF-67 and the membrane formation mechanism.The presence of ZIF-67 in the polymeric solution increased the number of lean-polymer phase zones and increased the rate of mass transfer between solvent and non-solvent, which promoted pore formation (Safarpour et al. 2015;Sotto et al. 2015;Karimi et al. 2019).On the other hand, during the NIPS process, the hydrophilic SPSf liquid film required more water for phase separation, which slowed the solidification rate and formed small surface pores and a low surface porosity (Widjojo et al. 2013).After filling with ZIF-67, the acid-base crosslinking between SPSf and ZIF-67 consumed some sulfonic acid groups, which may have promoted liquid-liquid phase separation and solidification (Wang & Lai 2013).Hydrophobic ZIF-67 reduced the interactions between NMP and the polymer, which significantly promoted the formation of large cavities (Tasselli et al. 2013).

Membrane ultrafiltration performance
The effect of ZIF-67 nanocrystal addition on the permeation and separation performance of the SPSf UF membranes was investigated by ultrafiltration experiments.As Figure 12 shows, increasing the ZIF-67 concentration in the casting solution significantly increased the pure water flux of all SPSf-Z67 mixed membranes to 57-111 L m À2 h À1 , which was up to 2 times higher than that of the SPSf membrane (55 L m À2 h À1 ).This was ascribed to a lower permeation resistance of water through the membranes due to a higher surface porosity and because ZIF-67 provided extra passageways for water (Yang et al. 2019a(Yang et al. , 2019b)). Figure 12 also displayed the BSA rejection rate of all membranes.The BSA rejection rate of SPSf-Z67-0.1, SPSf-Z67-0.3 and SPSf-Z67-0.5 were 93.9% + 2.2, 95.3% + 1.6, 95.8% + 0.1, respectively.These values are slightly higher than that of the pristine SPSf membrane having a BSA rejection of 93.2% + 1.5.These results indicated that ZIF-67 filled the SPSf matrix membrane and simultaneously enhanced the permeability and selectivity of the membranes, which overcame the common trade-off of polymer membranes.Due to the presence of ZIF-67 nanocrystals with a smaller pore size in the matrix membrane, ZIF-67 acts as a sieve that allows water to pass while blocking the passage of macromolecular organic pollutants (BSA) due to a size exclusion effect.

Membrane antifouling performance
The antifouling performance of SPSf and SPSf-Z67 MMMs were investigated by using them to filter a BSA solution.Figure 13 shows the flux changes of the membrane after three filtration cycles.When filtering the BSA solution, the MMMs flux decreased quickly due to the adsorption of BSA and pore blockage on the membrane surface.For example, the flux of the membrane loaded with 0.5 wt% ZIF-67 dropped from 111 L m À2 h À1 for pure water to 59 L m À2 h À1 for BSA at the first cycle.The bare SPSf membrane's flux decreased from 55 L m À2 h À1 to 52 L m À2 h À1 .These values indicate that the SPSf membrane has better fouling resistance.
FRR was used to represent the membrane antifouling ability because it reflects the water washing efficiency after three filtration cycles (Malaisamy et al. 2002;Ghiasi et al. 2019).Figure 14 shows that the FRR of pristine SPSf membrane, SPSf-Z67-0.1, SPSf-Z67-0.3, and SPSf-Z67-0.5 blend membranes after three BSA solution filtration cycles were 88.5%, 77.3%, 69.8%, and 47.7%, respectively.These results are consistent with the three BSA fouling cycles shown in Figure 13.The decrease in the FRR values of the MMMs could be ascribed to three factors: hydrophilicity, roughness, and pore size.By embedding the hydrophobic ZIF-67 in the SPSf matrix, the blended membranes have a larger contact angle, rougher surface, and larger surface pore size.Due to the hydrophobic interactions with BSA and the ridge-valley morphology of the blended membrane's surface, BSA was prone to accumulation and deposition on the surface to form a filter cake layer (Alizadeh Tabatabai et al. 2014;Nidhi Maalige et al. 2019;Zhang et al. 2019b).On the other hand, the larger pore size of the blend membranes allowed BSA to more easily enter and block the pores of the membrane to cause irreversible fouling (Polyakov & Zydney 2013;Zhu et al. 2020), reducing the water flux, which decreased the FRR.

CONCLUSIONS
In this study, using ZIF-67 as a basic additive to constructed an acid-base interaction between the sulfonic acid groups of SPSf and imidazole ligands in ZIF-67.On one hand, improving the dispersibility and compatibility of ZIF-67 with SPSf membrane.Others, promoting the formation of larger average surface pore sizes and higher porosities on the SPSf matrix membrane's surface to improve the permeability of blended membranes.Besides, the small pore size distribution of ZIF-67 could reject BSA organic molecules to simultaneously improve the permeability and selectivity of SPSf mixed-matrix membranes.However, it's worth noting that FRR values of SPSf-Z67 blend membranes were lower than that of the pristine SPSf membrane because the addition of hydrophobic ZIF-67 reduced the surface hydrophilicity and antifouling ability of membrane.Among the various blended membranes, the MMM with 0.3 wt% ZIF-67 displayed the best performance, with an initial

Figure 1 |
Figure 1 | Schematic illustration of the acid-base crosslinking structure formed between SPSf and ZIF-67.

Figure 8
Figure 8 | The 2D and 3D AFM images of UF membrane surface.

Figure 9 |
Figure 9 | Water contact angles of UF membranes.

Figure 10 |
Figure 10 | SEM images of membrane top surface and cross-section.

Figure 11 |
Figure 11 | EDS mapping (Co) of the surface (a) and cross-section (b) of membranes.

Figure 13 |
Figure 13 | Time-dependent flux change using BSA as a pollutant at 1 bar.

Figure 12 |
Figure 12 | Pure water flux and BSA rejection of UF membrane.

Table 1 |
The surface porosity (P s ), and average surface pore size (r s )