The impact of ZIF-8 particle size and heat treatment on CO₂/CH₄ separation using asymmetric mixed matrix membrane

Abstract

In this study, zeolitic imidazole framework 8 (ZIF-8) particles of different sizes were synthesized in aqueous media by varying the concentration of the base-type additive, triethylamine (TEA). ZIF-8 with particle sizes of ∼134 nm and ∼288 nm with surface areas of 418.44 m² g⁻¹ and 491.54 m² g⁻¹ were obtained without altering the crystalline structure. Synthesized ZIF-8 was further heat treated at 100 °C for a minimum of 12 hours, which led to an enhancement of its phase crystallinity and a surface area of 981.1 m² g⁻¹. Mixed matrix membranes (MMMs) were prepared via the dry–wet phase inversion method by dispersing as-synthesized ZIF-8s, heat-treated ZIF-8s and commercial ZIF-8 (∼493 nm) into a polysulfone (PSf) matrix. The thermal stability and mechanical strength of the membranes showed significant improvement after the incorporation of ZIF-8s. The MMMs were further subjected to the permeation experiments of CO₂ and CH₄. Although the majority of MMMs showed less selectivity than the neat PSf membrane, the incorporation of heat-treated ZIF-8 of the smallest size, exhibited CO₂/CH₄ selectivity of 28.5, which is significantly higher than the 19.43 obtained for the neat PSf membrane. Therefore, different ZIF-8 treatment protocols and particle sizes affect the MMMs performance significantly.

---

1 Introduction

The depletion of natural gas source as well as low quality gas wells accompanied by the wide range of impurities is an emerging concern. The reported impurity composition in natural gas wells is up to 80%,¹ making efficient natural gas purification crucial. Among the impurities, the removal of carbon dioxide (CO₂) is regarded as the most vital process due to its corrosive nature to the pipelines and that its presence can decrease the calorific value. Failure to remove CO₂ effectively would directly reflect on the processing cost. Conventional CO₂ capturing techniques, such as chemical-based absorption and cryogenic separation, have been implemented because they are reliable, operational on a large scale and adaptable for wide applications.^2–4 The conventional processes, however, still suffer from high solvent losses, high energy consumption and large space occupation.^2,4,5

Polymeric membranes have emerged as a potentially superior separation process over conventional techniques. Owing to its ability to discriminate gaseous species based on the molecular size, low capital cost, modest energy requirement and ease of fabrication, research on polymeric membranes has gained a lot of attention over the last few decades.^6–11 Despite its advantages, Robeson's upper bound between selectivity and permeability has limited their applications.¹² These challenges have encouraged researchers to find new classes of membrane materials. Inorganic materials, such as mesoporous silica, zeolite, carbon molecular sieves and ceramics, have emerged as promising membrane materials because they offer superior separation properties over polymeric membranes. Compared to polymeric membranes, the mechanical and thermal stability of inorganic membranes allow them to operate at high temperatures and pressures without deterioration of their structures.^13–17 Nevertheless, fabricating a defect free inorganic membrane remains a challenging task owing to its brittleness, high temperature and longtime requirement for synthesis, which has hindered its widespread applications. Hence, recent membrane development has focused on the mixed matrix membrane (MMM), a new class of membranes in which inorganic particles are incorporated into the polymer matrix. A MMM provides enhancement in chemical and thermal stability, mechanical strength, permeability and selectivity, relative to polymeric membranes. Over the years, inorganic fillers, such as zeolite, carbon molecular, silica, and metal organic framework (MOF) as the dispersed phase in MMM have been demonstrated to boost membrane performance.^18,19

Developing defect free MMM is a challenging task because MMM fabrication commonly suffers from polymer-filler incompatibility. The formation of non-selective voids, sieve-in-cage morphology, polymer rigidification, and pore blockage are the results of poor polymer-filler interactions, which leads to a deterioration of the membrane performance. Thus, the selection criteria for polymeric and inorganic filer are crucial for MMM development. Compared to other classes of fillers, MOFs demonstrate good interactions with the polymer because of its organic linkers, and have emerged as potential fillers for MMM.^20,21 MOFs also possess large surface areas, high adsorption capacity, and are easy to modify. In addition, the high affinity towards certain gases gives the edge for MOFs to be implemented as fillers.^22,23 Among the MOFs, zeolitic imidazole framework-8 (ZIF-8) is one of the most investigated MOFs. ZIF-8 has a porous crystalline structure with a M–Im–M angle (M = metal) close to 145°, which is coincident with the Si–O–Si angle found in many zeolites with a β-cage opening of 11.6Å and a 6-ring window aperture of 3.4 Å. ZIF-8 shows good chemical stability against polar and nonpolar solvents,²⁴ reorientation of its structure at high pressures²⁵ and high mechanical strength.²⁶ Other spectacular properties of ZIF-8 are the controllability of its crystal size. The crystal size of ZIF-8 is controlled by the synthesis temperature, mixing rate, type of solvent, metal salt–ligands–solvent ratio, and base-type additive.^21,27–31 The addition of base-type additives into the system is a common approach to achieve nano-sized MOF. Base-type additives, such as triethylamine (TEA), would increase the nucleation rate by deprotonation of the organic ligands, which provide more reactive sites and act as buffering ligands for easy release metal sites. As a result, a nanoscale particle size can be obtained through this approach.^32–34

The influence of the filler size on MMM has a significant impact on the membrane performance. A smaller particle size would enhance the polymer-filler interface and provide more free volume through better disruption of the polymer chain. Hashemifard et al.³⁵ showed that significant enhancement of the membrane separation properties can be achieved, even at low loadings. In the study, 0.5 wt% (total solid) of clay mineral (Cloisite 15A) showed CO₂ permeance and a 24% and 28% increase in CO₂/CH₄ selectivity, respectively, due to the combination of polymer chain disruption and the tortuous path for CH₄ to permeate. In other words, layered clay minerals are exfoliated into nano layers,³⁶ providing higher external surface area to volume ratios and an ease to interact with the polymer chain, which lead to permeability and selectivity enhancement.^20,37–40 Therefore, significant improvement in the membrane separation properties can be achieved using a nano filler, even at low filler loadings.

The influence of the particle size of fillers is an important factor in the development of defect-free MMM. Therefore, this study aims to develop defect-free MMMs using fillers with different particle sizes. Asymmetric MMMs were synthesized using the dry–wet inversion method by loading ZIF-8 fillers with different sizes with and without heat treatment. The effect of the particle size on the gas separation performance and membrane characteristics was evaluated. The influence of the particle size on the thermal stability, glass transition temperature (T_g) and mechanical strength of the prepared membranes was also investigated.

2 Experiment

2.1 Materials

Zinc nitrate hexahydrate (Zn(NO₃)₂·6H₂O) was purchased from Alfa Aesar Chemicals. 2-Methylimidazole (2-MeIM) and triethylamine (TEA) were obtained from Sigma Aldrich. Polysulfone (PSf Udel® P-1700) was supplied by Solvay Plastic. N,N-Methylpyrrolidone (NMP) and tetrahydrofuran (THF) were procured from Merck. For the coating solution, polydimethylsiloxane (PDMS) was purchased from Sigma Aldrich and n-hexane was purchased from Merck. All chemicals were used without further purification.

2.2 ZIF-8 synthesis

ZIF-8 was prepared using following the same procedure described elsewhere⁴¹ with a Zn(NO₃)₂ : 2-MeIM : H₂O ratio of 1 : 6 : 500. Briefly, a metal salt solution was prepared by dissolving 2 g of Zn(NO₃)₂ (6.72 mmol) in 12.11 g of deionized water (20 % of total deionized water). For the ligand solution, 2-MeIM (3.312 g, 40.34 mmol) was dissolved in 48.45 g of deionized water before adding the different TEA concentrations (2.0 mL for Z300 and 3.0 mL for Z100). The metal salt solution was added to the ligand solution, resulting in a cloudy solution mixture. The solution was stirred vigorously for 30 minutes followed by centrifugation of the reaction product, which was further washed several times with deionized water to remove the excess reactants and subsequently dried in an oven at 60 °C for 12 hours. The product powder was collected and weighed to calculate the mass yield. The product was called as Z100 and Z300 according to the particle size (Z100, ∼100 nm and Z300, ∼300 nm). The powder was further heat treated in an oven at 100 °C for 12 h to remove the guest molecules in the ZIF-8 pores. The heat treated ZIF-8s are called Z100a and Z300a. Commercial ZIF-8 (Basolite® Z1200) with a particle size of ∼500 nm was used for comparison and is denoted as Z500.

2.3 Asymmetric Flat Sheet Membrane Preparation

An asymmetric flat sheet MMM was prepared from the solution consisting of PSf (25 wt%), NMP (60 wt%), THF (15 wt%) and ZIF-8 (5 wt% of the solid). ZIF-8 was dispersed into a NMP/THF solvent before adding approximately 10% of polymer with stirring for the priming purpose. The remaining polymer was gradually added and the mixture was kept stirred until the solution became homogeneous. The casting dope was then hand-cast using a casting bar to a thickness of 120–190 μm. After standing on a glass plate for approximately 3 min, the cast film together with the glass plate was immersed in a coagulation bath (water at 27 °C), where the membrane was solidified. The membrane was then transferred to fresh water and kept there for 1 day to remove the residual solvent completely. Finally, the membrane was solvent-exchanged by immersing progressively in methanol and n-hexane, each for 2 h, before being air-dried for 24 hours in the ambient atmosphere. The pristine PSf membrane was prepared using the same process without adding ZIF-8.

The surface of the membrane was then brought into contact with 3 wt% PDMS/n-hexane solution for 10 min to seal possible pinholes before they were “cured” at 60 °C overnight. MMM for Z500, Z500a, Z300, Z300a, Z100 and Z100a were denoted as M500, M500a, M300, M300a, M100 and M100a, respectively.

2.4 Characterizations

Differential scanning calorimetry (DSC, Mettler Toledo DSC 822e) was used to determine the glass transition temperature (T_g) of the prepared membranes. The membrane sample was cut into small pieces, weighed and placed into a pre-weighed aluminum crucible. The sample was then heated from 50 to 400 °C at a heating rate of 10 °C min⁻¹ in the first cycle to remove the thermal history and cooled from 400 to 30 °C at a rate of 10 °C min⁻¹. The same heating protocol was repeated in the next heating cycle. T_g of the sample was determined as the midpoint temperature of the transition region in the second heating cycle.

Thermogravimetric analysis (TGA) was used to characterize the thermal stability of prepared ZIF-8 and MMM samples. TGA records the weight changes in the sample when heated continuously. The samples were heated from 50 to 900 °C at a heating rate of 10 °C min⁻¹ under a nitrogen atmosphere with a nitrogen flow rate of 20 mL min⁻¹.

Single point N₂ adsorption was undertaken at 130 °C for 1 h. The apparent surface area of the ZIF-8 particles was calculated using the BET equation.

X-ray diffraction (XRD) was carried out using a Siemens D5000 Diffractometer to measure the crystal size of the ZIF-8 samples and identify their structure. The crystal sizes of the prepared ZIF-8s were estimated using the Debye–Scherrer equation:

where d represents the crystal size (Å), k is a constant of 0.9, λ is the X-ray wavelength of CuKα radiation (1.54 Å), B is the full width at half maximum (FWHM) of the peak, and θ is the Bragg's angle in degrees.

The tensile strength and elongation at break of the membranes were evaluated according to the ASTM D3039 standard using an LRX 2.5 SKN Lloyd Instrument. The membranes were cut into 5 cm long and placed in the equipment.

Transmission electron microscopy (TEM) (JEOL, JSM-6701FJEOL 1230) was used to observe the microstructures of the ZIF-8. The samples were prepared by dispersing ZIF-8 powder into methanol. A drop of methanol was used to disperse the ZIF-8 onto carbon-coated copper grids, which were observed by TEM operating at 300 kV. The particle sizes of the prepared ZIF-8s were measured using at least 50 particles from the TEM images.

Scanning electron microscopy (SEM) was used to observe the membrane structure and morphology. Membrane samples were fractured cryogenically in liquid nitrogen. The samples were coated with gold before being imaged and photographed by a scanning electron microscope (TM3000, Hitachi) equipped with an energy dispersive X-ray spectrometer (EDX) (XFlash® 430H Detector, Bruker).

2.5 Gas Permeation

Gas permeation tests were performed using a constant pressure variable volume system described elsewhere.⁴² The membranes with an effective permeation area of 13.5 cm² were placed into the permeation cell and exposed to pure CH₄ or CO₂ gas. The feed pressure was controlled to 4 bar and the temperature was 27 °C. The pressure-normalized flux (permeance) of gas i was calculated as follows:where i represents the gas penetrant, V_i is the volume of gas permeated through the membrane (cm³, STP), A is the effective membrane area (cm²), t is the permeation time (s), and Δp is the transmembrane pressure drop (cmHg). The permeance is expressed as gas permeation units, GPU, as follows:

1 GPU = 1 × 10⁻⁶ cm³ (STP) cm⁻² s⁻¹ cmHg⁻¹.

The selectivity was obtained using Eq. (3):

where α_i/j is the selectivity of gas penetrant i over gas penetrant j, and P_i/l and P_j/lare the permeance of gas penetrant i and j, respectively.

3 Results and Discussion

3.1 ZIF-8 Characterizations

3.1.1 XRD. The synthesized ZIF-8s showed high mass yield of ∼90% (calculated based on the zinc salt mass), which is in good agreement with the previous reported.³⁰ Fig. 1a shows the XRD pattern of the as-synthesized Z100 and Z300, which was compared with the literature³⁴ to identify the crystal structure of the samples. Both prepared samples show strong peaks at 2θ = 7.30°, 10.35°, 12.70°, 14.80°, 16.40° and 18.00° for the {110}, {200}, {211}, {220}, {310}, and {222} planes, respectively. Overall, the consistency of the XRD pattern for the prepared ZIF-8s indicates that pure ZIF-8s were formed without the influence of TEA on its crystallinity. Our previous work, which focused on the formation of ZIF-8 at different TEA loadings, showed that an insufficient TEA loading into the low ligands–metal salts molar ratio would hinder ZIF-8 formation and excess loading does not alter the crystallinity of ZIF-8 because it achieves stable crystal form.⁴¹ Hence, the role of TEA addition to the synthesis solution is to assist the formation of the ZIF-8 crystal by inducing a deprotonate state of the 2-MeIM and altering the particle size.

Fig. 1 XRD pattern for ZIF-8s prepared (a) at various TEA loadings and (b) after heat treatment.

After heat treatment at 100 °C overnight, no significant changes in the planes for heat-treated Z100a, Z300a and Z500a from those of Z100, Z300 and Z500 were observed (Fig. 1b), which indicates that the structure was not changed by heat treatment. However, intensity of XRD peaks increased from Z100 and Z300 to Z100a and Z300a, respectively. This is due to the removal of guest molecules by heat treatment, which caused destructive interference of the XRD pattern.²¹ Interestingly, compared to Z500, the peak intensities of Z500a decreased significantly after heat treatment, which indicates a decrease in crystallinity. This has never been reported in the literature. The different behavior of Z500 from the prepared ZIF-8s is related to different preparation procedures; i.e., ZIF-8 synthesized in an aqueous environment exhibits higher tolerance towards elevated temperatures.³⁰

The crystal size of the ZIF-8 samples was calculated using eqn (1) and the results are listed in Table 1. The table shows that the crystal size decreases with increasing TEA loading. Because nano particles tend to agglomerate and form larger particles, the value obtained by XRD does not necessarily represent the actual particle sizes. Therefore, the ZIF-8s were further observed by TEM. Fig. 2 shows TEM images of Z100, Z300 and Z500. All samples possessed a rhombic dodecahedron morphology, showing good agreement with the literature.⁴³ The particle size of the prepared samples was estimated from the TEM images and tabulated in Table 1. A comparison of Z100 with Z300 showed that the particle size decreased significantly with increasing TEA loading. The presence of TEA eases the deprotonation of 2-MeIM, thus providing more sites to react with Zn²⁺. Consequently, nucleation is induced and crystallization between Zn²⁺ and MeIM⁻ is facilitated, which enables the formation of ZIF-8 at relatively low reactant concentrations. Furthermore, a high TEA loading leads to the rapid crystallization of ZIF-8 to produce smaller particle size compared to the low TEA loading.^28,44 The trend of a decrease in particle size with increasing TEA loading coincides with the decrease in crystal size calculated using eqn (1).

Table 1 Properties of the prepared ZIF-8s and Basolite® Z1200 (Z500)

Samples	TEA loading (mL)	Particle size (nm)	Scherrer crystal size estimation (nm)	BET surface area (m^2 g^−1)
				Before heat-treatment	After heat-treatment
a Z500 represented is commercially procured ZIF-8 and the TEA loading is not reported.
Z100	3.0	∼134	69.77	418.44	981.10
Z300	2.0	∼288	82.74	491.54	566.67
Z500	NA^a	∼493	90.05	1223.19	990.91

---

Fig. 2 TEM images of (a) Z100 (b), Z300, and (c) Z500.

3.1.2 Surface properties. The surface properties of Z100, Z300 and Z500 before and after heat treatment are presented in Table 1. The surface areas of the as-synthesized ZIF-8s (<500 m² g⁻¹) are significantly lower than the values previously reported. The low surface areas observed might be due to guest molecule entrapment in the ZIF-8 pores, which reduced the BET surface area.⁴¹ The samples were subjected to heat treatment to evacuate the guest molecule. The high surface area of the heat-treated ZIF-8s indicates improved the pore availability after removal of the guest molecules. The surface area of Z100 shows significant improvement from 418.44 m² g⁻¹ to 981.10 m² g⁻¹, whereas Z300 showed slight variations from 491.54 m² g⁻¹ to 566.67 m² g⁻¹ after heat-treatment. In our previous work, ZIF-8 was treated at 300 °C to evacuate the guest molecule, which resulted in an increase in BET surface to 1182.87m² g⁻¹.⁴¹ The relative difference between the BET surface area of the as-synthesized ZIF-8 with the sample treated at 100 °C and 300 °C was relatively small, suggesting that a treatment at 100 °C is adequate to improve the surface area. The different BET enhancements are related to the particle sizes of ZIF-8, i.e., a smaller particle (Z100) would result in greater enhancement compared to the larger particle. Even so, both Z100 and Z300 have lower BET surface areas than commercial ZIF-8 (Z500), even though its particle size is smaller. This phenomenon is commonly attributed to the different preparation history, i.e. ZIF-8 prepared in room temperature media has been reported to possess lower surface areas than the ZIF prepared in a thermally-induced synthesis media.^27,30,45,46 However, the mechanism involved regarding this phenomenon remains an open question.

3.1.3 Thermal properties. The TGA curves of the ZIF-8s prepared under different conditions are presented in Fig. 3. Minimal first weight loss occurred at 100 °C for the as-synthesized ZIF-8s due to the evaporation of a trapped solvent. The second weight loss began at approximately 230 °C, which can be associated with the carbonization of unreacted 2-MeIM,²⁵ while the third weight loss at 580 °C resulted from the decomposition of organic linkers and ZIF-8 structure. Less weight loss was observed after heat treatment throughout the entire temperature range, suggesting the evacuation of guest molecules during heat treatment, which coincides with the XRD pattern (Fig. 1) and BET analysis (Table 1).

Fig. 3 TGA curve of the as-synthesized and heat-treated ZIF-8.

3.2 Membrane characterization

The TGA curves of the prepared membranes showed two step weight loss (Fig. 4). The first weight loss at 100 °C was attributed to the evaporation of residual water trapped in the membrane during wet inversion. The weight loss of MMM was less than that of neat PSf because hydrophobic ZIF-8 reduces the amount of trapped water. The second weight loss at approximately 500 °C was attributed to degradation of the polymer matrix. The lack of significant differences between MMMs indicate that the ZIF-8 particle size and pore occupancy do not affect the thermal stability. Unlike the ZIF-8 samples, no weight loss occurred at 230 °C, which suggests that (1) guest molecules were flushed out when the ZIF-8 particles were dispersed in NMP and further removed during wet phase inversion, and/or (2) only a small amount of ZIF-8 (5 wt% of total solids) was incorporated into the PSf matrix.

Fig. 4 TGA curves of the prepared membranes.

The glass transition temperatures (T_g) of the prepared membranes are given in Fig. 5. The T_g of the neat PSf membrane was lower than the value previously reported⁴⁷ due to the different membrane preparation procedure. Incorporating as-synthesized ZIF-8 has led to a significant decrease in T_g, indicating an increase in polymer chain flexibility caused by poor polymer-filler interaction, especially for smaller particles. In contrast, the incorporation of heat-treated ZIF-8 increased T_g, suggesting that good interaction between ZIF-8 and PSf matrix has restricted polymer chain mobility. In addition, the increase in T_g with increasing particle size suggests that the good dispersion of smaller particles with a good filler-polymer interaction leads to less restriction of the polymer chain mobility.

Fig. 5 Influence of the different filler properties on the glass transition temperature (T_g) of the prepared membranes.

Fig. 6 shows the mechanical properties of the prepared membranes. The incorporation of untreated ZIF-8 into the PSf matrix decreased the tensile strength of MMMs, indicating poor compatibility between the filler and polymer matrix. Moreover, the smaller particle size (Z100) provided a large contact area to the polymer matrix; thus a severe decrease in tensile strength compared to the larger particles (Z300, Z500) was observed. In contrast, a slight increase in tensile strength was observed for the heat-treated ZIF-8s, indicating good filler-polymer interaction. The occupancy of ZIF-8 pores by the guest molecules has limited the interaction of particles with the polymer matrix, while the evacuation of pores via heat treatment resulted in a good interaction. On the other hand, the elongation at break of all prepared MMMs is significantly lower than the neat PSf membrane, indicating that the incorporation of ZIFs increases the membrane rigidity, regardless of the ZIF-8s treatment protocol.

Fig. 6 Mechanical stability of the prepared membranes. (a) Tensile strength (MPa) and (b) elongation at break (%).

SEM images of the as-synthesized and heat-treated ZIF-8 MMMs are presented in Fig. 7 and 8, respectively. The cross-section of the neat PSf membrane shows an asymmetric structure, with an apparent active dense layer on the top of a sponge-like sub-layer (Fig. 7a). The active layer formed by dry phase inversion has delayed the solvent exchange between NMP and water during the wet phase inversion, resulting in a sponge-like sub-layer. A similar cross-sectional morphology was also observed for all prepared MMMs, regardless of the particle size and heat treatment because the same membrane preparation protocol was used. However, the as-synthesized ZIF-8 severely agglomerated and large filler clusters were observed, as shown in Fig. 7b–d. EDX analysis of the both neat PSf and MMM confirmed that the clusters were indeed ZIF-8 particles (Fig. 9). Cluster formation is less apparent for the heat-treated ZIF-8 MMM (Fig. 8a and b), indicating well-dispersed ZIF-8 particles throughout the PSf matrix. On the other hand, commercial Basolite® Z1200 (M500a) exhibited partial agglomeration across the membrane cross-section even after heat-treatment (Fig. 8c).

Fig. 7 Cross-section morphology of the as-synthesized ZIF-8 MMM, with (a) Neat PSf, (b) M100, (c) M300, and (d) M500 (Red circle represent the presence of ZIF-8 cluster).

Fig. 8 Cross-section morphology of the heat-treated ZIF-8 MMM, with (a) M100a, (b) M300a, and (c) M500a (Red circle represent the presence of ZIF-8 cluster).

Fig. 9 EDX analysis of cross-section morphology of (a) neat PSf membrane, and (b) M100a.

3.3 Gas separation performance

The gas permeation properties of all samples are presented in Table 2. The performance of the neat membrane was lower than the value previously reported^20,42,48 due to the different processing and preparation procedures. The incorporation of 5 wt% ZIF-8s with different particle sizes and pore occupancy had a significant impact on the performance of MMM. The literatures suggests that the ZIF-8 incorporated polymer matrix enhances gas permeance due to the porous nature of ZIF-8, which provokes faster gas diffusivity compared to the non-porous filler through the porous networks. ^49–51 In contrast, the decrease in CO₂ permeance was observed after the incorporation of ZIF-8 (except for M300a). Most notably, the permeance of CO₂ decreased from 25.72 GPU of neat PSf to 15.43 GPU of M100. Similar observations were demonstrated by Ordoñez et al.³³ when the loading of nano-scale ZIF-8 was low. This suggests that an increase in the tortuous permeation path length induced by polymer-filler interaction overwhelmed the enhancement of the polymer chain flexibility, which led to a decrease in CO₂ permeance.

Table 2 CO₂ and CH₄ permeance for neat PSf and PSf/ZIF-8s membranes^a

Samples	Permeance (GPU)^a		CO2/CH4 Selectivity^b
	CH4	CO2
a Gas permeation was conducted at room temperature, 4 bar. 1 GPU = 1 × 10^−6 cm^3 cm^−2 s^−1 cmHg^−1.b The selectivity presented based on the average of the modules tested, not necessarily equates to the ratio of average permeance.
Neat PSf	1.33 ± 0.35	25.72 ± 6.34	19.43 ± 0.56
M100	1.14 ± 0.27	15.43 ± 1.82	13.52 ± 0.43
M100a	0.55 ± 0.07	15.60 ± 0.79	28.50 ± 1.42
M300	1.78 ± 0.46	19.35 ± 3.58	10.87 ± 1.38
M300a	4.46 ± 0.04	25.88 ± 1.64	5.8 ± 0.25
M500	3.72 ± 0.43	21.47 ± 8.43	5.77 ± 0.45
M500a	4.82 ± 0.41	28.05 ± 3.62	5.82 ± 2.06

---

It was reported that utilizing ZIF-8 as a filler would be beneficial towards the CO₂/CH₄ selectivity due to the interaction between the quadrupole moment of CO₂ and the electrostatic field of ZIF-8, while no such interaction is possible for CH₄.^22,52,53 In contrast, a significant drop of CO₂/CH₄ selectivity was evident after the embodiment of ZIF-8s into the PSf matrix. The unforeseen deterioration of the CO₂/CH₄ selectivity was the result of a significant increase in CH₄ permeance (M300a, M500), or significant decrease in CO₂ permeance (M100, M300). These phenomena are related to a deterioration of the membrane selective layer induced by the incorporated ZIF-8, as will be discussed in a later section.

3.3.1 Influence of ZIF-8 particle size on gas separation. The influence of different ZIF-8 particle sizes on the CO₂ permeance is presented in Fig. 10a. As discussed earlier, incorporating the ZIF-8 into the polymer matrix resulted in a significant drop in CO₂ permeance for most prepared membranes. Most notably, the 100 nm ZIF-8 loaded MMMs exhibited a 40% drop in CO₂ permeance compared to the neat PSf membrane. The CO₂ permeance steadily increases with increasing particle size and eventually M500 surpassed that of the neat PSf matrix. This trend could be attributed to the ZIF-8 dispersion and polymer chain disruption. A larger number of smaller particles in the polymer matrix compared to larger particles at the same mass load would provide more tortuous permeation paths.⁵⁴ Consequently, smaller particles would lower the mass transfer coefficient and hinder gas permeation across the membrane. In contrast, more polymer chains are disrupted by larger particles, creating larger free volumes compared to the smaller particles. Gas diffusion increases due to the less hindrance to permeation, resulting in an increase in the permeance of both CO₂ and CH₄. A similar observation was reported for different particle sizes of zeolite and fumed silica.^37,54

Fig. 10 Contribution of the different ZIF-8 particle sizes on the membrane performance (a) relative CO₂ permeance, and (b) relative CO₂/CH₄ selectivity.

Despite the restriction on the polymer chain mobility with dispersed ZIF-8, the prepared MMMs suffered from low CO₂/CH₄ selectivity because larger ZIF-8 particles was used. Fig. 10b shows that the relative CO₂/CH₄ selectivity systematically worsens as larger ZIF-8 particles are incorporated. Eventually, the selectivity of M500a becomes 70% less than the neat PSf. This is because larger particles present in the membrane provide fewer particles and limit the contact area between the particle and polymer matrix. In contrast, smaller particles would provide a large number of active sites for gas/filler particle interaction. As a result, restricted polymer chain mobility of 100 nm ZIF-8 provides better CO₂/CH₄ selectivity than the 300 nm and 500 nm ZIF-8.

3.3.2 Influence of heat treated ZIF-8 on gas separation. Overall, the CO₂ permeance increased after incorporating larger ZIF-8 particles. In addition, the permeance increased after heat-treatment, i.e. the permeance of M100a, M300a and M500a were 1.74%, 33.73% and 30.63% higher than M100, M300 and M500, respectively. Heat-treated ZIF-8s possess larger surface areas with more exposed metal sites to promote higher CO₂ uptake,⁵⁵ providing more active sites for gas to pass through. NMP occupation after solvent exchange between NMP–water during the ‘priming’ procedure is believed to have led to pore blockage of the as-synthesized ZIF-8 MMMs. This phenomenon is likely to occur because the β-cages of ZIF-8 are 11.6 Å and NMP can migrate easily into the pores. Hence NMP may cause pore blockage. On the other hand, there was no evidence of the residual NMP according to the TGA results. The selectivity of the as-synthesized ZIF-8 MMMs was less than the neat PSf membrane without exception. This is because there were fewer active sites available for CO₂ passage due to severe particle agglomeration (Fig. 7).

It should be noted that, although the majority of the prepared membranes demonstrated an increase in permeance and a decrease in CO₂/CH₄ selectivity relative to the neat PSf membrane, M100a performed distinctly well. The CH₄ and CO₂ permeances of M100a were 58% and 38%, respectively, which are less that the neat PSf membrane. As a result, M100a showed CO₂/CH₄ selectivity of 28.5, which is 47% higher than that of the neat PSf membrane. The pores of the 100 nm ZIF-8 were evacuated by heat treatment, leaving a surface area that enabled good interactions between the polymer and CO₂. The increased rigidity evidenced by high T_g and low elongation at break suggests that the chain mobility of M100a was restricted.⁵⁶ In addition, the pores of Z100a were partially blocked by the surrounding polymers. The penetration of larger molecules (CH₄ whose kinetic diameter is 3.8 Å) across the membrane was hindered by the partial pore blockage; thus the larger molecules are forced to travel a longer path. Hence, its permeance decreased much more than the smaller penetrant (CO₂ whose kinetic diameter is 3.3 Å).

3.3.3 Overall MMMs performance. The contributions of the filler embedded throughout polymer matrix to the gas separation performance are rationalized by the morphological diagram. According to Fig. 11 the MMM performance relative to that of the neat polymer membrane are classified as Case 1 (ideal morphology), Case 2 (unselective voids), Case 3 (polymer chain rigidification), and Case 4 (filler pore blockage). The “Ideal morphology” represents an increase in both permeance and selectivity. “Unselective voids” leads to increased permeance and diminished selectivity. “Filler pore blockage” leads to a decrease in both permeance and selectivity, whereas an increase in selectivity is generally caused by “polymer rigidification and/or partial pore blockage”. The performance of the MMMs prepared in this study are located schematically in Fig. 11. Overall, the prepared membranes underperformed compared to the neat PSf membrane, even though previous studies suggested that the incorporation of ZIF-8 into the polymer matrix would amplify both the permeance and selectivity.^22,49,57,58 The poor membrane performances are likely to be related to the ZIF-8 clusters observed at the porous substructure that could not be fully utilized to facilitate gas transport.

Fig. 11 Morphological diagram of PSf/ZIF-8.

The effect of the particle size on the membrane performance is less for the as-synthesized ZIF-8 MMMs than the heat-treated ZIF-8 MMMs. The permeance increased while the selectivity decreased from M100 to M500, but all M100, M300 and M500 belonged to Case 4 (pore blockage). Considering that the particle size increased with increasing ZIF-8 loading, the chain mobility increased with increasing particle size while the particles acted as impermeable fillers (Fig. 12a). Hence, the penetrant was forced to travel through the PSf matrix. In addition, the number of large particles is less than the smaller particles and the tortuous path length becomes shorter (Fig. 12d). This is why the CO₂/CH₄ selectivity decreases as the impermeable filler particle becomes larger.

Fig. 12 Illustration of gas permeation through (a) as-synthesized ZIF-8, (b) unselective voids, (c) heat-treated ZIF-8, and (d) different ZIF-8 particle size.

M300a and M500a belong to the Case 2 region (unselective voids). Remember that the pores of these membranes were evacuated by heat treatment, leaving large vacant voids. As a result, the penetrant preferred to travel through the voids (Fig. 12b), especially when the particle sizes were large, increasing the permeance while decreasing the selectivity. The filler of the smallest size (M100a) behaved differently. The small pore size allowed partial blockage of the pores, bringing M100a into the Case 3 region (rigidification/partial pore blockage). Thus, the permeance of M100a decreased while the selectivity increased (Fig. 12c). Similar effects have also been reported.^33,59

4 Conclusions

ZIF-8s of different particle sizes were synthesized successfully in aqueous media with TEA as an additive. The pore evacuation was successfully carried out by heat-treatment to improve the phase crystallinity and increase the surface area up to 981.1 m² g⁻¹. When the laboratory synthesized ZIF was further incorporated to PSf to fabricate MMMs, the change in particle size and the heat treatment of ZIF-8 had a significant impact on the MMM properties. The thermal stability and mechanical strength of the membranes showed significant improvement by the incorporation of ZIF-8s.

As for the effect of the particle size, incorporating smaller particles reduced the permeance and increases the selectivity. This is because a larger number of fillers interacted with the polymer matrix, forcing the penetrant molecules to travel through longer diffusion paths.

The presence of guest molecules in the pores of the as-synthesized ZIF-8 has limited its interaction towards the PSf matrix. Both the permeance and selectivity decreased by the incorporation of the as-synthesized ZIF-8 in MMM because it behaved as an impenetrable filler in the PSf matrix. Heat treatment of ZIF-8, however, showed an increase in permeance and a decrease in selectivity. This is because heat treatment evacuated the guest molecules from the pores, through which the penetrant molecules could pass. When the pore size is small, however, the pores were partially blocked, which hindered CH₄ permeation and enhanced CO₂/CH₄ selectivity to 28.

Acknowledgements

The authors gratefully acknowledge the Ministry of Higher Education (MOHE) for the scholarship and Long-Term Research Grant Scheme (LRGS) Program under Universiti Teknologi Malaysia with the grant number Q.J130000.2452.04H71.

References

1. T. Cnop, D. Dortmundt and M. Schott, UOP Continued Development of Gas Separation Membranes Tech Paper [Online], http://www.uop.com/?document=uop-continued-development-of-gas-separation-membranes-technical-paper%26download=1.
2. P. Chiesa, S. Campanari and G. Manzolini, CO₂ cryogenic separation from combined cycles integrated with molten carbonate fuel cells, Int. J. Hydrogen Energy, 2011, 36, 10355–10365.
3. D. Wappel, S. Joswig, A. A. Khan, K. H. Smith, S. E. Kentish, D. C. Shallcross and G. W. Stevens, The solubility of sulfur dioxide and carbon dioxide in an aqueous solution of potassium carbonate, Int. J. Greenhouse Gas Control, 2011, 5, 1454–1459.
4. U. E. Aronu, H. F. Svendsen and K. A. Hoff, Investigation of amine amino acid salts for carbon dioxide absorption, Int. J. Greenhouse Gas Control, 2010, 4, 771–775.
5. L. Zhao, E. Riensche, L. Blum and D. Stolten, How gas separation membrane competes with chemical absorption in postcombustion capture, Energy Procedia, 2011, 4, 629–636.
6. A. G. Wonders and D. R. Paul, Effect of CO₂ exposure history on sorption and transport in polycarbonate, J. Membr. Sci., 1979, 5, 63–75.
7. A. Y. Houde and S. A. Stern, Permeability of ethyl cellulose to light gases. Effect of ethoxy content, J. Membr. Sci., 1994, 92, 95–101.
8. A. Bos, I. G. M. Pünt, M. Wessling and H. Strathmann, CO₂-induced plasticization phenomena in glassy polymers, J. Membr. Sci., 1999, 155, 67–78.
9. C. Cao, Chemical cross-linking modification of 6FDA-2,6-DAT hollow fiber membranes for natural gas separation, J. Membr. Sci., 2003, 216, 257–268.
10. L. Shao, T. Chung, S. Goh and K. Pramoda, The effects of 1,3-cyclohexanebis(methylamine) modification on gas transport and plasticization resistance of polyimide membranes, J. Membr. Sci., 2005, 267, 78–89.
11. A. A. Shamsabadi, A. Kargari, M. B. Babaheidari and S. Laki, Separation of hydrogen from methane by asymmetric PEI membranes, J. Ind. Eng. Chem., 2013, 19, 1680–1688.
12. L. M. Robeson, The upper bound revisited, J. Membr. Sci., 2008, 320, 390–400.
13. R. S. A. de Lange, J. H. A. Hekkink, K. Keizer and A. J. Burggraaf, Permeation and separation studies on microporous sol–gel modified ceramic membranes, Microporous Mater., 1995, 4, 169–186.
14. C. Barboiu, B. Sala, S. Bec, S. Pavan, E. Petit, P. Colomban, J. Sanchez, S. de Perthuis and D. Hittner, Structural and mechanical characterizations of microporous silica–boron membranes for gas separation, J. Membr. Sci., 2009, 326, 514–525.
15. S. Battersby, T. Tasaki, S. Smart, B. Ladewig, S. Liu, M. C. Duke, V. Rudolph and J. C. Diniz da Costa, Performance of cobalt silica membranes in gas mixture separation, J. Membr. Sci., 2009, 329, 91–98.
16. M. Asaeda and S. Yamasaki, Separation of inorganic/organic gas mixtures by porous silica membranes, Sep. Purif. Technol., 2001, 25, 151–159.
17. K. Sato, K. Sugimoto, Y. Sekine, M. Takada, M. Matsukata and T. Nakane, Application of FAU-type zeolite membranes to vapor/gas separation under high pressure and high temperature up to 5 MPa and 180 °C, Microporous Mesoporous Mater., 2007, 101, 312–318.
18. D. Q. Vu, W. J. Koros and S. J. Miller, Effect of condensable impurity in CO₂/CH₄ gas feeds on performance of mixed matrix membranes using carbon molecular sieves, J. Membr. Sci., 2003, 221, 233–239.
19. Y. C. Hudiono, T. K. Carlisle, J. E. Bara, Y. Zhang, D. L. Gin and R. D. Noble, A three-component mixed-matrix membrane with enhanced CO₂ separation properties based on zeolites and ionic liquid materials, J. Membr. Sci., 2010, 350, 117–123.
20. B. Zornoza, A. Martinez-Joaristi, P. Serra-Crespo, C. Tellez, J. Coronas, J. Gascon and F. Kapteijn, Functionalized flexible MOFs as fillers in mixed matrix membranes for highly selective separation of CO₂ from CH₄ at elevated pressures, Chem. Commun., 2011, 47, 9522–9524.
21. E. V. Perez, K. J. Balkus, J. P. Ferraris and I. H. Musselman, Mixed-matrix membranes containing MOF-5 for gas separations, J. Membr. Sci., 2009, 328, 165–173.
22. S. Basu, A. Cano-Odena and I. F. J. Vankelecom, MOF-containing mixed-matrix membranes for CO₂/CH₄ and CO₂/N₂ binary gas mixture separations, Sep. Purif. Technol., 2011, 81, 31–40.
23. J.-R. Li, Y. Ma, M. C. McCarthy, J. Sculley, J. Yu, H.-K. Jeong, P. B. Balbuena and H.-C. Zhou, Carbon dioxide capture-related gas adsorption and separation in metal–organic frameworks, Coord. Chem. Rev., 2011, 255, 1791–1823.
24. K. S. Park, Z. Ni, A. P. Cote, J. Y. Choi, R. Huang, F. J. Uribe-Romo, H. K. Chae, M. O'Keeffe and O. M. Yaghi, Exceptional chemical and thermal stability of zeolitic imidazolate frameworks, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 10186–10191.
25. S. A. Moggach, T. D. Bennett and A. K. Cheetham, The effect of pressure on ZIF-8: increasing pore size with pressure and the formation of a high-pressure phase at 1.47 GPa, Angew. Chem., Int. Ed., 2009, 48, 7087–7089.
26. J. C. Tan, T. D. Bennett and A. K. Cheetham, Chemical structure, network topology, and porosity effects on the mechanical properties of Zeolitic Imidazolate Frameworks, Proc. Natl. Acad. Sci. U. S. A., 2010, 107, 9938–9943.
27. J. Cravillon, S. Münzer, S.-J. Lohmeier, A. Feldhoff, K. Huber and M. Wiebcke, Rapid Room-Temperature Synthesis and Characterization of Nanocrystals of a Prototypical Zeolitic Imidazolate Framework, Chem. Mater., 2009, 21, 1410–1412.
28. J. Cravillon, R. Nayuk, S. Springer, A. Feldhoff, K. Huber and M. Wiebcke, Controlling Zeolitic Imidazolate Framework Nano- and Microcrystal Formation: Insight into Crystal Growth by Time-Resolved In Situ Static Light Scattering, Chem. Mater., 2011, 23, 2130–2141.
29. Y. Pan, D. Heryadi, F. Zhou, L. Zhao, G. Lestari, H. Su and Z. Lai, Tuning the crystal morphology and size of zeolitic imidazolate framework-8 in aqueous solution by surfactants, CrystEngComm, 2011, 13, 6937.
30. Y. Pan, Y. Liu, G. Zeng, L. Zhao and Z. Lai, Rapid synthesis of zeolitic imidazolate framework-8 (ZIF-8) nanocrystals in an aqueous system, Chem Commun., 2011, 47, 2071–2073.
31. D. Yamamoto, T. Maki, S. Watanabe, H. Tanaka, M. T. Miyahara and K. Mae, Synthesis and adsorption properties of ZIF-8 nanoparticles using a micromixer, Chem. Eng. J., 2013, 227, 145–150.
32. O. M. Yaghi, C. E. Davis, G. Li and H. Li, Selective Guest Binding by Tailored Channels in a 3-D Porous Zinc(II)-Benzenetricarboxylate Network, J. Am. Chem. Soc., 1997, 119, 2861–2868.
33. M. J. C. Ordoñez, K. J. Balkus, J. P. Ferraris and I. H. Musselman, Molecular sieving realized with ZIF-8/Matrimid® mixed-matrix membranes, J. Membr. Sci., 2010, 361, 28–37.
34. A. F. Gross, E. Sherman and J. J. Vajo, Aqueous room temperature synthesis of cobalt and zinc sodalite zeolitic imidizolate frameworks, Dalton Trans., 2012, 41, 5458.
35. S. A. Hashemifard, A. F. Ismail and T. Matsuura, Effects of montmorillonite nano-clay fillers on PEI mixed matrix membrane for CO₂ removal, Chem. Eng. J., 2011, 170, 316–325.
36. H. R. Dennis, D. L. Hunter, D. Chang, S. Kim, J. L. White, J. W. Cho and D. R. Paul, Effect of melt processing conditions on the extent of exfoliation in organoclay-based nanocomposites, Polymer, 2001, 42, 9513–9522.
37. Z. He, I. Pinnau and A. Morisato, Nanostructured poly(4-methyl-2-pentyne)/silica hybrid membranes for gas separation, Desalination, 2002, 146, 11–15.
38. F. Li, Y. Li, T.-S. Chung and S. Kawi, Facilitated transport by hybrid POSS®–Matrimid®–Zn2+ nanocomposite membranes for the separation of natural gas, J. Membr. Sci., 2010, 356, 14–21.
39. Y. Li and T.-S. Chung, Molecular-level mixed matrix membranes comprising Pebax® and POSS for hydrogen purification via preferential CO₂ removal, Int. J. Hydrogen Energy, 2010, 35, 10560–10568.
40. S. Basu, A. Cano-Odena and I. F. J. Vankelecom, Asymmetric Matrimid®/[Cu₃(BTC)₂] mixed-matrix membranes for gas separations, J. Membr. Sci., 2010, 362, 478–487.
41. N. A. H. M. Nordin, A. F. Ismail, A. Mustafa, P. S. Goh, D. Rana and T. Matsuura, Aqueous room temperature synthesis of zeolitic imidazole framework 8 (ZIF-8) with various concentrations of triethylamine, RSC Adv., 2014, 4, 33292–33300.
42. A. F. Ismail and P. Y. Lai, Effects of phase inversion and rheological factors on formation of defect-free and ultrathin-skinned asymmetric polysulfone membranes for gas separation, Sep. Purif. Technol., 2003, 33, 127–143.
43. P. Y. Moh, Crystal Growth of Metal–Organic Framework ZIF-8, School of Chemistry, The University of Manchester, 2012.
44. J. Gascon, S. Aguado and F. Kapteijn, Manufacture of dense coatings of Cu₃(BTC)₂ (HKUST-1) on α-alumina, Microporous Mesoporous Mater., 2008, 113, 132–138.
45. Y.-S. Bae, O. K. Farha, A. M. Spokoyny, C. A. Mirkin, J. T. Hupp and R. Q. Snurr, Carborane-based metal–organic frameworks as highly selective sorbents for CO₂ over methane, Chem. Commun., 2008, 4135–4137.
46. H.-Y. Cho, D.-A. Yang, J. Kim, S.-Y. Jeong and W.-S. Ahn, CO₂ adsorption and catalytic application of Co-MOF-74 synthesized by microwave heating, Catal. Today, 2012, 185, 35–40.
47. G. C. Kapantaidakis, S. P. Kaldis, X. S. Dabou and G. P. Sakellaropoulos, Gas permeation through PSF-PI miscible blend membranes, J. Membr. Sci., 1996, 110, 239–247.
48. A. F. Ismail and W. Lorna, Suppression of plasticization in polysulfone membranes for gas separations by heat-treatment technique, Sep. Purif. Technol., 2003, 30, 37–46.
49. K. Díaz, M. López-González, L. F. del Castillo and E. Riande, Effect of zeolitic imidazolate frameworks on the gas transport performance of ZIF8-poly(1,4-phenylene ether-ether-sulfone) hybrid membranes, J. Membr. Sci., 2011, 383, 206–213.
50. Y. Dai, J. R. Johnson, O. Karvan, D. S. Sholl and W. J. Koros, Ultem®/ZIF-8 mixed matrix hollow fiber membranes for CO₂/N₂ separations, J. Membr. Sci., 2012, 401–402, 76–82.
51. C. Zhang, Y. Dai, J. R. Johnson, O. Karvan and W. J. Koros, High performance ZIF-8/6FDA-DAM mixed matrix membrane for propylene/propane separations, J. Membr. Sci., 2012, 389, 34–42.
52. Z. Bao, L. Yu, Q. Ren, X. Lu and S. Deng, Adsorption of CO₂ and CH₄ on a magnesium-based metal organic framework, J. Colloid Interface Sci., 2011, 353, 549–556.
53. Z. R. Herm, J. A. Swisher, B. Smit, R. Krishna and J. R. Long, Metal–organic frameworks as adsorbents for hydrogen purification and precombustion carbon dioxide capture, J. Am. Chem. Soc., 2011, 133, 5664–5667.
54. Ş. B. Tantekin-Ersolmaz, Ç. Atalay-Oral, M. Tatlıer, A. Erdem-Şenatalar, B. Schoeman and J. Sterte, Effect of zeolite particle size on the performance of polymer–zeolite mixed matrix membranes, J. Membr. Sci., 2000, 175, 285–288.
55. Y.-S. Bae, K. L. Mulfort, H. Frost, P. Ryan, S. Punnathanam, L. J. Broadbelt, J. T. Hupp and R. Q. Snurr, Separation of CO₂ from CH₄ Using Mixed-Ligand Metal–Organic Frameworks, Langmuir, 2008, 24, 8592–8598.
56. B. Zornoza, C. Téllez and J. Coronas, Mixed matrix membranes comprising glassy polymers and dispersed mesoporous silica spheres for gas separation, J. Membr. Sci., 2011, 368, 100–109.
57. K. Díaz, L. Garrido, M. López-González, L. F. del Castillo and E. Riande, CO2 Transport in Polysulfone Membranes Containing Zeolitic Imidazolate Frameworks As Determined by Permeation and PFG NMR Techniques, Macromolecules, 2010, 43, 316–325.
58. V. Nafisi and M.-B. Hägg, Development of Dual layer of ZIF-8/PEBAX-2533 mixed matrix membrane for CO₂ capture, J. Membr. Sci., 2014, 459, 244–255.
59. T. Pechar, S. Kim, B. Vaughan, E. Marand, M. Tsapatsis, H. Jeong and C. Cornelius, Fabrication and characterization of polyimide–zeolite L mixed matrix membranes for gas separations, J. Membr. Sci., 2006, 277, 195–202.

---