Structure-dependent CO 2 separation in PEBAX membranes incorporating polystyrene nanoparticles: freestanding versus thin-film composite configurations

Polystyrene (PS) nanoparticles are typically regarded as low-permeability fillers in gas separation membranes; however, this study demonstrates that their incorporation into PEBAX matrices can enhance both permeability and selectivity through configuration-dependent transport mechanisms. A systematic comparison between freestanding and thin-film composite (TFC) membranes reveals a clear shift in dominant gas transport behavior arising from structural differences. In ∼100 µm-thick freestanding membranes, PS nanoparticles act as bulk microstructural modifiers that partially disrupt crystalline packing, resulting in reduced crystallinity. This structural perturbation enhances chain mobility and increases the fractional free volume, enhancing gas diffusivity. At an optimal loading of 10 wt%, CO 2 permeability and CO 2 /N 2 selectivity simultaneously increased to 198.6 Barrer and 41.4 (54% and 31% improvements over pristine PEBAX), indicating partial mitigation of the permeability–selectivity trade-off through controlled free-volume enhancement. In contrast, ∼250 nm TFC membranes exhibit shear-assisted nanoparticle alignment and confinement within the ultrathin selective layer, producing an interfacial diffusion-controlled structure. While permeance remains relatively stable at moderate loadings, preferential suppression of N 2 transport enhances selectivity, increasing CO 2 /N 2 selectivity from 26 to 39 at 20 wt% PS with only minor permeance reduction. These findings demonstrate a mechanistic transition from bulk free-volume enhancement in freestanding membranes to interfacial diffusion control in TFC configurations, highlighting the critical role of nanoparticle organization in governing gas transport performance.

The continuous increase in carbon dioxide (CO 2) emissions has become a critical environmental concern because of its direct contribution to severe climate change and global warming (Mollahosseini et al., 2025; Chen et al., 2023; Oschatz and Antonietti, 2018). In response, the development of efficient and sustainable CO 2 capture technologies has attracted significant global attention. Conventional CO 2 capture techniques, including absorption (Kim et al., 2019), adsorption (Lee et al., 2018), and cryogenic separation (Asghari et al., 2024), have been widely applied in industrial processes. However, these methods typically suffer from inherent disadvantages such as high capital and operational costs, large energy consumption, and complex process requirements. In contrast, membrane-based separation has emerged as a promising alternative due to its inherent advantages of low energy demand, simple process design, scalability, and high operational efficiency (Kim et al., 2024a; Han et al., 2023).

Among various membrane materials, polymers have gained considerable attention because of their molecular diversity, mechanical flexibility, and ease of processing into thin, defect-free films (Dai et al., 2023; Hwang J. S. et al., 2025; Ahmadijokani et al., 2023). The gas transport behavior of dense polymer membranes is commonly described by the solution-diffusion mechanism (Wijmans and Baker, 1995), wherein gas molecules are first dissolved into the polymer matrix, followed by diffusion through the polymer chains and subsequent desorption at the downstream interface (Yong and Zhang, 2021). Consequently, the gas permeability can be expressed as the product of the solubility coefficient and diffusivity coefficient. The solubility coefficient is primarily governed by the thermodynamic affinity between gas molecules and the polymer matrix, while the diffusivity coefficient is mainly affected by the free volume, chain mobility, and packing density of the polymer (Hwang J. et al., 2025). Therefore, achieving high membrane performance requires a synergistic optimization of both diffusivity and solubility to maximize permeability without compromising selectivity.

Among the various polymer systems investigated, the poly (ether-block-amide) copolymer (PEBAX) has been recognized as one of the most promising candidates for CO 2 separation membranes (Yang et al., 2023; Wang et al., 2022). PEBAX consists of alternating soft and hard segments: the poly (ethylene oxide) (PEO) block provides CO 2 -philic properties, while the polyamide (PA) block contributes mechanical robustness and film-forming ability (Oh et al., 2025a). The presence of polar ether oxygens in the PEO segment promotes strong dipole-quadrupole interactions with CO 2 molecules, leading to high CO 2 solubility and excellent CO 2 /N 2 selectivity compared to nonpolar gases such as N 2, CH 4, and H 2 (Kim et al., 2024b). However, the crystalline PA domains serve as rigid, low-permeability regions, thereby restricting gas diffusion through the membrane (Guan et al., 2023; Jafarzadeh et al., 2025). As a result, PEBAX membranes exhibit the classical permeability-selectivity trade-off behavior that limits their large-scale industrial application (Liang et al., 2025).

To mitigate this trade-off, extensive research efforts have been devoted to the incorporation of various fillers into PEBAX matrices to form mixed-matrix membranes (MMMs) (Setiawan and Chiang, 2023a; Wang et al., 2024; Chen et al., 2025; Saini et al., 2021; Min et al., 2023). Inorganic porous fillers, such as graphene oxide (GO), metal-organic frameworks (MOFs) and covalent organic frameworks (COFs), have been used to introduce molecular sieving channels that preferentially facilitate CO 2 transport, enhancing both permeability and selectivity (Hua et al., 2025; Hassan et al., 2023). For instance, the incorporation of ionic liquid-modified ZIF-8 nanoparticles into PEBAX created tailored pore environments that improved CO 2 permselectivity relative to neat PEBAX membranes (Kang et al., 2024). Alternatively, nonporous inorganic fillers, including SiO 2 and TiO 2 nanoparticles, have been introduced to disrupt polymer chain packing and increase the fractional free volume, thereby promoting faster gas diffusion (Hassanzadeh et al., 2025). More recently, organic fillers have emerged as promising alternatives owing to their processability, chemical compatibility with polymer matrices, and cost-effectiveness.

Despite these advancements, the majority of studies on organic or polymeric additives for gas separation membranes have predominantly focused on the design of fillers with intrinsically high CO 2 affinity, such as poly (ethylene glycol)-based materials or amine-functionalized polymers (Oh et al., 2025b; Kim et al., 2017). These approaches are largely driven by the goal of enhancing CO 2 solubility through favorable molecular interactions, and improvements in separation performance are often attributed to solubility-selective effects. In contrast, polystyrene (PS), a hydrophobic polymer with a rigid phenyl-substituted backbone and inherently low gas permeability, has rarely been explored as a membrane additive (Guo et al., 2025; Ghaly et al., 2024). Owing to its limited CO 2 affinity and low intrinsic permeability compared to rubbery polymers, PS has traditionally been regarded as unsuitable for gas transport applications, and its incorporation into separation membranes has largely been avoided. However, recent advances in nanoscale synthesis, particle size control, and surface functionalization provide an opportunity to reconsider PS from a different perspective. When engineered as nanoparticles with tailored surface chemistry, PS can function not as a transport-active phase, but as a morphology-modulating additive capable of perturbing polymer chain packing, crystallinity, and free-volume distribution within the host matrix. In particular, surface sulfonation introduces polar functionalities that enhance interfacial compatibility with PEBAX, suppress particle aggregation, and promote uniform dispersion. Such interfacial engineering enables controlled disruption of polymer packing and an increase in free volume without generating nonselective defects. This morphology-driven strategy represents a distinct approach beyond conventional CO 2 -affinity-based designs, suggesting that PS nanoparticles may offer an underexplored pathway for tuning membrane microstructure and gas separation performance.

In this study, we explore PS nanoparticles as an unconventional morphology-directing additive in PEBAX-based membranes for CO 2 separation. Although PS is typically regarded as a low-permeability polymer unsuitable for gas transport applications, we demonstrate that surface-modified PS nanoparticles can induce configuration-dependent transport behavior through controlled microstructural reorganization of the PEBAX matrix. By systematically comparing ∼100 µm-thick freestanding membranes and ∼250 nm-thick thin-film composite (TFC) membranes, we reveal a mechanistic transition in gas transport behavior governed by nanoparticle organization. In thick freestanding membranes, PS primarily functions as a bulk free-volume modifier, whereas in ultrathin TFC layers, nanoparticle confinement and interfacial alignment give rise to diffusion-controlled transport characteristics (Kang et al., 2022). The microstructural, morphological, and interfacial characteristics of PEBAX/PS blend membranes with varying nanoparticle loadings were systematically examined using Fourier transform infrared (FTIR) spectroscopy, differential scanning calorimetry (DSC), and scanning electron microscopy (SEM). Collectively, these findings provide new insights into the role of polymeric nanoparticles as morphology-modulating additives and offer a promising strategy for overcoming the long-standing permeability-selectivity trade-off in polymeric gas separation membranes.

Materials

Polystyrene (PS) particles with an average diameter of approximately 165 nm were obtained from Chemowave (Korea) and used without further purification. PEBAX® 1,657 (Arkema, France), consisting of 40 wt% polyamide (PA) and 60 wt% poly (ethylene oxide) (PEO) segments, served as the polymer matrix and was used as received. Polysulfone (PSf) supports were provided by Toray Advanced Materials Korea and poly (1-trimethylsilyl-1-propyne) (PTMSP) was obtained from Gelest, Inc.

Membrane preparation

PEBAX pellets were dissolved in a mixed solvent of ethanol and deionized water (70:30 v/v) under magnetic stirring at 60 °C for 2 h to obtain a homogeneous polymer solution. Separately, PS particles were dispersed in the same solvent mixture via ultrasonication for 30 min to achieve a uniform suspension. The polymer solution was then slowly added to the particle dispersion under continuous stirring, followed by an additional 10 min of ultrasonication to prevent particle aggregation. For the fabrication of freestanding membranes, the resulting PEBAX/PS solution (10 wt% in solvent) was cast onto glass dishes and dried at 50 °C overnight to remove residual solvent. The dried films were subsequently placed in a vacuum oven at 50 °C for 18 h to ensure complete solvent removal. For TFC membranes, a gutter layer of PTMSP in cyclohexane (1.5 wt%) was first coated onto a microporous PSf support using an RK Control Coater (Model 101, Control RK Print-Coat Instruments Ltd., UK) and dried at 25 °C. The PEBAX/PS solution was then applied onto the PTMSP-coated PSf support using the same coating procedure (Bazhenov et al., 2016).

Characterization

The molecular interactions between PEBAX and PS were analyzed using a Fourier transform infrared (FT-IR) spectrometer (Spectrum Two, Perkin Elmer, United States). The thermal behavior of the membranes was analyzed using a differential scanning calorimeter (DSC, Discovery DSC, TA Instrument, United States) operated under a nitrogen atmosphere at a heating rate of 20 °C∙min -1. The surface and cross-sectional morphology of the PEBAX/PS membranes were observed using a field-emission scanning electron microscope (FE-SEM, 7610F-plus, JEOL Ltd., Tokyo, Japan). The mechanical properties of the membranes were measured using a universal testing machine (UTM, MultiTest2.5-I, Mecmesin) at a crosshead speed of 20 mm/min. The thermal stability of the membranes was characterized using a thermogravimetric analyzer (TGA, TGA 55, TA Instruments) operated at a ramping rate of 20 °C∙min -1 from 50 °C to 800 °C under a nitrogen atmosphere.

Gas permeation measurements

For freestanding membranes, gas permeation measurements were conducted using the time-lag method with a constant-volume/variable-pressure apparatus (Airrane Co., Ltd., Korea) operated at 760 Torr (1 bar) and 35 °C, with an effective membrane area of 15.2 cm 2. Prior to each measurement, the downstream chamber was evacuated to below 0.1 Torr to ensure negligible initial permeate pressure. The upstream side was then pressurized to 760 Torr with the test gas, and the transmembrane pressure difference (Δp) was maintained constant during the experiment. The pure gas permeabilities (P) of CO 2, N 2 and CH 4 (1 barrer = 10 −10 cm 3 (STP)?cm∙cm -2 ∙s -1 ∙cmHg -1) were determined from the steady-state permeation rate according to: P = J L/Δp. where J is the steady-state molar flux calculated from the linear increase in downstream pressure with time, L is the membrane thickness, and Δp is the transmembrane partial pressure difference. The downstream pressure increase rate (dP d /dt) was converted to molar flow using the ideal gas law and the known downstream volume. The ideal gas selectivities for CO 2 /N 2 and CO 2 /CH 4 were calculated as the ratio of CO 2 permeability to the permeability of the respective gas. For TFC membranes, gas permeation properties were evaluated at 25 °C and 1 bar using a constant-pressure/variable-volume apparatus (Airrane Co., Ltd., Korea) equipped with a flat-sheet permeation cell (effective area: 10.2 cm 2) (O'Brien et al., 1986). The gas permeances of CO 2, N 2 and CH 4 were reported in gas permeation units (GPU), where 1 GPU = 10 −6 cm 3 (STP)∙cm -2 ∙s -1 ∙cmHg -1. The diffusivity coefficients were determined using the time-lag method from single-gas permeation experiments performed in a constant-volume/variable-pressure apparatus. The time lag (θ) was obtained from the intercept of the linear steady-state permeation region with the time axis. The diffusion coefficient (D) was calculated according to D = L 2 /6θ, where L is the membrane thickness and θ is the time lag. The permeability coefficient (P) was calculated from the steady-state permeation rate, and the solubility coefficient (S) was subsequently obtained using the solution–diffusion relationship.

Interactions and morphology

PEBAX block copolymers are promising materials for CO 2 separation owing to their phase-separated structure, in which rigid PA domains provide mechanical strength while CO 2 -philic PEO segments enable selective CO 2 transport. However, the high crystallinity of the PA phase restricts chain mobility and free-volume availability, ultimately limiting the overall gas permeability of PEBAX-based membranes (Kim et al., 2024b; Piotrowska et al., 2024). To overcome this intrinsic limitation, PS nanoparticles functionalized with sulfonic acid (PS–SO 3 H) were incorporated into the PEBAX matrix as a diffusivity-modulating additive. The inclusion of PS particles was designed to disrupt the ordered packing of the PEBAX chains, effectively lowering the degree of crystallinity and increasing the fractional free volume available for gas diffusion. Moreover, the sulfonic acid groups grafted onto the PS nanoparticles enhanced interfacial compatibility with the polar PEBAX matrix through hydrogen bonding and dipole-dipole interactions, leading to improved filler dispersion and mechanical integrity of the composite films. As a result, flexible, freestanding dense PEBAX/PS composite membranes containing up to 15 wt% PS nanoparticles were successfully fabricated without noticeable defects or phase separation The incorporation of PS–SO 3 H particles not only promoted uniform dispersion within the polymer network but also provided an effective means to tailor gas diffusivity and permeability, thereby advancing the design of high-performance polymer membranes for CO 2 separation applications.

First, the intermolecular interactions between PEBAX matrix and PS filler in the membranes were characterized by FT-IR spectroscopy. Figure 1a shows the FTIR spectra of pristine PEBAX and PEBAX/PS composite membranes containing different loadings of PS particles (5, 10, and 15 wt%). The characteristic absorption peaks of PEBAX are clearly observed at 3,296 cm -1 (N–H stretching of the amide group), 3,026 and 2,888 cm -1 (C–H stretching of–CH 2 – groups), 1731 cm -1 (O–C=O asymmetric stretching), 1,637 cm -1 (C=O stretching of the amide linkage), and 1,100 cm -1 (C–O–C stretching of the ether segment in PEO) (Kim et al., 2024b; Murali et al., 2010; Setiawan and Chiang, 2023b). These peaks confirm the coexistence of the hard PA and soft PEO segments in the PEBAX block copolymer. Upon incorporation of PS nanoparticles, the characteristic weak peak near 757 cm -1 gradually intensified, corresponding to the out-of-plane bending vibration of the benzene ring, an identifying feature of PS. This progressive increase in peak intensity confirms the successful incorporation and increasing content of PS particles within the PEBAX matrix. Notably, the N–H and C=O stretching bands associated with the PA segments exhibited no significant shift, indicating minimal interaction between PS and the PA domains. In contrast, a slight shift of the C–O–C stretching band toward 1,105 cm -1 was observed, suggesting moderate interactions between the PS nanoparticles and the PEO segments of PEBAX. These interactions are likely mediated by hydrogen bonding between the sulfonic acid groups on the PS surface and the ether oxygen atoms in the PEO chains. Such interfacial interactions play a crucial role in enhancing compatibility between the PS fillers and the PEBAX matrix, promoting homogeneous dispersion and effectively suppressing macrophase separation. The improved interfacial adhesion not only strengthens the mechanical integrity of the composite membranes but also facilitates more continuous gas transport pathways, contributing to the overall enhancement of membrane performance. These FT-IR analysis verifies the successful fabrication of PEBAX/PS composite membranes, demonstrating both physical blending and molecular-level interfacial interactions between the polymer matrix and the functionalized PS nanoparticles. The systematic evolution of the benzene ring vibration peak intensity with increasing PS loading provides clear spectroscopic evidence of composite formation and improved interfacial compatibility within the PEBAX/PS system.

DSC analysis was conducted to examine the thermal transitions of the PEBAX/PS composite membranes, with particular attention to the glass transition temperature (T g) and melting temperature (T m), as shown in Figure 1b. The neat PEBAX membrane exhibited a T g at −49.66 °C, corresponding to the amorphous PEO phase, along with two distinct melting endotherms at 18.84 °C and 205.44 °C, which are associated with the crystalline domains of the PEO and PA segments, respectively (Nobakht and Abedini, 2023). These well-defined thermal transitions reflect the inherent semi-crystalline block copolymer architecture of PEBAX, in which the flexible PEO domains provide chain mobility while the rigid PA blocks contribute to mechanical integrity. It is worth noting that no distinct T g corresponding to PS (typically observed near 95 °C–105 °C for bulk PS) was detected in the DSC thermograms. This absence can be attributed to the relatively low PS loading (5–15 wt%) and its nanoscale dispersion within the continuous PEBAX matrix. The confined nature of the PS nanoparticles and their strong interfacial interactions with PEBAX chains likely broaden or suppress the characteristic T g transition of PS, rendering it calorimetrically indistinguishable. Moreover, the thermal region where PS T g would appear partially overlaps with baseline variations associated with the PA segment transitions, further limiting its resolution. Upon incorporation of PS nanoparticles, gradual yet systematic changes in both T g and T m were observed, depending on the filler content. At low to moderate PS loadings (5–10 wt%), the T g of the PEO phase shifted slightly downward from −49.66 °C to approximately −51.44 °C, indicating a modest enhancement in local segmental mobility within the amorphous regions. Concurrently, the melting peaks associated with the PEO and PA crystalline domains became progressively broader and less intense, while their peak temperatures remained largely unchanged. This trend suggests that PS nanoparticles primarily disrupt the ordered packing of PEBAX chains and suppress crystalline development, rather than fundamentally altering the crystalline phase identity. Such disruption of crystallinity and enhancement of amorphous-chain mobility are expected to facilitate gas diffusion, providing a clear thermal rationale for the pronounced increase in CO 2 permeability observed at PS loadings of 5 and 10 wt%. In contrast, further increasing the PS content to 15 wt% led to minimal additional reduction in T g and a partial stabilization of the thermal transitions, implying the emergence of local chain confinement or restricted segmental motion caused by excessive filler incorporation. Consequently, the beneficial balance between reduced crystallinity and enhanced chain mobility is partially diminished, in agreement with the observed saturation or slight decline in gas permeability at higher PS contents. These DSC results provide compelling thermal evidence that PS nanoparticles function as effective modulators of polymer diffusivity by tuning the microstructure of PEBAX through simultaneous modification of glass transition behavior and crystalline packing, rather than through a single dominant thermal effect. This interpretation aligns well with the gas permeation trends discussed in the following sections.

The crystalline structure of Pebax and Pebax/PS composite membranes was investigated using XRD, as shown in Figure 1c. All samples exhibit a characteristic diffraction peak centered at approximately 2θ ≈ 23.9°, which is attributed to the crystalline domains of the PA segments in Pebax. Notably, the peak position remains nearly unchanged with increasing PS content, indicating that the fundamental crystal lattice structure of the Pebax matrix is largely preserved upon incorporation of PS. However, a gradual reduction in peak intensity and slight peak broadening are observed as the PS loading increases, suggesting a decrease in crystallinity and reduced crystalline domain size. This behavior can be attributed to the incorporation of PS nanoparticles, which interfere with the regular packing of polymer chains. In particular, the sulfonated surface groups of PS are likely to interact with the ether segments of Pebax, disrupting interchain interactions and hindering the formation of well-ordered crystalline domains. As a result, the overall crystalline packing becomes less perfect, leading to increased amorphous content. The increase in amorphous regions is expected to enhance chain mobility and create additional free volume within the membrane. This structural modification plays a critical role in facilitating gas transport, particularly by increasing diffusivity across the polymer matrix. Therefore, although the crystal structure of Pebax is maintained, the partial disruption of crystalline packing and the corresponding increase in amorphous fraction contribute to the improved gas permeation properties observed in the composite membranes.

To further elucidate the effect of PS incorporation on the crystalline structure of Pebax, the crystallite size was estimated using the Scherrer equation based on the PA diffraction peak centered at 2θ ≈ 23.9°, as shown in Figure 1d. As PS content increases, the diffraction peak exhibits noticeable broadening, indicating a reduction in crystallite size. Despite the negligible shift in peak position, the decrease in crystallite size suggests that the incorporation of PS does not alter the crystal lattice but disrupts the growth and perfection of crystalline domains. This behavior can be attributed to the interfacial interactions between the sulfonated PS surface groups and the Pebax matrix, which hinder regular chain folding and crystal growth. As a result, smaller and less perfect crystallites are formed. The reduction in crystallite size is accompanied by an increase in amorphous regions, which enhances chain mobility and generates additional fractional free volume. The Scherrer analysis confirms that PS incorporation leads to nanoscale disruption of crystalline domains, contributing to improved gas transport properties without significantly altering the fundamental crystal structure.

To clarify the effect of PS nanoparticle incorporation on membrane morphology, surface structures were systematically examined by SEM, as shown in Figure 2. The neat PEBAX membrane exhibits a characteristic heterogeneous morphology in which nanofibril-like crystalline domains of the PA segments are widely dispersed within a continuous amorphous matrix formed by the PEO phase. This morphology reflects the intrinsic block copolymer nature of PEBAX: the rigid PA domains provide mechanical reinforcement and dimensional stability, while the amorphous PEO regions enable gas transport due to their high segmental mobility and strong CO 2 affinity. Nevertheless, because the crystalline PA domains are relatively impermeable, their widespread distribution can disrupt continuous diffusion pathways, thereby limiting overall gas permeability. The introduction of PS nanoparticles alters the crystallization behavior of PEBAX during solvent evaporation and film formation. Rather than inducing macrophase separation, the nanoparticles act as heterogeneous nucleation sites, leading to localized reorganization of PA crystalline domains. As a result, more defined microphase-separated regions composed of nanofibrillar crystalline domains embedded within smoother amorphous areas are observed. With increasing PS content (up to 10 wt%), the density of nucleation sites increases, producing more pronounced but spatially confined crystalline clusters. This morphology reflects crystallization-induced heterogeneity rather than particle aggregation. The sulfonic acid groups on the PS surface enhance interfacial compatibility with the PEO-rich domains, stabilizing the polymer-particle interfaces and preventing large-scale agglomeration. Consequently, the overall dispersion of PS nanoparticles remains stable, while the modified crystalline organization contributes to the redistribution of amorphous regions within the membrane. At PS loadings up to 15 wt%, the tendency toward nanoparticle aggregation became increasingly evident, leading to localized densification and partial macrophase separation on the membrane surface. These morphological changes indicate that, although sulfonic acid functionalization improves interfacial compatibility, it does not completely prevent aggregation at higher filler concentrations due to intrinsic mechanical and thermodynamic constraints in dense freestanding membranes. Excessive incorporation of rigid PS nanoparticles can introduce stress concentration and disrupt structural continuity, potentially resulting in defect formation. Nevertheless, at moderate loadings (≤10 wt%), PS nanoparticles effectively regulate the microphase-separated structure of PEBAX membranes. The incorporation of well-dispersed PS particles suppresses PEBAX crystallinity and promotes a more homogeneous distribution of amorphous domains, thereby increasing fractional free volume and enhancing gas diffusivity. This balanced microstructural modification explains the observed improvement in gas permeability and selectivity discussed in the following sections.

Thermal and mechanical properties

TGA was carried out to evaluate the thermal stability of the PEBAX/PS composite membranes, and the results are presented in Figure 3a. The neat PEBAX membrane exhibited a single-step decomposition pattern, indicating its intrinsic thermal stability. No significant weight loss was observed up to approximately 300 °C, confirming that the polymer matrix maintains structural integrity under typical membrane operation conditions. The major thermal degradation step occurred beyond this temperature range, corresponding to the breakdown of both the PA and PEO segments of PEBAX. The PEBAX/PS composite membranes demonstrated a similar one-step degradation profile but exhibited slightly enhanced thermal stability compared with the pristine PEBAX (Jafarzadeh et al., 2025; Eljaddi et al., 2022). This improvement can be attributed to the incorporation of PS nanoparticles, which possess inherently high thermal resistance. Notably, above 450 °C, a small increase in the residual weight was observed due to the carbonization of PS nanoparticles, leading to the formation of thermally stable carbonaceous residues. The absence of any noticeable decomposition below 300 °C across all compositions indicates that the surface-functionalized PS fillers, even though containing sulfonic acid groups, were present in relatively small quantities and did not compromise the overall thermal integrity of the membranes. Additionally, all membranes showed negligible weight loss up to 300 °C, confirming that no residual solvent or moisture remained after the drying process. This observation supports the effective removal of ethanol and water during membrane fabrication and highlights the compact and stable morphology of the composite structures. The high-temperature resistance of the PS fillers likely reinforces the polymer matrix, delaying the onset of thermal degradation by acting as thermally robust physical barriers to chain scission and volatilization. The PEBAX/PS composite membranes demonstrate excellent potential as high-performance materials for gas separation applications under elevated temperature conditions, combining superior processability with remarkable thermal robustness.

The mechanical performance of the freestanding PEBAX/PS composite membranes was evaluated using a UTM, and the results are summarized in Figure 3b. The neat PEBAX membrane exhibited outstanding mechanical flexibility and strength, with a tensile stress at break of approximately 21 MPa and an elongation at break reaching about 1,200%. These values reflect the inherent elastomeric nature of the PEBAX block copolymer, which arises from its unique microphase-separated structure consisting of hard PA crystalline domains dispersed within soft and ductile PEO segments. The PA domains act as physical crosslinking sites that provide mechanical reinforcement, while the amorphous PEO regions impart elasticity and extensibility to the membrane. Upon the incorporation of PS nanoparticles, the overall mechanical response of the membranes exhibited a moderate change. Specifically, the tensile stress at break showed a slight decrease, and the elongation at break was notably reduced compared to the pristine PEBAX membrane (Hwang J. et al., 2025; Mohammed et al., 2019). This reduction in strain is primarily attributed to the inherently rigid and brittle nature of the PS nanoparticles, which restrict the mobility of polymer chains and act as localized stress concentrators during deformation. Additionally, the incorporation of PS leads to partial disruption of the crystalline packing within the PEBAX matrix, increasing the amorphous fraction but simultaneously decreasing chain entanglement and elasticity. The resulting microstructural changes thus yield membranes that are slightly stiffer and less stretchable, while still retaining satisfactory toughness and cohesion. Despite these modest reductions, all composite membranes maintained excellent mechanical strength and flexibility, sufficient to withstand the operational stresses encountered during gas permeation testing, including exposure to high vacuum conditions in the time-lag permeation apparatus. The mechanical robustness of the PEBAX/PS blend can be attributed to the great interfacial adhesion between the PS nanoparticles and the PEBAX matrix, which ensures efficient stress transfer and prevents interfacial delamination under tension. These results demonstrate that the incorporation of PS nanoparticles into the PEBAX matrix effectively balances mechanical stability and flexibility, allowing the membranes to retain high structural integrity while benefiting from the improved diffusivity and gas transport properties conferred by the modified morphology. The PEBAX/PS composite membranes therefore meet the essential mechanical requirements for practical gas separation applications and can be considered reliable candidates for durable, high-performance separation systems.

Gas separation performance

The pure-gas permeability and selectivity of the ∼100 µm-thick freestanding PEBAX/PS composite membranes containing different loadings of PS nanoparticles were systematically investigated at 35 °C and 1 bar using the time-lag method, as summarized in Figure 4 and Table 1. The neat PEBAX membrane exhibited a CO 2 permeability of 128.6 Barrer, together with CO 2 /N 2 and CO 2 /CH 4 selectivities of 31.6 and 11.2, respectively, which are consistent with the typical performance of PEBAX-based membranes reported in literature. It is noteworthy that CH 4 permeability was consistently higher than that of N 2, despite its slightly larger kinetic diameter. This trend reflects the solubility-dominated transport behavior typical of rubbery polymers. According to the solution–diffusion mechanism (P = D × S), permeability depends on both diffusivity and solubility. Although N 2 may exhibit marginally higher diffusivity due to its smaller molecular size, CH 4 possesses higher condensability and therefore greater solubility in the polymer matrix. As a result, the solubility contribution outweighs the diffusivity difference, leading to higher CH 4 permeability compared to N 2 and consequently CO 2 /N 2 selectivity exceeding CO 2 /CH 4 selectivity. This behavior is consistent with gas transport characteristics widely reported for PEO-based and other rubbery polymer membranes. It should be emphasized, however, that the gas transport properties of PEBAX can vary depending on the membrane preparation parameters, such as solvent type, polymer concentration, casting temperature, solvent evaporation rate, and post-drying conditions. These factors strongly influence the crystalline-amorphous balance and the chain alignment within the microphase-separated structure, leading to differences in gas diffusion and solubility behaviors. In the PEBAX block copolymer, the polar PEO segments exhibit a strong affinity for CO 2 molecules through dipole-quadrupole interactions, resulting in high CO 2 solubility and, consequently, high CO 2 selectivity over nonpolar gases such as N 2 and CH 4. Conversely, the PA blocks are relatively rigid and crystalline, serving as dense domains that restrict molecular diffusion and therefore reduce overall permeability. The delicate interplay between these soft and hard segments governs the overall transport characteristics of the PEBAX matrix. Interestingly, despite the intrinsically low permeability of PS, the introduction of PS nanoparticles led to a gradual increase in gas permeability for all tested gases. This enhancement can be attributed to the increase in free volume and molecular mobility within the polymer matrix, which promotes gas diffusivity. The interfacial interactions between the hydrophilic functional groups of PS nanoparticles and the polar segments of PEBAX are likely to disturb the crystalline ordering of the PA domains and reduce chain packing density, thereby generating additional pathways for gas transport. Among the tested compositions, the PEBAX/PS10 membrane demonstrated the most balanced and superior separation performance. Specifically, it achieved a CO 2 permeability of 198.6 Barrer, representing an approximate 54% improvement compared with the neat PEBAX membrane. At the same time, the CO 2 /N 2 selectivity increased by 31%, reaching 41.4, and the CO 2 /CH 4 selectivity slightly improved to 12. These results indicate that the addition of a moderate amount of PS nanoparticles not only enhances gas diffusivity but also maintains favorable CO 2 solubility, leading to a synergistic improvement in both permeability and selectivity. Such a simultaneous enhancement suggests that the PS nanoparticles contribute to optimizing the microstructure of the PEBAX matrix by creating a more open yet selective diffusion network without sacrificing the intrinsic polymer-gas affinity balance. At 15 wt% PS loading, CO 2 permeability slightly decreased compared to the 10 wt% membrane, suggesting a non-linear filler effect. Excessive nanoparticle content likely induces partial aggregation and interfacial rigidification, reducing chain mobility and disrupting the continuity of CO 2 -philic PEO domains. Consequently, effective sorption sites for CO 2 diminish, whereas N 2 and CH 4 permeabilities remain relatively unaffected.

To decouple the individual contributions of diffusivity and solubility to gas transport, the diffusion coefficients (D) were determined using the time-lag method, and the solubility coefficients (S) were calculated from the solution–diffusion relationship (P = D × S). (Nobakht and Abedini, 2023). It should be noted that both diffusivity and solubility coefficients were directly derived from experimentally measured permeation data obtained via the time-lag method, without the use of additional fitting models or indirect estimation approaches. All diffusivity and solubility coefficients were obtained from pure-gas permeation experiments conducted at 35 °C and an upstream pressure of 1 bar using the time-lag method. Figure 5 presents the diffusivity and solubility values derived from these measurements. The CO 2 diffusivity increased steadily with increasing PS loading, indicating that the addition of PS nanoparticles facilitated molecular diffusion through the polymer matrix. This enhancement can be attributed to the formation of additional free volume and looser chain packing induced by the interfacial interactions and reduced crystallnity. The CO 2 solubility also exhibited a moderate increase up to 10 wt% PS, suggesting that the PS surfaces, likely functionalized with hydrophilic sulfonic acid moieties, contributed to maintaining favorable CO 2 sorption sites. However, beyond a certain PS concentration, a slight decrease in solubility was observed, which could result from partial particle aggregation and a corresponding reduction in accessible sorption domains. For N 2, a nonpolar gas with weak affinity toward PEO, the diffusivity increased gradually with PS loading, following a similar trend to CO 2 but with lower magnitude. This again supports that the primary influence of PS nanoparticles is the structural loosening and free-volume generation rather than changes in polymer-gas affinity. In contrast, the N 2 solubility remained nearly constant or slightly decreased with higher PS content, consistent with its low condensability and limited solvation capability in the polymer matrix. The CH 4 diffusivity also increased with PS addition, but to a lesser degree compared to CO 2, indicating that larger or more spherical gas molecules experience less benefit from microstructural expansion. The CH 4 solubility displayed a non-monotonic trend: it slightly increased at moderate PS contents (up to 5 wt%), likely due to improved molecular accommodation in expanded amorphous regions, and decreased thereafter as the particle aggregation reduced the available sorption volume. Overall, the introduction of PS nanoparticles promoted diffusivity enhancements for all gases through microstructural relaxation and free-volume enlargement, while maintaining or slightly improving CO 2 solubility due to favorable polymer-filler interactions. The combined diffusivity and solubility trends explain the observed increase in CO 2 permeability and selectivity at optimal PS contents (around 10 wt%), as discussed in the previous section. These results confirm that PS nanoparticle incorporation can effectively tune the microstructure of PEBAX membranes to balance diffusivity and solubility contributions, yielding superior gas separation performance.

Figure 6a plots CO 2 permeability against CO 2 /N 2 selectivity for freestanding neat PEBAX membrane and PEBAX/PS composite membranes with varying PS nanoparticle loadings, alongside literature-reported data for other PEBAX-based all-polymeric membranes (Kim et al., 2024b; Guan et al., 2023; Jafarzadeh et al., 2025; Martínez-Visus et al., 2025; Kim et al., 2020; Yu et al., 2023; Chen et al., 2024; Kabir et al., 2024; Park et al., 2021; Tan et al., 2025). MMMs incorporating highly porous fillers such as MOFs and COFs were excluded because their distinct structures lead to fundamentally different gas transport mechanisms. The neat PEBAX membrane is positioned within the typical performance range for PEBAX 1657, showing moderate permeability (∼128 Barrer) and CO 2 /N 2 selectivity (∼31). Upon incorporation of PS nanoparticles, both CO 2 permeability and selectivity improved, particularly at 10 wt% PS loading. The PEBAX/PS10 membrane exhibited a CO 2 permeability of approximately 198.6 Barrer and a selectivity of 41.4, representing a 54% and 31% increase, respectively, compared to the pristine membrane. This simultaneous enhancement in permeability and selectivity is significant because polymeric membranes typically suffer from the trade-off relationship described by the Robeson upper bound. The observed performance improvement suggests that the addition of PS nanoparticles effectively modifies the microstructure of PEBAX by introducing extra free volume while preserving the selective CO 2 solubility of the PEO segments. The uniform distribution and moderate interfacial interaction between PS and PEBAX chains disrupt crystalline regions and promote more efficient CO 2 diffusion without increasing non-selective voids. Consequently, the PEBAX/PS10 membrane approaches the upper performance limit of PEBAX-based membranes reported in the literature, demonstrating the efficacy of PS nanoparticles as microstructural modifiers for enhanced gas separation.

Figure 6b presents the correlation between CO 2 permeance and CO 2 /N 2 selectivity for the PEBAX/PS composite membranes fabricated as TFC layers on porous PSf supports. The data reveal that both CO 2 permeance and selectivity are affected by the incorporation of PS nanoparticles, showing a distinct composition-dependent trend (Sutrisna et al., 2017; Chou et al., 2025; Oh et al., 2024). As the PS nanoparticle content increased from 0 to 10 wt%, the CO 2 permeance remained nearly constant in the range of 760–800 GPU, indicating that the overall gas transport resistance through the selective layer was not significantly altered within this composition window. In contrast, the CO 2 /N 2 selectivity improved markedly from 26.4 for the neat PEBAX TFC membrane to 32.9 for the PEBAX/PS10 membrane, demonstrating that the addition of PS nanoparticles enhances the molecular discrimination capability of the membrane without compromising gas throughput. Beyond the optimal PS loading of 10 wt%, a gradual decrease in both CO 2 and N 2 permeance was observed. This reduction is attributed to the formation of larger transport regions dominated by the less permeable PS domains at higher filler concentrations. The inherently low gas permeability of PS can increase the tortuosity of the diffusion pathway and partially hinder gas transport. However, despite the overall decrease in permeance at higher loadings, the CO 2 /N 2 selectivity continued to increase steadily. This trend indicates that the denser interfacial structure and phase-separated morphology introduced by PS nanoparticles create more restrictive diffusion pathways that preferentially favor the smaller and more condensable CO 2 molecules over the larger and less soluble N 2. The relatively stable CO 2 permeance observed up to 10 wt% PS suggests that the effective thickness of the selective layer and the intrinsic diffusivity of the PEBAX matrix are well maintained, owing to the uniform dispersion of nanoparticles and the strong interfacial compatibility between PS and the PEBAX polymer chains. The PS nanoparticles are believed to facilitate microphase separation by promoting a heterogeneous interface between the soft PEO and rigid PA segments of PEBAX, generating regions of increased free volume that enable faster CO 2 diffusion. At the same time, the nanoparticles act as physical barriers that suppress non-selective gas transport and reduce the permeation of the less permeable gases, particularly N 2. This dual effect, enhanced CO 2 diffusivity and suppressed N 2 transport, leads to the observed increase in selectivity. The continuous improvement in selectivity with PS content also implies that the interfacial interactions between PEBAX and PS prevent the formation of large-scale voids or structural defects that could otherwise compromise the membrane integrity and selectivity. Instead, the dispersed PS nanoparticles act as rigid, impermeable fillers that refine the diffusion pathway and enhance molecular sieving effects. Furthermore, the reduced chain mobility in the vicinity of the PS interfaces may contribute to the selective exclusion of larger gas molecules, improving the overall separation efficiency. Overall, the PEBAX/PS10 TFC membrane exhibits an outstanding combination of high CO 2 permeance and enhanced CO 2 /N 2 selectivity, positioning it near the upper boundary of performance among polymeric gas separation membranes. This result clearly demonstrates that a moderate amount of PS nanoparticle incorporation is highly beneficial for balancing the trade-off between permeability and selectivity in thin-film membrane systems. The findings underscore that the controlled integration of inert, low-permeability nanofillers such as PS can serve as an effective morphological tuning strategy to optimize both the microstructure and interfacial properties of PEBAX-based membranes. The synergistic improvement achieved here confirms the broad applicability of this design concept for scalable CO 2 separation processes in industrial and environmental applications.

Figure 7 presents the cross-sectional (top) and surface (bottom) SEM images of the TFC membranes fabricated from neat PEBAX and PEBAX/PS composites coated onto PSf supports. The progressive microstructural evolution of the selective layer with increasing PS nanoparticle loading provides key insights into how filler incorporation influences morphology, interfacial adhesion, and coating uniformity. The neat PEBAX TFC membrane exhibits a smooth, compact, and defect-free selective layer approximately 200–250 nm thick, firmly attached to the underlying PSf substrate (Sutrisna et al., 2017; Fam et al., 2017). The cross-sectional view shows a homogeneous and dense morphology typical of high-quality PEBAX coatings, while the surface image reveals a featureless texture, confirming the absence of voids or micro-defects. With the addition of 5 wt% PS nanoparticles, the surface becomes slightly rougher, and the cross-sectional morphology more heterogeneous, though the selective layer remains continuous and well adhered to the support, implying good interfacial compatibility. Bright contrast spots corresponding to well-dispersed PS nanoparticles are observed on the surface, confirming their uniform distribution within the polymer matrix without notable aggregation. At higher PS loadings (up to 20 wt%), the selective layer thickness remains nearly unchanged, yet the nanoparticle density on the surface increases markedly, indicating successful incorporation of PS within the top layer. Remarkably, even at such high filler content, the nanoparticles remain evenly dispersed with minimal aggregation, maintaining a continuous and defect-free coating. This behavior contrasts sharply with that of freestanding membranes and with previous reports, where high filler loadings typically cause particle agglomeration, nonuniform morphology, and partial phase separation (Zheng et al., 2019; Meshkat et al., 2020). Interestingly, in our TFC membranes, the uniform nanoparticle alignment is attributed to shear-induced self-organization during the coating process (Liu et al., 2019), as illustrated in Figure 8. This unique structure allows the PS nanoparticles near the surface to act as a diffusion-controlling layer: gas permeance decreases overall, but the reduction is more pronounced for the less permeable N 2 than for CO 2, thereby enhancing CO 2 /N 2 selectivity. These morphological findings correlate strongly with the observed gas-transport behavior, demonstrating that controlled nanoparticle dispersion and interfacial organization are critical for optimizing the permeance-selectivity trade-off in TFC membranes, unlike the behavior observed in micron-thick freestanding membranes.

A direct comparison between freestanding and TFC configurations reveals distinct gas transport behaviors arising from structural differences. In freestanding membranes (∼tens of micrometers thick), PS nanoparticles primarily act as microstructural modifiers, increasing free volume and enhancing bulk diffusivity throughout the membrane thickness. Consequently, both permeability and selectivity improve simultaneously at moderate filler loadings. In contrast, in TFC membranes (∼200–250 nm selective layer), gas transport is strongly influenced by the coating-induced morphology and interfacial organization near the surface. The shear-assisted nanoparticle alignment and confinement within the ultrathin selective layer lead to a more diffusion-restrictive structure. As a result, permeance remains relatively stable at moderate loadings, while selectivity increases due to preferential suppression of N 2 transport. Therefore, the dominant mechanism shifts from bulk free-volume enhancement (freestanding) to interfacial diffusion control (TFC configuration).

In this study, we systematically investigated the role of PS nanoparticles in regulating the structure, morphology, and gas transport behavior of PEBAX-based membranes in both freestanding and TFC configurations. Despite the intrinsically low permeability of PS, its incorporation effectively enhanced CO 2 separation performance through controlled interfacial interactions and microstructural reorganization of the polymer matrix. FTIR analysis confirmed specific interfacial interactions between the sulfonic acid groups of PS nanoparticles and the ether oxygen atoms of the PEO segments in PEBAX, which improved filler-matrix compatibility and suppressed macrophase separation. Consistently, DSC and XRD results showed a reduction in PEBAX crystallinity upon PS incorporation, indicating increased segmental mobility and expanded fractional free volume. A comparison between freestanding and TFC configurations highlights distinct gas transport behaviors resulting from their structural differences. In freestanding membranes, PS nanoparticles act mainly as bulk microstructural modifiers, increasing fractional free volume and promoting diffusivity across the entire membrane thickness. Consequently, both permeability and selectivity are simultaneously enhanced at moderate filler loadings. Specifically, an optimal PS loading of 10 wt% produced a uniform microstructure and significantly enhanced CO 2 separation performance, achieving a CO 2 permeability of 198.6 Barrer and a CO 2 /N 2 selectivity of 41.4, representing increases of 54% and 31%, respectively, compared with pristine PEBAX. These concurrent enhancements demonstrate that PS nanoparticles can partially mitigate the permeability-selectivity trade-off by introducing controlled free volume without generating nonselective defects. In contrast, in TFC membranes (∼200–250 nm selective layer), gas transport is dominated by coating-induced morphology and interfacial organization near the surface. Shear-assisted nanoparticle alignment and confinement within the ultrathin selective layer create a diffusion-modulating structure. Consequently, permeance remained relatively stable at moderate PS loadings, while selectivity increased due to preferential suppression of N 2 transport. At higher loadings (up to 20 wt%), the well-dispersed and aligned nanoparticles formed a more diffusion-restrictive pathway, increasing CO 2 /N 2 selectivity from 26 to 39 with only a minor reduction in overall permeance. This behavior contrasts with conventional mixed-matrix membranes, where high filler loadings often lead to aggregation and performance deterioration. Therefore, the dominant transport mechanism shifts from bulk free-volume enhancement in freestanding membranes to interfacial diffusion control in TFC configurations. The incorporation of PS nanoparticles offers a versatile and scalable strategy to engineer PEBAX membranes with tunable microstructures and performance. The structure-property-performance relationships elucidated here provide broadly applicable design principles for polymer/nanoparticle membrane systems and introduce shear-induced nanoparticle alignment as a promising paradigm for advanced TFC membranes targeting efficient CO 2 capture and selective gas separations.