Introduction

Membrane-based gas separation processes have gained prominence as CO2 capture technologies due to their simplicity, low energy requirements, modular design, cost-effectiveness, and high efficiency1. Gas diffusion in dense polymeric membranes is typically described by the solution-diffusion mechanism, in which gas molecules first dissolve in the membrane material and then diffuse across it driven by a chemical potential gradient, often approximated by a partial pressure gradient in practice2,3. In the case of nonporous membranes such as PDMS, the flux (J) is proportional to the pressure difference across the membrane. The diffusion step within the membrane is governed by Fick’s law, and is driven by concentration gradients, but the boundary condition across the membrane reflects partial pressure differences2. However, the separation efficiency of polymer membranes is constrained by the Robeson upper bounds, which represent a well-established trade-off between gas selectivity and permeability4,5. This suggests that excellent selectivity and great permeability polymeric membranes are difficult to make. Different approaches, such as thin film composite fabrication (thin film architectures can be paired with selective materials or functional fillers), blending of two polymers having distinct chemical and physical characteristics, and producing mixed matrix membranes (MMMs) have been proposed to address the mentioned problem6,7. MMMs, which are solid fillers embedded within a host polymer matrix attracted attention because of its ease of manufacture and potential for improving separation performance3,8.

PDMS polymer has become one of the most appealing materials to fabricate MMMs for CO2 capture because of its outstanding chemical and physical characteristics6,9. PDMS is intrinsically flexible due to its alternating siloxane (Si-O-Si) backbone, rendering it permeable to gases like CO29,10. PDMS exhibits relatively high thermal stability among rubbery polymers11, so it can be utilized effectively in gas separation operations at higher temperatures. However, the CO2/CH4, and CO2/N2 selectivity through pristine PDMS is not great9,12,13,14. Researchers have introduced a variety of fillers into PDMS to enhance the CO2/CH4 and CO2/N2 selectivity of PDMS membranes. However, the formation of defect-free MMMs with PDMS is highly challenging because higher filler loading leads to polymer-filler agglomeration, creating non-selective voids between the fillers and the polymer matrix6,12,13. These voids allow gas to flow non-selectively through the gaps between the filler and polymer, thereby decreasing the membrane’s gas selectivity15,16.

Compared to alternative materials, silica is a widely used and reasonably priced functional filler in the production of MMMs9. Silica can be categorized as nonporous or porous, depending on its porosity17. Incorporating nonporous FS into the PDMS matrix can potentially increase the CO2 permeability by enhancing the fractional free volume of membranes through interaction with the polymer chain fragments and the nano-fillers, altering the polymer-chain packing13. However, FS exhibits strong hydrophilicity due to an abundance of silanol units present on its outermost layer. The tendency of silica particles to be hydrophilic obstructs their dispersion within the polymer matrix due to the heightened attraction among individual silica particles, surpassing the interaction between the polymer and silica18. This leads to the agglomeration of SiO2 particles within the PDMS matrix and often results in the formation of non-selective voids, which reduce the gas selectivity of the MMMs when the weight% of silica exceeds a certain threshold19. Surface treatment of FS before blending enhances matrix-filler interactions and uniform filler distribution throughout the matrix, thus improving the gas separation capability of MMMs20. This motivates us to fabricate defect-free MMMs by incorporating pre-treated FS into the PDMS polymer matrix. Our goal is to enhance CO2/N2 and CO2/CH4 selectivity, as well as CO2 permeability, surpassing the performance of MMMs developed using silica or other fillers. Gutierrez et al.21 fabricated a PDMS-silica composite membrane, reporting a high CO2 permeability of 8794 Barrer but a relatively low CO2/N2 selectivity of 11.64. Another study22 on dimethyl silane-modified silica/PDMS MMMs demonstrated an increase in CO2 permeability, reaching approximately 5000 Barrer. However, the CO2/CH4 and CO2/N2 selectivities were lower than those of the pristine PDMS membrane. Cao et al. incorporated hollow polyimide nanoparticles into a PDMS matrix, achieving a CO2 permeability of 6639 Barrer at a filler loading of approximately 3%, though the CO2/N2 selectivity remained low at 9.113. Similarly, in a study using magnesium oxide nanosheets in PDMS-based MMMs, the highest recorded CO2/N2 and CO2/CH4 selectivity values were 12.7 and 3.7, respectively, at 2 bar pressure. However, this came at the expense of reduced CO2 permeability, which dropped to 1929 Barrer at an optimal filler loading of 1%6. Gas permeation studies conducted by Silva et al. demonstrated that incorporating multi-walled carbon nanotubes (MWCNTs) into PDMS resulted in a maximum permeability of 1500 Barrer, with CO2/N2 and CO2/CH4 selectivity values of 11.83 and 2.73, respectively, at 1% filler loading23. Similarly, Ahmad et al.24 investigated the effects of incorporating D-MXene and ML-MXene into PDMS, achieving promising CO2 permeability of 13,917 Barrer, and 12,556 Barrer respectively. The enhanced gas transport was attributed to the nanochannels and interfacial free volume created by D-MXene and ML-MXene fillers, as well as favorable interactions at the filler–polymer interface, which facilitated selective CO2 diffusion over N2. However, despite the high permeability, the CO2/N2 selectivity remained relatively low, with values of 13.6 and 12.5 at filler loadings of 1% and 3%, respectively.

In our previous research, we incorporated commercially available PDMS-treated FS (Aerosil R202) into the PDMS matrix to fabricate MMMs for the separation of O2 from N220. In this study, we prepared MMMs by incorporating Aerosil R202 at different loading percentages into the PDMS matrix to separate CO2 from CH4 and N2. To the best of our knowledge, no existing published research documents the use of this MMM for CO2 separation. We aim to bridge the trade-off between permeability and selectivity by optimizing the pre-treated FS filler loading, achieving increased permeability and selectivity of CO2 over N2 and CH4 compared to most other reported studies on PDMS-based MMMs.

Materials and methods

Materials

Sylgard 184 PDMS elastomer, and analytical grade toluene was procured from Sigma-Aldrich St. Louis, MO. PDMS pre-treated FS, Aerosil R202 were obtained from Evonik Industries, Greensboro, NC. The specific surface area (BET), pH in 4% dispersion, and SiO2 content of Aerosil R202 were 80–120 m2/g, 4.0–6.0, and ≥ 99.8% respectively25.

Fabrication of FS/PDMS MMMs

MMMs were fabricated by incorporating varying loading percentages (0, 2.5, 5, 7.5, and 10%) of PDMS-treated FS (Aerosil-R202) into the PDMS matrix using the solution casting method, with slight modifications based on our previous publication (Ogbole et al., 2017). The appropriate amount of FS was mixed with toluene, stirred for 60 min, and degassed for an additional 45 min in an ultrasonic bath to ensure thorough molecular dispersion. Part A of the PDMS elastomer was then added to the dispersion, shaken for 40 min, and degassed again for 30 min. Afterward, elastomer part B (with a weight distribution ratio of A: B = 10:1) was added, followed by 40 min of stirring and 30 min of degassing. The final mixture was dispensed onto stainless steel disks, coated using a spin coater to form thin layers, and dried at room temperature for 24 h. The membranes were subsequently dried under vacuum at 30 °C for 24 h and annealed at 90 °C under vacuum for another 24 h. Finally, the membranes were peeled off and preserved for characterization and gas separation applications. The thickness of the membrane was kept within the range of 0.10 mm to 0.15 mm.

Characterization of FS/PDMS MMMs

FT-IR spectroscopy of the fabricated membranes was performed using Shimadzu IRT racer-100 FT-IR Spectrometer equipped with an ATR to detect the functional groups in the membrane.

The surface morphology of the prepared membranes was examined employing JEOL JSM-IT800 Schottky FESEM Scanning Electron Microscope (SEM) analyzer.

The structure of FS and membranes was studied using X-ray diffractometer (PXRD, SmartLab SE, Japan).

The thermal stability of fabricated membranes was assessed by utilizing TA SDT 650 Thermogravimetric analyzer (TGA) under N2 environment. The heating rate was 10 °C/min, and the temperature range was 25 °C to 950 °C.

Gas transport properties

The CO2 gas separation performance was evaluated by measuring the permeability of CO2, CH4, and N2 gases through the fabricated membranes (Fig. 1), along with the selectivity of CO2/N2 and CO2/CH4 gas pairs. The single gas permeability tests were carried out on membrane areas of 13.8 cm² at room temperature under feed pressures of 20 psig. Gas flow rates were recorded using a mass flow meter from Alicat Scientific Company. The gas flux and permeability were calculated using the following equations20.

The gas flux, \(\:\text{Q}=(\text{V}/\text{A}\text{t})\)

Where, \(\:\text{Q}\) represents the flux of the gas (cm3/cm2*s), \(\:\text{V}\) denotes the permeation gas’s volume (cm3), \(\:\text{A}\) indicates the membrane active area (cm2), and \(\:\text{t}\:\)refers the time interval (s).

The permeability coefficient, \(\:\text{P}=\:\text{Q}\text{l}/{\Delta\:}\text{P}=\text{D}\text{S}\)

Where, \(\:\text{P}\) represents the gas’s permeability coefficient in the polymer in Barrer ((10–10 cm3(STP)-cm /cm2 s cm Hg), \(\:\text{l}\) signifies the membrane’s thickness (cm), \(\:{\Delta\:}\text{P}\) indicates the Difference in pressure across the membrane (psig), \(\:\text{D}\) refers the diffusivity, \(\:\text{S}\) denotes the solubility of the gas in the polymer.

The selectivity of gases is determined by the ratio of the permeabilities of the two gases.

Selectivity of gas A and B, \(\:{\alpha\:}_{A∕B}={P}_{A}∕{P}_{B}\)7,8.

Where, \(\:{P}_{A}\) represents the permeability of gas A and \(\:{P}_{B}\) refers the permeability of gas B through the membrane.

Fig. 1
figure 1

Schematic representation of single gas permeation test.

Full size image

Results and discussion

Characterization of membranes

FT-IR analysis

Fig. 2
figure 2

FT-IR spectra of of FS, neat PDMS, and MMMs containing different wt% of FS loading.

Full size image

The consequence of the incorporation of FS on the functional groups of the mixed matrix membrane was analyzed using FT-IR spectroscopy (Fig. 2).

The pristine PDMS membrane and MMMs exhibit a small peak at 2962 cm− 1, represents asymmetric stretching vibration band arising for the -CH324. The intense peaks at 1257 cm− 1 is for deformation of the -CH3 group of PDMS6. The two peaks between 1100 and 1000 cm− 1 Si(CH3) attributed to bending vibration and Si-O stretching vibration, respectively26. Moderate peaks at 783 cm− 1 are corresponding to stretching vibration of Si-C24. PDMS treated FS shows peaks at 807 cm− 1 and 1035 cm− 1, which are due to stretching vibration of Si − O−Si bonds19,27,28,29.

Pure FS typically exhibits broad hydroxyl (–OH) stretching vibration peaks in the FTIR spectrum between 3200 and 4000 cm⁻¹, as reported by Lin et al.30 and Li et al.27. In this study, we used FS that had been pre-treated with PDMS by the manufacturer. The absence of –OH stretching peaks in the FTIR spectrum of the treated FS confirms the successful reaction of surface hydroxyl groups with the silicon atoms of the PDMS siloxane backbone20. Surface –OH groups on untreated FS promote SiO2–SiO2 interactions, leading to particle agglomeration20. However, PDMS surface treatment effectively removed these groups, thereby reducing agglomeration and promoting uniform dispersion of FS particles within the PDMS matrix. This improved dispersion was confirmed by FESEM analysis.

After incorporating the surface-treated FS into the PDMS matrix during membrane fabrication, the resulting FTIR spectrum closely resembled that of neat PDMS, showing no significant spectral changes. This is expected, as both the matrix and the filler share the same PDMS chemical structure, resulting in overlapping signals and minimal distinguishable features from the filler.

Investigation of membrane morphology

Fig. 3
figure 3

FESEM images of surface of (a) neat PDMS, (b) 2.5% FS/PDMS, (c) 5% FS/PDMS, (d) 7.5% FS/PDMS, and (e) 10% FS/PDMS.

Full size image

The surface and cross-sectional morphologies of the prepared membranes were analyzed using FESEM (Figs. 3 and 4) to examine the distribution of PDMS-treated FS within the polymer matrix. The neat PDMS membrane exhibited a smooth, homogeneous, and compact surface and cross-section (Figs. 3 and 4a). In the MMMs with 2.5, 5, 7.5, and 10 wt% FS loading, both surface and cross-section appeared uniform and free of visible defects, indicating a homogeneous dispersion of the filler throughout the PDMS matrix (Figs. 3 and 4b–e). Such uniform distribution is advantageous for enhancing gas selectivity and permeability24. The pretreatment of FS by the manufacturer using PDMS reduced filler–filler (FS–FS) interactions while promoting stronger filler–matrix interactions, contributing to the improved compatibility and dispersion of FS in the PDMS phase. Some foreign particles are visible in the cross-sectional images of all membranes (Fig. 4), which may have been introduced inadvertently during sample preparation or handling.

Fig. 4
figure 4

FESEM images of cross-section of (a) neat PDMS, (b) 2.5% FS/PDMS, (c) 5% FS/ PDMS, (d) 7.5% FS/PDMS, and (e) 10% FS/PDMS.

Full size image

XRD analysis

Figure S1 shows the XRD patterns of neat PDMS, MMMs, and PDMS treated FS. A polymer’s X-ray diffraction peak is often sharp and intense when it comprises a significant crystalline region, whereas an amorphous polymer’s peak is more broadly distributed7. The XRD patterns of the FS nanoparticles displayed a broad peak near 21.1o, which suggested the nanoparticles’ amorphous nature31. Similarly, the broad peaks observed in the XRD patterns of the PDMS membrane and all the MMMs suggest that these materials are largely amorphous. Since gas molecules typically cannot diffuse through crystalline regions of a polymer, a higher proportion of amorphous regions is advantageous for enhancing gas permeability32.

Thermogravimetric analysis of membranes and PDMS treated FS

Fig. 5
figure 5

Thermogravimetric analysis (TGA) of surface treated FS, neat PDMS, and MMMs containing different wt% of FS loading.

Full size image

TGA was done to assess the thermal stability of neat PDMS, and FS/PDMS MMMs by analyzing thermal decomposition (Fig. 5). The total weight loss of FS was 11.02%, possibly due to the presence of PDMS on the SiO2 surface during surface treatment by the manufacturer and trace contamination from the atmosphere22. All membrane TGA thermographs exhibited a single significant weight reduction. In neat PDMS, thermal degradation initiated around 230 °C, with substantial weight loss observed between 400 and 580 °C. In contrast, for the MMMs, thermal degradation started at approximately 300 °C, with the most significant decomposition occurring between 400 and 650 °C. The weight loss of neat PDMS was 81.47% with 18.53% char residue, which can be attributed to the breakdown of methyl groups on the Si-O skeleton20. MMMs with 2.5, 5, 7.5 and 10 wt% of FS exhibited a weight loss of 77.86, 65.46, 64.01 and 53.02%, respectively.

The residual weights of all MMMs were significantly higher than the sum of the original silica content added to the PDMS matrix and the residual weight of the neat PDMS membrane. In the absence of any interaction between the components, the organic portion of the FS/PDMS composite would be expected to decompose entirely, leaving only this combined residue. The higher observed residual weights indicate that silica influences the thermal degradation behavior of the PDMS matrix. Due to its excellent thermal insulation properties, silica acts as a barrier to mass transfer, thereby retarding the release of volatile decomposition products and enhancing the thermal stability of the composite33. This results in improved thermal stability of the MMMs, indicating a meaningful interaction between the PDMS chains and the silica particles.

Analysis of gas separation performance for CO2

Effect of PDMS treated FS loading

Table 1 Single gas (CO2, N2, and CH4) permeability and selectivity (CO2/N2, and CO2/CH4) of FS/PDMS membranes at 20 Psig pressure and at room temperature at different wt% of filler loading.
Full size table
Fig. 6
figure 6

Permeability of different gases (CO2, CH4, and N2) and selectivity of CO2/CH4 and CO2/N2 through the membranes at different weight percentages (wt%) of FS loading.

Full size image

Single gas permeation experiments for CO2, CH4, and N2 were conducted across the neat PDMS membrane and mixed matrix membranes (MMMs) containing FS loadings of 2.5, 5, 7.5, and 10 wt%. The tests were carried out at ambient temperature and a constant pressure of 20 psig to evaluate the effect of incorporating of filler (PDMS-treated FS) into the PDMS matrix on the permeability behavior of the gases. The permeability order of gases within the membranes is CO2 > CH4 > N2, which is comparable to the majority of polymeric gas separation membranes3,22. The permeability test results and selectivity of CO2/N2, and CO2/CH4 are shown in Table 1, and Fig. 6. The permeability ascended with the condensability of the penetrant gas (i.e. CO2) and decreased with increasing the kinetic diameter (i.e., N2)3,34. The permeability of CO2 (2889 Barrer) across neat PDMS in this study is silimar to the literature15, to some extent higher than that reported in literatures6,12,35,36 and slightly lower in comparison to literature13. This discrepancy may occurr because of the variation in solvents, amount of crosslinker, and the curing temperature6,24,37. The selectivity values of CO2/N2 and CO2/CH4 through neat PDMS are 10.66 and 3.47, respectively, which is very close to most published articles. The permeability of CO2 and selectivity of CO2/N2 and CO2/CH4 improved, upon the incorporation of filler up to 7.5 wt%. The permeability of CO2 increased by 1.42 times, from 2889 Barrer to 4105 Barrer, while the selectivity of CO2/N2 and CO2/CH4 increased by 1.47 times (from 10.66 to 15.67) and 1.39 times (from 3.47 to 4.84), respectively, when PDMS was filled with 7.5 wt% filler compared to the neat PDMS membrane. After the addition of 10 wt% filler, the permeability of CO2 further increased to 4431 Barrer; however, the selectivity of CO2/N2 and CO2/CH4 were slightly lower than that observed at 7.5 wt% filler loading.

The surface treatment of hydrophilic FS with PDMS by the manufacturer renders it hydrophobic, thereby enhancing its compatibility with the nonpolar PDMS matrix. This modification reduces interfacial voids and promotes uniform dispersion of the filler. The hydrophobic interactions, along with partial penetration of PDMS chains into the porous structure of the FS, alters the molecular packing of the PDMS matrix, increasing the fractional free volume of the membrane, thereby boosting the permeability of CO2 gas6,13,19. These structural features facilitate the faster diffusion of small, condensable gases such as CO228. Typically, untreated FS can interact preferentially with CO2 due to the presence of polar surface hydroxyl groups, which enhances its solubility selectivity over less soluble gases like N2 or CH4. Although surface treatment reduces the number of hydroxyl groups, the modified filler can still retain some affinity for CO2 through physisorption38. Furthermore, at optimal loading levels, FS increases the free volume within the polymer matrix without compromising the continuity of the polymer chains, thereby improving CO2 diffusivity. Simultaneously, well-dispersed or densely packed silica particles create a more tortuous pathway for larger or less soluble gases, enhancing diffusivity-driven selectivity39.

In many MMMs, gas selectivity tends to decrease at higher filler loadings due to agglomeration and clustering of nanoparticles within the polymer matrix, which can lead to the formation of non-selective voids3,12. These voids increase the permeability of all gases, not just CO2, thereby reducing overall selectivity. However, in our case, FESEM images of 10 wt% FS/PDMS membranes confirmed the absence of agglomeration or clustering, indicating that the membranes remained nearly defect-free. Therefore, the slight decrease in CO2/N2 and CO2/CH4 selectivity observed at 10 wt% filler loading compared to 7.5 wt% is likely due to experimental variation rather than structural defects.

Overall, the 7.5 wt% FS loading on MMM exhibits improved gas separation efficiency concerning permeability and selectivity than all other MMMs. The other MMMs showed higher permeability, and selectivity for CO2/N2 and CO2/CH4 gas pairs compared to neat PDMS, but it was lower than the 7.5wt.% FS/PDMS MMM. Table 2 provides a comparative evaluation of the CO2 separation performance of the developed membrane (optimum loading of FS) over N2, and CH4, in relation to the results reported in other published studies. Our test results demonstrated higher selectivity for both CO2/N2 and CO2/CH4 gas pairs compared to the cited studies in Table 2, while maintaining satisfactory CO2 permeability.

Table 2 Comparative analysis of the separation performance of various PDMS membranes for CO2/N2 and CO2/CH4 gas pairs.
Full size table

Consequences of variable feed pressure and time

Based on the preceding section, the 7.5 wt% filler loading, which exhibited the most favorable separation performance, was selected to investigate the influence of feed pressure on gas separation performance in this section. Figure 7(a) presents the CO2 permeability and the ideal selectivity for CO2/N2 and CO2/CH4 of the neat PDMS membrane under varying upstream pressures (20, 22, 26, and 30 psig). The CO₂ permeability of the neat membrane remained stable at approximately 2900 Barrer, and the selectivities for both CO2/N2 and CO2/CH4 exhibited negligible changes with increasing pressure. In contrast, Fig. 7(b) illustrates the performance of the membrane containing 7.5 wt% filler. The CO2 permeability remained consistently within the range of 4105 to 4126 Barrer across the tested pressure range. While the CO2/CH4 selectivity remained nearly constant, a slight enhancement in CO2/N2 selectivity was observed with increasing feed pressure. These results indicate that the mixed matrix membrane maintains stable gas separation performance under varying operating pressures, particularly at the optimal filler loading. Permeability is an inherent characteristic of the membrane and does not change unless the membrane undergoes plasticization6. Since PDMS does not exhibit plasticization characteristics, the permeability of these membranes remained unchanged.

We also performed a single-gas permeability test on the membrane containing 7.5 wt% FS loading over a duration of 80 h at 20 psig and room temperature (Fig. 8). For comparison, Wang et al.12 previously conducted a 100-hour single-gas permeability test using a metal-organic framework/PDMS mixed matrix membrane (MMM) at 2 bar and room temperature, where their membrane exhibited stable CO2 and N2 permeability. Similarly, Xu et al.40 evaluated the long-term performance of a layered LDH/FAS–PDMS hybrid membrane over 120 h at 27 kPa and 298 K. Their study reported high and stable separation performance for CO2/H2, CO2/N2, and CO2/CH4 without any significant degradation. Consistent with these findings, our membrane maintained nearly stable permeability for CO2, N2, and CH4 gases throughout the 80 hour testing period, confirming its long-term structural stability.

Fig. 7
figure 7

The permeability of different gases (CO2, CH4, and N2) and the selectivity of CO2/N2 and CO2/CH4 through (a) neat PDMS membrane and (b) PDMS membranes with 7.5 wt% FS loading at different feed pressures.

Full size image
Fig. 8
figure 8

The permeability of different gases (CO2, CH4, and N2) and the selectivity of CO2/N2 and CO2/CH4 through FS/PDMS membranes containing 7.5 wt% FS loading at different times at 20 psig pressure and room temperature.

Full size image

Conclusions

In this work, mixed matrix membranes (MMMs) were synthesized using PDMS-treated FS at varying loading percentages in the PDMS matrix for CO2 separation from CH4, and N2. FT-IR spectroscopy confirmed the disappearance of surface hydroxyl groups on FS, as they were effectively utilized to form bonds with the atomic silicon of the siloxane unit in PDMS during surface treatment. This indicates the proper distribution of FS particles within the PDMS matrix. SEM studies of the prepared MMMs confirmed the effective incorporation of FS into the PDMS matrix. TGA results demonstrated that FS/PDMS MMMs exhibited better thermal stability than neat PDMS, indicating strong interactions between the polymer and FS. The gas permeability characteristics of MMMs were analyzed using CO2, CH4, and N2 to evaluate the effect of FS inclusion. Incorporating PDMS-treated FS into the PDMS matrix up to 7.5 wt% resulted in an increase in CO2 permeability and improved selectivity for CO2/N2 and CO2/CH4. With the addition of 7.5 wt% filler, CO2 permeability increased by 1.42 times, while the selectivity of CO2/N2 and CO2/CH4 improved by 1.47 and 1.39 times, respectively. At 10 wt% filler loading, CO2 permeability further increased; however, the selectivity for CO2/N2 and CO2/CH4 slightly decreased, though it remained higher than that of the neat PDMS membrane. Overall, the 7.5 wt% FS/PDMS MMM exhibited the best gas separation performance in terms of both permeability and selectivity. This study also indicates that change of feed pressure and time had no significant effect on separation performance.