Ether-functionalized ionic liquid based composite membranes for carbon dioxide separation

Abstract

The efficient separation of CO₂ from other light gases has received growing attentions due to its importance in reducing greenhouse gas emissions and applications in gas purification. In this work, we developed a series of composite membranes composed of ether-functionalized pyridinium-based ionic liquids ([E_nPy][NTf₂]) and cellulose acetate (CA) polymer matrices to improve CO₂ separation performance. CA + [E_nPy][NTf₂] and CA + [C_nPy][NTf₂] composite membranes were fabricated by a casting method. The CO₂, N₂ and CH₄ permeabilities of the CA + IL composite membranes were measured, and the CO₂/N₂ and CO₂/CH₄ permselectivities were further calculated. The results showed that the CA + 40 wt% [E₁Py][NTf₂] composite membrane exhibits approximately a seven-fold increase in CO₂ permeability with CO₂/N₂ and CO₂/CH₄ permselectivities of 32 and 24, respectively. The characterization results showed that the mechanical properties and thermal stabilities of the CA + [E₁Py][NTf₂] composite membranes are affected by both plasticizing effect and affinity of the ILs for the gases, which also lead to the changes in the CO₂/N₂ and CO₂/CH₄ permselectivities. Compared with membranes containing the non-functionalized analogues [C_nPy][NTf₂], the addition of [E_nPy][NTf₂] improves the ideal permselectivities of CA + IL composite membranes, whereas it decreases slightly the gas permeabilities.

---

1. Introduction

There are growing concerns that increase of greenhouse gas levels (mainly CO₂, followed by CH₄ and N₂O) in the atmosphere is the primary cause for climate change. The CO₂ concentration in air has increased from the preindustrial level of 280 ppm ¹ to the present level of 400 ppm,² as a result of human activities. This growth has led to an increase in sea level and exacerbated natural disasters, e.g. drought, floods and hurricanes. There are many ways to reduce CO₂ emission, including improving energy conversion efficiency, replacing fossil energy with renewable energy, and CO₂ capture and sequestration (CCS). Since fossil fuels are predicted to be the major energy source for the next several decades, the most promising way to decrease CO₂ concentration is CCS.³ In CCS, the most critical part is CO₂ separation, which is very important not only for reducing greenhouse gas emission but also for industrial gas purification, such as CO₂ removal from natural gas, biogas and syngas. At present, CO₂ separation methods mainly include absorption, cryogenic distillation, adsorption, and membrane separation. Among these methods, membrane separation is a promising technology for CO₂ capture due to lesser equipment space required, lower energy costs and more environmentally friendly; however, the trade-off between permeability and permselectivity of gas separation membranes, known as the “Robeson upper bound”, limits large-scale applications of membranes.^3,4 As another prospective technology, ionic liquids (ILs) have attracted substantial attentions owing to their special properties, such as good thermal stability, negligible vapor pressure, tunability for targeted applications, and in particular the inherent solubility for CO₂.^5–9 However, the high viscosity and high cost of ILs limit their use in future industrial applications. To combine the advantages of membranes and ILs and overcome their disadvantages for CO₂ capture, immobilization of ILs in membranes has been extensively investigated.^{5,6,8,10–14}

Initially, supported ionic liquid membranes (SILMs) were mostly investigated due to their high permselectivities and permeabilities.¹² However, a main drawback is the leaching of ILs from the support membranes under a high cross-membrane pressure difference.^12,15 Furthermore, several researchers reported that classical SILMs have almost reached their gas separation performance limits.^16,17 Therefore, other IL membranes have been developed, such as p(RTIL) dense membranes,^18,19 mixed-matrix membranes^20,21 and IL-based composite membranes.^{2,6,10,11,22,23} In IL-based composite membranes, ILs are confined in polymer matrices to prevent the loss of ILs under high pressure operation. The addition of ILs in composite membranes has been applied to CO₂/N₂ and CO₂/CH₄ separation system for enhancing the separation performance of polymeric membranes. The results indicated that the gas separation performance of IL-based composite membranes improves compared with that of polymer membranes. Chen et al.⁶ reported that the permeability of CO₂ in a PVDF+[emim][BF₄] composite membrane showed a 4-fold increase with higher CO₂/N₂ permselectivity, compared with that of pure PVDF membrane. Hong et al.²³ prepared a PVDF-HFP+[emim][BF₄] gel membrane, which exhibited excellent gas separation performance over the Robeson upper bound. Jansen et al.²⁴ used [emim][NTf₂] and PVDF-HFP to fabricate a gel membrane, and the permeability rapidly increased with the increase of [emim][NTf₂] content. It was found that with the increasing of IL content, the gas separation process became solubility-controlled process instead of diffusion-controlled. Noble and his group^19,25 studied poly(RTIL)–RTIL membranes and observed that the CO₂ permeability of composite membranes is 3–6 times higher than that of the corresponding poly(RTIL) membranes.

However, these reported IL-based composite membranes were mainly prepared using conventional imidazolium-based ILs, and their gas separation performances were still relatively low. Task-specific ionic liquids (TSILs), also called functionalized ionic liquids, can improve the CO₂ absorption capacity and the absorption selectivity by introducing suitable moieties (like amine,^26,27 ether,²⁸ and nitrile groups²⁹) to ILs. The results^28,30 have indicated that ether-functionalized ILs could improve the CO₂/N₂ and CO₂/CH₄ selectivities for CO₂ separation. Moreover, ILs with pyridinium cations are relatively cheaper, more biodegradable, and less toxic compared with imidazolium-based ILs.^30–32 Considered the hydroxyl groups and the ether groups in cellulose acetate (CA), ether-functionalized pyridinium-based ILs should be compatible with CA, which can be beneficial for improving CO₂ separation performance.

In this work, a series of composite membranes were prepared by CA and four ILs, including two ether-functionalized pyridinium-based ILs ([E_nPy][NTf₂], n = 1, 2) and their analogues ([C_nPy][NTf₂], n = 1, 2). CA was chosen as the host polymer for membrane fabrication due to its low cost, good strength, and ease of handling.³³ The physicochemical properties and microstructure of the composite membranes were characterized, including the mechanical property and thermal stability. To determine the performance of the CA + IL composite membranes for gas separation, N₂, CH₄ and CO₂ permeabilities were tested, and the CO₂/N₂ and CO₂/CH₄ permselectivities were calculated.

2. Experimental

2.1 Materials

Four ILs, i.e., 1-butylpyridinium bis(trifluoromethane)sulfonamide ([C₄Py][NTf₂]), 1-[2-methoxy-ethoxyethyl]-pyridinium bis(trifluoromethane)sulfonamide ([E₁Py][NTf₂]), 1-heptylpyridinium bis(trifluoromethane)sulfonamide ([C₇Py][NTf₂]), and 1-[[2-(2-methoxyethoxy)-ethoxyethyl]pyridinium bis(trifluoromethane)sulfonamide ([E₂Py][NTf₂]), were synthesized in our laboratory using reported methods³¹ and were dried in a vacuum oven prior to use. CA (39.8 wt% of acetylation content) was purchased from Sigma-Aldrich. Acetone (99.5%) was purchased from Beijing Chemical Company. Scheme 1 shows the chemical structures of [E_nPy][NTf₂], [C_nPy][NTf₂], and CA. CO₂ (99.99%), CH₄ (99.99%), and N₂ (99.99%) were supplied by Beijing Beiwen Gas Factory.

Scheme 1 The chemical structures of (a) [E_nPy][NTf₂], (b) [C_nPy][NTf₂] and (c) cellulose acetate.

2.2 Membrane fabrication

The IL composite membranes were prepared using the following procedures. CA was first dissolved in acetone. A predetermined amount of IL was then added to the above solution and stirred at room temperature until the solute was dissolved. The solute ratio of the solution was maintained at 1 : 10 (g mL⁻¹). Subsequently, the solution was casted onto a glass plate, and the solvent was allowed to evaporate under constant temperature and humidity (25 °C, 15% RH). The membrane was then transferred to an oven and dried at 30 °C under vacuum for 48 h to remove the remaining solvent completely. The dried membrane was employed for further characterization. The IL content was from 0 wt% to 40 wt% because the composite membranes became too flexible to be used when the IL content exceeded 40 wt%. At least 3 membranes of each type were prepared. The thicknesses of all membranes were between 40 and 50 μm, as measured by a thin film thickness measuring instrument (CH-1-ST, Wuxi Qianzhou Measuring Instrument Factory).

In this study, the composite membranes are referred to as CA + XX IL, where IL denotes the ionic liquid, and XX refers to the mass fraction of IL in the composite membrane.

2.3 Membrane characterization

FTIR spectra were obtained in the range of 500−4000 cm⁻¹ on a Thermo Nicolet 380 spectrometer. Mechanical properties were carried out by the module of a Dynamic Mechanical Analyzer (DMA Q800), using a tension membrane geometry under isothermal temperature condition (35 °C). These experiments were performed using a control rate force (2.0 N min⁻¹) with an upper force limit of 20.0 N in films with rectangular shapes. The decomposition temperatures of membranes were tested by thermogravimetric analysis (Q5000 V3.15 Build 263) from room temperature to 700 °C at a heating rate of 10 °C min⁻¹ under N₂ atmosphere. Differential scanning calorimetry was carried out on a Mettler Toledo DSC1 instrument between −100 and 400 °C at a heating rate of 10 °C min⁻¹ under N₂ atmosphere. The morphology of membranes was examined using a JEM-7001F. Dried membranes were broken under liquid nitrogen to obtain a membrane cross-section. The samples were coated with a platinum layer using a JEC-3000FC auto fine coater. Picture of atomic force microscope was conducted using Bruker FastScan™.

2.4 Gas permeability measurement

The pure gas permeability (P), diffusivity coefficient (D) and solubility coefficient (S) of all membranes were determined by a permeability instrument (VAC-V2, Labthink), using a constant volume and variable pressure method. The temperature of the membrane cell was 23 °C maintained by a water bath. The feed pressure was approximately 100 kPa, and permeate pressure was in the range of 0 to 10 Pa. The tested samples were degassed before each experiment to remove residual gas. The tested gases were examined in the following order: N₂, CH₄ and CO₂. All membranes were measured at least 3 times. The permeability coefficient was calculated by the equation:where P (Barrer) is the permeability of a membrane (1 Barrer = 1 × 10⁻¹⁰ cm³ (STP) cm (cm² s⁻¹ Pa)), dp/dt (Pa s⁻¹) is the pressure increase in time t at the steady state, V (cm³) is the downstream volume, A (cm²) is the effective membrane area, T (K) is the temperature, L (cm) refers to the thickness of a membrane, T₀ and p₀ is the temperature and pressure at standard conditions (273.15 K, 1.0133 × 10⁵ Pa), and the Q (cm² (m² d⁻¹ Pa)) is the permeation flux.

The D was determined by the time-lag θ method:

D = L²/6θ (2)

According to the solution-diffusion mechanism, the S can be calculated:

S = P/D (3)

The ideal permselectivity, α_ij, was calculated by the ratio of the permeabilities of gas i and gas j.

α_ij = P_i/P_j (4)

3. Results and discussions

3.1 Membrane characterization

3.1.1 FTIR characterization. Fig. 1 shows the FTIR curves of the pure CA membrane, pure [E₁Py][NTf₂] and CA + 40[E₁Py][NTf₂] composite membrane. The spectrum of the pure CA membrane prepared in this work shows good agreement with a previous study.³⁴ For the CA + 40[E₁Py][NTf₂] membrane, several peaks are observed corresponding to the IL: 1637 cm⁻¹, 1491 cm⁻¹, 1352 cm⁻¹ (pyridine ring stretch) and 1136 cm⁻¹ (C–N bending). The results indicated that no peak is shifted or generated, which means there are no chemical interaction between CA and [E₁Py][NTf₂]. Therefore, ILs might be physically dispersed in the CA matrix.

Fig. 1 The FTIR spectra of the pure IL, the pure CA membrane and composite membranes.

3.1.2 Dynamic mechanical thermal analysis (DMA). Considering that the composite membranes are free-standing, the mechanical properties of the membranes are critical for industrial applications. The packing of CA chains will be changed by the introduction of ILs into membranes during the membrane preparation process.³⁵ The stress–strain curves obtained from DMA contain important information on mechanical properties, such as the Young's modulus (slope of the linear region of the plot) and stress versus strain at break. Fig. 2 shows the stress–strain curves for the composite membranes and the pure CA membrane, and Table 1 summarizes the Young's modulus and the tension at break of composite membranes and the pure CA membrane calculated from Fig. 2. It can be seen that the stress–strain curves are similar for the composite membranes with different IL content compared with the pure CA membrane. However, the Young's modulus and tension at break of the composite membranes decrease significantly compared with the pure CA membrane. For example, the CA + 40[E₁Py][NTf₂] composite membrane shows a low Young's modulus of 426.2 MPa and tension at break of 18.94 MPa, whereas the pure CA membrane displays a Young's modulus of 3181 MPa and tension at break of 68.87 MPa. These results suggest that [E₁Py][NTf₂] acts as a plasticizer in the composite membranes due to the increase of elasticity and decrease of tenacity and the Young's modulus.

Fig. 2 The stress–strain curves of the composite membranes and the pure CA membrane.

Table 1 Mechanical properties of the pure CA membrane and the composite membranes

Membrane	Young's modulus (MPa)	Tension at break (MPa)
CA	3181	68.87
CA + 20[E1Py][NTf2]	627.9	20.48
CA + 40[E1Py][NTf2]	426.2	18.94

---

3.1.3 Differential scanning calorimetry (DSC). The glass transition temperature (T_g) is a critical parameter of polymeric material, since it represents the transition between the glassy to the rubbery. The T_g values of pure CA membrane, the CA + [E₁Py][NTf₂] composite membranes and pure [E₁Py][NTf₂] are presented in Fig. 3. It is shown that the T_g decreases monotonically with the increase of [E₁Py][NTf₂] content. This decrease suggests that in CA + [E₁Py][NTf₂] composite membrane, the CA chain segment becomes more flexible and the intermolecular forces of CA chains are much weaker than those in the pure CA membrane. This proves the [E₁Py][NTf₂] is a plasticizer in the composite membrane, which is consistent with the DMA results.

Fig. 3 The relationship between T_g and [E₁Py][NTf₂] content in the CA + [E₁Py][NTf₂] composite membranes.

3.1.4 Thermogravimetric analysis (TGA). Fig. 4 shows the TGA curves of the pure CA membrane, pure [E₁Py][NTf₂] and CA + [E₁Py][NTf₂] composite membranes. Here, the pure CA membrane and [E₁Py][NTf₂] display one-step thermal decomposition behavior, as shown by the decomposition temperature at approximately 300 °C and 370 °C. However, the CA + [E₁Py][NTf₂] composite membranes show two-steps of decomposition: the first one is at approximately 300 °C based on CA, and the second is at 360 °C based on IL. Because there is no chemical interaction between CA and [E₁Py][NTf₂], the physicochemical property of the composite membranes, e.g. the decomposition temperature, should be in between that of the pure CA membrane and pure [E₁Py][NTf₂]. However, the decomposition temperatures T_d (5 wt% loss), of the composite membranes with low IL content are slightly lower than those of CA and [E₁Py][NTf₂], as shown in Table 2. Considering the [E₁Py][NTf₂] acts as a plasticizer, the reason for the decrease of T_d seems to be the plasticizing effect. When the membranes contain more ILs, the T_d increases because [E₁Py][NTf₂] has a higher decomposition temperature. Nevertheless, all composite membranes are stable at a temperature up to 300 °C, which is higher than the temperature in most CO₂ separation processes.

Fig. 4 Thermogravimetric curves of CA and CA + [E₁Py][NTf₂] composite membranes.

Table 2 The decomposition temperature of pure CA membrane, CA + [E₁Py][NTf₂] composite membranes and pure [E₁Py][NTf₂]

IL content (wt%)	CA	10	20	30	40	[E1Py][NTf2]
Td (°C)	299.8	294.0	292.5	290.9	291.3	370.6

---

3.1.5 Scanning electron microscope (SEM). The SEM images of the CA + [E₁Py][NTf₂] membranes and the pure CA membrane are presented in Fig. 5. It is shown that the pure CA membrane has a homogeneous and smooth surface, and more wrinkles are observed in the composite membranes as the IL content increased. In the cross-sectional SEM images, composite membranes are smoother compared with the CA membrane, suggesting that the IL may wet the polymer chains and fill the voids in CA.

Fig. 5 SEM images of (a) the surface of pure CA membrane (magnification ×8500); (b) the cross-section of pure CA membrane (magnification ×1800); (c) the surface of CA + 20[E₁Py][NTf₂] composite membrane (magnification ×8500); (d) the cross-section of CA + 20[E₁Py][NTf₂] composite membrane (magnification ×1200); (e) the surface of CA + 40[E₁Py][NTf₂] composite membrane (magnification ×7500); and (f) the cross-sections of CA + 40[E₁Py][NTf₂] composite membrane (magnification ×1200).

3.1.6 Atomic force microscope (AFM). The surfaces of the composite membranes and the pure CA membrane are observed by AFM, as shown as Fig. 6. The surface roughness parameter, R_q (the root mean square average roughness), of the membranes were calculated according to the AFM results. The curve in the AFM image indicates the fall and rise of the line. Here, it is observed that the smoothness of the membrane surface changes with [E₁Py][NTf₂] content increasing. When the membrane contains a low ratio of IL, the composite membranes become rougher with a higher R_q. And when the IL content is more than 20 wt%, the roughness of the composite membrane decreases with increasing IL content. The fluctuation of the curve is consistent with the roughness. It can be inferred that the incorporation of IL increases the surface roughness, whereas [E₁Py][NTf₂] smoothens the surface due to the liquid nature of ILs in high IL content membranes.

Fig. 6 AFM images of (a) pure CA membrane, (b) CA + 20[E₁Py][NTf₂] composite membranes and (c) CA + 40[E₁Py][NTf₂] composite membranes with a scanning area of 1 μm × 1 μm, (d) bar graph of roughness of membranes.

3.2 Gas separation performance

3.2.1 Effect of ionic liquid content. Pure gas permeabilities of the pure CA membrane and CA + [E₁Py][NTf₂] composite membranes were measured, and ideal gas permselectivities were calculated. Fig. 7 illustrates the diffusivity coefficient, solubility coefficient, and pure gas permeability of the composite membranes as a function of the IL content. The N₂, CH₄, CO₂ permeabilities of pure CA membranes are 0.2081, 0.1685 and 5.294 Barrer, respectively, which is consistent with the result of Li.³⁶

Fig. 7 The influence of [E₁Py][NTf₂] content on (a) diffusivity coefficient, (b) solubility coefficient, and (c) pure gas permeability of CA + [E₁Py][NTf₂] composite membranes.

The diffusivity coefficients substantially rise with the increase of IL content. The increase of the gas diffusivity is probably related to the increase of voids in the CA + [E₁Py][NTf₂] composite membranes. It is supposed that ionic bonds in the IL molecules may have some negative impacts on the intermolecular forces of CA chains and lead to the increase of more voids, thus causing the gas molecule diffusion to be enhanced in the composite membranes. On the other hand, as the IL content increases, the solubility coefficients of all gases decrease. For example, the CO₂ solubility in the CA + 40[E₁Py][NTf₂] membrane is approximately 81% lower than in pure CA membrane. This is probably because the CA is plasticized by well-dispersed ILs, which leads to the unrelaxed volume of CA. However in Fig. 6(c), gas permeabilities increase significantly with the increase of IL content, especially for CO₂, showing approximately a 7 fold increase, which suggests that the permeabilities of CA + [E₁Py][NTf₂] composite membranes depend mainly on the gas diffusivity instead of the solubility. Similar results have been observed by Kanehashi⁵ and Chen.⁶

The CO₂/CH₄ and CO₂/N₂ ideal permselectivities of the CA + [E₁Py][NTf₂] composite membranes are shown in Fig. 8. When the [E₁Py][NTf₂] content increases to 20%, the ideal permselectivities for CO₂/N₂ and CO₂/CH₄ decrease to 23.83 and 24.74, respectively. Considering the CO₂ permeability is an order of magnitude larger than the N₂ and CH₄ permeabilities, [E₁Py][NTf₂] has a minor impact on the CO₂ permeability compared with the N₂ and CH₄ permeabilities when the [E₁Py][NTf₂] content is low. This leads to a proportional increase of N₂ and CH₄ permeabilities which is more remarkable than that of the CO₂ permeability eventually resulting in the decrease of CO₂/N₂ and CO₂/CH₄ permselectivities. When the [E₁Py][NTf₂] content increases, the CO₂/CH₄ permselectivity decreases, but CO₂/N₂ permselectivity increases. This difference may be due to the different affinity of [E₁Py][NTf₂] for CH₄ and N₂. Bara et al.²⁸ reported that ether-functionalized ILs has a better affinity for CH₄ than N₂, which led to a higher CO₂/N₂ selectivity compared with CO₂/CH₄ selectivity. And the CO₂/CH₄ selectivity of [E₁Py][NTf₂] is 16.9 at 313.15 K, which is much lower than that of pure CA membrane.³⁰ When [E₁Py][NTf₂] content increases in the composite membranes, the ideal permselectivities for CO₂/N₂ and CO₂/CH₄ would be similar to those of [E₁Py][NTf₂].

Fig. 8 The influence of ionic liquid content on the ideal permselectivity of CA + [E₁Py][NTf₂] composite membranes.

3.2.2 Gas separation performance of CA + IL composite membranes. CO₂ permeabilities of ether-functionalized IL composite membranes are compared with their non-functionalized pyridinium analogues, i.e., CA + [C_nPy][NTf₂] composite membranes, as shown in Fig. 9. It can be seen that the gas permeabilities of all CA + IL composite membranes noticeably increase compared with the pure CA membrane. For instance, CA + 40[C₇Py][NTf₂] and CA + 40[E₂Py][NTf₂] composite membranes show higher CO₂ permeabilities of 71.35 Barrer and 35.69 Barrer, respectively, vs. 5.294 Barrer for the pure CA membrane. It is observed that the composite membrane with a longer alkyl chain tethered to IL has a higher CO₂ permeability. This may be because the IL with a longer alkyl chain has a more steric hindrance effect on the polymer chains, leading to more voids in composite membranes, which is beneficial for gas diffusion in membranes. The results also indicate that the increase in CO₂ permeability of CA + [E_nPy][NTf₂] composite membranes is relatively less than their analogues CA + [C_nPy][NTf₂] composite membranes. Moreover, the more ether groups the IL has, the less increase amount in the CO₂ permeability. This is probably attributed to the good compatibility between [E_nPy][NTf₂] and CA.

Fig. 9 Comparison of CO₂ permeability of CA + [E_nPy][NTf₂] membranes and CA + [C_nPy][NTf₂] membranes.

The CO₂ permselectivities of four CA + IL composite membranes are shown in Fig. 10. It can be seen that the CO₂/N₂ permselectivity in the CA + IL composite membranes is in the following order: CA + [E₁Py][NTf₂] > CA + [C₄Py][NTf₂] ≈ CA + [E₂Py][NTf₂] > CA + [C₇Py][NTf₂], and for the CO₂/CH₄ permselectivity: CA + [E₁Py][NTf₂] > CA + [E₂Py][NTf₂] > CA + [C₄Py][NTf₂] > CA + [C₇Py][NTf₂]. This suggests that the CO₂ permselectivities of all of that CA + IL composite membranes in this work are governed by both plasticizing effect and the affinity of ILs for the gases. This result demonstrates that though ether-functionalized IL composite membranes have lower permeabilities, they have better permselectivities compared with their equivalent length analogues, i.e., CA + [C_nPy][NTf₂] composite membranes. This is consistent with the gas absorption tendency when used ILs with ether groups as reported by Bara.²⁸

Fig. 10 Comparison of the permselectivities in CA + [E_nPy][NTf₂] membranes and CA + [C_nPy][NTf₂] membranes: (a) CO₂/N₂; (b) CO₂/CH₄.

3.2.3 Comparison of gas separation performance. As presented in Fig. 11, the CA + IL composite membranes in this work shows better or at least comparable CO₂ separation performance than the pure CA membrane or other modified CA membrane.^36–39 However, the gas separation performance of the CA + IL composite membranes are still below the “Robeson upper bound”, which means a big room for the further improvement. There are several ways to improve the CO₂ separation performance of CA + IL membrane. For example: the CA can be blended with other TSILs, such as the amine-functionalized ILs, to improve the CO₂ separation performance of pure CA membrane. Besides, use inorganic/organic particle (e.g., zeolites, metal–organic frameworks (MOFs), carbon nanotubes, carbon molecular sieves (CMS) etc.) to compose the CA + IL + nanoparticle mixed matrix membranes with better CO₂ separation performance.

Fig. 11 Comparison of CO₂ separation performance between CA + IL membranes in this work and other modified CA membranes.^36–39

4. Conclusions

In this work, a series of CA composite membranes composed of ether-functionalized pyridinium-based ILs ([E_nPy][NTf₂]) and the non-functionalized analogous ILs ([C_nPy][NTf₂]) were fabricated by a casting method. CO₂, N₂ and CH₄ permeabilities of the composite membranes with up to 40 wt% IL were measured and the permselectivities were determined by the ratios of the permeabilities of two gases. The results showed that the CO₂ permeability of the CA + 40[E₁Py][NTf₂] composite membrane exhibits a seven-fold increase with ideal CO₂/N₂ and CO₂/CH₄ permselectivities of 32 and 24, respectively. Various characterization methods were used to study the effect of the IL on the physiochemical properties of the composite membranes. The results showed that the physical properties of the CA + [E₁Py][NTf₂] membranes are jointly influenced by plasticizing effect and the affinity of ILs for the gases, which also probably influences the CO₂/N₂ and CO₂/CH₄ permselectivities. When the IL content is low in the membranes, plasticizing effect is dominant, causing the declines of the permselectivities. When the IL content increases, the CO₂/N₂ permselectivity improves, and the CO₂/CH₄ permselectivity decreases, which is mainly attributed to the affinity of ILs for the gases. Moreover, compared with the analogous CA + [C_nPy][NTf₂] composite membranes, the introduction of functional ILs [E_nPy][NTf₂] to composite membranes improves the ideal permselectivities of CO₂/N₂ and CO₂/CH₄, whereas decreases slightly gas permeabilities.

Acknowledgements

This work was financially supported by the National Natural Science Fund (No. 21425625 and 21506219), the External Cooperation Program of BIC, Chinese Academy of Sciences (Grant No. l22111KYS820150017, GJHZ201306), and the key program of Beijing Municipal Natural Science Foundation (No. 2141003).

References

1. M. Pera-Titus, Chem. Rev., 2014, 114, 1413–1492.
2. S. Yoo, J. Won, S. W. Kang, Y. S. Kang and S. Nagase, J. Membr. Sci., 2010, 363, 72–79.
3. S. D. Kenarsari, D. Yang, G. Jiang, S. Zhang, J. Wang, A. G. Russell, Q. Wei and M. Fan, RSC Adv., 2013, 3, 22739–22773.
4. Q. Song, S. K. Nataraj, M. V. Roussenova, J. C. Tan, D. J. Hughes, W. Li, P. Bourgoin, M. A. Alam, A. K. Cheetham, S. A. Al-Muhtaseb and E. Sivaniah, Energy Environ. Sci., 2012, 5, 8359–8369.
5. S. Kanehashi, M. Kishida, T. Kidesaki, R. Shindo, S. Sato, T. Miyakoshi and K. Nagai, J. Membr. Sci., 2013, 430, 211–222.
6. H. Z. Chen, P. Li and T. S. Chung, Int. J. Hydrogen Energy, 2012, 37, 11796–11804.
7. A. L. Miller, T. K. Carlisle, A. L. LaFrate, B. A. Voss, J. E. Bara, Y. C. Hudiono, B. R. Wiesenauer, D. L. Gin and R. D. Noble, Sep. Sci. Technol., 2012, 47, 169–177.
8. D. L. Gin and R. D. Noble, Science, 2011, 332, 674–676.
9. X. P. Zhang, X. C. Zhang, H. F. Dong, Z. J. Zhao, S. J. Zhang and Y. Huang, Energy Environ. Sci., 2012, 5, 6668–6681.
10. S. H. Lee, B. S. Kim, E. W. Lee, Y. I. Park and J. M. Lee, Desalination, 2006, 200, 21–22.
11. Y. I. Parka, B. S. Kima, Y. H. Byuna, S. H. Leeb, E. W. Leea and J. M. Leea, Desalination, 2009, 236, 342–348.
12. P. Uchytil, J. Schauer, R. Petrychkovych, K. Setnickova and S. Y. Suen, J. Membr. Sci., 2011, 383, 262–271.
13. Z. D. Dai, R. D. Noble, D. L. Gin, X. P. Zhang and L. Y. Deng, J. Membr. Sci., 2016, 497, 1–20.
14. J. E. Bara, C. J. Gabriel, T. K. Carlisle, D. E. Camper, A. Finotello, D. L. Gin and R. D. Noble, Chem. Eng. J., 2009, 147, 43–50.
15. W. Zhao, G. H. He, H. J. Liu, F. H. Li and Y. Zhang, Chem. Ind. Eng. Prog., 2014, 33, 3292–3308.
16. P. Scovazzo, J. Membr. Sci., 2009, 343, 199–211.
17. M. Kárászová, M. Kacirková, K. Friess and P. Izák, Sep. Purif. Technol., 2014, 132, 93–101.
18. T. K. Carlisle, J. E. Bara, A. L. Lafrate, D. L. Gin and R. D. Noble, J. Membr. Sci., 2010, 359, 37–43.
19. J. E. Bara, E. S. Hatakeyama, D. L. Gin and R. D. Noble, Polym. Adv. Technol., 2008, 19, 1415–1420.
20. L. Hao, P. Li, T. X. Yang and T. S. Chung, J. Membr. Sci., 2013, 436, 221–231.
21. R. Shindo, M. Kishida, H. Sawa, T. Kidesaki, S. Sato, S. Kanehashi and K. Nagai, J. Membr. Sci., 2014, 454, 330–338.
22. L. Liang, Q. Gan and P. Nancarrow, J. Membr. Sci., 2014, 450, 407–417.
23. S. U. Hong, D. Park, Y. Ko and I. Baek, Chem. Commun., 2009, 7227–7229.
24. J. C. Jansen, K. Friess, G. Clarizia, J. Schauer and P. Izak, Macromolecules, 2011, 44, 39–45.
25. J. E. Bara, D. L. Gin and R. D. Noble, Ind. Eng. Chem. Res., 2008, 47, 9919–9924.
26. B. E. Gurkan, J. C. d. l. Fuente, E. M. Mindrup, L. E. Ficke, B. F. Goodrich, E. A. Price, W. F. Schneider and J. F. Brennecke, J. Am. Chem. Soc., 2010, 132, 2116–2117.
27. Y. Q. Zhang, S. J. Zhang, X. M. Lu, Q. Zhou, W. Fan and X. P. Zhang, Chemistry, 2009, 15, 3003–3011.
28. J. E. Bara, C. J. Gabriel, S. Lessmann, T. K. Carlisle, A. Finotello, D. L. Gin and R. D. Noble, Ind. Eng. Chem. Res., 2007, 46, 5380–5386.
29. T. K. Carlisle, J. E. Bara, A. L. Lafrate, D. L. Gin and R. D. Noble, Ind. Eng. Chem. Res., 2008, 47, 7005–7012.
30. S. J. Zeng, J. Wang, L. Bai, B. Q. Wang, H. S. Gao, D. W. Shang, X. P. Zhang and S. J. Zhang, Energy Fuels, 2015, 29, 6039–6048.
31. J. Wang, S. J. Zeng, L. Bai, H. S. Gao, X. P. Zhang and S. J. Zhang, Ind. Eng. Chem. Res., 2014, 53, 16832–16839.
32. S. J. Zeng, H. S. Gao, X. C. Zhang, H. F. Dong, X. P. Zhang and S. J. Zhang, Chem. Eng. J., 2014, 251, 248–256.
33. S. Farrukh, S. Javed, A. Hussain and M. Mujahid, Asia. Pac. J. Chem. Eng., 2014, 9, 543–551.
34. D. Y. Xing, N. Peng and T. S. Chung, J. Membr. Sci., 2011, 380, 87–97.
35. L. C. Tomé, M. A. Aboudzadeh, L. P. N. Rebelo, C. S. Freire, D. Mecerreyes and I. M. Marrucho, J. Mater. Chem. A, 2013, 1, 10403–10411.
36. S. W. Jintong Li, K. Nagai, T. Nakagawa and A. W-H. Mau, J. Membr. Sci., 1998, 138, 143–152.
37. J. Chen, J. M. Zhang, Y. Feng, J. S. He and J. Zhang, Chin. J. Polym. Sci., 2014, 32, 1–8.
38. W. G. Kim, J. S. Lee, D. G. Bucknall, W. J. Koros and S. Nair, J. Membr. Sci., 2013, 441, 129–136.
39. J. Chen, J. Zhang, Y. Feng, J. Wu, J. He and J. Zhang, J. Membr. Sci., 2014, 469, 507–514.

---