The Influence of Polycation and Counter-Anion Nature on the Properties of Poly(ionic liquid)-Based Membranes for CO2 Separation

The current investigation is focused on the development of composite membranes based on polymeric ionic liquids (PILs) containing imidazolium and pyridinium polycations with various counterions, including hexafluorophosphate, tetrafluoroborate, and bis(trifluoromethylsulfonyl)imide. A combination of spectroscopic methods was used to identify the synthesized PILs and characterize their interaction with carbon dioxide. The density and surface free energy of polymers were performed by wettability measurements, and the results are in good agreement with the permeability and selectivity obtained within the gas transport tests. It was shown that the membranes with a selective layer based on PILs exhibit relatively high permeability with CO2 and high ideal selectivity CO2/CH4 and CO2/N2. Additionally, it was found that the type of an anion significantly affects the performance of the obtained membranes, with the most pronounced effect from bis-triflimide-based polymers, showing the highest permeability coefficient. These results provide valuable insights into the design and optimization of PIL-based membranes for natural and flue gas treatment.


Introduction
Modern society is moving towards the transition to the sixth technological paradigm, sustainable technologies, and ESG (Environmental, Social and Corporate Governance) [1]. In this regard, chemical engineering is actively searching for economically beneficial and environmentally friendly approaches to meet the current needs of holistic chemical and petroleum production. Membrane technologies are in a large part considered the main alternative to traditional methods of purification and separation of substances due to their high energy efficiency. This is because they enable continuous processes without the need to replace or regenerate the chemical sorbent [2]. Moreover, membrane processes' hardware is steric hindrance of the polymer chains [34]. Since the pioneering work was published by Nobel's group [35], PILs have been extensively studied as a membrane material in a gas separation field [10,[36][37][38][39]. Although some PILs in combination with ILs exhibit extraordinary selectivity for CO 2 /CH 4 and CO 2 /N 2 gas pairs and overcome the Robeson upper bound [40,41], they also have some disadvantages as a membrane material. Most PILs proposed as membrane matrix for CO 2 separation form brittle films incapable of withstanding the pressure drop of the gas flow. Another problem is the drastic decrease of CO 2 permeability in the direct ratio to polymer molecular weight [42]. However, a wide range of possibilities for polymer design through a reasonable choice of monomeric units, as well as polymer structure, may significantly improve a polymer's performance as a membrane material for CO 2 separation.
Although some advantages of applying styrene-based PILs in membrane design have been demonstrated, there remains a need for a deeper understanding of how ionic composition affects CO 2 separation performance for this class of polyelectrolytes. This work focuses on the synthesis and characterization of PILs with polystyrene as the polymer backbone, utilizing either methylimidazolium or pyridinium cationic substitutes and a range of counter-anions: chloride [Cl], tetrafluoroborate [BF 4 ], hexafluorophosphate [PF 6 ], and bis(trifluoromethylsulfonyl)imide [Tf 2 N]. Composite gas separation membranes based on the synthesized PILs were prepared and tested for individual gases (CO 2 , CH 4 , and N 2 ) permeation to investigate the influence of polycation functionality and anion type on the CO 2 separation performance of styrene-based PILs.

Poly(vinylbenzyl chloride) (pVBCl) Synthesis
pVBCl was synthesized via free radical polymerization in mass according to a wellestablished technique. The procedure was as follows: initiator AIBN (0.05 g) was dissolved in VBCl (20 g) and placed into an ampule. The ampule was frozen in liquid nitrogen, exposed to vacuum and then defrosted. The freezing-vacuuming treatment was repeated three times. Then, the ampule was sealed under vacuum. Polymerization was carried out for 12 h at 70 • C in an oil bath. The product of polymerization was purified by dissolution in chloroform followed by precipitation with isopropanol. Dissolving-precipitation treatment was repeated three times. The resulting pVBCl was dried under vacuum at 60 • C to constant weight. The yield of the product was 91%.

Poly(vinylbenzylpyridinium chloride) (pVBPyCl) and Poly(vinylbenzylmethylimidazolium chloride) (pVBmimCl) Synthesis
The obtained pVBCl was used as a precursor for pVBPyCl, pVBmimCl synthesis by Menshutkin reaction. For pVBPyCl synthesis, pVBCl (4 g) was dissolved in pyridine (10 mL). For pVBmimCl synthesis, pVBCl (4 g) was dissolved in toluene (10 mL) and a 10 mol% excess of 1-methylimidazole was added under constant stirring. The reaction was carried out for 12 h at 60 • C with reflux in an oil bath. The products were purified by decantation followed by threefold washing of the precipitate in diethyl ether. The resulting products were dried under vacuum at 50 • C to constant weight. The yield of pVBPyCl and pVBmimCl was 96% and 92%, respectively. Conductometric titration was performed to determine the functionalization degree (FD) as it is described in Section 2.5.

Anion Exchange Reaction
A synthetic route for PIL synthesis is demonstrated in Scheme 1. A series of PILs (pVBPyBF 4 , pVBPyPF 6 , pVBPyTf 2 N, pVBmimBF 4 , pVBmimPF 6 , pVBmimTf 2 N) with various counterions was synthesized from pVBPyCl and pVBmimCl by ion exchange reactions with one of the following salts: NaBF 4 , KPF 6 , or LiTf 2 N. The procedure for all PILs was as follows: a weighted amount of polyelectrolyte with Cl-anion was dissolved in water in a flask, and a 10 mol% excess of salt solution in water was gradually added under constant stirring. The reaction products were insoluble in water and precipitated immediately. Due to the polymeric nature of cations, the reaction was carried out for 24 h at 25 • C with stirring to ensure the maximum possible exchange. The product was isolated by filtering using a Buchner filter funnel with fine frit (G3) and purified by washing three times in distilled water. Following, drying under vacuum at 40 • C to constant weight was performed. Conductometric titration was performed to determine the anion exchange degree (ExD) as it is described in Section 2.5. distilled water. Following, drying under vacuum at 40 °C to constant weight was performed. Conductometric titration was performed to determine the anion exchange degree (ExD) as it is described in Section 2.5. Scheme 1. General synthetic route for PIL synthesis.

Conductometric Titration
Conductometric titration was used to confirm and quantify pVBCl functionalization as a functionalization degree (FD), the ratio of the experimentally determined amount of substituent groups to the calculated amount expressed in percentage, and the anion exchange degree (ExD), the ratio of the experimentally determined amount of exchanged anions to the calculated one expressed in percentage. The measurements were performed using a Aquasearcher benchtop meter (OHAUS, NJ, USA) equipped with STCON3 conductivity electrodes (OHAUS, NJ, USA). FD was determined by pVBPyCl or pVBmimCl solution in water (0.0001 M, 100 mL) titration with AgNO3 0.0001 M solution. The measurements were carried out in intervals of 1 mL and 0.2 mL near the equivalence point. Titration was performed 3 times; the average value was calculated. To evaluate ExD, solutions of Cl-containing inorganic salts, formed as a result of an ion exchange reaction, were titrated with AgNO3 0.0001 M solution in water. The measurements were carried out in intervals of 1 mL and 0.2 mL near the equivalence point. Titration was performed 3 times; the average value was calculated.

Nuclear Magnetic Resonance Spectroscopy (NMR)
The 1 H NMR spectra were recorded in DMSO-d6 solution (99.9% atom D, Sigma-Aldrich) on a DD2 400NB NMR spectrometer (Agilent, CA, USA) at 400 MHz. The residual solvent peak was used as an internal standard. The chemical shifts (δ) are reported in parts per million (ppm); J values are given in hertz (Hz).

Gel Permeation Chromatography (GPC)
Molecular weight (number-average molecular weight (Mn) and weight-average molecular weight (Mw)), as well as polydispersity (Mw/Mn) of pVBCl were determined by means of GPC on Prominence LC-20VP (Shimadzu) equipped with a RI-detector. The analyses were performed at a flow rate of 0.7 mL·min − 1 and at 40 °C using THF (≥99.9%, HPLC quality, Sigma-Aldrich) as an eluent. The GPC system was calibrated using the narrow-dispersed polystyrene (PS) ranging (Fluka).

Attenuated Total Reflectance Fourier Transforms Infrared Spectroscopy (ATR-FTIR)
IR spectra of the samples were recorded using an FTIR spectrophotometer IRTracer-100 (Shimadzu, Kyoto, Japan) equipped with modified HATR accessory with a ZnSe crystal plate (PIKE, NC, USA) in transmittance mode at ambient temperature. The mirror system of the accessory was constantly exposed to a N2 flow in order to exclude atmospheric Scheme 1. General synthetic route for PIL synthesis.

Conductometric Titration
Conductometric titration was used to confirm and quantify pVBCl functionalization as a functionalization degree (FD), the ratio of the experimentally determined amount of substituent groups to the calculated amount expressed in percentage, and the anion exchange degree (ExD), the ratio of the experimentally determined amount of exchanged anions to the calculated one expressed in percentage. The measurements were performed using a Aquasearcher benchtop meter (OHAUS, NJ, USA) equipped with STCON3 conductivity electrodes (OHAUS, NJ, USA). FD was determined by pVBPyCl or pVBmimCl solution in water (0.0001 M, 100 mL) titration with AgNO 3 0.0001 M solution. The measurements were carried out in intervals of 1 mL and 0.2 mL near the equivalence point. Titration was performed 3 times; the average value was calculated. To evaluate ExD, solutions of Cl-containing inorganic salts, formed as a result of an ion exchange reaction, were titrated with AgNO 3 0.0001 M solution in water. The measurements were carried out in intervals of 1 mL and 0.2 mL near the equivalence point. Titration was performed 3 times; the average value was calculated.

Nuclear Magnetic Resonance Spectroscopy (NMR)
The 1 H NMR spectra were recorded in DMSO-d 6 solution (99.9% atom D, Sigma-Aldrich) on a DD2 400NB NMR spectrometer (Agilent, CA, USA) at 400 MHz. The residual solvent peak was used as an internal standard. The chemical shifts (δ) are reported in parts per million (ppm); J values are given in hertz (Hz).

Gel Permeation Chromatography (GPC)
Molecular weight (number-average molecular weight (M n ) and weight-average molecular weight (M w )), as well as polydispersity (M w /M n ) of pVBCl were determined by means of GPC on Prominence LC-20VP (Shimadzu) equipped with a RI-detector. The analyses were performed at a flow rate of 0.7 mL·min − 1 and at 40 • C using THF (≥99.9%, HPLC quality, Sigma-Aldrich) as an eluent. The GPC system was calibrated using the narrow-dispersed polystyrene (PS) ranging (Fluka).

Attenuated Total Reflectance Fourier Transforms Infrared Spectroscopy (ATR-FTIR)
IR spectra of the samples were recorded using an FTIR spectrophotometer IRTracer-100 (Shimadzu, Kyoto, Japan) equipped with modified HATR accessory with a ZnSe crystal plate (PIKE, NC, USA) in transmittance mode at ambient temperature. The mirror system of the accessory was constantly exposed to a N 2 flow in order to exclude atmospheric CO 2 from registered spectra. As it was previously reported [52], HATR accessory was modified with inlet and outlet gas valves in the top cover enabling in situ experiments of gas (CO 2 ) sorption by the sample in the cell. The sample of polymer for FTIR analysis was prepared by polymer solution (1 mass% in DMSO) casting on a ZnSe crystal plate followed by solvent evaporation under vacuum at 40 • C. Three types of spectra were recorded for all polymer samples: (1) pure polymer, (2) in situ CO 2 sorption, (3) polymer after CO 2 desorption. A minimum of 30 scans was signal-averaged with a resolution of 4 cm −1 within the 4000-500 cm −1 range.

Density of Polymers
The density of the polymer films obtained by polymer solution (2 mass% in DMSO) casting on inert glass support was measured by implementing a flotation method [53,54]. To conduct an experiment, a graduated cylinder (50 cm 3 ) with a ground-in glass stopper was filled to half with a mixture of two miscible liquids with different densities at a 1:1 volume ratio. Particular liquids were chosen based on the following requirements: (1) liquids should be chemically inert to the polymer material, (2) liquids should not cause swelling or solvation of the polymer, (3) the mixture should cover the supposed density range for the polymer. Mixtures of liquid used for different PILs are listed in Table S1. A sample of the polymer film was immersed in the mixture of liquids and brought to an equilibrium middle position, according to the graduation, by adding one of the liquids dropwise. Then, the density of the obtained mixture of liquids was measured using a pycnometer (10 cm 3 ). Density was measured three times on each membrane to obtain the average density values of the polymers. The experiments were conducted at 20 • C.

Membrane Preparation
Composite polymer membranes with a selective layer based on synthesized PILs were obtained by casting corresponding solutions (2% in DMSO) onto a commercial microfiltration fluoropolymer membrane MFFK using an automatic casting knife MemcastTM Plus (POROMETER, Nazareth, Belgium) with a 100 µm blade followed by solvent evaporation in enclosed area under ambient temperature. After major solvent evaporation membranes were dried under vacuum to a constant weight.

Scanning Electron Microscopy (SEM)
Topography and morphological characteristics of the obtained composite membranes were studied by scanning electron microscopy (SEM) using an JSM-IT300LV electron microscope (JEOL, Peabody, MA, USA) with an electron probe diameter of about 5 nm and a probe current of less than 0.5 nA (operating voltage 20 kV). SEM scanning was performed using low-energy secondary electrons and backscattered electrons under a low vacuum to eliminate the charge. Supplementary Information for SEM images is listed in Table S3.

Wettability Measurements and Surface Free Energy Calculation
Wettability tests and surface energy calculation were performed according to a previously published technique [55]. The contact angle of wetting (θ) with three test liquids with different surface tensions (Table 2): water, glycerol, and diiodomethane was measured at an equilibrium state in a closed beaker and calculated using ImageJ software with a contact angle plugin based on experimental data. The measurements were conducted at 293 K. The results were collected for a series of five drops with contact angle deviations that did not exceed ± 1 • . Total surface free energy and its components were calculated using the Owens-endt method. According to the Owens-Wendt method, the surface free energy (γ s ) can be calculated as a sum of its dispersive (γ d s ) and polar (γ p s ) components: These values can be graphically obtained from the Owens-Wendt equation using the results of surface wettability measurements with three different testing liquids obtained previously: Here, γ l is the surface tension of the wetting liquid (mJ·m −2 ), γ d l is a dispersive component of the surface tension of the wetting liquid (mJ·m −2 ), and γ p l is a polar component of the surface tension of the wetting liquid (mJ·m −2 ).

Gas Permeation Tests
For the gas permeation test, three individual gases, nitrogen (N 2 ), carbon dioxide (CO 2 ), and methane (CH 4 ), were chosen to characterize the separation performance of the obtained membranes in natural gas and flue gas treatment. The presence of water influences the performance of ILs and related substances as well as the mechanism of CO 2 solubility and transport [56][57][58][59]. In this regard, the presence of water in membranes was measured using infrared moisture determination balance FD-610 (Kett, Tokyo, Japan). The membrane sample (5.5 g) was placed in the balance and dried at 120 • C to a constant weight. The presence of water in all membranes was less than 0.1%. All gases used were 99.9% purity with content of water in CO 2 at 0.001%; in N 2 , 0.007%; and in CH 4 , 0.0001%. The pure gas permeabilities of N 2 , CO 2 , and CH 4 through the polymeric membranes were measured by an experimental setup (Figure 1) supported by an automatic computing system based on a software-logic controller (Unitronix, Israel) at the initial transmembrane pressure of 130 kPa and ambient temperature (293 K) in a constant volume mode [15].

Diiodomethane
49.5 1.3 50.8 The measurements were conducted at 293 K. The results were collected for a serie of five drops with contact angle deviations that did not exceed ±1°. Total surface free en ergy and its components were calculated using the Owens-endt method. According to th Owens-Wendt method, the surface free energy can be calculated as a sum of its dis persive ( ) and polar ( ) components: (1 These values can be graphically obtained from the Owens-Wendt equation using th results of surface wettability measurements with three different testing liquids obtaine previously: is the surface tension of the wetting liquid (mJ·m −2 ), is a dispersive com ponent of the surface tension of the wetting liquid (mJ·m −2 ), and is a polar componen of the surface tension of the wetting liquid (mJ·m −2 ).

Gas Permeation Tests
For the gas permeation test, three individual gases, nitrogen (N2), carbon dioxid (CO2), and methane (CH4), were chosen to characterize the separation performance of th obtained membranes in natural gas and flue gas treatment. The presence of water influ ences the performance of ILs and related substances as well as the mechanism of CO2 sol ubility and transport [56][57][58][59]. In this regard, the presence of water in membranes wa measured using infrared moisture determination balance FD-610 (Kett, Tokyo, Japan The membrane sample (5.5 g) was placed in the balance and dried at 120 °C to a constan weight. The presence of water in all membranes was less than 0.1%. All gases used wer 99.9% purity with content of water in CO2 at 0.001%; in N2, 0.007%; and in CH4, 0.0001% The pure gas permeabilities of N2, CO2, and CH4 through the polymeric membranes wer measured by an experimental setup (Figure 1) supported by an automatic computing sys tem based on a software-logic controller (Unitronix, Israel) at the initial transmembran pressure of 130 kPa and ambient temperature (293 K) in a constant volume mode [15]. Each single-gas test was repeated at least three times. The selected experimenta curve of CO2 pressure time course in a permeation cell obtained for a membrane with selective layer based on pVBmimTf2N is represented in Figure S7. Permeability coeffi cients were calculated according to the following formula [60]: Each single-gas test was repeated at least three times. The selected experimental curve of CO 2 pressure time course in a permeation cell obtained for a membrane with a selective layer based on pVBmimTf 2 N is represented in Figure S7. Permeability coefficients were calculated according to the following formula [60]: where β is a geometric setup parameter (m −1 ), p f eed is the pressure in a high-pressure compartment (Pa), p perm is the pressure in a low-pressure compartment (Pa), P is the permeability coefficient (Barrer), t is time (s), and l is the selective layer thickness in a composite membrane (m). The ideal selectivity of the polymeric membranes was calculated as the ratio of the single gas permeability coefficients (gases A and B): The ideal selectivity gives some idea of the membrane separation properties. However, this value is based on a simplified model of gas adsorption and diffusion in the masstransfer process. In this regard, a true selectivity might differ when comes to a gas mixture.

PILs Identification and Properties
It was previously shown by L. C. Tomé [42] that PIL properties strongly depend on the polymer's molecular weight. Medium-molecular-weight polymers (~200 kDa) demonstrated optimal thermal, mechanical, and transport properties. In the current case, PILs were synthesized by the functionalization of a polymer precursor-pVBCl. The results of pVBCl molecular weight and polydispersity determined by GPC are represented in Table 3. pVBCl synthesized by radical polymerization is characterized by a high polydispersity expected for a radical process without the use of chain transfer agents. Since functionalization does not significantly influence the length of polymer chains, PILs' molecular weight could be defined as medium (~200 kDa).
PILs synthesis includes stages of tertiary amine quaternization and anion exchange. Both reactions are complicated by a sterical hindrance related to the polymeric nature of the substrate. This leads to incomplete functionalization of the initial pVBCl. FD and ExD, determined conductometrically, are listed in Table 4.  6 . For both pVBmimTf 2 N and pVBPyPyTf 2 N, ExD is much higher, which might be caused by the formation of a polymer with a loosened structure that facilitates an ion exchange reaction.
The formation of PILs was confirmed by a combination of 1 H NMR and ATR-FTIR spectroscopy. The data is represented in Supplementary Materials in Table S2 for 1 Figures S1-S6 for ATR-FTIR. ATR-FTIR spectra for PILs containing Tf 2 N anion are represented in Figures 2 and 3. After exposure to the CO 2 atmosphere, a strong CO 2 asymmetric stretch (ν 3 ) bond is observed for PIL spectra in the region 2333-2341 cm −1 . and in Figures S1-S6 for ATR-FTIR. ATR-FTIR spectra for PILs containing Tf2N anion are represented in Figures 2 and 3. After exposure to the CO2 atmosphere, a strong CO2 asymmetric stretch (ν3) bond is observed for PIL spectra in the region 2333-2341 cm −1 .   Table 5 compares the position of the ν3 CO2 bond after interaction with PILs with those of free CO2 of solid and gas states. It is worth noting that ν3 CO2 in the PILs spectra (2339 cm −1 ) shifts for ~10 cm −1 in comparison to free carbon dioxide in a gas phase (2349 cm −1 ), which confirms the emergence of interaction between CO2 and PILs. A small value of shift is in favor of a physical nature of interaction and not chemical. The disappearance of the ν3 CO2 band caused by the quick desorption during vacuum treatment without a and in Figures S1-S6 for ATR-FTIR. ATR-FTIR spectra for PILs containing Tf2N anion are represented in Figures 2 and 3. After exposure to the CO2 atmosphere, a strong CO2 asymmetric stretch (ν3) bond is observed for PIL spectra in the region 2333-2341 cm −1 .   Table 5 compares the position of the ν3 CO2 bond after interaction with PILs with those of free CO2 of solid and gas states. It is worth noting that ν3 CO2 in the PILs spectra (2339 cm −1 ) shifts for ~10 cm −1 in comparison to free carbon dioxide in a gas phase (2349 cm −1 ), which confirms the emergence of interaction between CO2 and PILs. A small value of shift is in favor of a physical nature of interaction and not chemical. The disappearance of the ν3 CO2 band caused by the quick desorption during vacuum treatment without a Figure 3. FTIR spectra of pure pVBPyTf 2 N, pVBPyTf 2 N after CO 2 adsorption, and pVBPyTf 2 N after CO 2 desorption. Table 5 compares the position of the ν 3 CO 2 bond after interaction with PILs with those of free CO 2 of solid and gas states. It is worth noting that ν 3 CO 2 in the PILs spectra (2339 cm −1 ) shifts for~10 cm −1 in comparison to free carbon dioxide in a gas phase (2349 cm −1 ), which confirms the emergence of interaction between CO 2 and PILs. A small value of shift is in favor of a physical nature of interaction and not chemical. The disappearance of the ν 3 CO 2 band caused by the quick desorption during vacuum treatment without a temperature increase is in the line with abovementioned observations. Thus, it could be concluded that the interaction between carbon dioxide and PILs is that of a physisorption type. However, ILs with Tf 2 N and carboxylate anion demonstrate chemosorption of CO 2 [56,58]. The position of the bond varies depending on the nature of the anion in polyelectrolyte. The frequency of the ν 3 CO 2 band demonstrates the most significant shift (14 cm −1 ) when PILs with Cl-anion are considered. Apparently, it proceeds due to the higher polarizing effect of small Cl-anion in comparison to other investigated anions with a larger radius. In the case of gas separation membranes, weaker interaction is preferable because it provides faster sorption-desorption acts in a mass-transfer process acting by the solution-diffusion mechanism. Further, the density of polymer films obtained in the same conditions as a selective layer of composite membranes on a glass support was evaluated.

H NMR and in
As seen in Figure 4 and in Table S1, the density of the nonporous films based on the synthesized PILs are higher than those for initial pVBCl (1.1889 g/cm 3 ) in all cases. The increase of the PILs' density compared to that of pVBCl can be addressed to the close package of macromolecular chains in the polymer matrix due to intermolecular electrostatic interactions. It was observed that density is insignificantly influenced by the type of polycation and, to a greater extent, depends on a type of counterion. The increasing of PILs' density values follows the consequence of PIL anion Cl −~T f 2 N − < BF 4 − < PF 6 − . Based on this correlation, it was supposed that the free volume in PILs has a likewise dependence. A relatively low density of polyelectrolytes with Tf 2 N − anion supports the assumption that ion exchange reaction in this case would result in a higher yield. Since gas separation processes in the case of nonporous polymeric membranes follow the solution-diffusion mechanism and the process is limited by a diffusion of penetrants in the polymer matrix, lower-density polymers are expected to be more permeable.

Membrane Characterization
Most previously reported PILs are characterized by low breaking strength and high fragility [28,42,46,51]. This drawback could be solved by developing a composite mem-

Membrane Characterization
Most previously reported PILs are characterized by low breaking strength and high fragility [28,42,46,51]. This drawback could be solved by developing a composite membrane with a porous support providing the required mechanical strength. Thereby, composite membranes composed of porous fluoropolymer support and a thin selective layer based on a synthesized polymers were developed. Membrane cross-sectional microstructure and surface topography registered by SEM are shown in Figures 5 and 6.

Membrane Characterization
Most previously reported PILs are characterized by low breaking strength and high fragility [28,42,46,51]. This drawback could be solved by developing a composite membrane with a porous support providing the required mechanical strength. Thereby, composite membranes composed of porous fluoropolymer support and a thin selective layer based on a synthesized polymers were developed. Membrane cross-sectional microstructure and surface topography registered by SEM are shown in Figures 5 and 6.  According to SEM, three well-defined layers are observed on a cross-sectional SEM microphotograph from top to bottom: (i) the PIL selective layer, (ii) the porous fluoroplastic microfiltration layer, and (iii) the nonwoven fabric. Layers (ii) and (iii) belong to According to SEM, three well-defined layers are observed on a cross-sectional SEM microphotograph from top to bottom: (i) the PIL selective layer, (ii) the porous fluoroplastic microfiltration layer, and (iii) the nonwoven fabric. Layers (ii) and (iii) belong to the commercial microfiltration membrane used as a support. PILs demonstrate good adhesion to the fluoroplastic layer. Penetration of PILs into the porous of the support is not observed. The values of selective layer thickness obtained from SEM cross-sectional microphotographs were used to calculate permeability coefficients for each PIL. The average thickness is about 50 micrometers, which is several times higher than in advanced reverse osmosis or pervaporation membranes. In this case, thick selective layer prevents mechanical defects that drastically influence membranes performance. Another way to provide mechanically stability to a thin PIL layer is mixing with compatible IL [63]. The obtained PILs form a homogeneous defect-free layer on the surface of porous fluoroplastic support. Comparison of SEM microphotographs of the surface for the initial pVBCl ( Figure 5) with PILs ( Figure 6) shows conformity. The flexible pVBCl backbone structure determines the appearance of the PIL surface in this case. Densely packed polymeric chains form uniform amorphous structures. Despite the PILs' selective layers having a similar topography, the nature of ionic substitute in a polymer significantly influences the energetic characteristics of the membrane surface.
The surface free energy and its polar and dispersive components were evaluated using the Owens-Wendt method based on wettability measurements. The results are shown in Figure 7. The surface free energy and its polar and dispersive components were evaluated using the Owens-Wendt method based on wettability measurements. The results are shown in Figure 7. PILs obtained by the functionalization of pVBCl combine properties of a hydrophobic polystyrene backbone with the hydrophilic properties of charged ionic moieties. Due to the amphiphilic nature of PILs, the surfaces of composite membranes are well wetted by the nonpolar test liquid diiodomethane as well as by water and glycerol; however, there are some differences in wettability as a result of different surface free energy, depending on the nature of the tertiary amine and counter ion. As was shown by Bogdanova and Dolzhikov [64] for glassy polymers the dispersive component is linearly correlated with the free volume of the polymer since both values are impacted by the density of pacing in the polymer matrix. The results of dispersive energy component calculation support PILs obtained by the functionalization of pVBCl combine properties of a hydrophobic polystyrene backbone with the hydrophilic properties of charged ionic moieties. Due to the amphiphilic nature of PILs, the surfaces of composite membranes are well wetted by the nonpolar test liquid diiodomethane as well as by water and glycerol; however, there are some differences in wettability as a result of different surface free energy, depending on the nature of the tertiary amine and counter ion. As was shown by Bogdanova and Dolzhikov [64] for glassy polymers the dispersive component is linearly correlated with the free volume of the polymer since both values are impacted by the density of pacing in the polymer matrix. The results of dispersive energy component calculation support the previously discussed results of the density measurements. The dispersive energy changes the same way as the density of PILs and follow the order: Tf 2 N − < BF 4 − < PF 6 − . The polar component of surface free energy is slightly lower for polycations with Tf 2 N − counter ion in comparison to other PILs. Overall, the results demonstrate that polycationic PIL surface free energy is insignificantly altered by the type of tertiary amine in the polymer structure and, to a greater extent, depends on the nature of the anion.
The density of the polymer and its surface properties are in a close correlation with gas transport properties. According to the results of density measurements and surface free energy calculations, a greater permeability of Tf 2 N − -containing PILs was expected and confirmed experimentally. The permeability coefficients of the obtained membranes are represented in Figure 8.  The data on the density and energy of the surface is consistent with the results of the gas separation. pVBPyTf2N and pVBmimTf2N have a lower density and, as a consequence, a larger free volume and lower surface energy, which turned out to be more permeable to all gases. While PILs containing PF6 anions are the least permeable, which agrees with the density measurements. Naturally, polymers with a high density (pVBPyPF6 and pVBmimPF6) have improved selectivity. Figure 9 shows the comparison of ideal selectivity for composite membranes composed of PILs. The data on the density and energy of the surface is consistent with the results of the gas separation. pVBPyTf 2 N and pVBmimTf 2 N have a lower density and, as a consequence, a larger free volume and lower surface energy, which turned out to be more permeable to all gases. While PILs containing PF 6 anions are the least permeable, which agrees with the density measurements. Naturally, polymers with a high density (pVBPyPF 6 and pVBmimPF 6 ) have improved selectivity. Figure 9 shows the comparison of ideal selectivity for composite membranes composed of PILs. gas separation. pVBPyTf2N and pVBmimTf2N have a lower density and, as a consequence, a larger free volume and lower surface energy, which turned out to be more permeable to all gases. While PILs containing PF6 anions are the least permeable, which agrees with the density measurements. Naturally, polymers with a high density (pVBPyPF6 and pVBmimPF6) have improved selectivity. Figure 9 shows the comparison of ideal selectivity for composite membranes composed of PILs.   PILs containing Tf 2 N − anions display an optimal combination of permeability and selectivity in comparison to other anions.

Conclusions
PILs have recently emerged as a promising platform for various separation applications due to a combination of their polymeric nature with a unique property of ILs-the ability to capture CO 2 . In this study, the design and performance of composite membranes based on imidazolium and pyridinium polycationic salts with different anions, such as hexafluorophosphate, tetrafluoroborate, and bis(trifluoromethylsulfonyl)imide, were investigated. PILs were synthesized using a simple ion exchange reaction from a basic polymer precursor, pVBCl. It was found that the functionalization degree and anion exchange degree depend on the nature of substitute. The formation of PILs was confirmed by a combination of NMR and ATR-FTIR spectroscopy. ATR-FTIR was also used to characterize PILs interaction with CO 2 . Several studies [56,58] have previously reported chemosorption of CO 2 by ILs containing Tf 2 N and carboxylate anions. Spectroscopic data on PILs interaction with CO 2 is more in favor of physisorption in the case of polymerized ionic liquids. Properties of the synthetized PILs and gas separation properties of PIL-based composite membranes are summarized in Table 6. In general, PILs showed properties comparable to the previously reported (Table 1) properties and could be further investigated as a membrane matrix material. PILs that contain Tf 2 N − anions exhibited superior separation performance compared to other polymers. The further investigation of pVBCl-based gas separation membranes could be focused on, combining polycations with carboxylate anions that also show a good CO 2 absorption [56,58], as well as the influence of water presence on the membrane performance.