Acidic Gases Separation from Gas Mixtures on the Supported Ionic Liquid Membranes Providing the Facilitated and Solution-Diffusion Transport Mechanisms

Nowadays, the imidazolium-based ionic liquids containing acetate counter-ions are attracting much attention as both highly selective absorbents of the acidic gases and CO2 carriers in the supported ionic liquid membranes. In this regard, the investigation of the gas transport properties of such membranes may be appropriate for better understanding of various factors affecting the separation performance and the selection of the optimal operating conditions. In this work, we have tested CH4, CO2 and H2S permeability across the supported ionic liquid membranes impregnated by 1-butyl-3-methylimidazolium acetate (bmim[OAc]) with the following determination of the ideal selectivity in order to compare the facilitated transport membrane performance with the supported ionic liquid membrane (SILM) that provides solution-diffusion mechanism, namely, containing 1-butyl-3-methylimidazolium tetrafluoroborate (bmim[BF4]). Both SILMs have showed modest individual gases permeability and ideal selectivity of CO2/CH4 and H2S/CH4 separation that achieves values up to 15 and 32, respectively. The effect of the feed gas mixture composition on the permeability of acidic gases and permeselectivity of the gas pair was investigated. It turned out that the permeation behavior for the bmim[OAc]-based SILM toward the binary CO2/CH4, H2S/CH4 and ternary CO2/H2S/CH4 mixtures was featured with high acidic gases selectivity due to the relatively low methane penetration through the liquid phase saturated by acidic gases.


Introduction
There is a high demand for the reduction of acidic gases across the world due to the strong need to deal with the global warming and the rapidly-increasing energy consumption. As long as petroleum, coal and natural gas are used as the primary global fuel, the production of CO 2 is inevitable. Acidic gases such as CO 2 and H 2 S contribute to pipeline corrosion and decrease the 260 days) and selective membrane for CO 2 separation has been obtained using SILM facilitated by the amine-terminated IL immobilized into the hydrophilic polytetrafluoroethylene (PTFE). At 2.5 kPa of CO 2 , bis(trifluoromethylsulfonyl)imide and trifluoromethanesulfone-based SILMs achieved the CO 2 /CH 4 selectivity of approximately 100 and 120, respectively. Recently CO 2 reactive amino acid ionic liquid [29][30][31][32][33] based FTM has been developed with exceptionally high levels of CO 2 permeability regardless of humidity which is a particularly novel feature. Besides amine-tethered ILs, another class of molten salts able to form N-heterocyclic carbenes [34] has been attracting wide attention due to the permanence of physical-chemical properties during the experiments, low viscosity, stability and high absorption capacity. Rogers [35] and Maginn [26] have reported that the basic anions such as acetate may remove the acidic proton on the C(2) position of cation with subsequent formation of N-heterocyclic carbine, which reversibly react with acidic gases (Figure 1). than 260 days) and selective membrane for CO2 separation has been obtained using SILM facilitated by the amine-terminated IL immobilized into the hydrophilic polytetrafluoroethylene (PTFE). At 2.5 kPa of CO2, bis(trifluoromethylsulfonyl)imide and trifluoromethanesulfone-based SILMs achieved the CO2/CH4 selectivity of approximately 100 and 120, respectively. Recently CO2 reactive amino acid ionic liquid [29][30][31][32][33] based FTM has been developed with exceptionally high levels of CO2 permeability regardless of humidity which is a particularly novel feature. Besides amine-tethered ILs, another class of molten salts able to form N-heterocyclic carbenes [34] has been attracting wide attention due to the permanence of physical-chemical properties during the experiments, low viscosity, stability and high absorption capacity. Rogers [35] and Maginn [26] have reported that the basic anions such as acetate may remove the acidic proton on the C(2) position of cation with subsequent formation of N-heterocyclic carbine, which reversibly react with acidic gases (Figure 1).
In this context, Santos et al. have obtained the results of CO2 separation on the acetate-based SILMs that was near the Robeson upper bound corresponding to the best polymeric materials [34]. The promising acidic gases separation performance was achieved on the Al2O3/TiO2 tubes with immobilized emim[OAc] [37]. To the best of our knowledge, no study has investigated the СО2/Н2S or СН4/СО2/Н2S mixed gas separations using the facilitated IL-based membranes. Therefore, there is high demand on future studies of the performance and stability of the membranes under СН4/СО2/Н2S mixed gas conditions which will be beneficial for the deployment of the membranes in real applications.
Due to the limited number of publications on the facilitated separation of acidic gases from gas mixtures using the SILMs, it is reasonable to investigate the gas transport behavior of task-specific carboxylate ionic liquids, such as 1-butyl-3-methylimidazolium acetate, toward CO2/CH4, H2S/CH4 and СН4/СО2/Н2S mixtures. In this work, the permeability of pure CO2, H2S and CH4 through the SILMs consisting of fluorinated polymer with immobilized 1-butyl-3-methylimidazolium tetrafluoroborate and 1-butyl-3-methylimidazolium acetate were measured in membrane gas separation setup with continuous sweeping gas supply. These measurements were performed in order to evaluate the distinctions in gas transport behavior of the ILs dissolving the acidic gases via the physical and chemical mechanisms. Permeabilities and selectivities of membranes for CO2/CH4, H2S/CH4, and CH4/CO2/H2S gas mixtures were compared with the respective single gas permeabilities and ideal selectivities; the stability of the SILMs under operational conditions was evaluated gravimetrically and via the contact angle data.

Materials and Methods
In this study, commercially available flat sheet porous microfiltration tetrafluoroethylenevinylidene fluoride composite membrane named as MFFK-1 with pores size 150 nm and thickness 100 μm, purchased from Vladipor JSC (Vladimir, Russia) was used as a polymeric support. The membrane was composed of tetrafluoroethylene-vinylidene fluoride copolymer of F42L (upper layer) and the nonwoven polypropylene bottom layer. To prepare the binary and ternary gas systems the high purity (99.9 vol.%) methane, carbon dioxide (99.99 vol.%) and hydrogen sulfide (99.9 vol.%) purchased from Monitoring (Russia) were used. Ionic liquids bmim[BF4] and bmim[OAc] were purchased from Sigma Aldrich Group (St. Louis, MO, USA). Their physical properties are given in  [35,36].
In this context, Santos et al. have obtained the results of CO 2 separation on the acetate-based SILMs that was near the Robeson upper bound corresponding to the best polymeric materials [34]. The promising acidic gases separation performance was achieved on the Al 2 O 3 /TiO 2 tubes with immobilized emim[OAc] [37]. To the best of our knowledge, no study has investigated the CO 2 /H 2 S or CH 4 /CO 2 /H 2 S mixed gas separations using the facilitated IL-based membranes. Therefore, there is high demand on future studies of the performance and stability of the membranes under CH 4 /CO 2 /H 2 S mixed gas conditions which will be beneficial for the deployment of the membranes in real applications.
Due to the limited number of publications on the facilitated separation of acidic gases from gas mixtures using the SILMs, it is reasonable to investigate the gas transport behavior of task-specific carboxylate ionic liquids, such as 1-butyl-3-methylimidazolium acetate, toward CO 2 /CH 4 , H 2 S/CH 4 and CH 4 /CO 2 /H 2 S mixtures. In this work, the permeability of pure CO 2 , H 2 S and CH 4 through the SILMs consisting of fluorinated polymer with immobilized 1-butyl-3-methylimidazolium tetrafluoroborate and 1-butyl-3-methylimidazolium acetate were measured in membrane gas separation setup with continuous sweeping gas supply. These measurements were performed in order to evaluate the distinctions in gas transport behavior of the ILs dissolving the acidic gases via the physical and chemical mechanisms. Permeabilities and selectivities of membranes for CO 2 /CH 4 , H 2 S/CH 4 , and CH 4 /CO 2 /H 2 S gas mixtures were compared with the respective single gas permeabilities and ideal selectivities; the stability of the SILMs under operational conditions was evaluated gravimetrically and via the contact angle data.

Materials and Methods
In this study, commercially available flat sheet porous microfiltration tetrafluoroethylenevinylidene fluoride composite membrane named as MFFK-1 with pores size 150 nm and thickness 100 µm, purchased from Vladipor JSC (Vladimir, Russia) was used as a polymeric support. The membrane was composed of tetrafluoroethylene-vinylidene fluoride copolymer of F42L (upper layer) and the nonwoven polypropylene bottom layer. To prepare the binary and ternary gas systems the high purity (99.9 vol.%) methane, carbon dioxide (99.99 vol.%) and hydrogen sulfide (99.9 vol.%) purchased from Monitoring (Russia) were used. Ionic liquids bmim [BF 4 ] and bmim[OAc] were purchased from Sigma Aldrich Group (St. Louis, MO, USA). Their physical properties are given in Table 1. All ILs were dried under vacuum for 24 h, at a point the water content found by Karl-Fischer titration (Coulometer 831 KF, Metrohm, Switzerland) less than 0.2% (w/w).  [38]; b [39].
For the preparation of supported ionic liquid membranes, a vacuum method was used. The polymeric support was fixed into a desiccator and was evacuated at 10 −5 mbar for at least for 2 h in order to remove air from pores. The ionic liquid was then introduced into a desiccator using a syringe, while maintaining the vacuum inside the desiccator, and spread on the surface of the membrane (for 2 h). Once samples were then taken out, the excess IL was removed by wiping with absorbing tissue. To determine the amount of IL immobilized (uptake), the membrane was weighed before and after immobilization procedure. The thickness was also determined before and after immobilization.
The membranes surface wettability was evaluated by contact angle measurements. The static contact angle was measured using an optical contact angle measurement system. A droplet of testing liquid was placed on the membrane surface and an image of drop shape was obtained. The optical system apparatus consisted of a light source, an adjustable stage, and a USB optical microscope. The microscope (Chuo Seiki, TS-H, Kumamoto, Japan) was fixed on an adjustable microscope mount. A digital image of the drop shape was made using a CCD camera interfaced to the microscope. Image J ® software with Dropsnake plugin was used for the calculation of the contact angle value. The contact angle value for each testing liquid was calculated as an average of 5 measurements of different positions for each sample. An electronic balance (Shimadzu AUW-220D, Kyoto, Japan) with the standard uncertainty of 0.01 mg was used for the gravimetric measurements.
The feed gas mixtures CH 4 /CO 2 , CH 4 /H 2 S and CH 4 /CO 2 /H 2 S were prepared in sealable gas container by static volumetric method. The CO 2 and H 2 S content in binary system were 15 ± 0.05 vol.% and 5 ± 0.05 vol.% (mixed with 85 ± 0.05 or 95 ± 0.05 of CH 4 ), respectively. The CO 2 and H 2 S content in ternary system were 18.2 ± 0.05 vol.% and 5.3 ± 0.05 vol.%, respectively. Gas mixture components were filled in the preliminary vacuumed container to reach the appropriate ratios. The gas mixtures composition verification was performed by gas chromatography (GC) method. The gases were introduced into the flow-through sampling valve of the gas chromatograph (Chromos GC-1000, Chromos Ltd., Dzerzhinsk, Nizhny Novgorod region, Russia) for the GC analysis. During the analysis the components of the permeate sample are separated in the chromatographic column (Porapak Q, 60/80 mesh, 353 K, 2 m × 3 mm i.d. stainless steel tube) in isothermal conditions and detected by the thermal conductivity detector (TCD, 100 mA, 100 • C). Carrier gas flow controller serves to supply the required flows of the carrier gas (Helium 99.9999+%). The low limit of detection was 4 ppmv for methane, hydrogen sulfide and carbon dioxide. The total analysis time was not more than 300 s.
The principal scheme of experimental setup is shown in Figure 2. The SILM with active area 4.9 cm 2 was placed on a PTFE disk support and sealed in a stainless-steel membrane module 1 using the Viton O-ring. The gas mixture is continuously supplied from the cylinder to the feed side of a membrane module (1) through a pressure regulator (2) with a constant pressure maintained at (200 ± 5) kPa. Pressure in the system was monitored by manometers (4). The gas permeated through the membrane is removed by a helium flow, which is also used as a carrier gas for GC system. Retentate flow rate is controlled by a needle valve (3) and monitored with a flow meter (5). Permeate content determination was performed by gas chromatography method. «Ex-situ» qualitative analysis of the permeate sample was carried out on the gas chromatograph-mass spectrometer (GCMS) QP2010Plus (Shimadzu, Kyoto, Japan) with a vacuum sample inlet system through automatic injection valve (Valco Instruments Co Inc., Houston, TX, USA). During the analysis the components of the permeate sample are separated on an Agilent capillary column Select for Permanent Gases/CO 2 with set of two parallel columns: CP-Molsieve 5 Å for permanent gases and PoraBOND Q for CO 2 analysis in accordance with the following temperature program: holding at 323 K (10 min), 323-423 K (20 K/min), holding at 423 K (5 min), carrier gas-helium (99.99999 vol.%). Quantitative analysis was performed by the method of absolute calibration. Reaction products were identified with the help of NIST-11 database of mass spectra and GCMS Real Time Analysis ® software. The stability of SILMs was determined by the gravimetric method. The weights of SILMs before and after N2 transport performances were compared. For the pressure values tested (1 bar), the distinctions in membranes weight are within the range of instrumental error.

Membranes Characterization
The hydrophobic tetrafluoroethylene-vinylidene fluoride copolymers (MFFK-1) with pore size equal to 150 nm was investigated as a supporting membrane material. This type of supporting material has been selected bearing in mind the results of the SILMs gas transport properties reported in our previous work [40]. According to those results, the highest values of acidic gases separation have been observed for the combination of MFFK-1 supporting material and bmim[BF4] liquid phase. It is noteworthy that the permeability of gases through the IL phase significantly exceeded that of through the dense polymeric support.
Compatibility of a supporting material with a liquid phase is a key factor determining the stability and operating properties of the SILMs. The interactions that may occur between the matrix and the IL contribute to the SILM stability, which can be estimated using contact angle measurements. The wettability investigations of hydrophobic MFFK-1 membrane by bmim[BF4] and bmim [OAc] have been carried out using the sessile drop method. The contact angles for the ILs are listed in Table 2 together with the calculated capillary pressures. According to this data, the fluorinated IL bmim[BF4] had remarkable affinity to the MFFK-1 surface composed of the fluorinated polymer. The contact angle of bmim[OAc] on MFFK-1 was slightly higher pointing to poorer interactions between the IL and the membrane. In general, it was found that the membrane has displayed a significant affinity toward both ILs.
In addition, the distribution of bmim[OAc] on the MFFK-1 surface was observed using scanning electron microscope (SEM) technique. Filling of the porous support by bmim[BF4] was examined in previous work [40]. As it can be seen from Figure 3, SEM micrographs of the membrane surfaces together with gravimetric measurements, contact angle data, and stabile gas transport properties revealed the compatibility of the supporting material and complete filling of the pores. The stability of SILMs was determined by the gravimetric method. The weights of SILMs before and after N 2 transport performances were compared. For the pressure values tested (1 bar), the distinctions in membranes weight are within the range of instrumental error.

Membranes Characterization
The hydrophobic tetrafluoroethylene-vinylidene fluoride copolymers (MFFK-1) with pore size equal to 150 nm was investigated as a supporting membrane material. This type of supporting material has been selected bearing in mind the results of the SILMs gas transport properties reported in our previous work [40]. According to those results, the highest values of acidic gases separation have been observed for the combination of MFFK-1 supporting material and bmim[BF 4 ] liquid phase. It is noteworthy that the permeability of gases through the IL phase significantly exceeded that of through the dense polymeric support.
Compatibility of a supporting material with a liquid phase is a key factor determining the stability and operating properties of the SILMs. The interactions that may occur between the matrix and the IL contribute to the SILM stability, which can be estimated using contact angle measurements. The wettability investigations of hydrophobic MFFK-1 membrane by bmim [BF 4 ] and bmim [OAc] have been carried out using the sessile drop method. The contact angles for the ILs are listed in Table 2 together with the calculated capillary pressures. According to this data, the fluorinated IL bmim[BF 4 ] had remarkable affinity to the MFFK-1 surface composed of the fluorinated polymer. The contact angle of bmim[OAc] on MFFK-1 was slightly higher pointing to poorer interactions between the IL and the membrane. In general, it was found that the membrane has displayed a significant affinity toward both ILs. In addition, the distribution of bmim[OAc] on the MFFK-1 surface was observed using scanning electron microscope (SEM) technique. Filling of the porous support by bmim [BF 4 ] was examined in previous work [40]. As it can be seen from Figure 3, SEM micrographs of the membrane surfaces together with gravimetric measurements, contact angle data, and stabile gas transport properties revealed the compatibility of the supporting material and complete filling of the pores.   The stability of the supported ionic liquid membranes depends on the capillary holding force, the supporting material pore sizes, and the viscosity of a liquid phase. This requires that during gas separation processes, the total amount of IL immobilized should remain constant inside the pores of the support. The minimal pressure required for desorption of the impregnated phase from support pores is calculated by the Young-Laplace equation at the steady state condition: where pc is the capillary pressure in Pascals, σ is the surface tension in N/m, θ is the contact angle in degree, and r is the average pore radius in meters [42]. Stability measurements have been carried out gravimetrically for 12 h at the N2 transmembrane pressure equal to 1 bar. The membrane weight together with the liquid phase loss as a function of time is depicted in Figure 4. It is observed that the membrane weight loss is only owing to the partial desorption of an IL immobilized within the supporting material. Both ILs have exhibited the similar stability behavior, so the gradual decrease of liquid phase has been noticed during the 10 h followed by a further stabilization. In case of bmim[OAc] the membrane weight diminished to approximately 93% accompanied by the IL loss equal to 14%. The SILM impregnated by bmim[BF4] loses 17.3% of the IL, while the membrane weight change was about 10%.
The obtained results have been compared with the data reported in literature [14,17,43]. Neves et al. compared the influence of the support hydrophobicity on the stability of the SILMs composed of fluorinated ILs [14]. The weight loss of hydrophilic membranes was more pronounced reaching 11-13%, while the hydrophobic ones were characterized by 1.5-7% decrease of the SILM weight. Alkhouzaam et al. evaluated the stability of polysulfone-supported ionic liquid membranes applying a pressure difference of 10 bars, for which the IL weight loss did not exceed 30% [17]. Zhao et al. The stability of the supported ionic liquid membranes depends on the capillary holding force, the supporting material pore sizes, and the viscosity of a liquid phase. This requires that during gas separation processes, the total amount of IL immobilized should remain constant inside the pores of the support. The minimal pressure required for desorption of the impregnated phase from support pores is calculated by the Young-Laplace equation at the steady state condition: where p c is the capillary pressure in Pascals, σ is the surface tension in N/m, θ is the contact angle in degree, and r is the average pore radius in meters [42]. Stability measurements have been carried out gravimetrically for 12 h at the N 2 transmembrane pressure equal to 1 bar. The membrane weight together with the liquid phase loss as a function of time is depicted in Figure 4. It is observed that the membrane weight loss is only owing to the partial desorption of an IL immobilized within the supporting material. Both ILs have exhibited the similar stability behavior, so the gradual decrease of liquid phase has been noticed during the 10 h followed by a further stabilization. In case of bmim[OAc] the membrane weight diminished to approximately 93% accompanied by the IL loss equal to 14%. The SILM impregnated by bmim[BF 4 ] loses 17.3% of the IL, while the membrane weight change was about 10%.

Single Gas Permeability
The results of the single gas (CH4, CO2 and H2S) permeabilities through the prepared SILMs are shown in Figure 5

Explanation of different gas transport behavior in the bmim[BF4] and bmim[OAc]
is based on two different mechanisms of acidic gases penetration across the SILMs. In case of bmim[BF4] the transport of CO2 and H2S undergoes in accordance with the physical absorption of gases in the liquid phase followed by the diffusion on the other side of the membrane. The permeation of the non-polar methane molecules through both ILs also takes place within the solution-diffusion mechanism. Therefore, the gas transport behavior of the SILMs containing bmim[BF4] and the methane transport through both ILs may be analyzed via the solution-diffusion mechanism, according to which the permeabilities of the individual gases are considered in terms of solution-diffusion model defined as a product of the solubility coefficient and the diffusivity coefficient: where P is permeability in mol·m/(m 2 ·s·Pa), S is the solubility coefficient in mol/(m 3 ·Pa) and D is the diffusivity coefficient in m 2 /s. The diffusivity coefficients for the SILMs were calculated by an equation proposed by Morgan et al. [44] for imidazolium-based ILs: The obtained results have been compared with the data reported in literature [14,17,43]. Neves et al. compared the influence of the support hydrophobicity on the stability of the SILMs composed of fluorinated ILs [14]. The weight loss of hydrophilic membranes was more pronounced reaching 11-13%, while the hydrophobic ones were characterized by 1.5-7% decrease of the SILM weight. Alkhouzaam et al. evaluated the stability of polysulfone-supported ionic liquid membranes applying a pressure difference of 10 bars, for which the IL weight loss did not exceed 30% [17]. Zhao et al. studied the effect of the pore size and the transmembrane pressures on the polysulfone SILMs stability and reported the 20% loss of IL at 1 bar pressure difference [43]. In general, the results on the SILMs stability studied in this work have correlated with literature, taking into account the distinctions in the membranes composition, pore sizes and transmembrane pressures.

Single Gas Permeability
The results of the single gas (CH 4 , CO 2 and H 2 S) permeabilities through the prepared SILMs are shown in Figure 5

Single Gas Permeability
The results of the single gas (CH4, CO2 and H2S) permeabilities through the prepared SILMs are shown in Figure 5 for both bmim[BF4] and bmim[OAc] ionic liquids. It can be observed that pure H2S shows the highest permeability (~380 Barrer) through the MFFK-1 doped by bmim[BF4] than pure CO2 and CH4 (~80 and 12 Barrer, respectively). In terms of MFFK-1 containing bmim[OAc] the permeabilities are much lower, taking on the values of ~110, 90 and 6 Barrer, respectively. Explanation of different gas transport behavior in the bmim[BF4] and bmim[OAc] is based on two different mechanisms of acidic gases penetration across the SILMs. In case of bmim[BF4] the transport of CO2 and H2S undergoes in accordance with the physical absorption of gases in the liquid phase followed by the diffusion on the other side of the membrane. The permeation of the non-polar methane molecules through both ILs also takes place within the solution-diffusion mechanism. Therefore, the gas transport behavior of the SILMs containing bmim[BF4] and the methane transport through both ILs may be analyzed via the solution-diffusion mechanism, according to which the Explanation of different gas transport behavior in the bmim [BF 4 ] and bmim[OAc] is based on two different mechanisms of acidic gases penetration across the SILMs. In case of bmim [BF 4 ] the transport of CO 2 and H 2 S undergoes in accordance with the physical absorption of gases in the liquid phase followed by the diffusion on the other side of the membrane. The permeation of the non-polar methane molecules through both ILs also takes place within the solution-diffusion mechanism. Therefore, the gas transport behavior of the SILMs containing bmim [BF 4 ] and the methane transport through both ILs may be analyzed via the solution-diffusion mechanism, according to which the permeabilities of the individual gases are considered in terms of solution-diffusion model defined as a product of the solubility coefficient and the diffusivity coefficient: where P is permeability in mol·m/(m 2 ·s·Pa), S is the solubility coefficient in mol/(m 3 ·Pa) and D is the diffusivity coefficient in m 2 /s. The diffusivity coefficients for the SILMs were calculated by an equation proposed by Morgan et al. [44] for imidazolium-based ILs: where D is the diffusivity of gases in ILs (m 2 ·s −1 ); µ 2 is the IL viscosity (Pa·s) at 298 K (Table 1); V 1 is the gas molar volume (m 3 ·mol −1 ); and ρ 2 2 is the density of the IL (kg·m −3 ). As shown in Table 3, among the gases the lowest diffusivity coefficient was in case of CO 2 penetration caused by the largest dimensions of the molecule (the Lennard-Jones diameter) whereas the CH 4 and H 2 S diffusion was almost identical. Owing to the higher viscosity of bmim[OAc], the calculated diffusivity coefficient for this IL was approximately 3-fold lower than this for bmim [BF 4 ]. Table 3. Diffusivities (D) of various gases in the ILs at 298 K, the gas solubility's based on the literature (S lit ) and calculated (S calc ) data, and the gas permeabilities (P). The solubilities of the different gases (S calc ) in the IL at the feed conditions (298 K, 0.2 MPa) were obtained using Equation (2), in which the permeabilities were preliminarily divided by the porosity of the membrane (80%), due to gas permeation occuring only through the pores filled with IL. Alternatively, known from the literature data [16] the solubilities of gases in the ILs (S lit ) given in mole fraction x were converted to the required units and compared with the calculated data:

Gas
where M IL is the molecular weight of the IL, ρ IL is the density of the IL and p is the feed pressure. The permeation test results for bmim [BF 4 ] were almost identical to those described in our study [40], but the remarkable differences were found compared with the solubility data known from literature. In particular, whereas the methane solubility coefficients were similar to the S lit , the CO 2 and H 2 S solubilities were in 4.25 and 4.45 times lower. Explanation of the dramatic diminishing of ILs sorption properties toward the acidic gases may be the consequence of water presence in bmim[BF 4 ], which is highly hydroscopic. Indeed, during the preparation of the SILM, it was exposed to air resulting in the presence of 0.168% (w/w) of water. This may lead to the reduction in gas solubility and ultimately to the reduction in the permeance. In the acetate-based IL the calculated solubility of methane differed from S lit insignificantly. On the other hand, the acetate-based IL interacts chemically with acidic gases (CO 2 and H 2 S) yielding the formation of the adducts [47]. Moreover, the facilitated transport mechanism of acidic gases across bmim[OAc] favors low concentrations and decreases drastically with increase of transmembrane pressure. In particular, Zhang et al. have reported the superior permeabilities of the acetate-based SILMs varying in a range about 1000-7000 Barrer at transmembrane pressures 0.1-0.5 bar. In the present work, the transmembrane pressure was equal to 1 bar, which, probably, dramatically diminished to the permeability of acidic gases. The distinctions in permeabilities of both acidic gases was less significant than in bmim [BF 4 ] indicating the chemical binding of gases. This observation is additionally proved by the comparable solubility values of carbon dioxide (0.272 mole fraction [48]) and hydrogen sulfide (0.255 mole fraction [49]) in bmim [OAc].
The chemisorption of acidic gases in bmim[OAc] has resulted in an unsubstantial difference between the selectivity of CO 2 /CH 4 and H 2 S/CH 4 . And vice versa, for the SILMs impregnated by bmim[BF 4 ] the distinctions concerning those two acidic gases separation were remarkable. The ideal selectivity as the ratio of the permeances of CO 2 and H 2 S pure gases over the CH 4 are presented in the Table 4. The measured difference between the ideal selectivity values can be attributed to the solubility selectivity of bmim [BF 4 ] IL, which is higher for hydrogen sulfide and is lower for CO 2 . The ideal selectivity for CO 2 /CH 4 and H 2 S/CH 4 will be compared next with the mixed gas selectivity.

Mixed Gas Permeability
The mixtures of CH 4 /CO 2 = 85/15% (v/v) and CH 4 /H 2 S = 95/5% (v/v) were also investigated for measuring the permeabilities and calculating the separation factor, as shown in the Table 4. Compared to the single gas permeation results, the CO 2 and CH 4 permeabilities of the MFFK-1 bmim[BF 4 ] membrane measured in the mixed gas test were much lower. The separation CH 4 /CO 2 and CH 4 /H 2 S binary mixtures across the SILMs impregnated by bmim [BF 4 ] was generally determined by the solubility selectivity and was comparable with the separation of ternary mixture (Table 4). This is reflected in the fact that the selectivities of CO 2 /CH 4 and H 2 S /CH 4 separation were equal to 8.7 and 23.4 in the binary mixtures and 8.1 and 15.9 in the ternary mixture.
Meanwhile in the case of bmim[OAc] the value of CO 2 permeability doubled to 180 Barrer, which led to a sharp increase of CO 2 /CH 4 selectivity (~96.9) in comparison with the ideal selectivity. As mentioned in [47], the facilitated transport membranes achieve the highest separation results at lowest concentrations of the active penetrants, therefore individual gases testing shows the modest results compared to the gas mixtures. Interestingly, the separation factors for both acidic gases in bmim [OAc] differ insignificantly as in case of individual gases. The results of the CH 4 /H 2 S mixed gas separation revealed the similar behavior as in the case of CH 4 /CO 2 . The effect of the incorporation of bmim [OAc] IL in the SILM is more pronounced, causing the double growth of the mixed H 2 S permeability up to 205 Barrer and consequently the sharp increase of the H 2 S/CH 4 selectivity up to 102.9. Hence, it is clear that the incorporation of bmim [OAc] in SILM is very promising for the removal of the both acidic gases from the gas streams.
In the separation of the ternary CH 4  The separation of the ternary mixture CH 4 /CO 2 /H 2 S has confirmed the previously noted patterns of the CO 2 and H 2 S penetration. In bmim[BF 4 ] the differences in CO 2 and H 2 S solubilities resulted in the remarkable permeability variance, whereas the chemisorption in bmim[OAc] almost equalized the permeabilities of them. From this point of view, the gas transport behavior of both ILs for CH 4 /CO 2 /H 2 S mixture duplicates the results of binary mixtures separation and pure gases. At the same time, the particular phenomena was observed when separating the binary and ternary gas mixtures through the membrane immobilized with bmim[OAc], such SILMs were almost impermeable for the methane despite its high concentrations in gas mixtures. Most probably, the changes in the IL structure led to an extremely low solubility of methane in the CO 2 -or H 2 S-saturated ILs resulted in a negligible permeability values. This tendency was also noticed for the ternary mixture permeation. This point stipulated the high values of selectivity of the proposed membranes toward acidic gases.

Effect of Temperature on the H 2 S Transport
The effect of temperature on the CO 2 penetration across SILMS has been widely addressed in the literature, incorporating reports on the membranes impregnated by either conventional ILs [5][6][7]10,[12][13][14][15] or providing the facilitated transport ones [6,27,47]; therefore, the further studies of the SILMs were carried out for H 2 S-containing binary mixture in the temperature range 298-333 K. The membrane impregnated by bmim [BF 4 ] was excluded from the study owing to a sufficient level of scrutiny of such SILMs.
As illustrated in Figure 6, the permeability of CH 4 and H 2 S increase with increasing temperature due to the acceleration of the diffusion rate. At the elevated temperatures, the viscosity of IL dramatically diminishes resulting in the enhanced gas diffusion. The permeabilities of CH 4 and H 2 S increase from 2 to 8 Barrer and from 205 to 573 Barrer, respectively. It is noteworthy that, despite smaller dimensions of the H 2 S molecule, the permeability of H 2 S is doubled in the temperature range 298-333 K, whereas that of CH 4 increases fourfold. This is most probably due to both decreasing of the physical solubility and weakening the chemical interaction between bmim[OAc] and H 2 S [47]. The selectivity of H 2 S/CH 4 separation has undergone a significant decrease from 102.9 to 72. Thermodynamic factors together with weakening the complexation of H 2 S with the IL contribute to a notable decline in the separation efficiency of the membrane. 298-333 К, whereas that of CH4 increases fourfold. This is most probably due to both decreasing of the physical solubility and weakening the chemical interaction between bmim[OAc] and H2S [47]. The selectivity of H2S/CH4 separation has undergone a significant decrease from 102.9 to 72. Thermodynamic factors together with weakening the complexation of H2S with the IL contribute to a notable decline in the separation efficiency of the membrane.

SILMs Performance Comparison
The gas separation performances of the SILMs containing bmim[BF4] or bmim[ace] reported in literature are represented in Table 4, including the results of this work. Those data correspond to both individual gases separation, and gas mixtures. It is worth drawing attention to the similarities between the SILMs containing bmim[BF4] studied in our work and the ones mentioned in [47]. Although the permeabilities of all gases in [47] were several times higher, the selectivities of CO2/CH4 and H2S/CH4 separation were consistent with our results. Other literature reference [50] reports the similar CO2/CH4 ideal selectivity, but notices the considerably higher results for H2S/CH4 separation. Owing to the approximately equal values of separation of gas mixtures and individual gases on the SILM impregnated by bmim[BF4], the relevant comparison of the pure gases is applicable to the gas mixtures as well. On the contrary, the permeabilities of the individual gases across bmim[OAc] as well as the selectivities were fairly low and were not comparable with the data given in work [49].

SILMs Performance Comparison
The gas separation performances of the SILMs containing bmim[BF 4 ] or bmim[ace] reported in literature are represented in Table 4, including the results of this work. Those data correspond to both individual gases separation, and gas mixtures. It is worth drawing attention to the similarities between the SILMs containing bmim [BF 4 ] studied in our work and the ones mentioned in [47]. Although the permeabilities of all gases in [47] were several times higher, the selectivities of CO 2 /CH 4 and H 2 S/CH 4 separation were consistent with our results. Other literature reference [50] reports the similar CO 2 /CH 4 ideal selectivity, but notices the considerably higher results for H 2 S/CH 4 separation. Owing to the approximately equal values of separation of gas mixtures and individual gases on the SILM impregnated by bmim [BF 4 ], the relevant comparison of the pure gases is applicable to the gas mixtures as well. On the contrary, the permeabilities of the individual gases across bmim[OAc] as well as the selectivities were fairly low and were not comparable with the data given in work [49]. Only in case of hydrogen sulfide removal the selectivity values that were found for the binary and ternary mixtures were consistent with the literature data. The binary mixture separation has been slightly higher than that reported in the aforementioned work as well as for the ternary mixture. Most probably, diminishing the selectivity for ternary mixture is the result of the competitive absorption phenomena of the both acidic gases, which is not possible in binary mixtures. Removal of carbon dioxide from gas mixtures on bmim[OAc] was found to be more pronounced in our work than in literature. And obtained membranes might be used for the separation of above mentioned mixtures in cascades for better separation efficiency [50,51].

Conclusions
In this work, the membranes consisting of fluorinated porous polymer with immobilized 1-butyl-3-methylimidazolium tetrafluoroborate and 1-butyl-3-methylimidazolium acetate have been tested for pure CO 2 , H 2 S and CH 4 as well for the binary mixtures of H 2 S/CH 4 , CO 2 /CH 4 and the ternary CH 4 /CO 2 /H 2 S gas mixture.
The permeability of carbon dioxide (single) through the bmim[OAc]-based SILM has been approximately 2 times greater as compared to the SILM doped by bmim [BF 4 ]. On the contrary, the ideal selectivity of H 2 S/CH 4 has been much higher than that for facilitated transport membrane (containing bmim[OAc]). The opposite trend has been observed for the mixed gas transport across the facilitated transport membrane. The latter was found to exhibit a negligible permeability of methane in CO 2 -or H 2 S-containing mixtures resulting in remarkable high selectivity of CO 2 or H 2 S separation. The gas transport behavior of bmim[OAc] for ternary mixture has duplicated the results of binary mixtures separation. In the case of bmim[OAc], the mass transfer of gases across SILMs may be enhanced by carrier-mediated reversible reactions.
From the results obtained in this work, it can be inferred that facilitated transport membranes containing the task-specific ionic liquids, which are able to form N-heterocyclic carbenes, is very promising media for the removal of the both acidic gases from the gas streams.

Conflicts of Interest:
The authors declare no conflict of interest.