Journal of Membrane Science 449 (2014) 109–118 Contents lists available at ScienceDirect Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci Gas transport properties of interfacially polymerized polyamide composite membranes under different pre-treatments and temperatures Jonathan Albo, Jinhui Wang, Toshinori Tsuru n Department of Chemical Engineering, Hiroshima University, 1-4-1 Kagayami-yama, Higashi-Hiroshima 739-8527, Japan art ic l e i nf o a b s t r a c t Article history: Received 10 July 2013 Accepted 21 August 2013 Available online 27 August 2013 Thin-film composite reverse osmosis membranes were dried under different membrane pre-treatment procedures and evaluated at increased temperatures by gas separation tests. The obtained permeance and selectivity values indicated the presence of highly-permeable regions in the dry samples of the commercial membranes. Treatment with ethanol–hexane in a solvent exchange process, as well as membrane immersion in t-butanol followed by freeze drying, increased the gas permeance by a factor of 1.8 to 9, and from 1.6 to 3.2, respectively, by comparison with room temperature and oven drying. Nevertheless, a Knudsendiffusion transport mechanism was dominant after both pre-treatments. The permeation temperature remarkably influenced gas selectivity and permeance, and a maximum He/N2 selectivity occurred at 150 1C with considerable high permeance results, which may suggest the use of polyamide membranes as alternative materials for high-temperature separation processes. The temperature-induced changes in the polymer structure and in the transport of compounds can be explained by Knudsen and activated diffusion mechanisms throughout a highly-permeable regions and a dense polyamide matrix, respectively. & 2013 Elsevier B.V. All rights reserved. Keywords: Gas separation RO membranes Polyamide Membrane pre-treatments Temperature influence 1. Introduction The reverse osmosis (RO) process, which uses polymeric semipermeable membranes to achieve molecular separation, is now an economic and universally accepted technique, with the major breakthrough in this field being achieved with the development of thin-film composites membranes (TFC) [1]. These TFC materials generally consist of three layers: a polyamide barrier skin, a porous support layer (often polysulfone), and a non-woven polyester. Both the polysulfone and polyester layers provide mechanical support for the polyamide layer, which generally provides selectivity to the membrane. The thin selective upper layer and porous support can be separately optimized to give high permeability and selectivity. Today, most commercial RO TFC membranes are formed in-situ by the interfacial polymerization of an aromatic polyamine such as m-phenylenediamine (MPD) with one or more aromatic polyacyl halides (for example, trimesoyl chloride (TMC)). These aromaticbased membranes exhibit good performance in many desalination and water purification applications and are already in mass n Corresponding author. Tel.: þ 81 824 24 7714; fax: þ 81 824 22 7191. E-mail address: tsuru@hiroshima-u.ac.jp (T. Tsuru). 0376-7388/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.memsci.2013.08.026 production [2,3] because of the high flux and high rejection provided by the thin separating layer in the composite structure. Aromatic polyamide materials, however, have not merited comparatively special attention for gas separation either in dense membranes [4–6] or composite structures [6–12] despite the mechanical and chemical properties that make them attractive for this separation process [13]. Since the active layer is extremely thin in TFC membranes, treatment prior to their use for gas separation is a very sensitive process. Usually this pre-treatment involves two general steps: (1) cleaning of the membrane surface by a solvent, and (2) drying of the residual solvent from within the membrane structure. The water-swollen hydrogel that fills the pores in the support membrane becomes a rigid glass when dried for use in gas separation, giving very low gas permeability, although relatively high selectivities may yet be achieved [2]. Therefore, membrane pretreatment can cause shrinkage and swelling, but can also cause the removal of residual monomers or additives and morphological changes that effectively varied membrane properties [7,8,14–21]. Additionally, RO membranes are generally perceived to have a non-porous separating layer where transport occurs via a solution– diffusion process, but an examination of the literature indicated that some dry RO membranes exhibits highly-permeable regions that are usually attributed to the presence of membrane defects [7,8,16–19] 110 J. Albo et al. / Journal of Membrane Science 449 (2014) 109–118 formed either during the membrane preparation or during drying treatments. For instance, Louie et al. [7] identified defects in commercial polyamide RO membranes that were eliminated when the membrane surface was coated with a polyether–polyamide block copolymer, suggesting that using n-butanol as a solvent for applying coatings negatively affects water and gas permeation. Kuehne et al. [19] studied the effect of varying processing conditions during TFC RO membrane fabrication and reported an enhanced flux after washing the membrane with glycerol and organic salts. They attributed rejection differences between membrane samples to membrane imperfections. Jezowska et al. tested RO polyamide membranes after a cleaning step concluding that membrane homogeneities could be partially caused by improper pre-treatment of the membrane material [16]. However, Sridhar et al. [11] tested interfacial polymerized TFC membranes with an ultrathin defect-free polyamide skin layer for gas separation. The limited selectivities obtained, in comparison to that of a dense film, suggested consideration of the presence of less cross-linked regions in the polyamide separation layer. Li et al. [12] recently compared defect and defect-free interfacially-polymerized TFC membranes, attributing the deterioration of gas selectivity not only to membrane defects, but to cross-linked polyamide layer density. As a consequence, a clear understanding of the TFC RO polyamide membrane pore structure has not yet been achieved. In general, the transport of molecules at different temperatures is dependent upon the available free volume in the polymer matrix, as well as sufficient energy of the molecules to overcome attractive forces between chains. In this sense, analysis of the change in membrane permeation due to polymer flexibility with temperature is useful for characterization of the TFC RO polyamide membrane pore structure. Because RO membrane pores are equal to, or only a few times the size of gas molecules, transport across these membranes can be expected to occur in the intermediate regions from Knudsen diffusion and activated diffusion. Ideally, this mechanism enables gas molecules to be separated, while the separation of those gas molecules is difficult if their diameter sizes are similar [22]. Since different conditions can influence membrane heterogeneity and the transport of compounds through the selective polyamide layer, the main objective of this work was to systematically explore whether drying pre-treatments and temperature can result in improvements in asymmetric RO membranes for gas separation. To the best of our knowledge, no studies of the effect of increasing permeation temperature on the separation performance of TFC RO polyamide membranes have been reported. The effects are discussed in terms of permeance and separation factor variations. An additional part of this work was devoted to determining whether such effects also produce the transformation of inhomogeneous regions into denser structures, thus defining the transport mechanism through the membrane. A clearer understanding of the effects of these processes on membrane properties will aid in the development of improved membranes with better separation performance for hightemperature processes. Table 1 Water permeability and salt rejection of RO membranes as listed by the provider. Membrane Test conditions Water permeability [L/(m2 h bar)] Salt rejection [%] SWC5 ESPA2 CPA5 32,000 ppm; 5.5 MPa; 25 1C 1500 ppm; 1.05 MPa; 25 1C 1500 ppm; 1.55 MPa; 25 1C 1.38 4.15 3.28 99.8 99.6 99.7 meta-bisulfite solution. Table 1 shows the water permeability and salt rejection in the RO membranes under the conditions listed in the provider information data sheet. The membranes consisted of a thin-film-composite with a topskin aromatic polyamide, PA, layer ( 200 nm), a middle microporous polysulfone, PSF ( 40 mm), and a bottom polyethylene terephthalate, PET, layer (  120 mm). The specific chemical composition of the PA layer is proprietary information of the supplier. PA membranes were additionally formed in-situ by interfacial polymerization [23] and applied in some sections of this work for comparison. Briefly, 1 ml of aqueous solution of 1,3-phenylenediamine (m-phenylenediamine, MPD) (2 wt%) and sodium lauryl sulfate (0.15 wt%) was poured on a PSF membrane, and then the excess of the solution was removed softly with filter paper. Subsequently, a 1 ml hexane solution of 1,3,5-benzenetricarbonyl trichloride (trimesoyl chloride, TMC) (0.1 wt%) was poured on the support. After a 1min polymerization reaction, the excess solution was drained, and the membrane was dried in air for 15 min. Finally, the membrane was rinsed with deionized water. All high purity chemicals of analytical grade applied in this study were provided by Sigma Aldrich (Japan). 2.2. Membrane pre-treatment methods Based on a review of the available literature, three different common membrane pre-treatments were selected [11,24,25]. The procedures were adapted and applied to membrane samples prior to their use in gas separation: Room Temperature–Oven (RTO): Membranes were washed several times in a pure-water bath, then dried at room temperature for 24 h, and finally placed in an oven at 120 1C for 30 min [11]. Ethanol–Hexane (EH): Membranes were washed several times in a pure-water bath, then immersed in ethanol for 5 min and afterwards soaked in a hexane bath for 1 min. Finally, the solvent was evaporated at room temperature for 15 min [11,24]. Freeze Drying (FD): Membranes were washed several times in a pure-water bath, then immersed in 50, 75, 90, 95, and 100 wt% t-butanol aqueous solutions for 15 min. Then, membrane samples were placed in pure t-butanol in freeze-dried equipment under vacuum for 2 h [25]. 2.3. Gas separation experiments 2. Experimental 2.1. Materials Three commercial RO membranes were provided by Niito Denko (Japan) and applied in this study: SWC5 (seawater membrane), ESPA2 (energy-saving RO membrane), and CPA5 (high-rejection RO membrane). Membrane samples were vacuum sealed in a polyethylene bag containing less than 1% sodium Membrane samples (2.21 cm2) were tested in a stainless permeation cell using He, H2, CO2, O2, N2, C3H8, and SF6 at temperatures that ranged from ambient (room temperature, 16 73 1C) to 200 1C in an oven. A schematic drawing of the experimental apparatus appears in Fig. 1. The feed gas pressure was set at 2.5 bar, and the permeate was at atmospheric pressure. The flow rate of the permeating gas was measured using a bubble flow meter. Prior to testing, membranes were under vacuum with a He flow for 1 h. Additionally, samples were under vacuum for 111 J. Albo et al. / Journal of Membrane Science 449 (2014) 109–118 He H2 CO2 O2 N2 C 3 H8 SF6 Fig. 1. Schematic drawing of the gas experimental setup. 5 min between different gas measurements in order to remove the gas from within the gas permeation equipment. Gas permeances were calculated using the following equation: PðiÞ ¼ jðiÞ ð1Þ 2 where P(i) is the permeance of i (mol/(m s Pa)), J(i) is the permeate flux (mol/(m2 s)), and Δp(i) is the partial pressure difference (Pa). The ability of the membrane to separate different gases, depending on the separating layer properties, was quantified by the membrane gas selectivity: PðiÞ PðjÞ 10-8 SWC5 ESPA2 CPA5 10 -9 2 3 4 5 6 Kinetic diameter [Å] ΔpðiÞ αði=jÞ ¼ Permeance [mol/(m2·s·Pa)] 10-7 Fig. 2. Single-gas permeation of the three commercial PA membranes after drying at RTO. Permeance was measured at room temperature (167 3 1C). 3. Results and discussion 3.1. Gas permeation properties of RO samples ð2Þ where α is the membrane selectivity for gas i relative to gas j. It is generally recognized that the gas transport mechanism through polymeric membranes is controlled by the solution and diffusion processes of the permeating gases, where the temperature dependence of permeance is decided by the relative magnitude of two counteracting contributions: kinetic (diffusion) and equilibrium (sorption). The number of molecules with kinetic energy larger than the activation energy for permeation, Ep, is proportional to exp( Ep/RT), assuming a Maxwellian velocity distribution of molecules. Therefore, the permeance P(i) for the activated diffusion mechanism can be expressed by an Arrheniustype relationship:   EpðiÞ ð3Þ PðiÞ ¼ CðiÞexp RT where C(i) is a constant depending on the system and Ep(i) (kJ/mol) is the activation energy for the permeation of gas, i, which is the difference between the sorption energy and substantial activation for diffusion. 2.4. Characterization a) Infrared spectroscopy The top surfaces of the membrane samples were pressed down against a 451 single-reflection internal element in a JASCO FTIR4100 Fourier Transform Infrared (FITR) with Attenuated Total Reflection (ATR-IR) spectrometer. Results were treated using a JASCO Spectra Manager Version 2.1. b) Thermal stability The polymer thermal stability was characterized by simultaneous thermogravimetric analysis (TGA) and by differential thermal analysis (DTA). Measurements were carried out in a Shimadzu TG/DTA-60 apparatus in nitrogen for a single heating cycle between room temperature and 500 1C at a constant heating rate of 10 1C/min. For TGA, the mass loss of the sample during heating was recorded. a) Scanning electron microscopy Surface images of membrane samples were obtained with a Hitachi S-4800 scanning electron microscope (SEM) using an accelerating voltage of 4 kV. Samples were sputter-coated with palladium/platinum to minimize charging in a JEOL JFC-1300 Auto Fine Coater. Fig. 2 shows the single-gas permeance, P, of all gases tested at room temperature as a function of kinetic diameter for three membrane types (SWC5, ESPA2 and CPA5) when pre-treated with RTO procedure. Samples were cut from the same membrane sheet, which minimized the membrane heterogeneity effect, and three samples were measured for each membrane type. Gas permeation variability for membrane samples of the same type varied by as much as 16.2%; these differences were attributed to uneven membrane density and to heterogeneities in the separating layer. N2 permeance, P N2 , in the SWC5 membrane resulted in P N2 ¼1.64 70.16  10 8 mol/(m2 s Pa), which was within the value range reported in the literature for the same membrane type, P N2 ¼1.6  10 8–3.4  10 8 mol/(m2 s Pa), and corresponded to the high permeance of 50–100 GPU (1 GPU ¼3.4  10 10 mol/ (m2 s Pa)) [7]. The membrane average selectivity (α), calculated using Eq. (2), for each gas with respect to nitrogen permeance is presented in Table 2. The gas selectivity value for all membranes under RTO pretreatment was consistent with Knudsen characteristics for CO2, O2, C3H8, and SF6 gases. Nevertheless, when compared with Knudsen diffusion, a slightly higher selectivity value was obtained for He and H2 permeance, which may indicate a mixed transport mechanism. The performance of these commercial membranes for the He/ N2 pair resulted in an unfavorable result when compared with the performance for the same pair in pure aromatic PA dense membranes synthesized from monomers bearing methyl substituents and p-phenylene hinge-like connecting linkages (α(He/N2) ¼ 67.39) [4] and for PA membranes bearing a pendent phenyl group, a hexafluoroisopropylidene (6F) linkage (α(He/N2)¼  40) and a sulfonyl (SO2) linkage (α(He/N2) ¼ 100) [26]. Additionally, the result was far from that of H2/N2 selectivity achieved with similar dense materials, such as poly(amide-imide), in Robeson's upper bound correlation, (α(H2/N2)¼72) [27]. The lower selectivities reported for the commercial membranes could be ascribed either to membrane defects or to the presence of a less crosslinked PA layer, where the higher polymer motion led to a decrease in gas separation. Moreover, N2 permeance in the PA membrane formed in-situ in this work resulted in P N2 ¼9.46 70.34  10 10 mol/(m2 s Pa), which was about a twoorder of magnitude lower permeance than that obtained by the TFC membranes. The relatively high permeances obtained may 112 J. Albo et al. / Journal of Membrane Science 449 (2014) 109–118 He Table 2 Gas selectivity values for three commercial RO membranes under RTO. SWC5 ESPA2 CPA5 Knudsen Gas selectivity, α [–] 1.647 0.16 1.30 70.14 1.117 0.18 He/ N2 H2/ N2 CO2/ N2 O2/ N2 C3H8/ SF6/ N2 N2 3.96 4.10 4.30 2.64 4.03 4.23 4.55 3.73 0.82 0.95 0.77 0.80 0.96 0.95 0.92 0.94 0.80 0.82 0.80 0.80 0.42 0.46 0.42 0.44 Permeance [mol/(m2·s·Pa)] Membrane N2 permeance, P N2  108 [mol/(m2 s Pa)] 10-5 10-6 Absorbance CPA5-PSF 1600 CPA5-RTO 10-7 10-8 2 3 4 5 6 Fig. 4. Single-gas permeation of CPA5-RTO and CPA5-PSF samples at room temperature (167 3 1C). (a) 1800 CPA5-PSF Kinetic diameter [Å] (d) (b) SF6 10-4 10-9 (c) CPA5-RTO CO2 N2 10-3 1400 1200 1000 Wavenumber [cm-1] Fig. 3. IR normalized spectra for the CPA5-RTO sample and membrane exposed to hypochlorite solutions (CPA5-PSF). (a) Amide bonding, C–N stretching; (b) amide bonding, C¼ O stretching; (c) PSF, C–O stretching; and, (d) PSF, symmetric SO2 stretching vibration. indicate that the PA separating layer of the commercial TFC membranes consists of two different structures: a dense matrix and highly-permeable regions in which gas flows primarily via Knudsen diffusion. 3.2. Estimation of highly-permeable regions in the separating layer Since the PA active layer of RO membranes is extremely vulnerable to chlorine compounds, a membrane sample was immersed in a sodium hypochlorite solution for 96 h [28]. This procedure was intended to chemically remove the PA layer, to which selectivity membrane characteristics are attributed, and therefore to evaluate the separation performance of the middle and support layers (PSF-PET). CPA5 membrane was primarily applied for the remainder of the present work due to the slightly higher He/N2 selectivity, as shown in Table 2. The partial removal of the PA top layer was confirmed in the ATR-FTIR analysis presented in Fig. 3. The FTIR spectra was normalized with 1148 cm 1 band absorbance, which can be attributed to the PSF symmetric SO2 stretching vibration [11], since PSF was expected to remain invariable after hypochlorite exposure. As observed, the spectra bands attributed to amide bonding at 1581 and 1488 cm 1 due to C–N and ¼O stretching of the amide group [11] reduced their intensity. The absorbance was reduced to 52.1% and 41.5% for 1581 and 1488 cm 1 bands, respectively, by comparison with CPA5-RTO, indicating the partial removal of the PA layer. This can be explained as the modification produced by chlorine exposure in the hydrogen bonding and ring chlorination during hypochlorite exposure, which led to the polyamide layer failure due to alteration [28]. However, bands at 1238 and 1148 cm 1 that are attributed to C–O and symmetric SO2 stretching vibration [11] confirmed that PSF remained after the treatment due to the lack of active hydrogen that confers a high resistance to chlorine exposure [29]. CPA5-RTO and CPA5-PSF samples were evaluated in terms of their permeance, and the results are presented in Fig. 4. The membrane permeance in CPA5-PSF was increased as high as three orders of magnitude for the studied gases, with selectivities below typical Knudsen values (i.e., α(He/N2)¼2.45, α(CO2/N2)¼0.69, α(SF6/N2)¼0.46). This reflects the increasing contribution of convective gas flow when the PA layer was partially removed, hence, a reduction in gas separation (gas selectivity) compared with the CPA5-RTO sample. Therefore, Fig. 4 confirms that the separation characteristics of the RO commercial membrane for gas separation can mainly be attributed to the top aromatic PA layer with a dense matrix and highly-permeable regions. 3.3. Gas permeation under different pre-treatments Values for the nitrogen permeation and gas selectivity of the samples treated under different drying procedures (RTO, EH and FD) are shown in Figs. 5 and 6, respectively. Gas permeance was increased by factors of 1.8 to 9 for EH samples, and from 1.6 to 3.2 for FD compared with that in RTO samples (Fig. 5), showing the degree to which the different pretreatments produced layer alterations with effect in gas permeance. The nitrogen permeance value for the ESPA2-EH sample was the highest, and was significantly higher than that of the SWC5-RTO sample. It is remarkable that the permeance was increased by factors of 9 and 7.1 for ESPA2-EH and CPA5-EH, respectively, when compared with that of the RTO sample. However, SWC5 showed no significant changes after the three procedures. The small changes in gas permeation indicated that SWC5 consists of a rigid material that it less affected by pre-treatments, while ESPA2 and CPA5, with higher water permeabilities than the SWC5 membrane (Table 1), present larger ridge-and-valley structures that cause gas permeance to be highly dependent on the pretreatment methodology applied to these materials. In the FD procedure membrane morphology is fixed with a minimized distortion that may occur during normal drying, and thus low membrane shrinkage is expected [30]. Besides, alcohols are known to swell polymers to various degrees, influencing the free volume between polymer chains and, as a consequence, the separation performance [14], which occurred in the EH pretreatment. The results showed that highly-permeable regions did not permanently shrink or collapse when the membrane was dried with ethanol–hexane or freeze drying, as such changes would have resulted in lower permeances and gas transport continued to be dominated by a Knudsen diffusion-based flow after both pretreatments (Fig. 6). 113 J. Albo et al. / Journal of Membrane Science 449 (2014) 109–118 Table 3 Gas permeance of CPA5 after RTO, EH and progressive solvent exchange (PSE) pretreatment. N2 permeance [mol/(m2·s·Pa)] 10-6 Membrane-treatment type Permeance, P  108 [mol/(m2 s Pa)] 10-7 He 10 CPA5-RTO CPA5-EH CPA5-PSE -8 10-9 RTO EH SWC5 FD RTO EH ESPA2 FD RTO EH FD CPA5 Fig. 5. N2 permeance of three commercial PA membranes after different pretreatments. H2 CO2 N2 4.79 7 0.33 5.077 0.51 0.86 7 0.06 1.03 7 0.14 20.8 7 1.3 30.1 72.2 6.58 7 0.45 7.92 7 0.87 20.6 29.7 6.18 8 mechanisms through which treatment with alcohols increase gas permeance in polyamide membranes are expected to be affected by other phenomena, such as the effective removal of plugging compounds from the membrane surface by pure ethanol or the interactions related to polymer–solvent solubility producing the swelling of the cross-linked polymer, on dependence of the ridgeand-valley membrane structures. 6 He / N2 3.4. Correlation of nitrogen and pure water permeability 4 Fig. 7 features a plot of the nitrogen permeance under RTO, EH and FD pre-treatments versus water permeability values from provider information (Table 1). Gas and water permeabilities did not generally correlate for FD and RTO procedures, except when there was EH pre-treatment. As discussed above, this could be attributed to preservatives or other compounds that were present initially in the samples that were not dissolved in water, but were dissolved during ethanol immersion, and the avoidance of pore shrinkage during the ethanol-hexane pre-treatment, probably preserving the original ridge-and-valley membrane structure and a good gas and water permeability correlation. 2 0 H2 / N2 6 4 2 0 3.5. Thermal characterization CO2 / N2 2 1 0 RTO EH FD Fig. 6. Gas selectivity values for RTO, EH and FD procedures in CPA5. Knudsen diffusion-based selectivity values are denoted by the black horizontal lines. Based on these results it is also possible that CPA5 and ESPA2 membranes had some permeable regions plugged by compounds (i.e., residual synthesis monomers or additives) that were removed only by ethanol immersion, but not in water contact. An enhanced permeance after treatment with a strong solvent was also reported in the literature for interfacially polymerized PA composite membranes, explained by the partial dissolution of small molecules fragments from the surface [31]. Consequently, a CPA5 membrane sample was additionally immersed in 50, 75, 90, and 95% aqueous solutions, and then in pure ethanol for 15 min in a progressive solvent exchange pretreatment (CPA5-PSE) to evaluate if the membrane was susceptible to an ethanol attack during pure ethanol immersion in EH. The results are presented in Table 3, together with permeances obtained in the CPA5-RTO sample for comparison. As shown in Table 3, the EH and PSE treatments showed similar permeances, indicating that the integrity of the membrane structure was preserved after immersion in pure ethanol. Therefore, the For the potential application of TFC PA membranes at elevated temperatures a thermal characterization is required. The glass transition temperature, Tg, is significant in determining the mechanical properties of the TFC membrane since structure and separation properties could sharply change at this point [32]. For DTA, the Tg is taken at the inflexion point of the heat capacity change, while the melting point, Tm, refers to the minimum of the exotherm peaks. The TGA and DTA curves under a nitrogen atmosphere are shown in Fig. 8 for a CPA5-RTO PA composite membrane and Table 4 summarizes the four inflexion temperatures detected in the DTA curve. In addition, Tg for different TFC layers after isolation were obtained from the respective DTA curves (see Supplementary information) for comparison. A thin blade was used to carefully remove the PET substrate from the composite membrane, while the PSF middle layer was isolated after hypochlorite immersion. Because of the difficulties in isolating the PA layer from the TFC, a thin blade was used to remove the pure PA from an interfacial polymerization membrane formed insitu and the Tg was included for discussion. The first inflexion point was related to the PA Tg, followed by crystallization and a second minimum at the exotherm peak, Tm ¼211.1 1C. This was in good agreement with the literature where a Tg and Tm at 190 1C and 200–221 1C, respectively, were observed for dense PA-6 [33]. In addition, the Tg was similar to that obtained for the formed pure PA material in the present study, Tg ¼ 186.6 1C. After isolation, the PSF glass transition temperature, Tg ¼ 183.3 1C, was in good accordance with the value obtained via differential scanning calorimeter measurements (Tg ¼ 185 1C) [34]. Note that the influence of the PSF substrate on the Tg value of the TFC membrane was hidden in the figure due to similar thermal 114 J. Albo et al. / Journal of Membrane Science 449 (2014) 109–118 10-6 10-7 10-8 SWC5 10-9 0 1 ESPA2 CPA5 2 3 0.5 2 3 -20 4 T = 192.9 ºC 150 200 250 100 200 300 400 He N He/N 0 RT 50 100 150 200 Fig. 9. Gas permeance and He/N2 selectivity for CPA5-RTO after increases in the permeation temperature. Temperature cycle 0 500 Temperature [ºC] Fig. 8. DTA (in black) and TGA (in gray) curves for CPA5-RTO composite membrane (heating constant rate of 10 1C/min) up to 500 1C. Table 4 DTA inflexion temperatures for the TFC membrane. Membrane T [1C] Notes CPA5-RTO 192.9 211.1 250.9 441.4 PA glass transition followed by crystallization PA melting point PET melting point Material decomposition conductivity that the PA layer. A third minimum exotherm peak for the TFC was detected at 250.9 1C and was ascribed to PET melting point [35]. Tg for PET after isolation was found to be Tg ¼61.3 1C. The final peak at 441.4 1C showed the material decomposition temperature, since it also brought a sharp decrease in the TGA curve (69% membrane weight reduction). Because of the high glass transition, Tg, and decomposition temperature, these PA-based membranes show promise as highperformance polymeric material for gas separation processes at elevated temperatures. 3.6. Effect of temperature on gas separation Temperature-permeance dependency at increasing temperatures (room temperature (RT), 50, 100, 150, and 200 1C) for He and N2 transport though a CPA5-RTO membrane sample is presented in Fig. 9. The temperature was progressively increased and controlled by an oven sensor placed on the flat module surface. As Fig. 9 shows, with an increase in temperature He permeance was importantly enhanced up to 150 1C, and then decreased. On the other hand, N2 permeance did not vary during exposure to temperatures reaching 150 1C. Relaxation is very temperaturedependent for polymeric materials [36] and so, membrane T [1C] Permeance, P  108 [mol/(m2 s Pa)] He Gas selectivity α(He/N2) [–] N2 First RT 50 100 4.86 7 0.6 1.137 0.15 4.3 70.2 5.98 7 0.9 1.03 7 0.12 5.8 70.4 17.8 7 0.21 1.08 7 0.13 16.6 70.3 Second 50 150 6.197 0.63 1.127 0.19 5.5 70.3 31.7 7 4.1 1.117 0.18 28.4 70.5 Third 50 200 50 6.86 7 0.55 1.107 0.13 0.85 7 0.33 0.08 7 0.04 0.737 0.12 0.09 7 0.03 0.25 300 -30 0 Weight [mg] Heat Flow [µW/mg] 0.75 1 10 10-9 Table 5 Permeance and separation factors for CPA5-RTO with increasing temperatures in cycles. 1 -10 10-8 Temperature [ºC] Fig. 7. N2 permeance after different pre-treatments versus pure water permeability for the three commercial membranes. RTO (○), EH (Δ) and FD (□). 0 20 5 Water permeability [L/(m2·h·bar)] 10 10-7 10-10 4 30 Gas selectivity [-] Permeance [mol/(m2·s·Pa)] N2 Permeance [mol/(m2·s·Pa)] 10-6 6.2 70.2 10 70.8 8.1 72.2 polymer chain mobility was enhanced for He permeation but not to the extent to facilitate the permeation for larger molecule sizes, such as N2, through the dense matrix. Hence, this phenomenon led to an enhanced He/N2 gas selectivity at 150 1C. This is in similarity to the highest salt rejection values for RO membranes heat-treated in a 120–150 1C temperature range [37]. Further temperature increases to 200 1C, beyond the glass transition point (Tg ¼192.9 1C), led to crystallization. This material reorganization brought a sharp decrease in He and N2 permeance, suggesting that a partial closing of membrane pores had occurred. To further explore the temperature effect, He and N2 were evaluated at increasing–decreasing temperatures. The cycles were defined as follows: (first) RT (16 73 1C), then 50 1C and 100 1C; (second) 50–150 1C; and (third) 50–200 1C to finally measure gas permeance at 50 1C. The results are presented in Table 5. He and N2 permeances remained practically invariable at 50 1C, in comparison with that at 50 1C after exposures of 100 1C (first cycle) and 150 1C (second cycle), which meant that CPA5 membrane suffered no irreversible structure alteration at temperatures up to 150 1C. Conversely, after exposures to 200 1C (third cycle), membrane performance (permeance) was drastically reduced for both gases. In fact, reductions of 95 and 92% for He and N2 permeance, respectively, were obtained, which led to an increase in gas selectivity, α(He/N2), from 5.8 to 8.1 due to the reduced number of highly-permeable membrane structures. If those regions had completely shrunk at the applied temperature, there would have been an important increase in gas selectivity because gas would have been forced to permeate through the dense polymer matrix. These results confirmed the irreversibility of membrane chain alterations beyond the Tg of the material, and the partial closing of membrane highly-permeable regions. In Fig. 10, surface images obtained by SEM revealed the roughness J. Albo et al. / Journal of Membrane Science 449 (2014) 109–118 of the membrane skin layer. The samples were placed in an oven for 1 h at 50, 150 and 200 1C before SEM. CPA5 images after exposure to 150 1C (c–d) show a rougher appearance than the membranes treated at 50 1C (a–b). This could have been related to a membrane densification process taking place at this temperature, as reported in the literature [37]. In addition, the appearance may have been related to an enhancement in dehydration and the removal of the preparation residual groups from the membrane surface after exposure to 150 1C, leading to a smoother top layer probably without affecting the core. A further temperature increase beyond the Tg at 200 1C (e–f), clearly showed physical alterations, in accordance with the aforementioned results. The membrane assumed a flatter surface, influencing the width and length of the highly permeable regions of the PA selective layer. To better understand membrane morphology and performance at high temperature, a CPA5 sample treated with hypochlorite solution (CPA5-PSF) was tested for gas permeation at 200 1C. This determined whether it was the PA membrane or the PSF/ PET support that were mainly responsible for the gas selectivities after Tg. Fig. 11 shows the obtained results. The values for permeation through the PSF/PET substrate treated at RT and at 150 1C were included for comparison, although no alterations are expected below the Tg point. Similar, or slightly higher, permeance values compared with those at RT and 150 1C, were obtained for a CPA5-PSF sample 115 treated at 200 1C. Also, the permeance results (103–104 mol/ (m2 s Pa)) were remarkably higher than those obtained for the CPA5 TFC membrane (108–109 mol/(m2 s Pa)), reflecting the increasing contribution of convective gas-flow in the PSF/PET sample. Consequently, the separation performance of CPA5-RTO after exposure at 200 1C may yet be attributable to the PA layer. 3.7. Transport model at increasing temperatures In the region of activated diffusion, molecules with size differences can be effectively separated by molecular sieving. When the activated diffusion transport mechanism is dominant, the permeation has a tendency to increase along with the temperature. Fig. 12 shows an Arrhenius plot (Eq. (3)) of the permeances observed for He, CO2, O2, N2, and SF6 below Tg of the TFC. He and CO2 permeances increased with increasing temperature. Actually, the permeance was enhanced by a factor of 7.4 for He and 2.2 for CO2 at 150 1C when compared with the performance at room temperature. Alternatively, O2, N2 and SF6 permeances remained practically constant with temperature changes, which possibly can be explained by the Knudsen transport mechanism. If the molecules permeate uniquely via Knudsen diffusion mechanism through the TFC membrane, the dependence of permeance on temperature should be independent of the gases species and pore size, and thus the experimental gas permeances, Fig. 10. SEM images of a CPA5 membrane sample under increasing temperatures. Images on the left side were taken at 10k magnification and on the right at 25k. (a–b) 50 1C, (c–d) 150 1C, (e–f) 200 1C. The scale bar at the lower right represents 5 mm (10 k magnification) and 2 mm (25k magnification). 116 J. Albo et al. / Journal of Membrane Science 449 (2014) 109–118 CO2 N2 SF6 10-4 He 200 ºC 10-4 150 ºC P · (MRT)0.5 [-] Permeance [mol/(m2·s·Pa)] He RT O N SF 10-5 10-5 10-6 10-6 2 3 4 5 Fig. 11. Single-gas permeation of CPA5-PSF sample at room temperature after RT, 150 and 200 1C temperature treatment. 10-6 He CO O 2 SF N 4 5 6 Kinetic diameter [Å] Fig. 13. Comparison of permeance, P, multiplied by (MRT)0.5 observed at RT, 50, 100, and 150 1C for the studied gases in CPA5-RTO. Larger markers indicate higher temperatures. Table 6 Activation energies for permeation, Ep, in CPA5-RTO. CPA5-RTO 10-7 Ep [kJ/mol] 10 3 6 Kinetic diameter [A°] Permeance [mol/(m2·s·Pa)] CO Gas He CO2 O2 N2 SF6 15.96 6.15 1.21 0.07 0.06 -8 10-9 2 2.5 3 3.5 1000/T [K-1] Fig. 12. Arrhenius plot for studied gases through CPA5-RTO sample. The results from three replicates were within 15.7% of the values shown. P, multiplied by (MRT)0.5 should be identical across all temperature range [38]. Fig. 13 shows the values of P(MRT)0.5 for all gases at the experimental temperatures. The highest P(MRT)0.5 values obtained for He and CO2 confirmed an activated diffusion transport mechanism. Aromatic PA materials are considered to have a high degree of chain rigidity provided by the presence of aromatic ring and the strong interchain forces, which accounts for an elevated cohesive energy density and molecular packing [4]. Thus, the limited increase in PA chain mobility with temperature also produced a limited increase in membrane free volume, resulting in sufficient energy for small gases to overcome the attractive forces between chains. Table 6 summarizes the activation energies for permeation, Ep (Eq. (3)). As expected, the values for He and CO2 were the highest among the studied gases species. The higher permeance at increasing temperature reported for He (2.6 Å) and CO2 (3.3 Å), when compared with that for O2 (3.46 Å), N2 (3.64 Å) and SF6 (5.5 Å), denoted an effective molecular sieving separation in the PA-dense region. Simultaneously, CO2 has a relatively large affinity to PA materials due to reactive functional groups of amine moieties [39–41]. Therefore, chemical reactions between permeating species and pores may also cause the separation performance to vary at increasing temperatures, in comparison with the increases observed in He permeation. The results confirmed that the dry PA layer consisted of two different pore structures. First, a dense matrix where chain mobility with temperature enabled permeation by the activated diffusion of small gases, such as He, and second, highly-permeable regions where larger species, such as N2, could permeate exclusively via a Knudsen mechanism. Finally, Table 7 summarizes the separation performance of polymeric materials in the Robeson's upper bound relationship for He/N2, H2/N2 and CO2/N2 in a 25–35 1C temperature range [27], in comparison with the values achieved in this work at 150 1C for a CPA5-RTO TFC membrane and for a PA membrane formed in-situ by interfacial polymerization (PA in-situ). The obtained permeance results for He (2.6 Å) in this work are compared with those for H2 (2.9 Å) in the literature, since they are expected to permeate in a similar manner by the same transport mechanism. The values are compared in permeance (mol/(m2 s Pa)). The gas selectivity values achieved in the present study at 150 1C were comparable to those found in the literature at room temperature, although the membrane did not outperform the previously reported highest separation factors for H2/N2 [46] and CO2/N2 [51]. Nevertheless, the PA membranes had a considerably high He and CO2 permeance. Actually, the He permeance at 150 1C was PHe ¼36.2 71.6  10 8 mol/(m2 s Pa) and PHe ¼41.8 75.3  10 8 mol/(m2 s Pa) for CPA5-RTO TFC membrane and for the PA membrane formed in-situ by interfacial polymerization (PA in-situ), respectively. These high He (2.6 Å) permeances outperformed the target set by the U.S. Department of Energy for membranes to be used in high purity H2 (2.9 Å) production P H2 ¼34  10 8 mol/(m2 s Pa) (1000 GPU) [52]. Therefore, these results endorse the use of PA membranes as alternative materials for H2 purification, as well as in high-temperature separation processes. Consequently, as far as we could ascertain, these results represent the first report of a bi-modal structure for interfacially polymerized PA TFC membranes. These structures were clearly identified after measuring the permeation-temperature dependency of PA-based membranes. The results will be useful in the development of efficient gas-separating membranes for hightemperature processes. 4. Conclusions In this work, polyamide-based thin-film composite reverse osmosis membranes were dried under different pre-treatment 117 J. Albo et al. / Journal of Membrane Science 449 (2014) 109–118 Table 7 Performance of polymeric materials for He, H2 and CO2 with respect to N2. Membrane type Permeance, P  108 [mol/(m2 s Pa)] He Polyimide (6FDA-6FpDA:DABA (2:1)) Polyarylate (TMHFBPA I/T) Hyflons AD60X Poly(trimethylsilylpropyne) CPA5-RTO TFC, 150 1C PA in-situ, 150 1C 13.6 0.10 0.28 5.44 36.2 41.8 Gas selectivity, 65 64.8 50.3 1.03 28.4 99.2 H2 Polybenzoxazinone imide (PBOI-2-Cu þ ) Polyimide (1,1-6FDA-DIA) Polyimide (NTDA-BAPHFDS(H)) PIM-7 PIM-1 Poly(trimethylsilylpropyne cophenylpropyne) (95/5) 0.01 0.03 0.04 1.03 0.95 8.40 960 165 141 20.5 14.1 2.5 CO2 Poly[bis(2-(2-methoxyethoxy)ethoxy) phosphazene] PIM-7 PIM-1 CPA5-RTO TFC, 150 1C PA in-situ, 150 1C 0.08 1.32 1.67 1.89 6.02 62.5 26.2 25 17.2 14.3 procedures and were evaluated at increasing temperatures by gas separation tests. As a result, the gas permeation properties through the membranes changed depending on the pre-treatment procedures applied to the membranes and the influence on the swelling/ shrinkage of the selective layer. In particular, an increase by a factor of 1.8 to 9 and from 1.6 to 3.2, was observed when the membranes were pre-treated with ethanol–hexane and freeze-drying, respectively, to samples dried at room temperature and oven. Additionally, gas and water permeability values did not correlate for freeze-drying and room-temperature–oven treatments, but there was correlation for the ethanol–hexane pre-treatment procedure. This was mainly attributed to the avoidance of the pore shrinkage occurred during the ethanol– hexane procedure, as well as the removal of plugging compounds from within membrane structure during alcohol immersion. Moreover, the permeation temperature importantly influenced gas permeability and selectivity. A maximum He/N2 gas selectivity was obtained at 150 1C, α(He/N2)¼ 28.4, explained by the activated diffusion of He through the dense polyamide region of the membrane, with an activation energy for permeation of Ep ¼ 15.96 kJ/mol. Further temperature increase to 200 1C, beyond glass transition temperature, led to membrane structure reorganization, and caused a sharp decrease in He and N2 permeance. The results were confirmed by SEM images. Consequently, gas permeation tests revealed that the polyamide layer of the composite membranes was not perfectly homogenous, but rather consisted of a dense matrix and highlypermeable structures in which gases permeated via Knudsen diffusion and determine the membrane separation performance in dry samples. The results of the present study suggest that the separation performance of polyamide RO membranes can be optimized by the application of drying pre-treatments and permeation temperatures, giving useful information for potential applications of polyamidebased thin-film composite reverse osmosis membranes. Acknowledgments The authors gratefully acknowledge the financial support from the Japan Society for the Promotion of Science, under the Postdoctoral Fellowship for Foreign Researchers FY2012. α (Gas/N2) [–] Ref. [42] [43] [44] [45] This work This work [46] [47] [48] [48] [49] [50] [51] [49] [49] This work This work Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.memsci.2013.08.026. References [1] J.E. Cadotte, Evaluation of composite reverse osmosis membranes, in: D. R. LIoyd (Ed.), Materials Science of Synthetic Membranes, American Chemical Society, Washington, DC, 1985, pp. 273–294. [2] R.W. Baker, Membrane Technology and Applications, second ed., John Wiley & Sons, Ltd., Chichester, 2004. [3] L.F. Greenlee, D.F. Lawler, B.D. Freeman, B. Marrot, P. Moulin, Reverse osmosis desalination: water sources, technology, and today's challenges, Water Res. 43 (2009) 2317–2348. [4] J. Espeso, A.E. Lozano, J. de la Campa, J. de Abajo, Effect of substituents on the permeation properties of polyamide membranes, J. Membr. Sci. 280 (2006) 659–665. [5] J. de Abajo, J.G. de la Campa, A.E. Lozano, J. Espeso, C. Garcia, Designing aromatic polyamides and polyimides for gas separation membranes, Macromol. Symp. 199 (2003) 293–305. [6] M. Ulbricht, Advanced functional polymer membranes, Polymer 47 (2006) 2217–2262. [7] J.S. Louie, I. Pinnau, M. Reinhard, Gas and liquid permeation properties of modified interfacial composite reverse osmosis membranes, J. Membr. Sci. 325 (2008) 793–800. [8] J.S. Louie, I. Pinnau, M. Reinhard, Effects of surface coating process conditions on the water permeation and salt rejection properties of composite polyamide reverse osmosis membranes, J. Membr. Sci. 367 (2011) 249–255. [9] J. Petersen, K.V. Peinemann, Novel polyamide composite membranes for gas separation prepared by interfacial polycondensation, J. Appl. Polym. Sci. 63 (12) (1997) 1557–1563. [10] M.O. Ekiner, G. Vassilatos, Polyaramide hollow fibers for hydrogen/methane separation—spinning and properties, J. Membr. Sci. 186 (2001) 71–84. [11] S. Sridhar, B. Smitha, S. Mayor, B. Prathab, M. Aminabhavi, Gas permeation properties of polyamide membrane prepared by interfacial polymerization, J. Mater. Sci. 42 (2007) 9392–9401. [12] S. Li, Z. Wang, C. Zhang, M. Wan, F. Yuan, J. Wang, S. Wang, Interfacially polymerized thin film composite membranes containing ethylene oxide groups for CO2 separation, J. Membr. Sci. 436 (2013) 121–131. [13] J. de Abajo, J.G. de la Campa, A.E. Lozano, J.C. Alvarez, Thermally stable polymers: novel aromatic polyamides, Adv. Mater. 7 (1995) 148–151. [14] V.K. Agrawal, P.S. Singh, Alcohol-induced change of “reverse osmosis” polyamide membrane surface, J. Appl. Polym. Sci. 124 (2012) 290–299. [15] Y.Y. Zhao, Q.P. Yuan, Effect of membrane pretreatment on performance of solvent resistant nanofiltration membranes in methanol solutions, J. Membr. Sci. 280 (2006) 195–201. [16] A. Jezowska, G. Schipolowski, G. Wozny, Influence of simple pre-treatment method on properties of membrane material, Desalination 189 (2006) 43–52. [17] B. Van der Bruggen, J.C. Jansen, A. Figoli, J. Geens, D. Van Baelen, E. Drioli, C. Vandecasteele, Determination of parameters affecting transport in polymeric membranes: parallels between pervaporation and nanofiltration, J. Phys. Chem. B 108 (2004) 13273–13279. 118 J. Albo et al. / Journal of Membrane Science 449 (2014) 109–118 [18] B. Cai, Q.T. Nguyen, J.M. Valleton, C. Gao, In situ reparation of defects on the skin layer of reverse osmosis cellulose ester membranes for pervaporation purposes, J. Membr. Sci. 216 (2003) 165–175. [19] M.A. Kuehne, R.Q. Song, N.N. Li, R.J. Petersen, Flux enhancement in TFC RO membranes, Environ. Prog. 20 (1) (2001) 23–26. [20] M. Nilsson, G. Trägårdh, K. Östergren, Influence of temperature and cleaning on aromatic and semi-aromatic polyamide thin-film composite NF and RO membranes, Sep. Purif. Technol. 62 (2008) 717–726. [21] M. Sabard, F. Gouanvé, E. Espuche, R. Fulchiron, G. Seytre, L.-A. Fillot, L. Trouillet-Fonti, Influence of film processing conditions on the morphology of polyamide 6: consequences on water and ethanol sorption properties, J. Membr. Sci. 415–416 (2012) 670–680. [22] R.R. Sharma, R. Agrawal, S. Chellam, Temperature effects on sieving characteristics of thin-film composite nanofiltration membranes: pore size distributions and transport parameters, J. Membr. Sci. 223 (2003) 69–87. [23] C. Kong, M. Kanezashi, T. Yamomoto, T. Shintani, T. Tsuru, Controlled synthesis of high performance polyamide membrane with thin dense layer for water desalination, J. Membr. Sci. 362 (2010) 76–80. [24] S.-H. Huang, C.-J. Hsu, D.-J. Liaw, C.-C. Hu, Effect of chemical structures of amines on physicochemical properties of active layers and dehydration of isopropanol through interfacially polymerized thin-film composite membranes, J. Membr. Sci. 307 (2008) 73–81. [25] Y. Wang, L. Yiang, T. Matsuura, T.S. Chung, S.H. Goh, Investigation of the fundamental differences between polyamide-imide (PAI) and polyetherimide (PEI) membranes for isopropanol dehydration via pervaporation, J. Membr. Sci. 318 (2008) 217–226. [26] A. Singh, K. Ghosal, B.D. Freeman, A.E. Lozano, J.G. de la Campa, J. de Abajo, Gas separation properties of pendent phenyl substituted aromatic polyamides containing sulfone and hexafluoroisopropylidene groups, Polymer 40 (1999) 5715–5722. [27] L.M. Robeson, The upper bound revisited, J. Membr. Sci. 320 (2008) 390–400. [28] H. Matsuyama, N. Kurata, T. Shintani, Development of a chlorine-resistant polyamide reverse osmosis membrane, Desalination 207 (2007) 340–348. [29] G.M. Geise, H.S. Lee, D.J. Miller, B.D. Freeman, J.E. McGrath, D.R. Paul, Water purification by membranes: the role of polymer science, J. Polym. Sci. Part B— Polym. Phys. 48 (15) (2010) 1685–1718. [30] C.-Y. Kuo, S.-L. Su, H.-A. Tsai, Y.-S. Su, D.-M. Wang, J.-Y. Lai, Formation and evolution of a bicontinuous structure of PMMA membrane during wet immersion process, J. Membr. Sci. 315 (2008) 187–194. [31] M.F.J. Solomon, Y. Bhole, A.G. Livingston, High flux membranes for organic solvent nanofiltration (OSN)—interfacial polymerization with solvent activation, J. Membr. Sci. 423–424 (2012) 371–382. [32] J. Rieger, The glass transition temperature Tg of polymers-comparison of the values from differential thermal analysis (DTA, DSC) and dymanic mechanical measurments (torsium pendulum), Polym. Test. 20 (2001) 199–204. [33] P. Samyn, P. de Baets, G. Schoukens, I. Van Driessche, Friction, wear and transfer of pure and internally lubricated cast polyamides at various testing scales, Wear 262 (2007) 1433–1449. [34] S.H. Maruf, D.U. Ahn, A.R. Greenberg, Y. Ding, Glass transition behaviors of interfacially polymerized polyamide barrier layers on thin film composite membranes via nano-thermal analysis, Polymer 52 (2011) 2643–2649. [35] R.M.R. Wellen, E. Canedo, M.S. Rabello, Nonisothermal cold crystallization of poly(ethylene terephthalate), J. Mater. Res. 26 (9) (2011) 1107–1115. [36] M. Nilsson, G. Trägårdh, K. Östergren, The influence of sodium chloride on mass transfer in a polyamide nanofiltration membrane at elevated temperatures, J. Membr. Sci. 280 (2006) 928–936. [37] T. Shintani, H. Matsuyama, N. Kurata, Effect of heat treatment on performance of chlorine-resistant polyamide reverse osmosis membranes, Desalination 247 (2009) 370–377. [38] T. Yoshioka, E. Nakanishi, T. Tsuru, M. Asaeda, Experimental studies of gas permeation through microporous silica membranes, AIChE J. 47 (9) (2001) 2052–2063. [39] Z. Wang, M. Li, Y. Cai, J.X. Wang, S.C. Wang, Novel CO2 selectively permeating membranes containing PETEDA dendrimer, J. Membr. Sci. 290 (1–2) (2007) 250–258. [40] L.Y. Deng, T.J. Kim, M.B. Hägg, Facilitated transport of CO2 in novel PVAm/PVA blend membrane, J. Membr. Sci. 340 (1–2) (2009) 154–163. [41] X. Yu, Z. Wang, J. Zhao, F. Yuan, S. Li, J. Wang, S. Wang, An effective method to improve the performance of fixed carrier membrane via incorporation of CO2selective adsorptive silica nanoparticles, Chin. J. Chem. Eng. 19 (5) (2011) 821–832. [42] J.H. Kim, W.J. Koros, D.R. Paul, Physical aging of thin 6FDA-based polyimide membranes containing carboxyl acid groups. Part I. Transport properties, Polymer 47 (2006) 3094–3103. [43] M.T. Guzmán-Gutierrez, F.A. Ruiz-Treviño, M. Zolutukhin, S. Hernández-López, U. Scherf, Gas transport properties of high free volume polyarylates based on isophthalic/terephthalic acid chloride mixtures, J. Membr. Sci. 305 (2007) 347–352. [44] M. Macchione, J.C. Jansen, G. De Luca, E. Tocci, M. Longeri, E. Drioli, Experimental analysis and simulation of the gas transport in dense Hyflon AD60X membranes: influence of residual solvent, Polymer 48 (2007) 2619–2635. [45] L.G. Toy, K. Nagai, B.D. Freeman, I. Pinnau, Z. He, T. Masuda, M. Teraguchi, Y. P. Yampolskii, Pure-gas and vapor permeation and sorption properties of poly [1-phenyl-2-[p-(trimethylsilyl)phenyl]acetylene](PTMSDPA), Macromolecules 33 (2000) 2516–2524. [46] G. Polotskaya, M. Goikhman, I. Podeshvo, V. Kudryavtsev, Z. Pientka, L. Brozova, M. Bleha, Gas transport properties of polybenzoxazinoneimides and their prepolymers, Polymer 46 (2005) 3730–3736. [47] M.E. Rezac, B. Schoberl, Transport and thermal properties of poly(etherimide)/ acetylene-terminated monomer blends, J. Membr. Sci. 156 (1999) 211–222. [48] K. Tanaka, M.N. Islam, M. Kido, H. Kita, K.-I. Okamoto, Gas permeation and separation properties of sulfonated polyimide membranes, Polymer 47 (2006) 4370–4377. [49] P.M. Budd, K.J. Msayib, C.E. Tattershall, B.S. Ghanem, K.J. Reynolds, N. B. McKeown, D. Fritsch, Gas separation membranes from polymers with intrinsic microporosity, J. Membr. Sci. 251 (2005) 263–269. [50] K. Nagai, A. Higuchi, T. Nakagawa, Gas permeability and stability of poly(1trimethylsilyl-1-propyne-co-1-phenyl-1-propyne) membranes, J. Polym. Sci.: Part B: Polym. Phys. 33 (1995) 289–298. [51] C.J. Orme, M.K. Harrup, T.A. Luther, R.P. Lash, K.S. Houston, D.H. Weinkauf, F. F. Stewart, Characterization of gas transport in selected rubbery amorphous polyphosphazene membranes, J. Membr. Sci. 186 (2001) 249–256. [52] A.Y. Ku, P. Kulkarni, R. Shisler, W. Wei, Membrane performance requirements for carbon dioxide capture using hydrogen-selective membranes in integrated gasification combined cycle (IGCC) power plants, J. Membr. Sci. 367 (2011) 233–239.