Synthesis, characterization, and C[O.sub.2] permeation properties of acetylene-terminated polyimide membranes.
Global warming is currently one of the most serious environmental problems. Global warming is closely related with increased C[O.sub.2] emissions from large fixed sources, such as power plants as well as iron and cement foundries. Polymer membrane-based C[O.sub.2] separation is one of the most attractive separation technologies. Membrane-based gas separation for C[O.sub.2] emission reduction is advantageous due to its low energy requirement and low cost compared with other techniques (1). Gas permeation and separation in polymer membranes use the pressure difference between the feed and downstream pressure as its driving force. Therefore, gas separation works well under highly pressurized gas streams for the acceleration of the membrane flux. However, many existing glassy polymer membranes are plasticized and swelled by high-pressure C[O.sub.2] (2-4). This phenomenon decreases the interaction among polymer segments and increases the mobility of polymer chains induced by sorbed C[O.sub.2] in the membranes. The gas separation performance of polymer membranes significantly decreases after membrane plasticization.
To date, numerous studies on the suppression of C[O.sub.2]-induced membrane plasticization in aromatic polyimide membranes, which are used in aerospace and electronics applications, have been investigated. The crosslinking technique between polymer chains is one of the efficient approaches for suppressing plasticization induced by C[O.sub.2] (5). Many types of reactions have been reported including the monoesterification and transesterification reactions of carboxylic acids (6-8), decarboxylation-induced crosslink reaction (9), imide ring-opening reaction (10-12), graft crosslink with etherification reaction (4), (13), ultra violet crosslink (14-16), acetylenes of a Diels-Alder-type cycloaddition reaction (17-19), and crosslink using dendrimer (20), (21). Although previous crosslinking techniques have been available in plasticization, gas permeability largely decreases because of the densification of crosslinked membranes attributed to a higher crosslink density. Plasticization may occur at polymer chain ends because the mobility at the polymer chain ends may be higher than that at the inner chains. Therefore, crosslinking reactions at polymer chain ends can be an effective approach for the suppression of membrane plasticization.
Cyclotrimerization involving acetylene groups is known as the cyclization-dimerization of end-capped acetylene groups and the formation of arene via cyclization-trimerization as shown in Fig. 1 (22). Crosslinking reactions at the ethynyl group in an imide oligomer end-capped with a phenyl ethynyl group have been widely investigated and reported (23-25). Because this crosslinking reaction requires thermal treatment at over 35[degrees]C, more moderate temperature will be needed. For gas separation membranes, the use of the crosslinking approach for polymer blends containing acetylene-terminated oligomer (17), (26), diacetylene-containing polyimide (27), and polyimide with internal acetylenes (19) has been reported. These crosslink reactions were prepared by imide oligomer or prepolymer, whereas acetylene-terminated polymer has not been investigated.
In this study, the novel crosslinking reaction of the cyclotrimerization of acetylene groups using tantalum (V) chloride (Ta[Cl.sub.5]) catalyst under thermal treatment to prepare an acetylene end-capped polyimide for the suppression of membrane plasticization was investigated. The physical and thermal properties, as well as the time-dependence on the C[O.sub.2] permeation property of the novel polyimide membranes were also discussed in terms of membrane plasticization.
The monomers used in this synthesis were 4,4-(hexa-fluoroisopropylidene) diphthalic anhydride (6FDA; Aldrich, purity > 98%), 4-(2-phenylethynyl) phthalic anhydride (PEPA; Wako Pure Chemical Industries, purity > 97%), and 3,4-diaminodiphenyl ether (DADE; Tokyo Chemical Industry Co., purity > 97%). All monomers were dried in a vacuum before use. N,N-Dimethylaceta-mide (DMAc; Junsei Chemical Co., purity > 99.0 %) and N,N-dimethylformamide (DMF; Junsei Chemical Co., purity > 99.5%) were dehydrated with well-dried molecular sieves before use. The transition catalyst Ta[Cl.sub.5] (Aldrich, purity > 99.9%) was used as received without further purification.
The 6FDA-DADE base polyimide was synthesized by polycondensation via the chemical imidization reaction according to our previous study (3): Fourier-transform infrared (FT-IR; KBr, [cm.sub.(-1)]), 1785 (C=0 asymmetric stretching), 1728 (C=O symmetric stretching), 1374 (C--N stretching), 721 (C=0 bending); and (1) IH-NMR (500 MHz, CD[Cl.sub.3]-d, [delta]), 7.93 (2 H), 8.03-8.01 (2 H), 7.88-7.84 (2 H), 7.40 (2 H), 7.18-7.14 (2 H), 7.51-7.49 (1 H), 7.22-7.19 (3 H).
Figure 2 shows the synthesis of the polyimide 6FDA-DADE-PEPA. Although an acetylene-terminated oligomer is usually synthesized based on the Carothers equation, the acetylene-terminated polymer was obtained via conventional polycondensation as follows. DADE diamine (2.261 g, 1.129 X [10.sup.-2] mol) was added to a DMAc solution and stirred for 30 min. 6FDA (5.016 g, 1.129 X [10.sup.-2] mol) equimolar amount of DADE) was added to the DMAc solution at a concentration of 20 wt% with stirring. After 12 h, DADE (0.251 g, 1.254 x [10.sup.-3] mol) was dissolved in the solution with stirring for 3 h to form an amine-terminated polymer. PEPA (0.615 g, 2.509 X [10.sup.-3] mol), the acetylene-containing anhydride monomer, was then added to the DMAc solution at a concentration of 15 wt% with stirring. The imidization reaction was performed using the chemical imidization catalysts pyridine and acetic anhydride. Pyridine (7 mol) and acetic anhydride (5 mol) of 6FDA were added. This reaction was continued for 5 h with stirring, and the viscous polymer solution was precipitated in a large amount of methanol solution. The polyimide was purified several times until all impurities were removed. Purification was performed using a solution-reprecipitation process with DMF/metha-nol: FT-IR (KBr, [cm.sup.-1]), 1785 (C=0 asymmetric stretching), 1727 (C=0 symmetric stretching), 1374 (C--N stretching), 721 (C=0 bending), 2212 (C[equivalent to]C stretching); and (1) H-NMR (500 MHz, CD[Cl.sub.3]-d, [delta]), 7.91 (2 H), 8.048.02 (2 H), 7.88-7.84 (2 H), 7.43-7.38 (2 H), 7.15-7.13 (2 H), 7.52-7.49 (1 H), 7.24-7.20 (3 H).
Isotropic dense noncrosslinked polyimide membranes were prepared by casting a filtered 2.0 wt% solution of the polymer in dichloromethane (DCM) onto polytetrafluoroethylene (PTFE) laboratory dishes. The dishes were covered with a glass dish to reduce the rate of solvent evaporation and prevent contamination. The membranes took approximately 24-48 h to evaporate, after which they were dried in a vacuum at room temperature to remove minute amounts of residual solvent. The protocol of thermal treatment was vacuum drying at room temperature for 24 h, and then at 60[degrees]C for 1 h, 100[degrees]C for 1 h, 150[degrees]C for 1 h, 200[degrees]C for 2 h, and 250[degrees]C for 24 h. To consider the influence of membrane thickness, the well-dried membrane (50-60 [micro]m thick; measurement uncertainty is [+ or -]2 [micro]m) was used.
Isotropic dense crosslinked (cyclotrimerization reacted) polyimide membranes were prepared by the casting method. The solution was filtered with 2.0 wt% polymer solution containing 5 wt% and 35 wt% Ta[Cl.sub.5] in DCM, and then casted onto PTFE laboratory dishes. For example, 0.011 g of Ta[Cl.sub.5] was dissolved in 0.200 g of 6FDA--DADE--PEPA, and 0.109 g (35 wt%) of Ta[Cl.sub.5] was dissolved in the polymer. The high-content Ta[Cl.sub.5] in the membrane served not only as a catalyst for the cyclotri-merization but also as a filler to improve the physical property of the polymer. The membrane containing more than 35 wt% Ta[Cl.sub.5] was inhomogeneous because of its excess Ta[Cl.sub.5] content. 6FDA--DADE--PEPA without Ta[Cl.sub.5] was also prepared for comparison. The membranes took approximately 24-48 h to evaporate, after which they were removed from the plates. These preparations were performed on a dry box with diphosphorous pentoxide under nitrogen to prevent the deactivation of Ta[Cl.sub.5]. The membranes were dried in a vacuum at room temperature until the removal of residual solvent was analyzed by (1) H-NMR. The membrane was then dried under thermal treatment to enable the cyclotrimerization of the acetylene groups at the polymer chain ends. The thermal treatment protocol of the noncrosslinked 6FDA-DADE-PEPA membrane (thermal) was same of 6FDA-DADE (thermal) as aforementioned.
All characterization data were determined in the membrane state for at least three samples to confirm the reproducibility of the experimental results. (1) H-NMR was performed using a JNM-ECA500 instrument (JEOL, Tokyo, Japan). FT-IR analysis using the KBr method was performed with an FT/IR-4100 instrument (JASCO Co., Tokyo, Japan) at 23 [+ or -] 1[degrees]C.
The average molecular weight number average molecular weight ([M.sub.w]), and molecular weight distribution ratio ([M.sub.w]/[M.sub.n]) of the polymers were determined using a gel permeation chromatography (GPC) system (HLC-8220, Tosoh Co., Tokyo, Japan) with TSK-gel columns (SuperHM-H) and a detector (RI-8220). The calibration was performed using polystyrene standards at 40[degrees]C in tetrahydrofuran (THF) solvent at a flow rate of 0.300 ml/min. The inherent viscosity of the polymers was measured using a Canon Fensuke viscometer (Yoshida Seisakusho Co.). The viscosity was calculated using 0.2, 0.3, 0.4, and 0.5 g/dl DMF solution at 30[degrees]C.
Membrane density (p) was measured by the flotation method at 23[degrees]C [+ or -] 1[degrees]C. The theoretical membrane density ([p.sub.theory]), which considers the catalyst content, was calculated using the following equation:
([p.sub.theory]= (([m.sub.polymer])/([m.sub.polymer]+[m.sub.catalyst])) x [p.sub.polymer] + (([m.sub.catalyst])/([m.sub.polymer]+[m.sub.catalyst])) x [p.sub.catalyst] (1)
[m.sub.polymer] and [m.sub.catalyst] are the weights of the polymer and catalyst, respectively. [P.sub.polyme] and [p.sub.catalyst] catalyst are the densities of the polymer and catalyst, respectively. The density of Ta[Cl.sub.5] used in this study was 3.68 g/[cm.sup.3] at 25[degrees]C. The void fraction of the membranes was calculated using experimental and theoretical membrane density values:
Void fraction = ([P.sub.theory])/p (2)
The gel content in chloroform at 23[degrees]C of the cross-linked membranes was determined by the following equation:
Gel content = (([m.sub.insoluble])/([m.sub.dry])) x 100 (3)
where [m.sub.dry] and [m.sub.insoluble] are the weights of the dry membrane before dissolution and the insoluble part after dissolution, respectively.
Wide-angle X-ray diffraction (WAXD) measurement was performed on a Rint 1200 X-ray diffractometer (Rigaku, Co., Tokyo, Japan) at a scanning speed of 2[degrees]/min using a Cu K[alpha] radiation source at 40 kV and 20 mA at a dispersion angle from 3[degrees] to 50[degrees].
Scanning electron microscopy was performed using a high-resolution field scanning emission electron microscope (S5200, JEOL, Tokyo Japan). Qualitative elemental analysis was performed using an energy dispersive X-ray analyzer (EDAX Japan Co., Genesis, Tokyo, Japan) with Mg K[alpha] radiation source (hv = 1253.6 eV). The accelerating voltage was 15 kV, the beam spot diameter was 100 pm, and the beam incidence angle was 10[degrees].
Thermogravimetric analysis (TGA) was performed using a Pyris 1 TGA thermogravimetric analyzer (PerkinElmer, Shelton). A polymer sample (ca. 1.0 mg) was heated from 50[degrees]C to 800[degrees]C in a platinum pan at a heating rate of 10[degrees]C/min under nitrogen atmosphere at a flow rate of 60 ml/min. The glass transition temperature ([T.sub.g]) was measured using a Diamond differential scanning calorimetry (DSC) system (PerkinElmer, Shelton). DSC measurements were performed between 100 and 300[degrees]C at a heating rate of 10[degree]C/min under a nitrogen atmosphere. [T.sub.g] was determined as the middle point of endothermic transition in the second scan.
The C[O.sub.2] permeability coefficient in polyimide membranes was determined using a constant-volume/variable-pressure method according to literature (3). The operating temperature was 35 [+ or -] 0.1[degress]C. The upstream maximum pressure was 40 atm, and the downstream pressure was maintained under a vacuum. This experimental condition is based on the actual separation environment. Each measurement time was 720 min at a given pressure. Plasticization was observed after 720 min; thus, a membrane was changed after each measurement as previously described (3). All permeation data were determined for at least three samples to ensure the reproducibility of the experimental results.
RESULTS AND DISCUSSION
Structure Analysis and Characterization
[.sup.1]H-NMR and FT-IR analyses confirmed the chemical structures of the synthesized polyimides, as shown in Fig. 2. [M.sub.n] and ([M.sub.w])/([M.sub.n]) determined by GPC analysis were 50,000 g/mol and 1.8 for 6FDA--DADE, as well as 75,000 g/mol and 1.9 for 6FDA-DADE-PEPA, respectively. The inherent viscosities of 6FDA--DADE and 6FDA-DADE-PEPA polyimides were 0.60 and 0.74 dl/g, respectively.
Figure 3 shows the photographs of the polyimide membrane. All polyimide membranes exhibited good membrane formation property. The noncrosslinked 6FDA-DADE membranes were yellow, and the crosslinked 6FDA-DADE-PEPA membranes became more yellow with increased Ta[Cl.sub.5] catalyst content. The solubility of these polyimide membranes are summarized in Table 1. The solubility of noncrosslinked 6FDA-DADE, 6FDA-DADE-PEPA (as-cast) without Ta[Cl.sub.5], and 6FDA-DADE-PEPA (thermal) without Ta[Cl.sub.5] were almost the same. They were able to be dissolved by the DCM, DMAc, DMF, NMP, and THF, C[H.sub.3]Cl, but not be methanol and water. On the other hand, the crosslinked membranes showed gel content in the good solvents of both noncrosslinked 6FDA-DADE and 6FDA-DADE-PEPA membranes. The gel contents of crosslinked 6FDA-DADE-PEPA (5 wt% and 35 wt%) were 9 [+ or -] 2% and 13 [+ or -] 1%, respectively. These results revealed that the cyclotrimerization of the acetylene groups at the polymer chain ends occurred and affected the solubility of the membranes. However, these gel contents of the crosslinked membranes were lower than those of conventional crosslinked membranes, which are almost insoluble in any solvent (14). This result can be attributed to the high molecular weight of the polymer and the reaction point at the polymer chain ends in this study. The crosslinked reaction in high-molecular-weight polymers at the polymer chain ends was found as able to prevent increased crosslink density, that is, membrane densification.
TABLE 1. Solubilities of the polyimide membranes. Polymer DCM DMAc DMF NMP THF C[H.sub.3]Cl MeOH Water 6 FDA-DADE ++ ++ + ++ ++ ++ - - 6FDA-DADE-PEPA ++ + + + + + - - (thermal) 6PDA-DADE-PEPA ++ + + + ++ ++ ++ - - (as-cast) Crosslinked +- +- +- +- +- +- - - 6FDA-DADE-PEPA (35 wt%) Crosslinked +- +- +- +- +- +- - - 6FDA-DADE-PEPA (5 wt%) ++, dissolved after 30 mm completely (good solubility); +, dissolved after 24 h completely; -, insoluble after 24 h; and +m partially insoluble after 72 h
The physical properties of the polyimide membranes are summarized in Table 2. The membrane density of noncrossfinked 6FDA-DADE (as-cast) was 1.434 [+ or -] 0.001 g/[cm.sup.3], which was consistent with that in literature (1.438 g/[cm.sub.3]) (28). The density of noncrosslinked 6FDA-DADE-PEPA (as-cast) was 1.433 [+ or -] 0.001 g/[cm.sup.3], which was almost the same as that of 6FDA-DADE. Therefore, the membrane density of noncrosslinked 6FDA-DADE-PEPA (as-cast) was not affected by the polymer chain end-capped acetylene groups. The membrane density of noncrosslinked 6FDA-DADE-PEPA (thermal) was 1.443 [+ or -]. 0.001 g/[cm.sup.3], which was higher than those of 6FDA-DADE and 6FDA-DADE-PEPA (as-cast). This result is considered as a typical annealing effect of polymer membranes (3), (29), (30). The membrane density of crosslinked 6FDA-DADE-PEPA increased with increased Ta[Cl.sub.5] content. For example, the density of 6FDA-DADE-PEPA (5 wt% and 35 wt%) were 1.504 [+ or -] 0.001 and more than 1.560 g/[cm.sup.3], respectively. These results suggested that the crosslink reaction at the polymer chain ends and the presence of the catalyst (density; 3.68 g/[cm.sup.3]) strongly affected the membrane density. However, the theoretical density of the crosslinked 6FDA-DADE-PEPA membrane calculated from Eq. 1 was 1.560 g/[cm.sup.3] for 5 wt% and 2.232 g/[cm.sup.3] for 35 wt%. Consequently, some voids formed between the polymer and catalyst of the cross-linked mebranes because the experimental membrane density values of the crosslinked membranes were smaller than the theoretical values. The void fractions of the membranes calculated from Eq. 2 were 3.6% for 5 wt% and 30.1% for 35 wt% Ta[Cl.sub.5] content. This result suggested that the void fraction of the membranes increased with increased catalyst content.
TABLE 2. Physical properties of polyimide membranes. Polymer [rho] [[rho].sub.theory] Void (g/[cm.sub.3]) (g/[cm.sub.3]) fraction (%) 6FDA-DADC 1.434 [+ or -] 1.434 0.0 0.001 6FDA-DADE-PEPA 1.433 [+ or -] 1.433 0.0 (as-cast) 0.001 6FDA-DADE-PEPA 1.443 [+ or -] 1.443 0.0 (thermal) 0.001 Crosslinked 1.504 [+ or -] 1.560 3.6 0.001 6FEDA-DA DE-PEP A (5 wt%) Crosslinked 1.560 < 2.232 30.1 6FDA-DADE-PEPA (35 wt%) Polymer (d-Spacing (a) Gel ([Angstrom]) content (wt%) 6FDA-DADC 5.9 [+ or -] 0 0.1 6FDA-DADE-PEPA 5.9 [+ or -] 0 (as-cast) 0.1 6FDA-DADE-PEPA 5.7 [+ or -] 0 (thermal) 0.1 Crosslinked 5.3 [+ or -] 9 [+ or 0.1 -] 2 6FEDA- DA DE-PEP 3.6 [+ or -] A (5 wt%) 0.1 Crosslinked 5.3 [+ or -] 13 [+ or 0.1 -] 1 6FDA-DADE-PEPA 3.6 [+ or -] (35 wt%) 0.1 (a) Values based on the polyimide.
Figure 4 shows the WAXD patterns of the polyimide membranes. Given that all patterns exhibited broad halo peaks, all polyimide membranes were completely in the amorphous state. Based on Bragg's condition, the d-spacing determined from the maximum intensity for the noncrosslinked 6FDA-DADE-based polyimide membrane was 5.9 [+ or -] 0.1 [Angstrom] whereas those for noncrosslinked 6FDA-DADE-PEPA (as-cast) and (thermal) were 5.9 0.1 and 5.7 [+ or -] 0.1 [Angstrom], respectively (Table 2). Based on the membrane density, the decreased d-spacing of noncros-slinked 6FDA-DADE-PEPA (thermal) can be attributed to the enhanced interaction among polymer segments. Therefore, these results also indicated that the thermal treatment of the membrane makes more dense structure as observed in the membrane density. On the other hand, new peaks were observed at 20 = 27[degress] in the crosslinked membranes. The halo in the small angle (around 20 = 16[degrees]) appeared at a distance between polymer segments, whereas that in the large angle (around 20 = 27[degrees] was attributed to the intermolecular distance. The peak intensity at 20 = 27[degrees] was increased with increasing Ta[Cl.sub.5] content. This result suggested that the crosslink reaction at the polymer chain ends produced the novel chemical structures as is evidenced from Fig. 4. The d-spacing values estimated from each peak for the crosslinked membranes are 5.3 [+ or -] 0.1 [Angstrom] and 3.7 [+ or -] 0.1 [Angstrom], regardless of the catalyst content. The distance between polymer segments (i.e., 5.3 [+ or -] 0.1 [Angstrom]) of the crosslinked membranes was lower than those of the noncrosslinked 6FDA-DADE and 6FDA-DADE-PEPA membranes (i.e., 5.9 [+ or -] 0.1 [Angstrom]) due to the crosslink reaction at the polymer chain ends.
Figure 5 shows the spectra of elemental analysis of noncrosslinked 6FDA-DADE-PEPA (as-cast) and cross-linked 6FDA-DADE-PEPA (5 wt%) membranes. For the crosslinked 6FDA-DADE-PEPA membrane, the analyses were performed in three directions, namely, top, middle, and bottom sides. Evaporated Pt was detected in all membranes. The peak attributed to the presence of Ta[Cl.sub.5] catalyst in the crosslinked 6FDA-DADE-PEPA membrane was observed in each region of the top, middle, and bottom sides. The elemental peak intensity of each region was almost the same in the crosslinked membrane. Hence, the crosslinked membrane was homogeneous with well-dispersed Ta[Cl.sub.5] catatyst. The same result was obtained for crosslinked 6FDA-DADE-PEPA (35 wt%).
Figure 6 shows the surface and the cross-sectional of SEM image of the crosslinked 6FDA-DADE-PEPA membrane (35 wt%). The homogeneous and smoothly surface was observed in the 6FDA-DADE and noncrosslinked 6FDA-DADE-PEPA membranes, whereas the crosslinked 6FDA-DADE-PEPA membrane showed interfacial voids between the polymer and the catalyst as presented in Fig. 6. In addition, the catalyst were well-dispersed without aggregation in the bottom side of the membrane. Based on this result, the crosslinked membranes formed voids between the polymer and the catalyst and were homogeneous with well-dispersed catatyst. The similar result was observed for crosslinked 6FDA-DADE-PEPA (5 wt%).
Figure 7 shows the TGA curves of the polyimide membranes. The thermal properties are summarized in Table 3. All polyimide membranes exhibited a single-step thermal decomposition behavior. The temperatures of thermal weight loss percentage, [T.sub.5] and [T.sub.10], which reflect thermal stability, had almost identical values of 247 [+ or -] 1 and 567 [+ or -] VC, respectively. However, the residual weight at 800[degrees]C ([W.sub.R](800)) of noncrosslinked 6FDA-DADE was 53 [+ or -] 1 wt%, whereas that of noncrosslinked 6FDA-DADE-PEPA (as-cast and thermal) was almost identical (56 [+ or -] 1 wt%), which was slightly higher than that of 6FDA-DADE. This result indicated that no cyclo-trimerization occurred at the acetylene groups of the polymer chain ends by thermal treatment without Ta[Cl.sub.5] catalyst. In comparison, WR(800) of the crosslinked 6FDA-DADE-PEPA (5 wt% and 35 wt%) were 57 [+ or -] 1 wt% and 61 [+ or -] 1 wt%, respectively. These values were higher than those of the 6FDA-DADE and 6FDA-DADE-PEPA (as-cast) membranes. Therefore, the cyclotrimerization in 6FDA-DADE-PEPA at the polymer chain end groups enhanced their thermal stability because of the increased interaction among polymer segments. Some aromatic ring structures also formed at the polymer chain ends after cyclotrimerization because single-step decomposition for the crosslinked 6FDA-DADE-PEPA membranes was observed. [T.sub.g] of noncrosslinked 6FDA-DADE was 260 [+ or -] 1[degrees]C, which well agreed with that in literature (260[degrees]C) (28). The [T.sub.g] values of noncrosslinked 6FDA--DADE--PEPA (as-cast and thermal) were 259 [+ or -] 1 and 261 [+ or -] 1[degrees]C, respectively. No obvious difference was observed between the Tg values of the noncrosslinked 6FDA--DADE and 6FDA-DADE-PEPA (as-cast) membranes. [T.sub.g] of the crosslinked 6FDA-DADE-PEPA (5 wt%) and 6FDA-DADE-PEPA (35 wt%) membranes were 262 [+ or -] 1 and 263 [+ or -] 1[degrees]C, respectively. This result indicated that [T.sub.g] in the crosslinked membranes was higher than those in the noncrosslinked 6FDA-DADE and 6FDA-DADE-PEPA membranes, that is, cyclotrimerization reaction occurred. As the cyclotrimerization at the polymer chain end groups progressed, the mobility of the polymer segments was gradually suppressed due to the enhanced interaction among polymer segments. However, no obvious increase in the thermal decomposition temperature and [T.sub.g] was observed. Consequently, this novel crosslink approach could prevent the membrane densification because a high-molecular-weight polymer and the crosslink reaction at the polymer chain ends were used in this study.
TABLE 3. Thermal properties of polyimide membranes. Polymer [T.sub.5] [T.sub.10] [W.sub.R] ([degrees]C) ([degrees]C) (800[degrees]C) (wt%) 6FDA-DADE 547 [+ or -] 3 568 [+ or -] 2 53 [+ or -] 1 6FDA-DADE-PEPA 544 [+ or -] 3 568 [+ or -] 2 55 [+ or -] 1 (as-cast) 6FDA-DADE-PEPA 546 [+ or -] 2 565 [+ or -] 2 56 [+ or -] 1 (thermal) Crosslinked 545 [+ or -] 3 567 [+ or -] 2 57 [+ or -] 1 6FDA-DADE-PEPA (5 wt%) Crosslinked 546 [+ or -] 2 568 [+ or -] 3 62 [+ or -] 1 6FDA-DADE-PEPA (35 wt%) Polymer [T.sub.g] ([degress]C) 6FDA-DADE 260 [+ or -] 1 6FDA-DADE-PEPA 259 [+ or -] 1 (as-cast) 6FDA-DADE-PEPA 261 [+ or -] 1 (thermal) Crosslinked 262 [+ or -] 1 6FDA-DADE-PEPA (5 wt%) Crosslinked 263 [+ or -] 1 6FDA-DADE-PEPA (35 wt%)
Figure 8 shows the time dependence on the C[O.sub.2] permeability ratio of each polyimide membrane at 35[degrees]C and 40 atm, considering the severe condition in the actual separation environment. In general, gas permeation phenomenon in a dense polymer membrane is obeyed the solution-diffusion mechanism (1). The permeability ratio of the noncrosslinked 6FDA-DADE-based poly imide membrane increased by 200% compared with the initial permeability at 40 atm, whereas those of the noncrosslinked 6FDA-DADE-PEPA membranes increased to 170% for as-cast and 120% for thermal, respectively. This increase in the C[O.sub.2] permeability is characteristic for the membrane-plasticization phenomenon (3). In comparison, the crosslinked 6FDA-DADE-PEPA membranes increased to 110% for 5 wt% and were almost constant for 35 wt%. Therefore, the order of C[O.sub,2]-induced plasticization in the polyimide membranes was noncrosslinked 6FDA--DADE > noncrosslinked 6FDA-DADE-PEPA (as-cast) > noncrosslinked 6FDA-DADE-PEPA (thermal) > crosslinked 6FDA-DADE-PEPA (5 wt%) > crosslinked 6FDA-DADE-PEPA (35 wt%). The thermal treatment of glassy polymer membranes improved the suppression of C[O.sub.2]- induced plasticization due to the enhancement of the polymer chain packing (3, 29). The crosslinked membranes suppressed the plasticization compared with noncrosslinked 6FDA-DADE-PEPA (thermal). The reason for the plasticization in the crosslinked membrane (5 wt%) is due to the low crosslink density based on the reaction at the polymer chain ends and low content of the catalyst. Therefore, the membrane plasticization was improved with increased Ta[Cl.sub.5]] catalyst content. Consequently, the cyclotrimerization of the acetylene groups in the polymer chain ends with Ta[Cl.sub.5] under thermal treatment is an effective approach for C[O.sub.2]-induced plasticization. It should be noted that the large amount of the catalyst seems to hinder the mobility of the polymer chains and work as an anchorage of the chains hindering their possible motion and plasticization as well as decrease gas permeability and diffusivity. However, the interfacial voids between the polymer and the catalyst as presented in Fig. 6 could help suppression for the reduction in the gas permeability and diffusivity. Based on these results, the crosslink at polymer chain ends was found to be more effective approach for gas separation to prevent the C[O.sub.2]-induced plasticization behavior without the membrane densification.
A novel fluorine-containing telechelic polyimide end-capped with acetylene group, 6FDA-DADE-PEPA for gas separation was synthesized via polycondensation. The novel crosslink reaction in this polyimide membrane by cyclotrimerization using Ta[Cl.sub.5] catalyst under thermal treatment was investigated. The crosslinked 6FDA-DADE-PEPA membranes were denser than the noncros-slinked 6FDA-DADE membrane, as evidenced by the increased membrane density and decreased mean distance between polymer segments (i.e., d-spacing) after cyclotrimerization. In addition, the crosslinked membranes showed gel content in THF, which is a good solvent for noncrosslinked ones. Therefore, this result can be attributed to the crosslinking reaction at the polymer chain ends. However, the theoretical density of the crosslinked membranes was smaller than the experimental one. This result indicated the voids between polyimide and the catalyst were formed. The C[O.sub.2] permeability ratio of noncrosslinked 6FDA-DADE and 6FDA-DADE-PEPA (as-cast) membranes increased with increased experimental time at 40 atm, suggesting that the typical C[O.sub.2]-induced plasticization was observed. In comparison, the crosslinked 6FDA-DADE-PEPA membranes exhibited resistance to membrane plasticization compared with noncrossl inked thermal treated 6FDA-DADE-PEPA. The cyclotrimeriza-tion of the acetylene groups at the polymer chain ends and the use of a transition metal catalyst were found to be more effective than conventional thermal treatments for the suppression of membrane plasticization without the membrane densification.
Correspondence to: K. Nagai; e-mail: email@example.com
Contract grant sponsor: Grant-in-aid 2005-2006 from the Research Institute of Innovative Technology for the Earth (RITE), Kyoto, Japan.
Published online in Wiley Online Library (wileyonlinelibrary.com).
[c] 2012 Society of Plastics Engineers
(1.) R.W. Baker, Membrane Technology and Applications, McGraw-Hill, New York (2000).
(2.) D.R. Paul and Y.P. Yampol'skii, Polymeric Gas Separation Membranes, CRC Press, Boca Raton, FL, USA (1994).
(3.) S. Kanehashi, T. Nakagawa, K. Nagai, X. Duthie, S. Kentish, and G. Stevens, J. Membr. Sc., 298, 147 (2007).
(4.) K. Nagai, N. Booker, A. Mau, J. Hodgkin, S. Kentish, G. Stevens, and A. Geertsema, Polym. Mater. Sci. Eng., 85, 91 (2001).
(5.) S. Kanehashi, S. Sato, and K. Nagai. Membrane Gas Separation, Wiley, Chichester (2010).
(6.) C. Staudt-Bickel and W.J. Koros, J. Membr. Sc., 155, 145 (1999).
(7.) J.D. Wind, C. Staudt-Bickel, D.R. Paul, and W.J. Koros, Macromolecules, 36, 1882 (2003).
(8.) J.H. Kim, W.J. Koros, and D.R. Paul, J. Membr. Sci., 282, 32 (2006).
(9.) M. Adam, M. Kratochvil, and W.J. Koros, Macromolecules, 41, 7920 (2008).
(10.) R.A. Hayes, U.S. Patent 4,981,497 (1991).
(11.) P.S. Tin, T.S. Chung, Y. Liu, R. Wang, S.L. Liu, and K.P. Pramoda, J. Membr. Sci., 225, 77 (2003).
(12.) C.E. Powell, X.J. Duthie, S.E. Kentish, G.G. Qiao, and G.W. Stevens, J. Membr. Sci., 291, 199 (2007).
(13.) X.J. Duthie, S.E. Kentish, C.E. Powell, G.G. Qiao, K. Nagai, and G.W. Stevens, Ind. Eng. Chem. Res., 46, 8183 (2007).
(14.) H. Kita, T. lnada, K. Tanaka, and K. Okamoto, J. Membr. Sci., 87, 139 (1994).
(15.) S. Matsui, T. lshiguro, A. Higuchi, and T. Nakagawa, J. Polym. Sci. Part B: Polym. Phys., 35, 2259 (1997).
(16.) S. Matsui and T. Nakagawa, J. Appl. Polym. Sci., 67, 49 (1998).
(17.) M.E. Rezac, E.T. Sorensen, and H.W. Beckham, J. Membr. Sci., 136, 249 (1997).
(18.) A. Bos, I.G.M. Punt, M. Wessling, and H. Strathmann, J. Polym. Sci. Part B: Polym. Phys., 36, 1547 (1998).
(19.) Y. Xiao, T.-S. Chung, H.M. Guan, and M.D. Guiver, J. Membr. Sci., 302, 254 (2007).
(20.) T.-S. Chung, M.L. Chng, K.P. Pramoda, and Y. Xiao, Langmuir, 20, 2966 (2004).
(21.) Y. Xiao, T.-S. Chung, and M.L. Chng, Langmuir, 20, 8230 (2004).
(22.) T. Masuda and T. Higashimura, Adv. Polym. Sci., 81, 121 (1987).
(23.) C. Gong, Q. Luo, Y. Li, M. Giotto, N.E. Cipollini, Z. Yang, R.A. Weiss, and D.A. Scola, J. Polym. Sci. Part A: Polym. Chem., 48, 3964 (2010).
(24.) F.W. Harris, A. Pamidimukkala, and R. Gupta, J. Macromol. Sci. Chem., 21, 1117 (1984).
(25.) C. Feger, Advances in Polyimide Science and Technology, Technomic, PA (1993).
(26.) M.E. Rezac and B. Schoberl, J. Membr. Sc., 156, 211 (1999).
(27.) N.T. Karangu, M.E. Rezac, and H.W. Beckham, Chem. Mater., 10, 567 (1998).
(28.) K. Tanaka, H. Kita, M. Okano, and K. Okamoto, Polymer, 33, 585 (1992).
(29.) X. Duthie, S. Kentish, S.J. Pas, A.J. Hill, C. Powell, K. Nagai, G. Stevens, and G. Qiao, J. Polym. Sci. Part B: Polym. Phys., 46, 1879 (2008).
(30.) H. Kawakami, M. Mikawa, and S. Nagaoka, J. Membr. Sci., 118, 223 (1996).
Shinji Kanehashi, (1) Masaki Onda, (1) Ryohei Shindo, (1) Shuichi Sato, (1) Shingo Kazama, (2) Kazukiyo Nagar (1)
(1) Department of Applied Chemistry, Meiji University, 1-1-1 Higashi-mita, Tama-ku, Kawasaki, Kanagawa 214-8571, Japan
(2) 2 Research Institute of Innovative Technology for the Earth (RITE), 9-2 Kizugawadai, Kizu-Cho, Soraku-Gun, Kyoto 619-0292, Japan
|Printer friendly Cite/link Email Feedback|
|Author:||Kanehashi, Shinji; Onda, Masaki; Shindo, Ryohei; Sato, Shuichi; Kazama, Shingo; Nagai, Kazukiyo|
|Publication:||Polymer Engineering and Science|
|Date:||Aug 1, 2013|
|Previous Article:||Effect of molding condition and pellets material on the weld property of injection molded jute/polylactic acid.|
|Next Article:||Microstructure, thermal, physico-mechanical and tribological characteristics of molybdenum disulphide-filled polyamide 66/carbon black composites.|