4270
J. Phys. Chem. B 2008, 112, 4270-4275
Characteristics of Polyethersulfone/Sulfonated Polyimide Blend Membrane for Proton Exchange Membrane Fuel Cell L. Wang,†,‡ B. L. Yi,*,† H. M. Zhang,† and D. M. Xing† Proton Exchange Membrane Fuel Cell Key Materials and Technology Laboratory, Dalian Institute of Chemical Physics, Chinese Academy of Science, 457 Zhongshan Road, Dalian, Liaoning 116023, People’s Republic of China, and Graduate School of the Chinese Academy of Sciences, Beijing 100039, People’s Republic of China ReceiVed: NoVember 8, 2007; In Final Form: January 4, 2008
Solution-cast membranes from sulfonated polyimide (SPI) and its blend were prepared from polyethersulfone (PES) and SPI. The water uptake and swelling were tested and compared between the SPI membrane and the four kinds of blend membranes. Through comparison of the stability of the membranes, we concluded that the PES could greatly increase the stability of the whole membrane and restrict the swelling. However, the PES did not decrease the water uptake very much. We also compared the fuel cell performance with different membranes. The performance was decreased when the content of the PES in the blend membrane increased. The loss of the fuel cell performance with the blend membranes did not decease very much before the content of the PES was exceeded 20%. It was prospected that the blend membrane could increase the stability of the SPI and, more importantly, even replace the commercial Nafion membranes.
1. Introduction Fuel cells are efficient devices that generate electricity via chemical reaction of fuels and oxygen and therefore have been attracting more and more attention as a clean energy system.1 One of the big challenges in the current fuel cell research is to develop an alternative membrane to the perfluorinated ionomers, for example, Nafion (DuPont). The perfluorinated ionomer membranes are highly proton conductive and are chemically and physically stable at moderate temperatures. However, these preferable properties are deteriorated above their glass transition temperature (Tg) ca. 110 °C. High fuel permeability, high cost, and environmental inadaptability of the fluorinated materials are serious shortage for the practical fuel cell applications. Alternative low-cost membranes with improved properties are thus needed, and many different kinds of membranes have been investigated during the past decade.2,3 These membranes are mainly sulfonated aromatic thermostable polymers such as polyetherketones, polysulfones, polyparaphenylenes, or polybenzimidazoles. Sulfonated polyimides (SPI) based on naphtahalenic moieties are considered to be promising material owing to their swelling, mechanical, and conducting properties.4 Despite the above-mentioned salient properties, the hydrolytic stability is a primary obstacle for the application as an electrolyte membrane in a fuel cell. Various approaches have been put forth to develop membranes with higher hydrolytic stability that include: (1) incorporating monomers having flexible linkages,5 (2) incorporating monomers not having both sulfonic acid group and amine group in the same ring,6 (3) diamine monomers having high nucleophilicity,7 (4) using aliphatic diamines,8 (5) using napthalenic dianhydrides,9 and (6) diamines having sulfonic acid group in side chain.10 For preparing more stable membranes, a method of reinforcement was developed to * To whom correspondence should be addressed. Phone: +86 411 84379536. Fax: +86 411 84379535. E-mail address:
[email protected]. † Dalian Institute of Chemical Physics, Chinese Academy of Sciences. ‡ Graduate School of Chinese Academy of Sciences.
increase the stability of hydrocarbon membranes. For example, Xing et al. used porous polytetrafluoroethylene (PTFE) to reinforce sulfonated poly(ether etherketone) (SPEEK).11 Yamaguchi used porous PTFE as a porous substrate and filled the pores with poly(vinylsulfonic acid/acrylic acid) crosslinked gel. The composite membrane had low methanol permeability and was thermally stable even up to 130 °C.12 In our previous work, we used the porous PTFE membrane to reinforce the Nafion.13 The mechanical stability and maximum strength of the membrane were improved. The good tensile strength made it possible for a thinner composite membrane to be used in fuel cell that consequently improves the fuel cell performance. Polymer blends wherein a sulfonated polymer with high proton conductivity is combined with a nonconductive engineering thermoplastic chosen to maintain mechanical integrity has become a popular contemporary approach to the design of improved PEM materials. A variety of different polymer pairs have been considered to be of relevance to some recently published papers, in particular SPEEK in combination with poly(ether imide) (PEI),14 nylon 6 (PA6),15 poly(ether sulfone) (PES),16 and poly(2,6-dimethyl phenylene oxide) (PPO).17 The improvement in mechanical stability can be attributed to the entanglement of these polymers and to possible (partial) mixing due to specific interactions, e.g., ion-dipole, dipole-dipole, and proton transfer.14-17 However, the stability of SPI within blend membranes has not been reported so far. PES is a kind of material with good stability. Whether or not the minority component of PES in the PES/SPI blend membrane has the ability to restrict hydrolyzation of the whole membrane has not been investigated. Here we report a blend membrane prepared from PES and SPI. The stability of the SPI was greatly improved. We selected the 20%PES/80%SPI (donated as PES/SPI-20) membrane to investigate on transmission electrony microscopy (TEM). The TEM results indicated that the PES and SPI could blend homogeneously. The stability of the SPI is also greatly improved. The fuel cell performance of the blend membranes
10.1021/jp710704v CCC: $40.75 © 2008 American Chemical Society Published on Web 03/19/2008
Characteristics of a PES/SPI Blend Membrane
J. Phys. Chem. B, Vol. 112, No. 14, 2008 4271
with different content of PES was tested. The way to modify the SPI is proved to be very effective. Besides, the blend membranes would be considered an alternate material to the commercial Nafion membranes. 2. Experimental Section 2.1. Membrane Preparation. The blend membranes were prepared by the solution-cast method. The SPI (NTDA-BDSA/ ODA (1/1), ion-exchange capacity (IEC) ) 1.98 mequiv g-1) was prepared by the same method reported previously.18 A calculated amount of SPI and PES (CHINA ANOW MICROFILTRATION CO., LTD) was added to the beaker, and then dimethyl sulfoxide (DMSO) was added to dissolve the blend polymers under stirring for 24 h. Then, the 3% (w/v) solution was poured onto glass plate. The glass plate was dried on a hot plate at 60 °C for 24 h and finally dried in a vacuum oven at 150 °C for 5 h. The thickness of the blend membranes was easily controlled by the amount of the PES/SPI solution. The membrane was soaked in 0.5 M H2SO4 at room temperature for 12 h. The membrane was thoroughly washed with deionized water. The membrane thickness was about 25 µm. The IEC was measured by means of a traditional titration method. A membrane sample, which is about 0.5 g, was soaked in 100 mL of 1.0 M NaCl solution for 2 days. Released proton concentration was titrated using 0.01 M NaOH solution. 2.2. TEM Characterizations. The blend membranes were stained with lead by ion exchange of the sulfonic acid groups by immersing them overnight in a large excess of 0.5 M Pb(ac)2 aqueous solution, rinsing with water, and drying at room temperature for 12 h.19 The stained samples were embedded in epoxy resin followed by ultra-microtomy with a diamond knife to obtain thin sections and placed on copper grids. Images were taken on a JEOL JEM-2011 transmission electron microscope using an accelerating voltage of 120 kV. 2.3. FTIR Spectrum. The infrared spectras of the membranes were recorded at a resolution of 2 cm-1 with a FTIR spectrometer (JASCO FT/IR-4100). 2.4. Water Uptake and Swelling. Samples of the membranes were weighed (W1) after immersion in deionized water for 12 h at controlled temperature. Then, samples were weighed (W2) after drying in a vacuum oven at 80 °C for 12 h. Water content (∆W) was calculated from eq 1
∆W (wt %) ) (W1 - W2)/W2 × 100
3. Results and Discussion
(1)
where W1 and W2 are the wet mass and dry mass of the sample expressed in grams, respectively. Membranes specimens (size 4 cm × 5 cm) were stored in a vacuum oven at 80 °C for 12 h, and the distance between specified positions was measured before (L1) and after (L2) the samples were soaked in deionized water controlled at 80 °C for 12 h. Dimensional change (∆L) was calculated by using the following equation
∆L (%) ) (L2 - L1)/L1 × 100
2.6. Tensile Strength. The maximum strength of the membranes was measured with a tension tester AG-2000A (Shimadzu, AUTO graph) at room temperature. Tensile conditions were based on Chinese Standard QB-13022-91, and samples were measured using a programmed elongation rate of 50 mm min-1. 2.7. Thermal Stability. Thermal analysis of the SPI and PES/ SPI membranes were performed via thermogravimetry (TGA) with a Pyn_sdiamond TGA-DTA-THERMOSTSR instrument at a heating rate of 10 °C min-1 under nitrogen. 2.8. Durability of the Fuel Cell with SPI and Blend Membranes. For the durability test of the cells with 25 µm blend membranes, the cells were operated at a constant current density of 500 mA cm-2 at 80 °C under fully humidified conditions. The pressure of the H2 and O2 was set at 0.12 MPa. The flow rates of the inlet gases were kept at H2 25 mL min-1 and O2 100 mL min-1. The voltage at 500 mA cm-2 and the open-circuit voltage (OCV) were recorded. 2.9. Membrane Electrode Assembly Preparation and Single Cell Test. The membrane electrode assembly (MEA) was prepared by a hot pressing process. The 20 wt. % Pt/Vulcan XC-72 (Pt/C) catalyst from E-TEK, carbon paper from Toray, PTFE suspension, and Nafion solution (DuPont) were used in the MEA preparation. The Pt/C catalyst loading of the anode and the cathode were both 0.5 mg Pt cm-2. Two electrodes with effective area 5 cm2 were hot-pressed onto a membrane to form a MEA. The MEA was mounted in a single cell with stainless steel end plates and stainless steel mesh flow field as current collectors. The flowing channel of the hardware was a stainless tube. The fuel cell performance was evaluated by I-V curves at 80 °C. The fuel and oxidant were fed in with co-flow. The flow rates of inlet gases were kept at H2 40 mL min-1 and O2 80 mL min-1. The fuel cell was started by increasing the cell temperature set point to 80 °C and raising the humidifier temperatures to their respective values. Similarly, the H2 and O2 pipeline temperatures were also maintained 5 °C higher than the humidifiers to prevent water condensation in the pipelines. All data were obtained at the following conditions: humidifier temperature TH2/TO2 ) 80/80 °C; cell temperature Tcell ) 80 °C; gas pressure PH2/PO2 ) 0.30/0.30 MPa. And the data were recorded after 12 h of fuel cell operation.
(2)
2.5. Water Stability. The stability in water of the SPI and PES/SPI membranes was examined by soaking the membranes in distilled water at 80 °C and was characterized by the expended time until the hydrated membranes lost their mechanical properties. The criterion for the judgment of the loss of mechanical properties was that the membrane was broken when lightly bent. Also the residual weight after being immersed in distilled water at 80 °C for 24 h was recorded for both the SPI and PES/SPI membranes. The PES/SPI-32 membrane before and after 600 h water stability test was examined by FTIR.
3.1. TEM Characterizations of the Membranes. SPI and homogeneous PES/SPI-20 blends were obtained by solutioncasting method from DMSO, as indicated in the TEM micrograph in Figure 1. The blend compositions used in the experiments, from 95/5 (w/w) to 70/30 (w/ w), are all completely transparent. The very fine black dots in the micrographs represent the hydrophilic (ionic) domains, and the brighter areas represent hydrophobic domains. All the membranes exhibited spherical ionic clusters of relatively uniform size. These ionic clusters were not connected with each other but were rather isolated. Figure 2 shows the distribution of the hydrophilic (ionic) domains’ size based in parts c and d of Figure 1. From the statistical result, the size of the domains in Figure 1c was mostly in the range between 20 and 30 nm. The size of the domains in Figure 1d was mostly in the range between 65 and 70 nm. The added PES causes a little aggregation of the hydrophilic phase and the larger of the black dots in the membranes. The size of the black dots increased from about 30 to 70 nm after 20% PES was added. However, there is still homogeneous morphology of the blend membrane. The single-phase morphology could
4272 J. Phys. Chem. B, Vol. 112, No. 14, 2008
Wang et al.
Figure 1. TEM images of SPI and PES/SPI-20 membranes: (a) SPI (X30000); (b) PES/SPI-20 (X30000); (c) SPI (X100000); (d) PES/SPI-20 (X100000).
arise from thermodynamic miscibility of the two polymers or from the solution casting procedure, which could produce a meta-stable morphology for the blend by freezing-in a singlephase state of the solution during solvent evaporation. 3.2. FTIR Spectrum of Membranes. Figure 3 is the FTIR spectrums of the blend membranes with different PES contents (0, 5, 10, 20, and 32%). The sulfone groups of PES give characteristic peaks at 1152 cm-1. Antisymmetric CsO stretching frequencies occur at 1250 and 1014 cm-1, while strong absorptions in the 1600-1475 cm-1 region are associated with the benzene ring stretching mode.20 The absorption density of these peaks increases with the PES content in the blend membrane increasing. This indicates that the PES was successfully introduced into the blend membranes. Similarly, the IR spectra of the SPI shows the peaks at 1714 and 1670 cm-1 are due to the naphthalimide absorption of the polyimides and peak at 1350 cm-1 represent the CsN vibration. The peak at 1030 cm-1 can be assigned to the symmetrical SdO stretching vibration, while the peak at 1255 cm-1 can be identified as an asymmetric OdSdO stretching vibration. The peak at 1499 cm-1 is due to the ether bridge absorption. These could prove
that the presence of SPI and PES in the blend membranes. There are also some frequency shifts between the characteristic groups of the pure SPI and PES/SPI blend membranes. The frequency shifts are observed at 1098-1105 cm-1 (Figure 4). These frequency shifts usually indicate specific interactions between the two polymers. The interactions could make the two polymers miscible. 3.3. Water Uptake, IEC, Swelling, and Water Stability of the SPI and Blend Membranes. Table 1 shows the comparison of the water uptake and swelling of SPI and blend membranes at 80 °C. The water uptake of SPI is obviously higher than blend membranes. This is because the addition of the PES in the blend membranes. The IECs calculated and measured of the SPI and blend membranes were listed in Table 1. The IEC values measured were close to the theoretical values. The IEC of the SPI polymer we used is about 1.98 mequiv g-1. It has very large dimensional change when immersed into hot water (80 °C).21 The swelling of the SPI and blend membranes show anisotropy of the SPI induced by the rigidity of the SPI polymer chains.22 The swelling of the blend membranes is reduced by the PES and the interaction between the PES and SPI by about
Characteristics of a PES/SPI Blend Membrane
J. Phys. Chem. B, Vol. 112, No. 14, 2008 4273
Figure 4. FTIR spectrums of the blend membranes with different PES contents. (a) PES/SPI-32, (b) PES/SPI-20, (c) PES/SPI-10, (d) PES/ SPI-5, (e) SPI.
Figure 2. Distribution of sulfonate groups’ particle size in the SPI and PES/SPI-20 blend membrane obtained from the cross-section image of TEM photomicrograph: (a) SPI; (b) PES/SPI-20.
Figure 3. FTIR spectrums of the blend membranes with different PES contents. (a) PES/SPI-32, (b) PES/SPI-20, (c) PES/SPI-10, (d) PES/ SPI-5, (e) SPI.
53.1-91.7% and 82.2-94.7% for the length and thickness, respectively. The extreme swelling and dimensional change is one of the reasons caused the fail of the membranes in fuel cell. On the other hand, the sorption of a large amount of water will bring polymer chain relaxation, leading to significant loss in tensile strength. Here the PES reduced the swelling and the dimensional change greatly. Figure 5 showed the water-stability test results of the SPI and PES/SPI membranes. The hydrolytic stability of the blend membranes is much better than the SPI membrane by measuring the residual weight of the membranes after being immersed in water at 80 °C for 24 h and by comparing the time of breaking of the membranes when being lightly bent at 80 °C in distilled water. The blend membranes exhibited higher hydrolytic stability than the pure SPI and the hydrolytic stability increased with the increase of PES content. The PES/SPI-32 membrane could endure up to 800 h water stability test. Figure 6 showed the FTIR results of the PES/SPI-32 membranes before and after
Figure 5. Water stability of the SPI and PES/SPI blend membranes: (a) weight after water stability test, (b) steady time for the membranes.
Figure 6. FTIR spectrums of the PES/SPI-32 membranes before and after the 600-h water stability test. (a) PES/SPI-32 before the test; (b) PES/SPI-32 after the test.
the 600-h water stability test. The peak at 1030 cm-1, which can be assigned to the symmetrical SdO stretching vibration was changed a little. The peak around 765 cm-1, which can be assigned to the out-of-phase bending of imides ring had no obvious change. All of these could prove that the backbone and the sulfonic acid group of the SPI did not degrade much after the 600-h water stability test. The blend membranes had very good water stability. This may be because with the increase of the PES content the water uptake of the blend membrane was markedly reduced and the interaction between the two polymers increased. These increase the mechanical stability of the polymer blends. In other words, the lower IEC and water uptake of the PES/SPI blends than the pure SPI can also be an additional reason to increase the hydrolytic stability. All of these are expected to improve the stability of the membrane and then increase the durability of the fuel cell. It is likely due to the interaction between the PES and the SPI.
4274 J. Phys. Chem. B, Vol. 112, No. 14, 2008
Figure 7. TGA traces of SPI, PES, PES/SPI-5, PES/SPI-10, PES/ SPI-20, and PES/SPI-32.
Wang et al.
Figure 9. Fuel cell durability test of a single cell with PES/SPI-20 membranes with fully humidified H2/O2 at 80 °C. PH2/PO2 ) 0.12/0.12 MPa.
TABLE 1: Water Uptake and Swelling of the SPI and PES/SPI Blend Membranes swelling (%) water IEC IEC uptake membranes calculated measured (%) SPI PES/SPI-5 PES/SPI-10 PES/SPI-20 PES/SPI-32
Figure 8. Performance of H2/O2 fuel cells with Nafion NRE-212, SPI, and PES/SPI blend membranes under fully humidified conditions at 80 °C.
3.4. Thermal Stability of SPI and PES/SPI Blend Membranes. The thermal stability of the SPI and PES/SPI blend membranes were analyzed by TGA. Figure 7 presents the TGA traces of the five membranes. All the membranes used to test are stored at ambient atmosphere. The figure shows that membranes exhibited a three-step weight loss when the content of the PES is below 20%. When the content of the PES is higher than 20%, the membranes exhibited a four-step weight loss. The first step was observed at around 100 °C due to the loss of water molecules absorbed by the highly hygroscopic SO3 groups.23,24 It can be seen that, when the PES content in the membrane becomes higher, the weight loss of the first step becomes less. This was due to the PES in the membrane decreases the water uptake of the membrane. The second step of weight loss was observed at around 270 °C due to the decomposition of sulfonyl groups.23-25 The third step indicates the decomposition of the polymer backbones. For SPI, PES/ SPI-5, and PES/SPI-10, the onset for third step of weight loss started around 550 °C. But for PES/SPI-20 and PES/SPI-32 the onset for the third step of weight loss started around 370 °C due to the decomposition of the PES backbones. Generally, all the membranes are stable under typical operating conditions of PEMFC. These results suggest that all the blend membranes can be used as polymer electrolyte membranes for medium and high-temperature operation. 3.5. Performance of PEMFCs Operated under Fully Humidified Conditions. Figure 8 shows the polarization curves of cells with SPI and PES/SPI blend membranes operated under fully humidified conditions with cell temperature at 80 °C. The cell performance with the pure SPI was a little better than that with the PES/SPI membranes. When the PES content in the blend membrane increased, the fuel cell performance with the blend membrane decreased. The performance loss of the PES/
2.00 1.88 1.78 1.54 1.34
1.98 1.85 1.75 1.50 1.27
45.20 41.63 34.68 20.29 16.27
tensile strength length thickness (MPa) 7.10 3.33 2.86 2.00 0.59
55.60 9.09 5.26 4.34 2.94
24.00 28.24 29.42 34.03 34.59
SPI blend membranes was due to the decrease of the proton conductivity of the membrane caused by the PES. However, the cell performances were of the blend membranes were still close to the cell performance of NRE-212 membrane. 3.6. Durability Test of the Fuel Cell. The voltage at the current density 500 mA cm-2 and the OCV of PEM fuel cell with the PES/SPI-20 membrane with time is shown in Figure 9. It is found that there is no obvious decrease of the cell voltage at 500 mA cm-2 and OCV of PES/SPI-20 membrane at the time scale of experiment with four times stop and start of the fuel cell. These results were greatly improved compared to the results of pure SPI in our previous work.18 It is suggested the PES blended SPI membrane has good stability under fuel cell conditions. Our data shows the increase in the stability of the membrane can resist the OCV decay rate of the fuel cell. These results provide a basis to the hypothesis that durability is also enhanced through improvements in dimensional stability and water uptake, presumably due to a reduction of the impact of the shrinkage stresses developed in the membrane. 4. Conclusions A kind of PES/SPI blend membrane was prepared using PES to blend with SPI. The characterization of the blend membranes were examined and verified by FTIR and TEM tests. The FTIR results proved the presence of the PES and SPI. The TEM results showed homogeneous structure of the blend membranes. Because of the addition of PES, the dimensional stability and the tensile strength were greatly improved. The swelling of the membrane was greatly reduced, and the stability of SPI was greatly improved. The cell performances with SPI and PES/ SPI membranes were with high fuel cell performance that is excellent for fuel cell applications. Both in situ and ex situ tests showed improved stability of PES/SPI blend membranes. Acknowledgment. We gratefully acknowledge the financial support for this work provided by National Natural Science Foundation of China (20476104).
Characteristics of a PES/SPI Blend Membrane References and Notes (1) Savadogo, O. J. Power Sources 2004, 127, 135-161. (2) Savadogo, O. J. New Mater. Electrochem. Syst. 1998, 1, 47-66. (3) Rozie`re, J.; Jones, D. Annu. ReV. Mater. Res. 2003, 33, 503-505. (4) Asano, N.; Aoki, M.; Suzuki, S.; Miyatake, K.; Uchida, H.; Watanabe, M. J. Am. Chem. Soc. 2006, 128, 1762-1769. (5) Fang, J.; Guo, X.; Harada, S.; Watari, T.; Tanaka, K.; Kita, H.; Okamoto, K. Macromolecules 2002, 35, 9022-9028. (6) Guo, X.; Fang, J.; Watari, T.; Tanaka, K.; Kita, H.; Okamoto, K. Macromolecules 2002, 35, 6707-6713. (7) Lee, C.; Sundar, S.; Kwon, J.; Han, H. J. Polym. Sci. Part A: Polym. Chem. 2004, 42, 3621-3630. (8) Miyatake, K.; Zhou, H.; Uchida, H.; Watanabe, M. Chem. Commun. 2003 368-369. (9) Genius, C.; Mercier, R.; Sillion, B.; Petiaud, R.; Cornet, N.; Gebel, G.; Pineri, M. Polymer 2001, 42, 5097-5105. (10) Yin, Y.; Fang, J.; Cui, Y.; Tanake, K.; Kita, H.; Okamoto, K. Polymer 2003, 44, 4509-4518. (11) Xing, D. M.; Yi, B. L.; Liu, F. Q.; Fu, Y. Z.; Zhang, H. M. Fuel Cells 2005, 3, 406-411. (12) Yamaguchi, T.; Miyata, F.; Nakao, S. J. Membr. Sci. 2003, 214, 283-292. (13) Wang, L.; Xing, D. M.; Liu, Y. H.; Cai, Y. H.; Shao, Z.-G.; Zhai, Y. F.; Zhong, H. X.; Yi, B. L.; Zhang, H. M. J. Power Sources 2006, 161, 61-67.
J. Phys. Chem. B, Vol. 112, No. 14, 2008 4275 (14) Kerres, J.; Ullrich, A.; Meier, F.; Haring, T. Solid-State Ionics 1999, 125, 243-249. (15) Cui, W.; Kerres, J.; Eigenberger, G. Separation Purification Technol. 1998, 14, 145-154. (16) Manea, C.; Mulder, M. Desalination 2002, 147, 179-182. (17) Jorissen, L.; Gogel, V.; Kerres, J.; Garche, J. J. Power Sources 2002, 105, 267-273. (18) Wang, L.; Yi, B. L.; Zhang, H. M.; Liu, Y. H.; Xing, D. M.; Shao, Z.-G.; Cai, Y. H. J. Power Sources 2007, 164, 80-85. (19) Kerres, J.; Ullrich, A.; Haring, Th.; Baldauf, M.; Gebhardt, U.; Preidel, W. J. New Mater. Electrochem. Syst. 2000, 3, 229-239. (20) Mohr, J. M.; Paul, D. R.; Pinnau, I.; Koros, W. J. J. Membr. Sci. 1991, 56, 77-98. (21) Wang, L.; Yi, B. L.; Zhang, H. M.; Liu, Y. H.; Xing, D. M.; Shao, Z.-G.; Cai,Y. H. J. Power Sources 2007, 167, 47-52. (22) Rollet, A. L.; Diat, O.; Gebel, G. J. Phys. Chem. B 2004, 108, 1130-1136. (23) Ye, X. H.; Bai, H.; Winston Ho, W. S. J. Membr. Sci. 2006, 279, 570- 577. (24) Guo, X.; Fang, J.; Watari, T.; Tanaka, K.; Kita, H.; Okamoto, K. Macromolecules 2002, 35, 6707-6713. (25) Blachot, J. F.; Diat, O.; Putaux, J. L.; Rollet, A. L.; Rubatat, L.; Vallois, C.; Muller, M.; Gebel, G. J. Membr. Sci. 2003, 214, 31-42.
