Solid State Ionics 201 (2011) 21–26
Contents lists available at SciVerse ScienceDirect
Solid State Ionics j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s s i
PVA nano composite membrane for DMFC application Jatindranath Maiti, Nitul Kakati, Seok Hee Lee, Seung Hyun Jee, Young Soo Yoon ⁎ Energy and Sensor Laboratory, School of Materials Science and Engineering, Yonsei University, 134 Shinchon Dong, Seoul 120-749, South Korea
a r t i c l e
i n f o
Article history: Received 22 July 2010 Received in revised form 15 July 2011 Accepted 18 July 2011 Available online 13 September 2011 Keywords: Polyvinyl alcohol (PVA) Sulfonated MWCNT Fluorinated MMT Polymer nano composite membrane DMFC
a b s t r a c t A new PEM composite membrane comprising of polyvinyl alcohol (PVA), sulfonic acid functionalized CNT and ﬂuorinated MMT has been fabricated. Composite polymer membrane has been prepared by simple solution casting method. Composite properties have been evaluated by using thermal gravimetric analysis (TGA), scanning electron microscopy (SEM), and FTIR techniques. The proton conductivity, methanol crossover and water uptake properties of newly fabricated membrane have been studied. The polymer membrane shows good thermal properties. The water content is in the range of 35–45%. Especially, it has been found that the ﬂuorinated MMT used in this study plays a decisive role in water uptake and acts as a hydrophobic surface for controlling the swelling. The proton conductivities and the methanol permeabilities of all the membranes are in the range of 10 − 3 to 10 − 2 S/cm and 2.08 × 10 − 6 cm 2/s at room temperature, respectively. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Fuel cell technology is a benign process and has the potential to become future generation green energy for portable electronics and vehicle propulsion [1–3]. Now a day, research on the improvement of polymer membrane for direct methanol fuel cell application is one of the challenging assignments to resolve the balance between proton conductivity and methanol crossover [4–6]. The large-scale commercial utilization of Naﬁon® in DMFC causes some issues such as low power density which is due to methanol crossover and dehydration at high temperature [4,7]. Low cost membrane is also required for DMFC commercial application without sacriﬁcing their properties. Reduction of membrane cost could be achieved by using non-ﬂuorinated polymer electrolytes with a cheaper polymer. From commercial point of view, PVA is a possible candidate to be used as a membrane for DMFC because of its low cost, good chemical stability, ﬁlm-forming ability, and high hydrophilicity and availability of cross-linking sites to create a stable membrane with good mechanical properties and selective permeability to water [8–11]. Furthermore, the PVA polymer used is biodegradable, nonhazardous, and environmentally benign . PVA based composite membranes if optimized may serve as a potential alternative protonconducting membrane for direct methanol fuel cell applications. Water uptake is also important in determining the ultimate performance of proton exchange membrane materials. In essentially all current polymer based membrane, water is required to facilitate proton conductivity. However, absorbed water also affects the mechanical properties of the membrane by acting as a plasticizer, lowering the Tg and modulus of the
⁎ Corresponding author. Tel.: + 82 2 2123 2847; fax: + 82 2 365 5882. E-mail address: [email protected]
(Y.S. Yoon). 0167-2738/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.ssi.2011.07.013
membrane . Careful control of water uptake is critical for reducing adverse effects of swelling and degradation of the mechanical properties of the membrane in humid environments, as well as inducing stresses between the membrane and the electrodes. Both conductivity and water uptake rely heavily on the concentration of ion conducting units (most commonly sulfonic acid) in the polymer membrane. Varying the ion content of the membrane can control both its water uptake and conductivity. While it is desirable to maximize the proton conductivity of the membrane by increasing its ion content (decreasing equivalent weight), other physical properties must be considered. Too many ionic groups will cause the membrane to swell excessively with water, which compromises mechanical integrity and durability. Polymer swelling can be reduced by crosslinking with suitable crosslinking agent, whereas proton conductivity can be increased by the formation of hybrid composites by the incorporation of proton conductors such as sulfonic acids . In the past few decades, nano clay (MMT) [15–17] and carbon nanotubes (CNTs)  were studied intensively as typical nano ﬁllers to incorporate into polymer matrices. The dispersion of nano clay in polymer can result in a reduction of moisture absorption, thermal stability, barrier properties, and ﬂammability as well as signiﬁcant enhancements of modulus, strength, and hence overall performance of nanocomposite . Furthermore, clay is inexpensive relative to traditional reinforcing materials and environmentally friendly. The sulfonic acid functionalized carbon nano tube based polymer composites has been revealed a remarkable improvement of proton conductivity and also capable of increasing the mechanical stability along with a decrease in methanol permeability . Herein we present a chemical strategy to increase the sulfonic acid content and better channel like network for proton transport of PVA membrane by incorporating sulfonic acid functionalized multi walled carbon nanotubes. The membranes with highly hydrophobic
J. Maiti et al. / Solid State Ionics 201 (2011) 21–26
ﬂuorinated surface exhibited improved proton conductivity and reduced methanol permeability at a relatively low water uptake . The hydrophobic blocks can serve as matrix for mechanical strength and limited swelling. We have integrated both hydrophobic surface and acidic group by using two different nano ﬁller. The strategy of using both sulfonated CNT and ﬂuorinated MMT to enrich proton conductivity, self humidiﬁcation, thermal and barrier properties of the membrane might be helpful in alleviating many signiﬁcant difﬁculties associated with fuel cell. In this work, composite membrane composed of polyvinyl alcohol (PVA), sulfonic acid functionalized CNT and ﬂuorinated MMT have been prepared by simple solution casting method. This is a new approach to add two different functionalized nano ﬁller into same matrix for studying the water uptake, proton conductivity, methanol permeability and thermal stability of composite membrane.
(SSA) (10 wt.%) and the mixture was stirred at room temperature for 24 h. After that, the homogeneous solutions were poured onto a plastic petridish. The cast polymer solutions were allowed to dry in air at room temperature for 24 h. The fully dried membranes were peeled off away from the petridish, and then heated in an oven at 120 °C for 1 h to make cross linking reaction. The membranes were stored in DI water before use. The membrane thickness was in the range of 100– 150 μm.
The membrane thermal stability was evaluated by using a thermogravimetric analyzer TA Q 50 system TGA. The samples were scanned at a heating rate of 10 °C/min under ﬂow of nitrogen.
2.1. Materials Polyvinyl alcohol (PVA) (molecular weight of 31,000–50,000), sulfosuccinic acid (SSA) (70 wt% in water solution), Hexaﬂuorophosphoric acid (65 wt.% solution in water), Montmorillonite K10, all the chemicals were purchased from Aldrich company. Multi walled CNT was purchased from EM Power, South Korea.
2.5. FTIR Fourier transform infrared (FT-IR) spectroscopic measurements were performed using a JASCO FT-IR 300E device. 2.6. Thermogravimetric analysis
2.7. Surface morphology and chemical composition characterization A scanning electron microscope with an energy dispersive X-ray spectroscopy system (FESEM JSM-6700F, JEOL coupled with INCA energy dispersive X-ray spectroscopy) was used to evaluate the membrane microstructure and chemical composition.
2.2. Sulfonation of MWCNT 2.8. Water uptake First, puriﬁcation of MWCNT (1 g) was carried out by reﬂuxing the CNT with 200 ml of 60% HNO3 at 120 °C for 4 h to remove the metal particles. The mixture was diluted, centrifuged and washed with excess DI water. The puriﬁed product was dried at 70 °C in vacuum oven for overnight. Second, the sulfonation of MWCNT was executed in presence of (NH4)2SO4 . 0.25 g of ammonium sulfate dissolved in 5 ml DI water was mixed with 0.25 g MWCNT. After that, the mixture was well agitated; it was heated at 235 °C for as long as 30 min. It is believed that at 235 °C, (NH4)2SO4 decomposes to generate SO3, and the formed SO3 reacts with carbon via its surface hydrogen atoms to have − SO3H groups linked onto it . ðNH4 Þ2 SO4 →2NH3 + H2 O + SO3
Carbon−H + SO3 →Carbon−SO3 H:
2.3. Fluorination of MMT MMT (1.2 g) was dispersed in 100 ml of deionized water using an ultrasonic bath. The suspension was stirred with 1 N H2SO4 (20 ml) at room temperature for 2 h in order to increase its surface activity and to remove impurities. The mixture was washed with excess deionized water and dried in vacuum oven at 70 °C for 12 h. 2 ml hexaﬂuorophosphoric acid (60 wt.% solution in water) and MMT (1 g) in 5 ml water were mixed and the mixture was stirred using a Teﬂon beaker at room temperature for 24 h . The product was washed in a mixture of isopropyl alcohol and water (1:1) for preventing agglomeration of MMT particle resulting in higher surface area product. Finally the material was dried in vacuum oven at 70 °C for 12 h.
The water uptake of the membranes was determined by measuring the change in the weight before and after the hydration. Pre-dried membranes were immersed in deionized water for 24 h, and then surface attached water onto the membrane was removed with ﬁlter paper. After that, the wetted membrane weight (Wwet) was determined as quickly as possible. The weight of dry membrane (Wdry) was determined after completely drying it in vacuum at 60 °C for 24 h. The water uptake (%) value of the membranes was calculated by using the following equation Water uptake ð% Þ =
Wwet −Wdry X 100: Wdry
2.9. Proton conductivity measurements Proton conductivity measurements were carried out at ambient temperature after equilibrating the membrane in de-ionized water for 1 day. The proton conductivity cell was composed of two 5 mm diameter platinum electrodes. The membrane sample was sandwiched between the platinum electrodes. Proton conductivity of the membranes was measured by an impedance spectroscopy using a Solartron 1260 gain phase analyzer, interfaced to a Solartron 1480 multistat. The measurement was carried out in a potentiostatic mode in the frequency range of 0.1 Hz to 10 MHz with 5 mV of oscillating voltage. The laboratory made four probe conductivity cell was used. The conductivity cell was placed in the head-space of a temperature controlled sealed vessel which was maintained at 100% relative humidity. Proton conductivity (σ) of the samples was calculated from impedance data using the following equation :
2.4. Membrane preparation 10 wt.% of PVA in water was stirred continuously at 90 °C for 6 h until the solution mixture reached a homogeneous solution. Then the PVA solutions were mixed with sulfonated MWCNT (1wt.%), ﬂuorinated MMT (1 wt.%) and cross linking agent sulfosuccinic acid
where R is the membrane resistance derived from the impedance value at zero phase angle, L is the distance between two potential
J. Maiti et al. / Solid State Ionics 201 (2011) 21–26
sensing platinum electrodes, W and D are the width and thickness of the membrane respectively. 2.10. Methanol crossover Methanol permeability measurement was carried out with a home-made permeation measuring cell that had two compartments. Compartment A was ﬁlled with 150 ml 20% (v/v) methanol solution in de-ionized water, and compartment B was ﬁlled with 150 ml deionized water. The membrane was mounted between the two compartments, and the diameter of the diffusion area was 3.0 cm. The solutions in both compartments were magnetically stirred. The methanol concentration in compartment B was monitored using a refractive index detector (RI BT600, Younglin Instrument Co., Korea) through a 1-mm diameter silicon tube with a 1.0 ml min − 1 constant ﬂow driven by a Master ﬂex pump. The output signal was converted by a data module (Autochro, Younglin Instrument Co., Korea) and recorded by a personal computer. Methanol permeability (P) was obtained by means of the following relationship : CB ðtÞ ¼
A P C ðt to Þ VBL A
where CA is the initial methanol concentration in compartment A; CB(t) the methanol concentration in compartment B at diffusion time t; VB the volume of de-ionized water in compartment B; L the thickness of the membrane; and A is the effective permeating area. 3. Results and discussion 3.1. FTIR The membrane was prepared by solution casting method as shown in scheme 1. FTIR studies have been carried out on samples containing PVA and sulfonated MWCNT, and composite membrane (Fig. 1). In
pure PVA, we observe a doublet peak around 1000–1300 cm− 1 and a broad region around 3000–3500 cm − 1. They are characteristic of PVA and have been assigned to C–O stretching and O–H stretching, respectively. In sulfonated MWCNT, two peaks at 1030 cm − 1 and around 700 cm − 1 have been assigned to S_O and S–O symmetric stretching, respectively. Upon blending sulfonated MWCNT and SSA with PVA, the S_O and S–O stretching characteristics of membrane grow up while the C–O stretching characteristic of PVA decreases. More direct evidence of the percentage of sulfur and ﬂuorine comes from EDX studies, as will be discussed later. 3.2. Thermal properties The thermal stability of the PVA nano composite membrane was evaluated through the TGA experiments. The thermogravimetric analysis (TGA) results for the PVA and composite membranes are shown in Fig. 2. Pure PVA sample exhibited two thermal decomposition stages. The ﬁrst occurred at 285 °C and the second at 400– 450 °C. These two stages reﬂect the breakage of the side and main backbone polymer chains, respectively . At the end of the analysis, at 800 °C, the PVA had 5% residue. The TGA curve of the PVA membrane showed three consecutive weight losses arising from the processes of thermal solvation, thermal desulfonation, and thermo oxidation and degradation of the polymer matrix. The ﬁrst weight loss of about 5 wt.% at 100 °C is closely associated with the loss of absorbed water molecules. Most of these absorbed water molecules are supposed to be in a bound state, rather than in the free molecular state . The second weight loss of about 35 wt.% at around 150– 380 °C corresponds to the loss of sulfonic acid group by the desulfonation and a breakage of some portion of polymer chains as well as breakage of the ester bonds. In the third weight loss of about 40 wt.% at temperatures N400 °C is due to the decomposition of the main chains of the PVA . The decomposition temperature of PVA was 285 °C. That temperature increased to 410 °C with chemical crosslinking and inorganic nano ﬁller addition. Many researchers have
Scheme 1. Membrane preparation.
J. Maiti et al. / Solid State Ionics 201 (2011) 21–26
Fig. 1. FTIR spectra of PVA, sulfonated MWCNT and membrane.
pointed out that incorporation of nanoparticles into a polymer matrix enhanced the membranes thermal stability for PVA. Mbhele et al.  claimed that the thermal degradation occurs from free radical formations at weak bonds and/or chain ends, followed by their transfer to adjacent chains via interchain reaction. The presence of the nanoﬁller restricted the mobility of the polymer chains, prohibited the free radical transfer and, therefore, suppressed the thermal degradation. It could therefore be concluded that the thermal stability was improved due to the additive effect of the MMT and CNT ﬁllers and the chemical crosslinking reaction between the OH group on the PVA and the COOH group on the SSA. 3.3. SEM and EDX The morphology and composition of composite membranes have been analyzed by a SEM microscope. SEM photographs for the PVA/S-MWCNT/F-MMT/SSA (membrane 1) and PVA/S-MWCNT/SSA (membrane 2) composite polymer membrane are shown in Fig. 3 (a) and (b), respectively. The basic difference between these membranes is that membrane 1 contains F-MMT and membrane 2 does not contain F-MMT except this all other compositions are same for both. No relevant morphological features have been noticed in these membranes although some difference on the surface morphology for membrane 1 has been detected. Many different sizes of aggregates or chunks that are randomly distributed on the top surface have been
PVA polymer PVA nano composite
Fig. 3. (a) SEM photograph of the PVA/S-MWCNT/F-MMT/SSA (membrane 1) and (b) PVA/S-MWCNT/SSA (membrane 2).
observed for membrane 1 (PVA/S-MWCNT/F-MMT/SSA). The compatibility of membrane 2 (without F- MMT) is still uniform and homogenous. This indicates that the nano F-MMT was not properly dispersed within the PVA polymer matrix, as shown in Fig. 3. However, it is clearly seen that cracks have been generated on the top of the surface in the both type (with or without MMT) of membrane. The degree of sulfonation and ﬂuorination has been calculated by EDX measurement as shown in Fig. 4, 0.16 wt.% of the sulfur is attached to the MWCNT by this method (Fig. 4 a). Indeed the amount of sulfonation is low and can be systematically varied by changing the concentration of ammonium sulfate. The degree of ﬂuorination on the MMT is 0.62 wt.% (Fig. 4b). 3.4. Water uptake
60 40 20 0 0
Temperature (oC) Fig. 2. TGA plot of PVA and nano composite membrane.
The water uptake of the PVA nano composite membranes is 37 wt.%. While the water uptake of pure PVA membrane (24.4 wt%) is comparable to that of Naﬁon®117 membrane, the PVA nano composite membrane exhibited a remarkably higher water uptake owing to the high hydrophilic nature of cross linking agent and the disruption of the highly ordered arrangement of pristine PVA chain individually . We have prepared PVA nano composite membrane without ﬂuorinated MMT to study the positive effect of the addition of hydrophilic SSA as well as sulfonated CNT and the negative effect of hydrophobic ﬂuorinated MMT function on the water uptake. The water uptake of PVA nano composite without ﬂuorinated MMT is
J. Maiti et al. / Solid State Ionics 201 (2011) 21–26
Fig. 4. (a) EDX analysis of S-MWCNT, (b) EDX analysis of F-MMT.
45 wt.% which is higher than that of the ﬂuorinated MMT based membrane. The ﬂuorinated MMT decreases the water content by increasing the hydrophobic surface property. This result shows that water uptake can be controlled by using hydrophobic–hydrophilic interaction to reach a balance.
crosslinking agent present in the membrane. Water uptake plays a critical role in proton conduction because it is the major carrier of protons. However, excess swelling in water reduces the membrane's mechanical strength. Typically, many polymer electrolyte membranes swell or even become soluble in water when the sulfonation level
3.5. Proton conductivity and methanol permeability The proton conductivity of the membrane is a key property that directly affects operational fuel cell voltage. The proton conductivity measurements of the membrane were run at RH 100% at room temperature in the longitudinal direction by AC impedance spectroscopy. The proton conductivities of the hybrid membranes measured at a temperature range between 25 and 80 °C. The proton conductivity value of membrane 1 (PVA/S-MWCNT/F-MMT/SSA) and membrane 2 (PVA/S-MWCNT/SSA) is 0.006 S/cm and 0.004 S/cm at 30 °C, respectively. We have added same amount of sulfonic acid source (1 wt.% sulfonated MWCNT and 10 wt.% crosslinking agent with respect to PVA) in both type of membrane. Both type of membrane show more or less same order of proton conductivity. Proton conductivity increases with increasing temperature in both cases (Fig. 5). It was reported that proton conductivity of PVA based membrane increased with increasing the content of crosslinking agent (SSA) [30,31]. In our case, observed proton conductivity is low compared to other membranes (reported proton conductivity in the order of 10 −2 S/cm) [31,32] due to low level of sulfonic group in MWCNT as well as
Fig. 5. Proton conductivity of membrane 1 (PVA/S-MWCNT/F-MMT/SSA) and membrane 2 (PVA/S-MWCNT/SSA) in the temperature range from 25 to 80 °C under 100% RH conditions.
J. Maiti et al. / Solid State Ionics 201 (2011) 21–26
increases in order to obtain high proton conductivity . In our case, crosslinked as well as hydrophobic–hydrophilic interaction both approach are applied to balance water uptake and proton conductivity. The methanol permeability of the membranes 1 and 2 has been calculated as 2.08 × 10 −6 cm 2/s and 4.13 × 10 −6 cm 2/s, respectively. It is known that methanol permeates through hydrophilic ionic channels; especially free water molecules and that proton are transported by hopping between ionic sites due to hydrogen bonding between bound water molecules as well as through ionic channels. Therefore, it is expected that the methanol permeability should be decreased due to the MMT particles acting as materials for blocking the methanol transport and/or for reducing free water. However methanol permeability of our membrane is high compared to other type of PVA membranes (methanol permeability in the range of 10 −8–10 −7 cm 2 /s) [31,32]. This is possibly due to the crack generation on the surface of the membrane. MMT does not act as a blocking material for methanol transport in our study owing to their agglomerate structure. For comparison purposes, methanol permeability and proton conductivity of Naﬁon 115 membranes were also measured using the same apparatus and testing conditions. The values of methanol permeability and proton conductivity obtained were 1.78 × 10−6 cm 2/s and 0.112 S/cm at 30 °C, respectively. In relation to this study, most of the membranes prepared herein have similar methanol permeability compared to Naﬁon 115 membrane. Proton conductivity values are, however, lower than that of the Naﬁon115. Further attempts have yet to be made to improve the proton conductivity as well as methanol permeability of the PVA nanocomposite membrane. This might be achieved by optimizing the degree of sulfonation and the degree of crosslinking of the membrane. 4. Conclusions In the present work, crosslinked PVA nanocomposite membrane containing sulfonic acid and ﬂuorine group has been prepared by simple solution techniques and also evaluated as a potential polymer electrolyte membrane in direct fuel cell application. Especially, sulfonic acid functionalized MWCNT and ﬂuorine functionalized MMT have been effectively introduced into the PVA matrix with sulfosuccinic acid as a crosslinked agent. The proton conductivity of the membrane is in the range of 0.004 to 0.01 S/cm. The low proton conductivity is due to low level of sulfonic acid group present in the membrane. The methanol permeability of the membrane is 2.08 × 10 −6 cm 2/s. It has been found that the water uptake of the membrane can be controlled by using ﬂuorine functionalized MMT. Our future attempt is to obtain a uniform, improved property and to
demonstrate the practical applicability of using this PVA composite membrane without cracks and nano-sized chunks or aggregate in DMFC. Acknowledgments This research was supported by the Pioneer Research Center Program through the National Research Foundation of Korea funded by the Ministry of Education, Science and Technology (20110001676). References                                
M. Winter, R.J. Broadd, Chem. Rev. 104 (2004) 4225. G. Alberti, M. Casciola, Solid State Ionics 145 (2003) 3. C.Y. Wang, Chem. Rev. 104 (2004) 4727. H. Ahmed, S.K. Kamarudin, U.A. Hasran, W.R.W. Daud, Int. J. Hydro. Energy 35 (2010) 2160. A.S. Arico, S. Srinivasan, V. Antonucci, Fuel Cells 1 (2001) 133. G.G. Kumar, P. Uthirakumar, K.S. Nahm, R.N. Elizabeth, Solid State Ionics 180 (2009) 282. Q. Li, R. He, J.O. Jensen, N.J. Bjerrum, Chem. Mater. 15 (2003) 4896. M. Krumova, d. Lopez, R. Benavente, C. Mijangos, J.M. Perena, Polymer 41 (2000) 9265. Y. Jin, J.C.D. Costa, G.Q. Lu, Solid State Ionics 178 (2007) 937. A. Martinelli, A. Matic, P. Jacobsson, L. Borjesson, M.A. Navarra, A. Fernicola, S. Panero, B. Scrosati, Solid State Ionics 177 (2006) 2431. J. Qiao, T. Okada, H. Ono, Solid State Ionics 180 (2009) 1318. A. Anis, A.K. Banthia, S. Bandyopadhyay, J. Power Sources 179 (2008) 69. M.A. Hickner, H. Ghassemi, Y.S. Kim, B.R. Einsla, J.E. McGrath, Chem. Rev. 104 (2004) 4587. D.S. Kim, H.B. Park, J.W. Rhim, Y.M. Lee, Solid State Ionics 176 (2005) 117. C.C. Yang, Y.J. Lee, J.M. Yang, J. Power Sources 188 (2009) 30. R.A. Vaia, H. Ishii, E.P. Giannelis, Chem. Mater. 5 (1993) 1694. P. Duangkaew, J. Wootthikanokkhan, J. Appl. Polym. Sci. 109 (2008) 452. R. Kannan, B.A. Kakade, V.K. Pillai, Angew. Chem. Int. Ed. 47 (2008) 2653. D.S. Kim, H.I. Cho, D.H. Kim, B.S. Lee, S.W. Yoon, Y.S. Kim, G.Y. Moon, H. Byun, J.W. Rhim, J. Membrane Sci. 342 (2009) 138. C.Y. Du, T.S. Zhao, Z.X. Liang, J. Power Sources 176 (2008) 9. F. Peng, L. Zhang, H. Wang, P. Lv, H. Yu, Carbon 43 (2005) 2405. A. Majid, S. Argue, D. Kingston, S. Lang, J. Fluorine Chem. 128 (2007) 1012. S.H. Park, J.S. Park, S.D. Yim, S.H. Park, Y.M. Lee, C.S. Kim, J. Power Sources 181 (2008) 259. R.Q. Fu, J.J. Woo, S.J. Seo, J.S. Lee, S.H. Moon, J. Power Sources 179 (2008) 458. S.J. Lue, J.Y. Chen, J.M. Yang, J. Macromol, Sci. Part B: Physics 47 (2008) 39. I. Honma, O. Nishikawa, T. Sugimoto, S. Nomura, H. Nakajima, Fuel Cells 2 (2002) 52. D.S. Kim, I.C. Park, H.I. Cho, D.H. Kim, G.Y. Moon, H.K. Lee, J.W. Rhim, J. Ind. Eng. Chem. 15 (2009) 265. Z.H. Mbhele, M.G. Salemane, C.G.C.E.V. Sittert, J.M. Nedeljkovic, V. Djokovic, A.S. Luyt, Chem. Mater. 15 (2005) 5019. Z. Jiang, X. Zheng, H. Wu, F. Pan, J. Power Sources 185 (2008) 85. J.W. Rhim, H.B. Park, C.S. Lee, J.H. Jun, D.S. Kim, Y.M. Lee, J. Membrane Sci. 238 (2004) 143. C.C. Yang, S.J. Lue, J.Y. Shih, J. Power Sources 196 (2011) 4458. Y.F. Huang, L.C. Chuang, A.M. Kannan, C.W. Lin, J. Power Sources 186 (2009) 22.