Polymer Blend Membrane with Charge-Transfer Complex for PEFC Application

Wednesday, 4 October 2017: 16:05
National Harbor 14 (Gaylord National Resort and Convention Center)
M. Nishihara (Kyushu University), S. Feng, L. Christiani (Graduate School of Engineering, Kyushu University), S. Kondo, T. Kaseyama, T. Nakazawa, T. Kikuchi (Nissan Chemical Industries, Ltd), and K. Sasaki (Department of Hydrogen Energy Systems, Kyushu University)
Major strategy for development of high performance polymer electrolyte membranes (PEMs) is to form PEMs with phase-separation structure in nano size like Nafion, and many studies have been reported on hydrocarbon-type PEMs with phase-separation structure. However, most of phase-separated PEMs were consisted of synthesized polymers with well-defined molecular structures, causing time-consuming synthesis process. Therefore, development of simple and easy preparation method for high performance PEMs is strongly required for generalization of PEFC. In order to solve this technical problems, we introduced new concept to develop high performance PEMs.

As new and unique PEM preparation method, we have reported charge-transfer (CT) complex method1–3. We reported that properties of polymers with electron-accepting unit could be controlled by adding electron-donating small molecules. Although CT complex method have some advantages, there is a limitation to control polymer property only using polymer-small molecule system. In order to expand applicability of CT complex method, CT complex hybrid membranes consisted of acceptor polymer and donor polymer were developed in this study.

One of the interesting advantages of CT complex in the membrane is that molecular geometry and assembled molecular structure in CT complex membranes can be designed because donor molecules always form CT complex with acceptor molecules by the expected molecular geometry. Using this self-assembling property of CT complex, we can design molecular geometry of PEM which is able to be spontaneously formed well-defined proton pathway in molecular level in the blend membrane. In order to confirm to prepare these well-designed polymer blend membranes, the polymer blend membrane with CT complex have been studied (Figure 1). First, we prepared CT complex blend membrane which was consisted of electron-accepting sulfonated polyimide (SPI) and electron-donating polyether (PE). In this study, we used a model PE which did not have a proton conductive unit in the molecular structure in order to evaluate about formation of the blend membrane with CT complex.

SPI was provided from Nissan Chemical Industries, Ltd. and PE was synthesized from dihydroxynaphthalene and epichlorohydrin. The reason why epichlorohydrin was selected as a monomer for PE was introduction of flexible spacer in donor polymer. As a control donor polymer, aromatic polyether (A-PE) which did not have flexiblity, was also prepared. CT complex blend membrane with SPI and PE have been prepared by simple solvent cast method. The obtained SPI/PE blend membranes showed dark brown color which was typical color of CT complex. On the other hand, the blend membrane consisted of SPI and A-PE did not show dark brown, indicating that SPI and non-flexible A-PE could not be formed CT complex. Only difference between PE and A-PE was main chain flexibility of polymer. Therefore, proper CT complex formation in polymer blend membrane would be required the certain flexibility to form CT complex effectively. In this paper, non-sulfonated PE was used as a model donor polymer. However, SPI/PE CT membrane was used for general property experiments for PEFC because SPI had sulfonic acid group as a proton conductor.

We evaluated mechanical strength, hydrogen permeability and proton conductivity of SPI/PE CT membrane (SPI : PE = 1 : 1 (mol)). SPI/PE CT membranes showed about 2.0 and 16 times higher tensile stress than original SPI and Nafion 212, respectively. Hydrogen permeability through SPI/PE CT membrane was 1.9 and 5.4 times lower than that of original SPI and Nafion 212 membrane, respectively. And proton conductivity of SPI/PE membrane was 4.1 mS/cm at 80 ˚C and 90 RH%, while that of Nafion 212 was 57 mS/cm.

Although proton conductivity of SPI/PE membrane was lower than Nafion 212, SPI/PE membrane showed higher mechanical strength and lower hydrogen permeability than Nafion. Applied these properties of SPI/PE membrane, we prepared thinner SPI/PE CT membrane (10 mm thickness) than Nafion 212 (50 mm). Thinner SPI/PE CT membrane was expected to reduce resistance of the blend membrane, allowing to use for PEFC test (Figure 1). From the result of fuel cell performance test of thin SPI/PE membrane, the thin membrane showed slightly lower OCV than Nafion 212 and similar resistance compared to Nafion 212, suggesting that thin SPI/PE membrane could be applied for PEFC application. In addition, SPI/PE membrane also showed more than 10-hour durability in fuel cell test.


1. R. Watari, M. Nishihara, H. Tajiri, H. Otsuka, and A. Takahara, Polym. J., 45, 839–844 (2012).

2. L. Christiani, S. Hilaire, K. Sasaki, and M. Nishihara, J. Polym. Sci. Part A Polym. Chem., 52, 2991–2997 (2014).

3. L. Christiani, K. Sasaki, and M. Nishihara, Macromol. Chem. Phys., 217, 654–663 (2016).