Synthesis, characterization and optimization of sulfonated polyether ether ketone composite membrane for direct methanol fuel cell

Abstract

The high costs of perfluorinated membranes have prompted research for alternative membranes based on hydrocarbon polymer and composite membrane. New composite membranes were prepared using sulfonated polyether ether ketone (SPEEK) polymer and inorganic proton conducting fillers developed from tungstosilicic acids (SiWA) loaded on silica-aluminium oxide (SiO2-Al2O3) composite. The SiWA was fixed on stable structure (composite oxide) so that it became insoluble in water. The SPEEK polymers were characterized using hydrogen nuclear magnetic resonance (H NMR) and fourier transformed infrared (FTIR). An X-ray diffraction (XRD) analysis was performed on the inorganic proton conducing fillers to reveal the existence of the interaction between SiWA and composite oxide. The membranes morphological structural were characterized using scanning electron microscope (SEM) and their performance were examined in terms of proton conductivity, water uptake and methanol permeability. The result showed that the presence of inorganic proton conducting fillers led to both high water uptake and proton conductivity (maximum value 6.1x10-2 Scm-1). Low methanol permeability values were recorded for the membranes which appeared as a very promising material to be used in direct methanol fuel cell (DMFC). In order to obtain the optimum membrane formulations that have optimum response, which are minimum water uptake, maximum proton conductivity and minimum methanol permeability; a central composite experimental design (CCD) combined with response surface methodology (RSM) was carried out. The factors considered were SiWA content (A) and SiO2 content in SiO2-Al2O3 (B) composite. The data collected were analyzed using Design Expert Version 6.0.8 (StatEase, USA). Two optimum membrane formulation generated were DS66/68.27SIWA/25SO/75AO and DS66/30SIWA/53.86SO/46.14AO. By performing the validation experiments, the models which were developed using CCD-RSM method appeared to be practically accurate and can be used for prediction within the range of the factors studied.