Design of experiments and modelling of the direct methanol fuel cell

Abstract

Environmentally friendly polymer electrolyte membrane fuel cells (PEMFCs) have the potential to revolutionise mobile power sources. One of the more promising PEMFC candidates is the direct methanol fuel cell (DMFC). Significant commercial interest has been expressed in the DMFC as a consequence of it becoming a possible replacement technology for batteries and internal combustion engines. The DMFC is a simple system that utilises liquid fuel and which requires minimal ancil lary equipment, and hence are more suited to the logistics of portable and vehicular applications than hydrogen fuel cells. However, significant technological challenges remain that must be addressed prior to the DMFC becoming more commercially exploitable. These challenges include improving the poor anode kinetics of methanol oxidation and reducing methanol crossover. To aid the understanding of the various factors limiting the widespread application of the DMFC, the statistical method of design of experiments was applied. A fractional factorial design was implemented to understand the main effects and interactions of a number of operating parameters on the overall performance of the DMFC, in which the effect of the crossover of methanol through the membrane was considered. The statistical models developed facilitated the detection of key two-factor interactions of temperature with methanol concentration, type of oxidant and cathode back pressure, which suggested that an improvement in DMFC performance was achievable by reducing the effect of methanol crossover. Based on the outcomes of the parametric study, response surface methodology was applied to optimise catalyst layer formulation. The response surface method highlighted the significance of high catalyst loading and the non-linear behaviour of the Nafion@ content. Furthermore, the advantage of adding PTFE in the anode catalyst formulation, to make the anode morphology favourable for carbon dioxide gas evolution, was demonstrated. Steady state semi empirical models for the anode based on methanol oxidation kinetics and cathode considering the effect of methanol crossover through the membrane were also developed. The kinetic models for the anode illustrated the significance of water and surface intermediates in the methanol oxidation reaction on a dual site Pt-Ru catalyst and highlighted the subtle balance between the methanol adsorption-dehydrogenation step and the subsequent oxidative removal step. The cathode model developed provided insight into the effect of methanol crossover on the cathode open circuit potential and helped in reliable estimation of the cathode polarisation curve. Finally a combination of these two models was used in the prediction of the cell polarisation characteristic as a function of cell potential, temperature and amount of methanol crossed over through the membrane