Reaction Kinetics-Based Modeling and Parameter Sensitivity Analysis of Direct Ethanol Fuel Cells

Abstract

Ethanol is considered an alternative fuel to power fuel cells, especially due to its ease of transport and storage and renewable production on a large scale. However, its use in direct ethanol fuel cells (DEFC) is still limited by incomplete electro-oxidation and slow reaction kinetics. Modeling approaches have focused on investigating different reaction mechanisms, but so far, no formal analysis of model parameter sensitivity has been conducted. This work modeled and identified sensitive parameters for different types of Pt–Sn catalysts previously prepared by our research group that displayed good performance in the 5–15 mW/cm2 range (relative to a performance of 12 mW/cm2 achieved by a commercial ETEK catalyst). Analyses to study the effect of these parameters on coverage fraction distribution, reaction rates and possible correlations were also performed. The model was developed based on Butler–Volmer kinetics and on a reaction mechanism previously reported in the literature. Statistical developments were considered to compute parameter uncertainties for a non-linear system with non-linear restrictions. The model achieved very good fits to experimental data, with low RMSE values between 0.22 × 10−4 and 4.2 × 10−4 A/cm2, while also showing surface coverage fraction distributions in agreement with other experiment-based works from the literature. All catalysts taken into account, the most sensitive parameters were the reaction rate constants associated with the formation of adsorbed CH3CO, and the direct and reverse water dissociative adsorption reactions, respectively. Additional analyses suggested that there is not much correlation between the parameters. The results from this work could contribute to the state-of-the-art DEFC models by providing insights into which variables may be assumed constant or which ones have the greatest impact on the model output, thus helping to reduce the model size and computational time for future broader DEFC models.