Analysis of water transport phenomena in thin porous media of a polymer electrolyte membrane fuel cell

Abstract

This thesis explores and quantifies water transport related to the desaturation of the thin porous layer known as the Gas Diffusion Layer (GDL) associated with Polymer Electrolyte Membrane (PEM) fuel cells. The proper management of water within this layer is critical to optimal fuel cell performance. If there is not enough water, the membrane can become dehydrated, which leads to poor cell performance, but if too much water accumulates or becomes flooded, gas transport is restricted, which also lowers performance and can potentially lead to total cell failure. Understanding the desaturation of this layer is thus key to obtaining and maintaining optimal fuel cell performance. This behavior is explored both at the macroscale, through the quantification of the removal of excess water from an active fuel cell, as well as at the micro-scale, through the use of synchrotron X-ray computed tomography (X-ray CT) to visualize and quantify the desaturation of an initially flooded GDL. The macro-scale investigation extends the previously developed qualitative Anode Water Removal (AWR) test, which functions to identify when poor PEM fuel cell performance is due to excess water, to a diagnostic protocol that quantifies the amount of water being removed by the test through an analysis of the anode pressure drop. Results show that the protocol can be applied to a variety of fuel cell setups and can be used to quickly quantify water management capabilities of novel GDL materials. The microscale investigations show that while both convection and evaporation play a role in the desaturation, evaporation is required to fully desaturate the GDL. Additionally, the microscale investigation allows for the spatial segmentation of the GDL to identify local desaturation rates and temporal saturation profiles, which show that the overall desaturation of the GDL is a heterogeneous process that depends on initial conditions, flow field geometry and the natural anisotropy of the material. Results show that future control strategies and modeling studies will need to expand their investigated domains in order to accurately capture the fully heterogeneous nature of this process.