Physical modelling of DMFC performance heterogeneities and the recovery of reversible cathode degradation

Abstract

Direct methanol fuel cells (DMFC), which are an alternative power source to batteries and diesel engines, exhibit a great potential for a locally heterogeneous cell performance. The DMFC anode is fed with a liquid methanol-water mixture while the cathode is supplied with air, which results in an even more complex fluid management in comparison with the structurally similar polymer electrolyte fuel cell (PEMFC) operated with hydrogen as fuel. The transfer of water and methanol from the anode through the polymer electrolyte membrane to the cathode side is an important factor for limits in the cell performance. The crossover of water from the anode to the cathode side, where water is also produced in the electrochemical reaction, increases the risk of liquid accumulation in the porous layers of the cathode and thus mass transport limitations in the cell. Methanol crossover leads to the formation of a mixed potential in the DMFC cathode, and the resulting high overpotential increase the development of oxide species on the platinum catalyst surface. These processes lead to a reduction of the cell performance, which is partially reversible. In this work, a physics-based DMFC model in 2D is developed in order to study the local cell performance along the channel with a focus on the two-phase flow as well as humidity-related properties of the ionomer. The model features a spatial resolution of the catalyst layers, which enables the examination of the local conditions' impact on the electrochemical reactions and on effects at the membrane interface. The model is verified against experimental data from a macro-segmented DMFC single cell for two different humidity levels in the cathode. The validation not only comprises the local cell performance, but also mass transport and the ohmic resistance of the membrane. Simulation results for the cell performance under varying operating conditions are shown in comparison with corresponding experimental data, proving the predictiveness of the model. The transient model is further used to study the processes inside the cell during the recovery of reversible degradation effects in the cathode. The formation of platinum oxide species during DMFC operation and their reduction during a refresh sequence including an OCV phase as well as a phase with air starvation is simulated and explored with respect to the local conditions inside the cathode catalyst layer. Moreover, the simultaneously occurring spontaneous evolution of hydrogen in the DMFC anode is examined. Several variations of the air stop sequence are simulated and evaluated with regard to their effectiveness in recovering the temporary performance losses within the DMFC cathode.