Modeling of Lithium-Flow Batteries with Aqueous Electrolyte

Abstract

Flow batteries with aqueous electrolyte at the cathode are safe, low-cost and ecofriendly energy storage systems. When the aqueous cathode is coupled with a Li-metal anode with organic electrolyte, these batteries have extremely high energy densities (over 4,000 Wh/Kg), which makes them attractive not only for grid applications (where flow batteries are often used) but also in maritime applications such long endurance surface vessel systems and wave gliders [1-4]. In this presentation, we derive the mathematical model for lithium-flow batteries and use this model to predict the performance of the batteries under different discharge conditions, cathode microstructures, water flow velocities, etc. The results are then compared to experimental data already available in the literature. The mathematical model is based on coupling the flow equations (Navier-Stokes) with electrochemical equations for reaction rates and using average microscopic models for the cathode. For instance, Fig. 1 shows the voltage of a flow battery as a function of time for different water velocities in a flow battery simulated by solving the above model. For low water velocities, the water is not able to provide enough oxygen in the cathode and the voltage of the battery collapses after a few seconds. However, as long as the water velocity is above approx. 1.2 cm/s, the battery will have enough oxygen to operate continuously; in this case, the higher the flow of the water, the higher the voltage of the battery in the stationary state will be because of lower overpotential losses. In these simulations, the anode contained Li-metal. The cathode had a length of 1 cm (measured in the direction of water flow) and a thickness of 100 μm. The cell was discharged at a small rate of 0.01 mA/cm2. The incoming water had an oxygen concentration of 5 mg/l (which is equal to the oxygen concentration in seawater) and was pumped through the cathode at different velocities. A detailed derivation of the mathematical model and more simulation results and predictions will be presented at the meeting.