Looking Deep inside the Cathode of Li-O2 Batteries: Unraveling the Local Distribution of Li2O2 with a Combined Experimental and Model-Based Approach

Abstract

Carbon-based, high surface area gas diffusion electrodes (GDEs) are commonly used as cathodes in Li-O2 batteries. These GDEs provide an sufficient amount of active sites to deposit the main discharge product, Li2O2. Despite the fast electron transfer at the electrode surface to form Li2O2, experimental results show that the very high theoretical energy densities of Li-O2 batteries cannot be achieved at the moment. Instead of the desired formation of high amounts of soluble LiO2, predominantly insoluble Li2O2 precipitates at the active sites of the cathode during discharge. Unfortunately Li2O2 is poorly conductive for electrons, which leads to a low amount of deposited Li2O2 per active site and a fast passivation of the cathode. Thus, only the cathode surface area and not the pore volume is utilized by Li2O2, which explains the discrepancy between theory and experiments. Furthermore, the precipitation leads to continuous changes in porosity, available active surface area and predominant reaction pathway. For this reason a detailed analysis of the distribution of Li2O2 at all points in the GDE and of the factors that influence the Li2O2 morphology is needed to overcome performance limitations. [1]In this work, we elucidate how the local distribution of Li2O2 inside the GDE and its particle size evolve as a function of discharge current density. We apply a powerful combination of experimental and model-based analysis. To the best of our knowledge, this is the first study on the cathode surface utilization by Li2O2 and its particle sizes performed for different locations and states of discharge (SOD). The impact of the distribution and the resulting changes in transport resistance and active surface area on the battery performance are derived thereof.In the experimental part, decreasing capacities and Li2O2 particle sizes are observed for increasing discharge current densities (see SEM analysis of discharged battery cathodes in fig. 1 a)). Furthermore, the experimental results show that particle precipitation starts mainly at the side of the cathode that faces the oxygen reservoir of the battery (see fig. 1 b)). Discharging a battery at low current densities leads to a uniform cathode surface coverage by Li2O2 even at low SOD whereas discharging at high current densities yields a gradient of blocked active sites through the cathode.Simulations based on a physical GDE model show that a two-step reaction mechanism with the soluble species LiO2 as reaction intermediate can quantitatively explain the experimental findings. Depending on the current density, either a chemical or an electrochemical particle growth process predominates, which leads to the distinct different particle size distributions (see fig. 1 c) and d)). The strong dependence of capacity on the current density is related to different capacities per carbon cathode surface area at the end of discharge. At high discharge rates and thereby lowered potential more nucleation takes place. As a result, the particle number is higher and the average size smaller. Due to the surface volume ratio, smaller particles entail lower capacities per cathode surface area. The results point out that even at moderate current densities the battery capacity is limited by the surface area of the GDE and not by the concentration of dissolved O2 in the liquid electrolyte. Therefore, the solubility and reaction kinetics of the reaction intermediate LiO2 play a crucial role to enhance Li2O2 particle size and with it the obtainable discharge capacity.The presented results provide detailed insight into the cathode surface utilization and the underlying processes that limit the performance of GDEs in Li-O2 batteries. In the end, the study will help to achieve higher discharge capacities for this type of battery and thus will propel the ongoing research and the efforts in commercialization of metal-oxygen batteries.Acknowledgement:The authors are thankful for financial support by the BMBF (Federal Ministry of Education and Research) within the project “MeLuBatt” (03XP0110A).