Virtual design of structured battery electrodes

点击翻译按钮获取中文摘要

To improve the performance of Li-ion batteries, a metallic network of very thin fibers was produced at the Max Planck Institute for Intelligent Systems (AK Spatz). These were successfully introduced into the active material layers of cells. First measurements on prototype cells with a metallic, fiber-like 3D network could show that the performance of the electrodes could be significantly improved. The capacity of the electrode with fiber material was improved by 25% (in the 50th charge/discharge cycle) compared to a reference electrode without fibers with the same active material loading. A micro-model was developed to simulate the battery cells. The spatially resolved geometry of the porous electrode structure and the embedded fibers required for this can be measured or virtually generated. Due to the high computational cost, the micro-model is not suitable for a virtual cell design and for extensive parameter studies. For this reason, it is imperative to derive a dimensionally reduced model that takes into account the influence of the microstructure on the process with less computing time compared to the detailed model. In the AK Nieken there is a lot of experience with the method of asymptotic homogenization (AH). This makes it possible to derive effective rates of transport processes from a detailed geometry model. Thus, it is possible to build a macroscopic process model for structured porous materials that intrinsically accounts for the microstructure in a rigorous manner. The challenge for the macroscopic process model is to extend the range of validity to an operation with high C-rates. At high C rates, large concentration gradients form on the microscopic scale, so that (sufficient) scale separation is not possible. This is a well-known problem in the application of AH. First approaches to solve this problem are formulated in the proposal. Based on these approaches, it should be possible to predict the electrochemical behavior of the batteries developed by AK Spatz at technically relevant C rates to a good approximation. The aim of the project is therefore to derive an effective, macroscale process model that is suitable for carrying out parameter studies and structure designs. For this purpose, the existing process models are to be extended by the influence of the metal fiber network and methodological improvements are to be developed for large local gradients.