Parametric microstructure modeling of compressed cathode materials for Li-ion batteries

The microstructure of electrodes significantly influences the electrochemical performance of lithium-ion batteries. Thus, a deeper understanding of the electrode microstructure provides valuable information for the design of optimized electrode morphologies. One promising approach is called virtual materials testing, where stochastic microstructure models are used for generating a large number of virtual, but realistic morphologies, which serve as input for numerical transport simulations. Doing so, relationships between the microstructure and functional properties of the electrodes can be investigated. In the present paper, we utilize a parametric stochastic microstructure model, which is calibrated to tomographic image data of eight differently manufactured cathodes, where the compaction load has been varied. Since one and the same model type is used for all compaction loads, we can predict the model parameters for an arbitrary compaction load. This allows us to perform predictive simulations, i.e., we are able to generate virtual microstructures that correspond to a compaction load which has not been observed experimentally. The goodness of fit of the microstructure model is validated by comparing phase-based as well as particle-based characteristics of model realizations and tomographic image data. In addition, the suitability of the stochastic microstructure model for predictive simulations is pointed out by cross-validation.