Future expectations on lithium ion batteries are, among others, increasing energy density, power density and lifetime. Because the battery performance is ruled by various microstructure characteristics, we need computational models considering this complex 3D microstructure.
In our model three-dimensional model geometry can be based on reconstructed microstructures obtained by FIB-SEM tomography or on detailed virtually generated microstructures. The poster considers the latter, generated by a newly developed microstructure model.
To avoid singularities during the model calculation due to non-percolated volumes, the microstructure is preprocessed by island removal operations. The resulting model geometry is meshed with the software Synopsys Simpleware ScanIP to a tetrahedral volume mesh using +FE freemeshing.
The 3D model with three spatially resolved phases i.e., the active material phase, the electrolyte, and the carbon/binder phase is based on the Doyle–Fuller–Newman model. The model is configured as a half‐cell of a lithium‐cobalt‐oxide (LCO) cathode against metallic lithium.
The model is applied to eight virtually generated cathode structures of 38.5 x 38.5 x 50.6 µm³ with different carbon black distributions but identical active material distributions. The carbon black contend is varied in four steps between 5% and 20%, for each of which the carbon black was first distributed homogenous and additionally depleted near the current collector. The latter aims to investigate the effect of a bottleneck of carbon black due to an inhomogeneous carbon black distribution. The results present the cathode structure at a SOC of 50% during a 3C discharge.
The shown three-dimensional distributions of the electrical overpotential in the active material reveal that the electrical conductivity clearly correlates with the carbon black distribution. However, more important than the amount of carbon black is its distribution. Areas of carbon black depletion lead to bottlenecks in electrical transport and to significant gradients in the electrical potential.
Additionally, the impact of (i) partially covered reaction surface inhibiting the charge transfer and (ii) reduced of the electrolyte space inhibiting electrolyte transport are investigated.
Since losses due to electron transport are decreasing with higher carbon black content whereas the losses due to ion transport and charge transfer reaction are increasing, a well-balanced carbon black content and carbon black distribution applies for each cathode structure and application.
Overall, the presented method is capable of investigating the influence of microstructural characteristics on the battery performance. In combination with detailed virtually generated microstructures, model-based optimisation enables the development of cathodes without the afford of processing real electrodes.