| A computational approach is applied for modeling charging and discharging in lithium ion batteries. Lithium transport in electrode materials is typically limited by interface mobility and lithium ion diffusion in single phases. A novel multi-scale multi-physics computational approach is used to predict time evolution in electrode materials, particularly the kinetics of the phase boundary. Two different electrode materials are studied in detail, layered Li1+xV3O8 and spinel Li1+xTi 2O4. The phase behavior and kinetic pathways of Li1+x V3O8 are investigated by density functional theory (DFT) and the cluster expansion (CE) method. Because of strong electron correlation in d bands of vanadium cations, the DFT calculation produces incorrect formation energies. We find that the DFT+U method yields various material properties that are qualitatively in agreement with experiment. For spinel Li1+x Ti2O4, formation energies and migration energy barriers are calculated by the ab initio method, and then approximated by a cluster expansion and a local cluster expansion, respectively. Metropolis Monte Carlo and kinetic Monte Carlo were further employed to yield the thermodynamics and kinetics, particularly homogeneous free energy, anisotropic interfacial energies and lithium mobility, which are then used to parameterize a phase field model. This model is used to study the charging and discharging of electrode particles in the lithium ion battery macroscopically. |
