Predicting Transport, Mechanical, and Microstructural Properties of Porous Li-Ion Battery Electrodes By a Particle-Based Simulation

Predicting Transport, Mechanical, and Microstructural Properties of Porous Li-Ion Battery Electrodes by a Particle-Based Simulation

Transport properties and performance of porous electrodes in Li-ion batteries are highly affected by microstructure, and in turn, the microstructure is influenced by the fabrication process. Therefore, a detailed knowledge of the relationships between transport properties, microstructure, and fabrication process variables is essential for battery researchers and manufacturers in order to model and design optimized batteries.

Factors that influence electrode microstructure include high-level variables such as composition and porosity, as well as detailed fabrication conditions. The fabrication process for Li-ion electrode films includes mixing a slurry of carbon, binder, solvent, and active material; coating the slurry onto a metallic current collector; drying the film; and calendering to the desired porosity. The resulting microstructure of commercially made battery electrodes is not necessarily optimal, and improvements could be made with a detailed physical understanding of each step.

In this work experiments and computer simulations are performed in order to elucidate fabrication-microstructure-performance relationships. A slurry made with a representative composition, namely with spherical active material particles (Toda NCM 523), carbon black, polymeric binder, and NMP solvent, was made in our laboratory. Slurry viscosity measurements at different shear rates, shrinkage ratio during the drying process, and dried electrode film elasticity (Young Modulus) were measured and used to parameterize the model.

 LAMMPS, a molecular simulation code, was adapted for the mesoscale particle simulations. Shifted force Lennard-Jones and granular Hertzian potential were used to represent the interaction between particles. Relatively soft particles were used to represent implicit solvent, carbon, and binder aggregates. Relatively hard particles were used to represent active material. Equations of motion coupled to inter-particle forces are solved to simulate particle motion and subsequent immobilization during drying steps.   

The microstructure of the dried film was validated by comparing to experimental results, namely FIB/SEM sequential cross-sections of real composite electrodes. Image processing algorithms were used to segment the images into three phases also used in the model: active material, carbon/binder domains, and macroscopic pores. In addition, effective electronic and ionic conductivities of the model structures were compared to experimental values.

This work is supported by the U.S. Department of Energy through the BMR program.

Figure SEQ Figure \* ARABIC 1: Microstructure comparison: (left) liquid simulation, (center) dried film simulation, (right) dried film experimental. Blue indicates active material,, green indicates additives (and solvent), white indicates macro-size pores.

.