Compression of packed particulate systems: Simulations and experiments in graphitic Li-ion anodes

Increased thermal conductivity, electronic conductivity,, and reversible capacity (i.e., reduced irreversible capacity loss, or ICL) have been demonstrably achievable by compression of anodes into higher volume fraction plates, though excessive compression can impair Li-ion battery performance. In our previous study, we correlated conductivity and compression of these materials. Here, we further investigated the effects of friction and deformability of particles on the compressibility of model carbons of Li-ion anodes. First, we implemented a statistically unbiased technique for generating a range of random particulate systems, from permeable to impermeable arrangements, along with a contact model for randomly arranged triaxial ellipsoidal particles, suitable for implementation in finite element analysis of compression of a random, porous system. We then quantified the relationship between interfacial friction and jamming fraction in spherical to ellipsoidal systems and applied these models to correlate maximum stresses and different frictional coefficients, with morphology (obtained by image analysis) of graphite particles in Li-ion anodes. The simulated results were compared with the experiments, showing that the friction coefficient in the system is close to 0.1 and that the applied pressure above 200 kg/cm(2) (200 MPa) can damage the materials in SL-20 electrodes. We also conclude that use of maximum jamming fractions to assess likely configuration of mixtures is unrealistic, at best, in real manufacturing processes. Particles change both their overall shapes and relative orientations during deformation sufficient to alter the composite properties: indeed, it is alteration of properties that motivates post-processing at all. Thus, consideration of material properties, or their estimation post,facto, using inverse techniques, is clearly merited in composites having volume fractions of particles near percolation onset.