In the current researches intended to improve the Lithium-ion batteries capacity, enriching anode with silicon is one of the more considered solutions. Nevertheless, silicon shows a volume change of +280% in the fully lithiated state, leading to a significant « breathing » of electrodes during cycling that raises safety and performance issues. In this sense, prediction of the volume changes of the anode is key to design the cells. Irreversible volume changes are observed and conjectured to be related to microstructure changes, but current publications addressing the modelling aspects mainly use analytical or continuous models that can hardly capture the properties of granular media microstructure. Thus, this study aims to apply Discrete Element Method (DEM) on a silicon-based anode in order to take into account the discrete microstructure in the breathing behaviour.

The core of this work is to use the DEM software LIGGGHTS® on a sample of anode, modelled as a 3D distribution of spherical particles mechanically interacting with each other. The breathing amount during charge and discharge cycles is evaluated via evolutions of thickness, void volume, porosity and coordination number. The global approach follows a progressive enrichment of contact laws, and successive sensitivity analysis of each upgraded parameter evaluate their capacity to reproduce more finely the observed behaviours.

In the parameters studied so far, the pressure applied on the anode, the adhesion and the bonds between particles are three major parameters that affect the breathing of the anode. Indeed, considering the change of anode thickness, the pressure analysed from 0.01 to 100 MPa decreases the breathing range with a factor of 1.7. Moreover, the adhesion analysed from 0 to 100 J/m² is able to create an irreversibility at first cycle of around 30% of breathing range. Similarly, the bonds between particles analysed with a maximum breaking stress from 0 to 40 MPa can generate an irreversibility of more than 50% of the first breathing amplitude.