(Invited) Optimization of Highly Ordered Laser-Patterned Electrode (HOLE) Design for Achieving Enhanced Fast Charging Performance in Graphite Anodes with > 3mAh/cm2 loading

Abstract

For widespread adoption of electric vehicles, we need Li-ion batteries (LIBs) that are both energy and power dense. In the state-of-art Li-ion batteries, there exists a tradeoff between the energy and power density. This tradeoff originates due to the electrode design used in conventional LIBs, where one needs to increase the loading of active material (either in terms of active material mass fraction or the electrode thickness) to achieve high energy density. However, such a design limits the mass transport of electrolyte, which leads to it poor performance at high-charging rates. It has been demonstrated that the rate performance of the energy dense electrodes can be improved by employing 3D architectures such as highly ordered laser-patterned electrode (HOLE),1 which alleviate the electrolyte mass transport limitations by providing rapid mass transport via laser ablated channels through the electrode thickness. In this study, we investigate how the geometric parameters of the HOLE design such as inter-channel spacing and channel radius affect the fast-charging performance of graphite anodes with > 3 mAh/cm2 loading and the HOLE architecture. We conduct this analysis using a fully parameterized continuum scale model based on the porous electrode theory. Our results show that for a constant volume retained (after the laser ablation), the smaller and closer channels exhibit better 4C charging performance than the channels that are larger and farther. The improved performance of a HOLE anode with smaller and closer channels is a result of the small value of the Damköhler number, Da, defined as the ratio between the characteristic diffusion time and reaction time, for the anode. Furthermore, we developed a semi-analytical framework to estimate Da for a given HOLE design, which includes the channel spacing, size, and arrangement. The results from the analysis match closely with the results from the 3D simulations. Thus, our framework can be used for optimization architectures similar to the HOLE architectures without the need of running computationally expensive 3D simulations. In addition to our simulation results, we will also discuss the automated process that was developed to parameterize our model.