Local variations of mechanical, structural, transport, and kinetic properties, referred to as heterogeneity (generated by a non-optimized manufacturing process [1]), can detrimentally affect battery life and performance, especially at high cycling rates and low temperatures, in a number of manners. The distribution of current density across an electrode surface can be non-uniform as a consequence of electrode heterogeneity. A high local charging rate (especially at local low temperature) may lead to local lithium plating, resulting in capacity fade and shorter cycle life. Additionally, deviation of local charging-discharging rate from the average rate results in variation in local state of charge (SOC, Li concentration in active material). For a given chemistry and application, there is an SOC operation range (lower and upper bounds) beyond which capacity loss and damage occur. Therefore, highly heterogeneous electrodes may cause “hot” and “cold” spots, in terms of SOC, leading to non-uniform aging within the cell. Another negative effect of local inhomogeneity is non-uniform utilization of active materials, resulting is decreased volumetric and gravimetric energy density of a Li-ion battery.

In this work, we introduce a computationally efficient Newman-type model to further understand and quantify the effects of local inhomogeneities, particularly non-uniform ionic and electronic impedance. As shown in Fig. 1, we imitate the real heterogeneity of the cell by a system consisting of three localized Li-ion domains, each with specific properties ion-transport associated with “cold” (high tortuosity and low porosity [2]), “middle” (moderate tortuosity and porosity), and “hot” (low tortuosity and high porosity) spots. The properties for each of the domains are obtained experimentally through a combination of SEM/FIB and localized conductivity measurements. The three simulated domains are treated as separate cells electrically connected in parallel. This approach can simplify mathematical treatment and significantly reduces intensive computation requirements needed for true 3D modeling.

[1] M.M. Forouzan, C.-W. Chao, D. Bustamante, B.A. Mazzeo, D.R. Wheeler, Journal of Power Sources, 312 (2016) 172-183.

[2] M.M. Forouzan, M. Wray, L. Robertson, D.R. Wheeler, J. Electrochem. Soc., 164 (2017) A3117-A3130.

Figure 1: Heterogeneity of the NMC electrode using SEM/FIB images and equivalent circuit used in the Newman-type model.