Accelerated Aging of Lithium-Ion Batteries: Insights from Nonlinear Frequency Response Analysis

Abstract

Lithium-ion batteries (LiBs) are considered the leading contender in facilitating the global shift from fossil-fuel based to electric mobility. Accurately evaluating and monitoring the state-of-health (SoH) of these batteries is essential to ensure safe operation, including the early identification of accelerated aging. In this work, accelerated aging is defined as the step drop in the capacity fade, which usually occurs at the late ageing stage. This ageing behaviour is detrimental and is influenced by various factors such as charging and discharging rate, depth-of-discharge (DoD), cycling range, and operating temperature.Electrochemical impedance spectroscopy (EIS) is a common method of aging characterization in LiBs. Complementary to EIS, nonlinear frequency response analysis (NFRA) has proven useful in aging studies, such as detecting lithium plating [1] and predicting SoH [2,3]. Via NFRA, a cell is subjected to sinusoidal signals of high current amplitudes, and the response voltage signal is analyzed [4]. By analyzing the nonlinear response of the cell, we can obtain information about processes that are not easily observable using conventional EIS. Here, we propose the combination of both frequency response analyses to study the aging behaviour in LiBs.This study investigates the impact of DoD and cycling range on commercial 18650 cells with NMC as cathode and Si-C as anode. The aging test was performed at 45 °C with 1.33C constant current – constant voltage (cut-off current at C/10) charging and discharge profile. At defined cycling intervals, we captured the instantaneous cell behavior (discharge curve, EIS, NFRA) at 23 °C. We observed larger capacity fade for cells cycled with larger DoD. For cells with the same DoD, SoH fade was more significant for cells cycled at a higher state-of-charge (SoC) range. Via frequency response analysis, we identified a mutual correlation between the kinetic losses of the charge transfer process and SoH fade, which is similar to our prior study [2]. In addition, we found out that the onset of accelerated aging can be pinpointed by the increase in ohmic resistance in EIS. Notably, analyzing the nonlinear response (second Y2 and third Y3 harmonic) of the cells showed that Y3/Y2 can capture the different aging stages (see Figure 1), indicating that not only the nonlinear behaviour but also the nature of the charge transfer process itself has constantly changed during ageing. This compelling finding highlights the potential of NFRA for low-cost offline SoH diagnosis of LiBs.