Nonlinear frequency response analysis on lithium-ion batteries: Process identification and differences between transient and steady-state behavior

Electrochemical Impedance Spectroscopy (EIS) is the most commonly applied method for electrochemical characterization of processes in batteries. Nevertheless, one of its limitations is the linear excitation of the system, denoting loss of information about the natural nonlinearities of the processes in batteries. Here, the potential of a nonlinear dynamic analysis method such as Nonlinear Frequency Response Analysis (NFRA) can be revealed. A reliable interpretation of the response signal of a LiB requires to understand its sensitivities to key parameters and processes. This work thus uses a Singleparticle battery model to systematically analyze the sensitivity of the response to input and cell parameters and to the evaluation time. Thereby, additional experimental evaluation reveals that an initially excited system shows higher nonlinearities than the same system in its steady-state excitation state; yet, the steady-state signal should provide for a more reliable interpretation, especially regarding experimental effects. Simulation thereby indicates similar time constants for the initial excitation and the EIS. NFRA is also shown to be an excellent method to study the nature of the Lithium-ion transport process across the SEI/electrolyte interface. In contrast to EIS, NFRA allows to discriminate whether the process is nonlinear or linear, as in most transport processes. Further, ageing effects caused by changes due to the SEI are shown to influence the spectrum in the low-, mid-and high-frequency domain and changes in the reaction and diffusion influence the mid to low and low frequency range, respectively. Individual higher harmonics show different sensitivities to the single nonlinear processes, which can be used for more in-depth studies of processes at its characteristic frequency range. The presented methodology to use NFRA promises to significantly advance electrochemical characterization and state estimation of Lithium-ion batteries. Combination of NFRA and EIS is further feasible to gain an even deeper characterization of processes than with the conventional EIS alone. (c) 2018 Elsevier Ltd. All rights reserved.