Knowledge of the internal state of lithium-ion batteries is crucial for the development of safe charging control algorithms. Most commonly, graphite is used as the negative electrode, possessing a high specific capacity and cycling stability. When such lithium-ion batteries are charged, lithium intercalates into the layered carbon structure, and the electrode material transitions through a series of phases differentiated by the number of carbon layers between each lithium layer. Each phase has different electrochemical properties, making it interesting to accurately track the phase content during lithiation. However, this quantity can only be measured through spectroscopic experiments, which cannot be included in any battery application. In this work, a method based on incremental capacity analysis and kernel smoothing is introduced to estimate the phase content of graphite electrodes from the electrode potential during constant-current charging. This kernel density function (KDF) method is validated at low currents using physics-based simulations of multiphase electrode dynamics, achieving an average phase estimation error below one percentage point per phase. Furthermore, we apply the KDF-method to experimentally measured coin cell data, for which the estimated phase content closely agree with simulations in the mid to high range state of charge. • ICA knee-points reliably map to graphite phase transitions at low current. • Kernel smoothing bridges low- and higher-current phase estimation. • The kernel standard deviation indicates the degree of lithiation inhomogeneity. • A first step towards BMS-compatible alternatives to XRD-based phase-tracking.
Borghed et al. (Fri,) studied this question.