A comparative study of impedance based modelling of lithium-ion batteries on half and full cell level

Summary:

The dynamic performance of lithium-ion cells is strongly dependent of the underlying loss processes such as contact resistances, charge transfer and ion diffusion. Electrochemical impedance spectroscopy (EIS) is a well-known characterisation method for quantifying the complex impedance of electrochemical systems such as lithium-ion cells. In combination with an equivalent circuit model, the EIS offers the possibility to determine the different loss processes as a function of temperature and state-of-charge (SoC).
In commercial cells, only the full cell behaviour can be studied and characterised. Therefore, no statement about the limiting electrode in certain operating situations, e.g. fast charging at different temperatures, is possible. Experimental cells on the other hand offer the opportunity to examine the two electrodes separately. They are therefore a tool to determine the limiting electrode in various situations and to predict new operating conditions of lithium-ion cells.
In this work a comparative study of impedance based modelling of experimental cells on half and full cell level is performed. The time-dependent behaviour of the lithium-ion cell is analysed, whereby different simulation approaches are used.
In the first approach, experimental cells in full-cell configuration are constructed and characterised. The data obtained is used to parameterise an impedance model and to simulate and validate the cell performance in the time domain. In another approach, the individual electrodes of the half cells are investigated in full-cell configuration in the experimental cells. The simulated overall cell performance results from the sum of the two half cells. In the third approach, experimental cells in half cell configurations are characterised and will also be used to simulate the cell performance.
The impedances of the two full-cell configurations are almost identical, there is only a slightly lower ohmic resistance in case of the addition of the both half-cells. The summation of both electrodes in the half cell configurations on the other hand, show higher ohmic resistances as well as higher polarisation resistances and therefore high deviations from the direct measured full-cell impedance.
The resulting simulated full-cell behaviour shows for both full-cell modelling approaches a very similar behaviour. The deviation to the validation measurement is for current pulse profiles less than for constant current discharging. However, the half-cell modelling approach shows an opposite behaviour. The simulation of the constant current discharging results in a better agreement with the experimental validation than for the dynamic current profile.