The ability to fast-charge Lithium-Ion batteries is a key requirement for the success of electric vehicles. One main problem when fast-charging Lithium-Ion batteries is Lithium-plating. There, the Lithium-Ions are not intercalated into the anode but are deposited on its surface. This leads not only to safety issues, but also reduces the capacity of the battery by reducing the amount of cyclable Lithium. While in experimental half cells the onset of Lithium-plating can be well correlated with the Anode potential crossing a value of 0 V, in commercial cells the Anode potential cannot be measured. Thus, extensive aging experiments are required to determine the optimal charging current profiles for a given battery cell, which charge the cell fast but do not cause Lithium-plating. As these profiles vary strongly with temperature and State-of-Charge (SOC), the effort for doing such an investigation is extremely high and not practicable.
Instead, we present an approach, where the anode potential of the cell Samsung INR21700-40T is predicted by a physical battery model. The model contains sub-models for both anode and cathode and covers all relevant loss processes. It is shown to be valid in the whole operational range of the cell and thus delivers a valid estimate for the anode potential. Using this model, we will show how to derive the fast-charging current profile for the Samsung INR21700-40T cell. The derived fast-charging current profile was applied to the cell for 200 times in our lab and the capacity loss over time was tracked. It will be shown that the aging was comparable to the aging during 200 cycles of slow charging, which validates the approach.