Long charging times of commercial lithium-ion batteries (LIBs) constitute one of the critical remaining bottlenecks in the widespread deployment of electric vehicles (EVs). Therefore, there is a global push towards extreme fast charging (XFC)1in 10–15 minutes for 80% of the pack capacity. However, existing LIBs cannot achieve these XFC goals without significant capacity fade during cycling over battery lifetime.2 One of the key XFC degradation mechanisms is “Li plating,” which occurs when Li metal forms on the graphite electrode due the liquid electrolyte and solid-phase (graphite) mass transport limitations and sluggish charge-transfer kinetics of the Li intercalation reaction.3 This is mostly attributed to the formation of dead Li,4 which is the irreversibly plated Li that has become electronically disconnected from the graphite electrode. Since dead Li constitutes most of the XFC-related capacity loss, it is extremely important to unravel the link between its morphological evolution/progression, and battery capacity fade during extended XFC cycling.In this work, we report non-destructive, in-situ three-dimensional (3D) characterization of dead Li on graphite electrodes in full-cell LIBs during extended XFC using high-resolution (pixel size: ~ 7.8 µm; effective spatial resolution: ~25-30 µm) neutron micro-computed tomography (µCT). Specifically, we imaged batteries in the discharged state at the following cycle numbers at 6C charging rate: 10, 25, 50, and 150 cycles. Neutron µCT is promising for in-situ 3D Li detection on graphite electrodes because it provides strong contrast between Li and C (~ 71.87: 5.55 barns5). To leverage high-resolution neutron µCT for this work, we designed a custom neutron-friendly LIB coin-cell that provided exceptional neutron transmission at the graphite-separator-electrolyte interface as well as standard XFC performance at 6C between 3.0 and 4.1 V. We performed these imaging experiments at the MARS (CG-1D) beamline6 at the High Flux Isotope Reactor, Oak Ridge National Laboratory.Our results reveal 3D evolution of dead Li morphology from isolated plated deposits on only one edge of the graphite electrode after 25 cycles at 6C charging, to increasingly mossy-like covering the entire circumference of the graphite electrode after 150 cycles. Based on our imaging data, we did not observe any dead Li after only 10 XFC cycles. However, we noticed that dead Li started to nucleate around a partial edge of the graphite electrode after 25 XFC cycles, and then grew to the other edges after 50 XFC cycles, eventually forming a complete ring after 150 XFC cycles. At 150 XFC cycles, we also observed the dead Li ring outgrowths away from the edges of the graphite electrode onto the separator. In the extended XFC-cycled samples, 50 and 150 XFC cycles, we found significant dried out salt deposits on the separator, spacers, spring, and casing, which may be due to other parasitic reactions occurring at increased XFC cycling. Lastly, we determined quantitative relationships between the amount of dead Li in 3D, XFC cycling number, and XFC-related capacity loss of the battery.In conclusion, this work contributed to the advanced understanding of the 3D progression of dead Li and its relationship with XFC-related capacity loss on graphite electrodes during extended XFC cycling. We believe insights from this work will help develop improved 3D graphite electrode architectures that will minimize Li plating during repeated XFC cycling in fast-charged LIBs.
Read full abstract