3D Morphological Evolution of Dead and Active Lithium at the Graphite-Separator-Electrolyte Interface during Fast Charging

Abstract

Lithium-ion batteries (LIBs) have profoundly advanced the development of electric vehicles (EV). However, one of the remaining bottlenecks in the widespread adoption of EVs is the long charging time required for commercial LIBs. There is a global push towards extreme fast charging (XFC) to reduce charging times to 10-15 minutes.1 However, XFC results in rapid degradation in the electrochemical performance of batteries, mainly due to Li plating. This plating refers to the loss of active Li, where it either becomes dead meaning electronically disconnected after plating on the anode, or it becomes inactive due to the irreversible reaction of Li with the electrolyte to form a solid electrolyte interphase. Dead Li contributes significantly to capacity loss during XFC;3 thus, differentiating dead and inactive from active Li is extremely important to unravel their link with battery capacity fade. However, owing to the complexity of these competing electrochemical reactions in addition to the intricacy of detecting Li at the buried graphite-separator-electrolyte interface, a comprehensive understanding of Li plating remains a challenge.In this work, we conducted non-destructive, in-situ three-dimensional (3D) characterization of dead and active Li on energy-dense graphite electrodes during XFC using high-resolution neutron micro-computed tomography (µCT). Thick or industry-relevant (negative and positive electrodes ~ 130 μm) graphite electrodes are promising battery materials because they store higher amounts of energy. However, they plate Li faster and at lower XFC cycle numbers, and hence experience higher loss in battery capacity due to Li plating.To characterize Li plating on thick graphite electrodes, we performed high-resolution neutron µCT4 (pixel size: ~ 5.74 µm; effective spatial resolution: ~10-15 µm) at the ICON beamline5 at the Swiss Spallation Neutron Source at the Paul Scherrer Institute. Neutron µCT was used because it provides high relative neutron contrast between Li and C (~ 71.87: 5.55 barns6) for in-situ 3D imaging. Therefore, to leverage high-resolution neutron µCT for in-situ Li detection, we designed a neutron-friendly LIB. Our designed battery enabled excellent neutron transmission at the graphite-separator-electrolyte interface while exhibiting similar XFC performance at a 6C charging-rate between 3.0 and 4.1V. We imaged three batteries in the charged and discharged states under the following conditions: 4 cycles at 1C charging; 4 and 6 cycles at 6C charging. Here, we note that since inactive Li was below the spatial resolution of the ICON beamline, only dead, active, and plated Li were investigated in this work.Our results reveal changes in dead Li morphologies from isolated deposits after 4 cycles at 1C charging to increasingly mossy-like dense deposits covering the entire circumference of the graphite electrode after 6 cycles at 6C charging. Furthermore, as the XFC cycle number increased from four to six at 6C charging, the amount of dead Li increased significantly. After 6 cycles at 6C charging, we also observed dead Li outgrowths in a circular geometry onto the separator around the graphite electrode. Using our imaging data, we determined a quantitative link between the amounts of dead and active Li with the capacity loss of the battery. Our analysis suggests that the amounts of plated and dead Li increase with the charging rate and cycle number. Finally, we developed quantitative relationships in 3D between the amounts of dead, active, and plated Li, and XFC-related capacity loss of the battery.In conclusion, this work contributed to the advanced understanding of the morphological evolution of dead, active, and plated Li and their relationship with XFC-related capacity loss on thick graphite electrodes. We believe these insights will help inform rational 3D designs of thick graphite electrodes that will minimize Li plating in fast-charged LIBs.