Study of electrode - electrolyte interface in lithium-ion battery using electrochemical imaging

Abstract

Lithium-ion batteries (LIB) have been widely used due to the proliferation of electric vehicle, energy storage system, intelligent robots, and biomedical devices. However, more extensive application of LIB might be limited if LIB requirements such as high-energy-density, high safety, and fast-charging materials are not met. Therefore, in order to develop the technology that can satisfy the requirements of LIB, we should use various advanced methodologies such as nanoscale electrochemical analysis, non-destructive interior analysis, and machine learning. In this study, the formation mechanism of the initial electrode-electrolyte interphase layer according to charge and discharge in the natural graphite composite electrode with Li-Si-Ti-P phosphate-based solid electrolyte was analyzed through electrochemical strain microscopy (ESM) and X-ray photoelectron spectroscopy (XPS). It was confirmed that the formation of the solid-electrolyte interphase (SEI) layer depends on the content of the solid electrolyte through the change in Coulombic efficiency during initial charging and discharging. This was further elucidated by ESM where the SEI layer first uniformly covered both the anode and the solid electrolyte resulting in high roughness with uniform lithium distribution while it filled the rough surface to decrease the roughness and increase the variation of lithium distribution in thickness direction at the nanoscale. This was further corroborated by the depth-resolved XPS results where the lithium and fluorine distribution changed as a function of solid electrolyte content in the thickness direction. Thus, the formation process of the initial electrode-electrolyte interphases, which determines the performance of a solid-state battery using a solid electrolyte was systematically analyzed by ESM and XPS methods. The second part is about all-solid-state battery that have a great merit suppressing the ignition and explosion risks of liquid electrolyte through solid electrolyte without liquid electrolyte. Specifically, we synthesize interconnected solid electrolyte, which controls the shape of a garnet type Li7La3Zr2O12 (LLZO) oxide solid electrolyte to well percolate lithium-ion conduction paths in the composite electrode. It was confirmed that the solid electrolyte shows better performance than the particle LLZO on the natural graphite composite anodes in higher C-rate (1C) performances. In particular, non-destructive nano X-ray computed tomography images were obtained to determine the internal porosity and spatial distribution of the all-solid-state battery electrode. The difference in porosity distribution in the graphite anode as functions of the cycle and the solid electrolyte content was analyzed, which validated the superior performance of the shape-controlled solid electrolyte. In addition, it was confirmed that the performance can be maintained even at a low slurry mixing rate, which is helpful in improving processability of composite electrodes for all-solid-state batteries. In conclusion, the meaning of this study is that the structure – property correlation related to the electrode and electrolyte interface for the improvement of lithium-ion battery performance was systematically analyzed from various aspects considered with electrochemical imaging and novel material synthesis, and it is important to understand the meaning of the fundamental electrode-electrolyte interface phenomenon. In supplementary study, the LIB cathode and electrolyte additive research platform is developed through a simple image tools of scanning electron microscopy (SEM) with machine learning. 12 samples of different composition (4 cases) and state (3 cases) were prepared for a transition metal oxide (NCM) cathode with different Ni contents (Ni = 0.3, 0.5, 0.6, 0.8) and different states of pristine, formation, and 100 times cycled. We trained NCM electrode samples through convolutional neural network by EfficientNet algorithms structure and evaluated the accuracy of the composition and state estimations. Furthermore, we conducted untrained NCM electrodes that examine the electrochemical tests with functional additives. Through these transfer learning, we provide LIB researchers insight into high-throughput cathode and additive developments that can screen the elements for electrolyte additive and optimize morphology of NCM cathode materials to maximize the LIB performance.