On the Cycling Performance and Stability of Silicon-Based Anodes in Lithium-Ion Batteries - Revealing Challenges and Failure Mechanisms Using Post Mortem Analyses
MetadataVis full innførsel
Li-ion batteries with improved energy density can be obtained by increasing the Si content in the graphite electrode, or replace it entirely with a Si based electrode. This may have the additional benefit of improving safety and cost. Two main approaches are reported to improve the cycling stability of a Si based electrode; 1) mixing SiO2 as active material to form a composite electrode, and 2) optimizing a Si electrode by conventional methods. The electrochemical characterization was carried out by assembling the active material into coin cells, followed by measuring the development of capacity with number of charge and discharges (cycles). Afterwards, some of the cells were disassembled and extensively characterized using scanning electron microscopy (SEM), focused ion beam (FIB), energy-dispersive X-ray spectroscopy (EDS) and Fourier transform infrared spectroscopy (FTIR), in order to investigate the failure mechanisms. The Si/SiO2 composite was manufactured by simple mixing/milling of Si and SiO2 as active material in the electrode manufacturing process. Despite the verified electrochemical activity of SiO2 and relatively good capacity retention of a pure Si electrode, the SiO2 was not found to be electrochemically active in the composite electrode and in fact worsen the capacity retention. Consequently, the SiO2 was not attributed an electrochemical role, nor a structural role, but rather dead weight with low conductivity. The conventional methods to improve the Si electrode cycling stability included milling of the Si particles, comparing Na-CMC with Na-alginate as binder material, adding fluoroethylene carbonate (FEC) and vinylene carbonate (VC) to the electrolyte, investigate two different formation cycling procedures, and explore the influence of two different lower cut-off voltages during cycling (0.05 and 0.12 V). By post mortem SEM, EDS, FIB and FTIR, two different failure mechanisms were observed; physical disconnection of Si from the current collector, and continuous electrolyte reduction that formed a dense and thick film throughout the porous electrode structure. A small effect on cyclability was observed for a high-energy ball milled Si, attributed to the verified reduced particle and crystallite size that reduces the disconnection of active material. Despite the proposedly enhanced Li diffusion, the capacity for the milled Si was lower during long term cycling and particularly at higher current densities. This was explained by a possibly thicker solid/electrolyte interphase (SEI) layer and thicker native oxide layer. The binder was shown to be of large significance for cycling performance and failure mechanism. Na-alginate was found to also benefit from replacing the water solvent with a citric acid potassium citrate buffer, which is commonly known for Na-CMC. This changed the dominant failure mechanism from physical disconnection to extensive film formation, consistent with the supposedly improved Si/binder interactions caused by the lower pH. Still, the Na-alginate was not found to outperform the Na-CMC binder. Adding 5 wt.% FEC and 1 wt.% VC to the ethylene carbonate:diethyl carbonate (EC:DEC) electrolyte improved the coulombic efficiency and initial SEI, by a slightly larger amount of polycarbonates and LiF. However, the initially improved SEI did not seem to be sufficient to prominently extend the cyclability, and the extensive film formation was still observed as failure mechanism, with LiPF6 decomposition products of alkyl phosphates and F-P-F compounds indicated by FTIR. Additionally, an in-house designed FTIR procedure enabled characterization of the air sensitivity of cycled electrodes, which was found to be significant. A 4-step gradual formation cycling with low current was found to reduce the maximum stress in the electrode compared to one formation cycle with low current. The findings were based on differential capacity plots and FIB cross-sections. The 4-step gradual formation cycle combined with a cut-off voltage of 0.12 V (compared to 0.05 V) during cycling, significantly improved the cyclability and the electrode showed small sign of degradation even after 100 cycles. The effect was attributed to a smaller volume expansion of Si, preventing some of the continuous electrolyte decomposition. The cycling stability was compared to currently outperforming Si based electrodes and commercial demands, followed by suggestions for further optimization. The findings and further suggestions will hopefully contribute to bringing electrodes with higher Si content to the market, enabling batteries with improved energy density, cost and safety, all of which are essential for a society based on renewable energy sources.