Silicon is considered a promising anode material for Li-ion batteries due to the very high theoretical capacity (ca 3600 mAh/g). In fact, commercial batteries today might have up to 5-10 wt% silicon added to the anode. The extensive expansion of the material is a challenge for the mechanical integrity of the electrode, and leads to continuous exposure of fresh surface of silicon to the electrolyte and thereby continuous formation of solid electrolyte interphase (SEI). Thus, the choice of electrolyte, i.e. an electrolyte with good passivating properties, is crucial for the performance of the anode. Common additives for electrolytes to be used in combination with silicon anodes, are fluoroethylene carbonate (FEC) and vinylene carbonate (VC), both known to reduce prior to the solvent (i.e. ethylene carbonate).

Furthermore, replacing the lithium hexafluorophosphate (LiPF₆) salt, known to decompose to PF₅, which again form HF upon reaction with trace amounts of H₂O, could improve the performance. More stable salts are particularly advantageous for silicon electrodes, as HF might attack the native surface oxide (SiO₂) [1]. Salts known to be more stable are lithium bis(trifluoromethanesulfonylimide) (LiTFSI) and lithium bis(fluorosulfonyl)imide (LiFSI), and silicon anodes have been shown to perform better with electrolytes containing LiFSI [2]. These salts are furthermore highly soluble in a wide number of solvents, and have therefore frequently been applied in concentrated electrolytes, which are showing great promise as Li-ion battery electrolytes [3]. In spite of increased viscosity and reduced conductivity, a significant change of interfacial reactions at high concentrations might outweigh these drawbacks. Improved performance of Silicon electrodes (nanowires) in highly concentrated electrolytes has been demonstrated [4].

In this work, the electrolyte was optimized by increasing the concentration of LiFSI salt (1M, 3M and 5 M), combined with the electrolyte additives FEC, as well as the anion receptors tris(hexafluoroisopropyl) (THFIPB) and tris(pentafluorophenyl) borane (TPFPB). The silicon based anodes were made of 73 wt% Si (Silgrain®, e-Si 400, a commercially available battery grade silicon from Elkem), with an average particle size of 3 µm, 11 wt% carbon black (Timcal, C-Nergy C65, CB) and 16 wt% Na-CMC binder (Sigma Aldrich Mw ~90000). Slurries were cast onto dendritic copper foil. The electrodes were cycled in 2016 coin cells in a half cell configuration with metallic lithium as the counter electrode.

The impact of the concentration on the solvation of the salt was investigated by FTIR. The surface of the cycled electrodes was investigated post mortem by FTIR, XPS and SEM/EDX as well as FIB-SEM. By increasing the electrolyte concentration, the initial capacity increased, and the SEI layers were found to contain more salt reduction products, and also the SEI appeared to be thinner than for 1M concentration. However, the capacity of the 5M was fading more rapidly than the other electrolytes. The best capacity over 200 cycles was obtained for cells cycled with 3M LiFSI in combination with 10 wt% FEC and 2 wt% THFIPB. Based on the post mortem investigations, the SEI layer of this electrode was found to be rich in LiF, and with only small amounts of salt reduction products, comparable to the 1M electrolyte.

[1] B. Philippe, R. Dedryvere, M. Gorgoi, H. Rensmo, D. Gonbeau, and K. Edström, Chemistry of Materials, 25(3) (2013) 394

[2] B. Philippe, R. Dedryvere, M. Gorgoi, H. Rensmo, D. Gonbeau and K. Edström, Journal of the American Chemical Society, 135 (2013) 9829

[3] Y. Yamada and A. Yamada, Journal of the Electrochemical Society, 162 (14) (2015) A2406

[4] Z. Chang, J. Wang, Z. Wu, M. Gao, S. Wu and S. Lu, Chem Sus Chem, 11 (2018) 1787