A CR2032 coin-type cell was used to evaluate the cycling performance of lithium metal anodes. We prepared a nickel foil (12 mm diameter, 99% purity) for use as a working electrode, and as-received lithium foil for use as a counter electrode. For the cycling test, 0.5 mAh cm-2 of lithium was initially electrodeposited (charged) on the working electrode at 0.2 mA cm-2 or 1.0 mA cm-2, and the same charge amount of lithium was stripped (discharged) from the working electrode until the voltage reached 1.0 V against a counter electrode’s potential. For the preparation of electrolytes, the dried lithium salts of lithium perchlorate (> 99.0% purity), lithium bis(trifluoromethanesulfonyl)imide (>99.0% purity), lithium bis(fluorosulfonyl)imide (>99.5% purity), or lithium nitrate (LiNO3) (>99.9% purity) was respectively dissolved into the DMSO solvent. The concentration of each lithium salt in a DMSO solvent was controlled from 1.0 M to 4.0 M.
One reason for the poor coulombic efficiency of lithium anodes in DMSO-based electrolytes is thought to be the unstable solid electrolyte interphase (SEI) on the lithium surface as a result of side reactions with the electrolyte. Another reason is thought to be the direct reduction of the Li ion-DMSO complex that penetrates the SEI layer since DMSO strongly interacts with Li ions owing to its high Gutmann donor number. In order to form a stable SEI on lithium surface, a selection of lithium salts is revealed to be important. Lithium nitrate which forms an inorganic compound (Li2O) instead of soluble product (Li2S) on lithium surface shows an enhanced coulombic efficiency which is over 75%, >20% higher than other types of lithium salts regardless of the loading current density. Moreover, it’s revealed that the amount of free DMSO molecules in the electrolyte should be reduced to obtain further improved cycling performance of a lithium anode. As shown in Fig. 1, a highly concentrated electrolyte of 4.0 M LiNO3/DMSO which has hardly any free DMSO molecules achieved enhanced cycling performance which is over one hundred cycles with an average coulombic efficiency of 84.3%, which had not previously been reported. This enhancement is due to the absence of free DMSO solvent in the electrolyte and the promotion of the desolvation of Li ions on the SEI surface.