Can a “Bad” Salt be a Good Salt for Li Metal Batteries?

Abstract

Ethylene oxides (EO) based electrolytes are commonly used for their relatively good stability face to lithium metal. These ethers allow a good dissociation and solvation of lithium salts such as the commonly used LiTFSI salt. However, the EO/LiTFSI complexes if they show quite good conductivities, their transport numbers are quite low, and the TFSi anions do not give very good Solid electrolyte interphases (SEI) which implies quite poor performance vs lithium metal anodes. By replacing the salt by LiNO3, we observed an outstanding improvement in the cyclability of symmetrical Li cells, even at rather high current density (i. e. above 1 mAcm-2). To understand this improvement, we first compared the ionic transport properties of LiNO3 versus LiTFSI salt in small glymes (molar mass = 240 g / mol). At 40°C, PEG laden with 1M LiNO3 present an ionic conductivity one order of magnitude lower than LiTFSI, but the Li+ transport number is very high at 0,87. From ab initio calculation, Raman spectroscopy and NMR, we demonstrate that LiNO3 is only partially dissociated and tends to form dimers (or more) with lithium ion. These anionic complexes are not very mobile, which explains the lower conductivity and higher lithium ion transport number cationic conduction. Then, by mean of XPS and EIS, the composition and electrical properties of the SEI formed on lithium metal were obtained and compared. Finally, the lithium electrodeposits morphologies were characterized by X-Ray micro-tomography and SEM, allowing us to have all the cards in hand to explain the far better performance of the LiNO3 salt vs LiTFSI salt.