Contrasting film formation reactions of ethereal and carbonate solvents on metallic lithium

A summary of studies by photoelectron spectroscopy of the reactions of prototypical ethereal and alkyl carbonate solvents with lithium metal in ultra-high vacuum (UHV) is presented. The Li/solvent interface is formed by condensing solvent vapor onto the surface of freshly vacuum evaporated Li at 130 K. The reaction is observed by obtaining core-level photoelectron spectra as a function of temperature. The interfacial reactions are distinctly different for the ethereal vs. the carbonate solvents. The ethers, both cyclic (THF and 1,3-dioxolane) and linear (glymes), react initially to form an LiOR or radical anion (1,3-dioxolane) precursor, followed by the formation of a polyether layer. In the carbonates, there is a significant difference between linear and cyclical forms. The linear carbonates (DMC and DEC) are more reactive, reacting even below the melting point, and produce both LiOCO2R and LiR (alkyl lithium). There are no unreacted solvent molecules left on the surface at room temperature. Since alkyl lithium is soluble in the bulk solvent, the lithium surface will not passivate in this solvent until considerable dissolution of lithium has occurred. The cyclic carbonate PC forms only insoluble LiOCO2R, but no unreacted PC (or PC-like molecule) remains at room temperature. Thus, the corrosion of lithium in DMC/DEC is predicted to be significantly greater than in PC. Since the ethers incorporate a significant amount of solvent molecule in the passive layer, e.g. by polymerization, the SEI layer formed in the ethers is predicted to be more highly conducting than in the carbonates. This prediction is based on the chemistry of the pure solvent molecules, absent real life impurities, e.g. water, or an anion, which may react preferentially with lithium.