(Battery Division Student Research Award Sponsored by Mercedes-Benz Research & Development) Quantitative SEI-Based Descriptors for Li metal Liquid Electrolyte Design

Abstract

Energy density requirements for next-generation batteries make the Li metal anode a key candidate to replace graphite in Li-ion batteries due to Li’s improved capacity (3,860 mAh/g vs. 372 mAh/g). However, Li anodes display lower Coulombic efficiency (CE, <99.9%) than graphite (>99.95%), resulting in faster loss of reversible Li over cycling. This shortfall derives from parasitic Li-electrolyte reactions that lead to accumulation of inactive Li0 and electrolyte-derived byproducts, which result in the formation of a native solid electrolyte interphase (SEI) on Li. Despite the importance of the SEI, its properties and composition are challenging to probe experimentally due to its exceedingly low amount (sub-μmolLi/cm2/cycle at >99% CE). Consequently, the framework on which the understanding of SEI composition and function was built has largely been qualitative, and descriptors to CE have overwhelmingly focused on electrolyte-, as opposed to SEI-based, metrics.To bridge this gap, this collective work focuses on advancing experimental methodologies to precisely quantify (1) functional properties of native SEIs on Li and (2) their chemical composition, both in order to explore their relationships with CE. First, we use a combination of self-consistent measurements involving EIS and CV to measure rates of Li+ exchange across these native interfaces, revealing that this functional property varies across electrolytes, increasing from low to high CE. These experiments further reveal an evolution of Li+ exchange with cycling unique to high-CE electrolytes and that is tightly linked to rate capability. In particular, we find that SEIs of high-CE electrolytes can support Li+ exchange rates in excess of 1 mA/cm2, and increasing to >>10 mA/cm2 after SEI formation and cycling. In order to (2) explore which SEI phases tend to be favored at high CE, we leveraged the quantitative power of analytical tools such as GC, NMR and ICP-AES to measure byproducts of SEI formation. We developed a custom operando GC experiment to measure and chemically-resolve gas evolution in situ during Li cycling. With sub-nmol/min resolution, these revealed, quantitatively, that SEI formation reactions that release CO or CO2, thus leaving behind a decarbonylated/decarboxylated byproduct in the SEI are associated with higher CE. We also expanded our chemical quantification studies to postmortem cells using a titration scheme that was capable of quantifying with ~nAh resolution SEI phases such as ROCO2Li, Li2C2, Li2C2, LiF and P-containing phases, in addition to the inactive Li0 that accumulates on the anode after cycling. In carbonate-based electrolytes, it was found that salt-derived phases (LiF and P) comprised only ~5% of the SEI, with solvent-derived products being the majority of the SEI. Furthermore, it was found that Li2C2, a previously-unmeasured phase on Li, shows a strong anti-correlation with CE. More recently, we further expanded our titration strategy to include other previously-unmeasured phases in leading high-CE electrolytes. Altogether, chemical quantification of the SEI reveals quantitative “SEI fingerprints” that are unique to each electrolyte. Taken together, these quantitative SEI-based descriptors enable a new avenue for electrolyte engineering, informed by a fundamental understanding of the “optimal” Li SEI and their associated capacity loss mechanisms at high CE.