Understanding the Effect of Lithium Nitrate As Additive in Carbonate Based Electrolytes for Silicon Anodes

Abstract

Owed to its high specific capacity and its natural abundance, silicon is one of the most promising anode materials for next generation lithium-ion batteries. However, large volumetric changes of silicon upon (de-)lithiation of ~300 % lead to severe degradation phenomena such as particle pulverization and rupture of the solid electrolyte interphase (SEI), resulting in continuous electrolyte consumption and irreversible loss of lithium upon cycling.[1-3] Electrolyte additives such as FEC can improve the cycling performance of silicon anodes significantly by forming a more stable and effective SEI, which can accommodate the morphological changes. However, FEC has only a limited thermal stability when used with LiPF6 as electrolyte salt[4] and shows higher gassing compared to an EC-based electrolyte.[5] Lithium nitrate (LiNO3), predominantly used to passivate the lithium metal electrode in Li-sulfur or Li-air battery systems, was recently demonstrated to be also an efficient additive for silicon anodes.[6-9] In contrast to FEC, the reaction mechanism of LiNO3 in carbonate-based electrolyte systems is not completely understood yet.In the present study, the reduction mechanism of LiNO3 and LiNO2 in 1 M LiPF6 in EC/DEC (1:2, v:v) is investigated using cyclic voltammetry on carbon black electrodes with a high surface area (C65/PvDF 5:5 w/w; 4.8 m2 BET cm-2 el) in half-cells, further supported by post mortem investigations on the harvested electrodes via X-ray photoelectron spectroscopy (XPS). First, the reductive features of LiNO3 and LiNO2 are identified: here, two distinct features with peaks at 1.56 V vs. Li+/Li and 1.39 V vs. Li+/Li are observed for the LiNO3-based electrolyte (see Figure 1). However, the electrolyte containing LiNO2 revealed only the second feature at 1.39 V vs. Li+/Li, indicating that this peak is the reduction feature of LiNO2 as compared to the feature at 1.56 V vs. Li+/Li corresponding to the reduction of LiNO3 to LiNO2. In addition, no EC reduction feature is perceived at 0.8 V vs. Li+/Li for both additives, suggesting the complete and effective passivation of the C65 particle surface by LiNO3 and LiNO2 before the EC reduction potential is reached. Combining these findings with the post mortem XPS analysis we deduced the reduction reaction mechanism of LiNO3. We found that the primary reduction product of LiNO3 – also reported in literature to be LiNO2 [10,11] – is soluble in carbonate-based electrolyte systems and therefore does not act as direct SEI component. Both, LiNO3 and LiNO2 are finally tested as electrolyte additives in long term cycling experiments in Si||NMC622 full-cells, comparing the performance with the additive-free electrolyte.