Li/Mn-rich layered oxide cathode active materials (LMR) have gained attention due to their exceptionally high practical specific discharge capacity (>250 mAh g-1) originating from both cationic and anionic redox. The latter is only achievable after electrochemical activation of oxygen redox at high potential (> 4.6 V vs. Li+|Li), though this is accompanied by the destabilization of lattice oxygen, resulting in large voltage hysteresis and redox couple shift upon cycling. To enable stable cycling of LMR, common stabilization strategies such as doping, coating, and electrolyte additives have been investigated, frequently in LMR || Li cells. The use of such a cell setup conceals the additional challenges associated with the use of graphite anodes, i.e. electrode crosstalk. In high voltage cells, transition metal (TM) dissolution from the cathode with subsequent deposition on the anode leads to inhomogenous lithium plating and rapid capacity fade, as known from state-of-the-art Li ion batteries at high voltage. Nevertheless, those relations remain unclear and need to be validated in LMR || graphite cells.

In this work, the impact of electrochemical conditions and electrolyte formulations is investigated in LMR || graphite cells via varying upper cut-off voltages. Utilizing the unstable O3-type LMR, which forms highly reactive molecular oxygen during charging, an extreme condition for the electrolyte is created. Up to 4.5 V, performance can be improved by the addition of 1 wt.% lithium difluorophosphate (LiDFP), likely via dissolved TMs scavenging effect,[7] though the impact at 4.7 V is severely limited. By comparing the solid-electrolyte interphase (SEI) and cycled electrolyte composition using energy dispersive x-ray spectroscopy and ion chromatography-conductivity detector, respectively, we found a significantly higher degree of LiDFP decomposition at 4.7 V. Although the scavenging of TMs is proposed to happen through their precipitation reaction with the decomposition products of difluorophosphate anion i.e. monofluorophosphate and phosphate, we show that the higher degree of LiDFP decomposition at 4.7 V is not followed by more effective TM scavenging effect. Considering the highly reactive nature of charged LMR and the electrolyte instability towards chemical oxidation, it is likely that the larger amount of H2O formation due to charging at 4.7 V exacerbates the hydrolysis of LiDFP and LiPF6, resulting in excessive HF formation. The presence of excess H2O might also accelerate the decomposition of monofluorophosphate to phosphate, which is less soluble, thus less concentrated to effectively scavenge higher Mn2+ amounts.