Since it was first introduced by Abraham and Jiang, the nonaqueous rechargeable Li-O₂ battery has recently attracted considerable attention as a possible energy storage device for electric vehicle (EV) application owing to its extremely high theoretical energy density, which far exceeds that of any other existing energy storage technologies. The theoretical specific energy of a non-aqueous Li-O₂ cell is approximately 11 kWh/kg, which approaches that of the gasoline, if the calculation is based on the lithium electrode alone.  A typical non-aqueous Li-O₂ cell, which consists of a Li anode, Li⁺ conducting aprotic electrolyte and porous carbon cathode, operates by reduction of O₂ to form lithium oxides within the pores of the cathode during discharge, with the process being reversed on charge ideally. It is based on the net electrochemical reaction (1) with a thermodynamic potential U₁ = 2.96 V, from the Nernst equation.

2Li+ + 2e- + O2 <---> Li2O2  (1)

Early experiments exhibited very high charge overpotenials (>1V) for Li-O₂ cells, mostly in the carbonate-based electrolytes (e.g. LiPF₆ in propylene carbonate), which result in large energy storage inefficiency. Although considerable electrolytic reduction of charging overpotential can be achieved when metal or metal oxide nanoparticles are incorporated into the porous carbon cathode, the carbonate or mixed ether-carbonate based electrolytes are known to severely decompose on cell discharge. Such decomposition leads to the discharged product consisting of a mixture of lithium propyl dicarbonate, lithium carbonate, HCO₂Li, CO₂, C₃H₆(OCO₂Li)₂, CH₃CO₂Li etc., rather than the desired lithium peroxide, Li₂O₂. Subsequent charging involved the partial decomposition of these Li-containing compounds with CO₂ and H₂O evolution, which correlated with high charge overpotential and capacity fading alongside electrolyte consumption. Considering the scientific and technological significance, it is important to understand the underpinning mechanisms and identify the products in the process of charging lithium oxides cathode in a Li-O₂ cell.

The goal of this work is to experimentally determine the reactions involved in charging Li-O₂ cathodes in different electrolytes, i.e. propylene carbonate (PC), tri(ethylene glycol)-substituted methyltrimethyl silane (1NM3) and Tetraethylene glycol dimethyl ether (TEGDME) with a Li containing salt. Several techniques were carried out to demonstrate electrochemical decomposition of lithium oxides and other side reactions involved, including in-situ high-energy X-ray diffraction (HE-XRD), X-ray photoelectron spectroscopy (XPS) and gas chromatography (GC). Figure 1 emphasizes the importance of salt selection on the performance. Based on previous reports, the decomposition of lithium salts used in the electrolytes could cause the detrimental electrolyte decomposition in Li–O₂ batteries. Therefore, the electrolyte decomposition may be responsible for increasing the charge capacities.. To further confirm that lithium oxides can be electrochemically decomposed, in situ X-ray diffraction patterns during the charge process of Li₂O₂ in  LiTFSI +1NM3 electrolyte were collected and displayed in figures 2 which clearly shows that the Li₂O₂ peaks totally disappeared during the charge process to 2750 mAh·g^-1. The peaks of Li are also confirmed by XPS analysis, and its intensity was enhanced gradually during charging. So we can preliminarily estimate that the Li₂O₂ does decompose to Li and O₂. More importantly, there were no other peaks appearing, especially for LiF, which means that there are no side reactions for the (LiTFSI + 1NM3) electrolyte. Therefore, it is of critically important to confirm that the recorded performance was indeed a measure of the electrochemical decomposition of Li₂O₂. Interestingly, for the charge capacity of 2750 mAh·g⁻¹ obtained using LiTFSI+1NM3, the Li₂O₂ had totally decomposed electrochemically, as shown in figure 2. Hence, it’s further confirmation that the side reactions from electrolyte decomposition are responsible for the capacity increase shown in figure 1.