Singlet Oxygen Formation during the Charging Process of an Aprotic Lithium-Oxygen Battery

Abstract

Aprotic lithium–oxygen (Li-O2) batteries have attracted considerable attention in recent years owing to their outstanding theoretical energy density.[1] Parasitic side reactions are the central problem for the development of a reversible Li-O2 cell chemistry. They mainly occur upon charge, causing the degradation of both the carbon electrode and the electrolyte upon battery cycling, leading to cell death within a few cycles. The exact electrochemical or chemical nature of the side reactions during charge remains largely unclear, but some form of “nascent” oxygen produced during charging has been suggested to contribute to the degradation of several cell components.[2] Singlet oxygen (term symbol 1Δg, thereafter 1O2) could potentially be this reactive nascent oxygen, as it is a strong oxidizing agent and known to form upon chemical oxidation of Li2O2, Na2O2 and a series of organic peroxides.[3,4] We present the first experimental investigation of 1O2 formation during the charge of an aprotic Li-O2 battery. The identification of 1O2 is based on its reactivity towards a specific spin trap (4-Oxo-TEMP), which is added to the electrolyte, forming a stable radical (4-Oxo-TEMPO) that can be detected by in-operando electron paramagnetic resonance (EPR) spectroscopy. The EPR experiments were carried out in a purpose-built spectro-electrochemical cell[5] , modified by adding a reference electrode and a gas purging unit, which can be cycled directly inside the cavity of the EPR spectrometer. Firstly, the electrochemistry in the presence of the spin trap is investigated. By a combination of UV/VIS experiments and On-line electrochemical mass spectrometry (OEMS) it is shown that the presence of the spin trap does not affect the charging processes up to 3.9 V. In the next step, a constant current charging experiment with the 4-Oxo-TEMP spin trap as electrolyte additive (100 mM) is carried out. Figure 1 shows the charging potential and the concentration of the 4-Oxo-TEMPO radical as determined by in-operando EPR spectroscopy. The charging processes can be separated into four distinct phases (s. Fig. 1): During phase I and II, a normal charging process with Li2O2 oxidation as the main electrochemical reaction is taking place, as also indicated by the oxygen evolution determined by OEMS. The 4-Oxo-TEMPO amount starts to increase once the electrode potential exceeds ∼3.55 V, which is close to the thermodynamic threshold for 1O2 formation (3.45 to 3.55 V), suggesting that this increase is caused by the reaction of the spin trap with 1O2. In phases III and IV electrochemical side reactions set in that are not related to the 1O2 cell chemistry. The presence of these four phases shows that a time and voltage resolved in-operando EPR experiment, as presented in this work, is critical for unraveling the different mechanisms for forming and decomposing the 4-Oxo-TEMPO radical. To provide further evidence that the increase of 4-Oxo-TEMPO during phase II is caused by 1O2, potential controlled experiments were carried out, which fully confirm the results of the constant current experiments. Lastly, based on a calibration procedure and a simple mechanistic consideration, the lower limit of the 1O2 fraction formed during cell charge was estimated to be 0.5%. The real fraction of 1O2 is most likely higher, as a significant part of the evolved 1O2 will be quenched by the solvent or other side reactions before being trapped by 4-Oxo-TEMP. However, this simple estimate already shows that 1O2 is formed in substantial quantities. We believe that the occurrence of the highly reactive singlet oxygen might be the long overlooked missing link in the understanding of the electrolyte degradation and carbon corrosion reactions that occur during the charge of Li-O2cells.