The most considerable push to higher energy density Li-ion batteries (LiBs) has been achieved by incremental improvements of the positive electrode cathode active materials (CAM).1 One approach has been to increase the nickel content of layered transition metal oxide CAMs like lithium nickel cobalt aluminum oxide (NCA; LiNixCoyAlzO2, with x+y+z=1) to above 80%, which has already been achieved on a commercial level. While this increase is accommodated by a higher achievable discharge capacity at a given upper cut-off potential, it also reduces the structural stability of the material at high degrees of delithiation (i.e. at high state-of-charge (SOC)).2,3 The structural instability of the CAM is associated with a release of lattice oxygen at high SOC, starting at approximately 80% SOC.2,3 The release of reactive oxygen species is reported to cause a cascade of degradation mechanisms resulting in capacity fade. During this process, the near-surface region of layered transition metal oxide particles is converted into an electrochemically inactive rock-salt phase, which results in the formation of a resistive surface layer.4 Furthermore, the released oxygen can stimulate the chemical oxidation of the typically employed electrolyte solvents, such as ethylene carbonate (EC), resulting in the formation of CO, CO2, and H2O, which in turn hydrolyzes the electrolyte salt.2,5 Consequently, the detection and quantification of released oxygen for different CAMs and cycling conditions is of high interest. Gas quantification methods such as differential electrochemical mass spectrometry (DEMS) and on-line electrochemical mass spectrometry (OEMS) are powerful tools for the qualitative and quantitative measurement of gaseous species such as O2 and CO2.2,6 However, such methods require an expensive mass spectrometer setup and a sophisticated interface between the LiB cell hardware and the mass spectrometer.Thus, in this study, we will present a simplified method that can detect and quantify released oxygen. The method is based on a setup consisting of a dual working electrode (WE) configuration, namely a primary WE with an NCA CAM (LiNi0.80Co0.15Al0.05O2, from BASF TODA Battery Materials LLC, Japan) and an auxiliary WE with only carbon (Vulcan-type carbon, XC-72, Tanaka, Japan), both being coated with a ; a lithium iron phosphate (LFP) electrode coated on aluminum serves as counter electrode (CE). The latter is delithiated to a degree of 90% and capacitively oversized, allowing the NCA working electrode to be cycled against a constant potential (with the LFP potential being at ~3.43 V vs. Li+/Li).8 While charging the NCA working electrode within the desired potential window, the carbon electrode is polarized to a constant potential of 1.21 V vs. the LFP counter electrode corresponding to ~2.20 V vs. Li+/Li. Since oxygen is reduced at that potential, a reductive current is detected once lattice oxygen is being evolved.9 By the precise knowledge of the number of electrons involved in the reduction of oxygen at the carbon electrode, we are able to convert the reductive current into an amount of released oxygen (e.g., in units of moles of oxygen per gram of cathode active material).We apply the developed cell setup to study the oxygen release characteristics of the NCA when being charged to 5.0 V vs. Li+/Li at 0.1 C and 25 °C, and compare these data to the gassing profile analyzed via OEMS. As will be shown, there is a good agreement between integrated and converted reductive current and the amount of detected O2 in the OEMS.
Read full abstract