Electrochemical Analysis of Charge Overpotentials in Non-Aqueous Lithium and Sodium Oxygen Batteries

Abstract

Aprotic metal-oxygen batteries, especially Li-O2 and Na-O2 batteries, could afford theoretical specific energies of more than 500 Wh/kg. However, while not compromising on rechargeability, the practical realization of the theoretically possible high specific energies has been elusive. A better understanding of the differences and similarities between Li–O2 and Na–O2 battery systems in terms of charge-discharge mechanisms and parasitic chemistry will be meaningful in solving these challenges. Here, we explore the differences between the two systems using a combination of galvanostatic charge-discharge measurements, Differential Electrochemical Mass spectrometry (DEMS), and Potentiostatic Electrochemical Impedance Spectroscopy (PEIS) analysis. The discharge profiles of these battery chemistries are similar in terms of overpotential and have been widely studied. But the origins of the differences in charge overpotentials are not clear. Using in-situ pressure decay measurements during discharge, we calculated the number of electrons/oxygen (e-/O2) to be 2.02 and 1.01 for Li-O2 and Na-O2 cells, respectively. Further, DEMS analysis during charge has shown that the ideal 2 e-/O2 is attained only during the initial phases of charging for Li-O2 cells. However, for Na-O2, the ideal 1 e-/O2 is achieved during a signification portion of the galvanostatic charge step. We corroborate these findings with PEIS and distribution of relaxation times (DRT) analysis to develop a mechanistic view of the processes that lead to poor coulombic and oxygen evolution efficiencies in metal-oxygen cells. Based on DRT analysis, we identify a lower time constant process that is common to both Li-O2 and Na-O2 cells. Interestingly, we observed that voltage polarization is preceded by the observation of this lower time constant peak. In the case of Li-O2 cells, this lower time constant process was observed at the end of discharge and throughout the recharge process, while for the Na-O2 cells, this was only observed after about 70% of recharge. We discuss the possible origins of this low time-constant process and its implications for recharging metal-oxygen batteries. We will also discuss the role of singlet oxygen, if any, in the recharge efficiencies of these two aprotic metal-oxygen batteries. Our results imply that the identification of the origins of this lower time constant process and possible suppression of this process will be key for rechargeable metal-oxygen batteries.