Effects of Fixed Redox Mediator in Air Electrode for Lithium-Air Batteries

Abstract

Because of higher theoretical energy density (3505 Wh kg-1) than that of commercial lithium-ion batteries, Lithium (Li)-air (O2) batteries (LABs) have been intensively investigated for electric vehicles, large energy storages, etc. For practical use, one of the big problems is a large charging overvoltage. Recently, redox mediators (RMs) in electrolyte solution have attracted much attention. RMs (I-[1] or Br-[2]) electrolyte system suppresses charging overvoltage. However, RMs have some challenges such as low solubility in electrolyte and redox shuttle effects, i.e. RM oxidant does not effectively react with Li2O2 at air electrodes but Li metal negative electrode (NE). To overcome the challenges (low solubility of RMs and shuttle effects), we introduced RMs into air electrode, positive electrode (PE). The fixation of RMs in PE is expected to concentrate RMs nearby PE to suppress the shuttle effects. We constructed LAB test cell with the normal PE without RM or with RM in electrolyte solution, or RM fixed PE without RM in electrolyte solution and Li metal NE for investigation of the effects on the suppression of shuttle effects, the reduction of charging overvoltage and the improvement of cycle performance. Ketjen black (KB) and a RM (e.g. LiBr) were ground with a mortal and pestle. To only KB or the prepared mixture, polyvinylidene difluoride (PVdF) and N-methylpyrrolidone (NMP) were added, as a binder and solvent, respectively. The slurry was coated on a carbon paper and dried at 110 oC for overnight. On the other hand, 0.2 M Lithium bis(trifluoro methanesulfonyl)imide/diethyleneglycol dimethyl ether (LiTFSI/G2) for electrolyte was prepared in an Ar-filled grove box. For comparison, 0.2 M LiTFSI+50 mM LiBr/G2 was prepared in the same procedure. LAB test cells were constructed with PE (normal or the RM fixed), electrolyte (with/without a RM) and Li metal NE. Discharge/charge cycle performance test was curried out at applied current density of 200 mA g(KB)-1, maximum capacity of 500 mAh g(KB)-1 and the range of cut off voltage of 2.0 – 4.5 V. All the LAB test cells showed the similar discharge and charging overvoltages at 1st cycle (Fig.1). At the 3rd cycle, the cell composed of the PE containing only KB showed sudden rise of charge voltage over 4.0 V. Even at several cycles, on the other hand, the cell composed of PE containing LiBr kept the charge voltage at around 3.5 V, which corresponds to the redox potential of Br-/Br3 - and the charge capacity was kept compared to that of PE containing only KB (Fig. 2). These indicate that the LiBr fixed in PE effectively works as a RM. Charge voltage of the cell composed of electrolyte containing LiBr also showed sudden rise at 3rd cycle. However, the change was smaller than that of PE composed of only KB. The Br- in electrolyte worked as a RM but did less efficiently than that fixed in PE. This indicates the effectiveness of the fixation of RM in PE; a part of the fixed Br- probably exists nearby Li2O2 passivation layer on PE, resulting the suppression of shuttle effect. The discharge voltage of electrolyte containing RM was lower, possibly because of shuttle effects, compared to that PE containing only KB and PE containing LiBr. These facts also indicate that PE containing RM reduced charging overvoltage and suppressed shuttle effects. PE containing other RMs and the details of RMs reaction is also discussed at the presentation. This study was supported by JST Project “ALCA-SPRING”, Japan. [1] H.D. Lim, et al., Angew. Chem., Int. Ed., 53, 3926 (2014). [2] W.-J. Kwak, et al., Energy Environ. Sci., 9, 2334 (2016). Figure 1