QCM Study of Li2O2 Formation in DMSO for Li-Air Batteries: Solution Vs. Surface Pathway and the Influence of a Redox Mediator on the Oxidation of Li2O2

To reduce the dependence on fossil fuels, the challenge for scientists is to come up with a variety of energy storage systems adapted to the needs of various applications. For the automotive industry, Li-O₂ batteries are one of the most promising systems due to their high theoretical energy density. This is reflected in the exponential growth of the scientific output on Li-O₂ batteries in the past few years. Although scientists do not always agree on the mechanism of oxygen reduction and Li₂O₂ formation during discharge. Two models divided the Li-O₂ battery research, one where Li₂O₂ is formed via oxygen reduction on the surface of the electrode. The second model describes the formation of Li₂O₂ in solution which subsequently precipitates on the electrode surface. Very recently though, Johnson et al. proposed a single unified mechanism, which explains the oxygen reduction across the whole range of solvents and for which the two previous models are limiting cases.¹ In other words, depending on the system parameters, such as electrolyte donor number or discharge potential, the Li₂O₂ formation goes via a surface or via a solution pathway.

In this work an electrochemical quartz crystal microbalance (EQCM) was used to investigate this unified model in DMSO. Both cyclic voltammetry and potentiostatic cycling was used to detect which products are formed during reduction, via which pathway (surface or solution pathway) and to evaluate the reversibility of the reactions, from a mass point-of-view during cycling. Especially the influence of applied cathodic potential, proves the usefulness of an unified deposition mechanism. We confirm by EQCM (and RRDE) that LiO₂ is soluble, which was concluded from the time delay we observed between the deposition of the expected mass (based on Faraday's law) and the measured mass. Ambiguity in reported literature values for the slope of the deposited mass per electron M/z is due to the negligence in considering this time delay and thus the solution pathway.

At a low overpotential, 2.6 V vs. Li⁺/Li, the average M/z value versus cathodic charge indicates that soluble LiO₂ is the first product of the ORR which reacts further to form Li₂O₂. Slightly increasing this overpotential to 2.4 V vs. Li⁺/Li shifts the deposition mechanism from the solution pathway to the surface pathway. Cycling experiments on the EQCM at a potential of 2.6 V vs. Li⁺/Li shows a very good reversibility for the formation and dissolution of Li₂O₂ from a mass point-of-view. The latter potential of 2.4 V vs. Li⁺/Li on the other hand, results in a mass increase cycle after cycle. From these results is was concluded that a moderate discharge potential of 2.6 V vs. Li⁺/Li, results in a solution pathway deposition for which the best reversibility is observed.

We also investigated the influence of a redox mediator tetrathiafulvalene (TTF) on the mass reversibility was investigated. The redox mediator, an electron-hole transfer agent, aids the oxidation of Li₂O₂ as has been reported by Chen et al.² Here the redox mediator was activated in situ and it has proven its usefulness on the mass reversibility of the deposition/dissolution of Li₂O₂ during cycling. Even at a the higher overpotential of 2.4 V vs. Li⁺/Li a good mass reversibility was observed during cycling. This means that thanks to the redox mediator recharge rates can be improved, which are impossible without the presence of a redox mediator.

To conclude, the mass of Li₂O₂ formation and dissolution was monitored by EQCM and it showed the elegant balance between the surface and solution pathway in DMSO, an electrolyte with a high donor number. From a mass point-of-view the reversibility was improved by applying a low overpotential, resulting in the solution pathway deposition or by adding the redox mediator, tetrathiafulvalene (TTF).

¹ L. Johnson, C. Li, Z. Liu, Y. Chen, S. A. Freunberger, J. M. Tarascon, P. C. Ashok, B. B. Praveen, K. Dholakia, and P. G. Bruce, Nature Chemistry, 6, 1091 – 1099 (2014).

² Y. Chen, S. A. Freunberger, Z. Peng, O. Fontaine, and P. G. Bruce, Nature Chemistry, 5, 489 – 494 (2013).