Computational studies on stability of components in an aprotic Li-O₂ battery

Abstract

Active research on the aprotic Li-air battery has been drawn by its high theoretical energy and power storage capacities of 11,000 Wh/kg and 3800 mAh/g, respectively. These values double those of the most advanced lithium ion batteries and close to those of gasoline. However, the production of a commercially viable battery has been hindered by the many challenges still waiting to be overcome. One of the biggest challenge is to find an aprotic solvent that can remain stable through many charge and discharge cycles of the battery. Common solvents used in Li-ion batteries such as propylene carbonate (PC), ethylene carbonate (EC), and dimethyl carbonate (DMC) have long been shown to degrade during discharge. As a result, the desired discharge products, lithium peroxide (Li2O2), is not formed. Researchers attribute the failure of these solvents to nucleophilic reactions with the superoxide radical ion (O¬2-). However, solvents such as acetonitrile (MeCN), 1-methyl-2-pyrrolidone (NMP), dimethoxy ethane (DME), and dimethyl sulfoxide (DMSO), that are shown to be stable in the presence of O2- and produce Li2O2 as the dominant discharge product, have subsequently been found to degrade slowly over many charge and discharge cycles. Recently, a more complete mechanism for discharge in an aprotic Li-air battery has been proposed, which accounts for the formation of solvated peroxides by disproportionation. In the first part of the present work, nucleophilic attacks by these solvated peroxides, O22-(solv), LiO2-(solv), and Li2O2(solv), on some commonly used solvents in aprotic Li-air batteries have been explored by calculating the reaction and activation free energies using density functional theory (DFT) method. The results show that despite some solvents (MeCN, NMP, DME and DMSO) demonstrating strong stability against nucleophilic attacks by O2-(solv), these solvents are susceptible to nucleophilic attacks by O22-(solv) and LiO2-(solv).

In the next part of the present work, the investigation is expanded to hydrogen abstraction reactions between these solvated superoxides and peroxides and those commonly used solvents. The computed reaction and activation free energies using DFT show that nearly all the investigated solvents are stable against hydrogen abstraction by O2-(solv) and LiO2(solv). However, most of them are susceptible to hydrogen abstraction reactions with O22-(solv) and LiO2-(solv), and these reactions can contribute to the degradations of these solvents.

Finally, attention is turned to another major challenge facing the Li-O2 battery, namely the degradation and erosion of the oxygen cathode. Many experimental results point to possible side reactions between LiO2 or Li2O2 and the carbon-based cathodic material as the main culprit for the degradations. Due to the primitiveness of the understanding of the formation mechanism of Li2O2, the thermodynamics of nucleation mechanism on reduced graphene oxide (RGO) are explored using DFT-based computational methods. The investigation is then extended to explore the thermodynamics for the formation of undesired degradation product lithium carbonate (Li2CO3) on these different RGO materials. The computational results suggest that the formation of Li2CO3 is thermodynamically favorable on RGO, which can be used to provide some insights towards designing a stable oxygen-cathode.