Computational Modelling of defects in Battery Materials

: (Alternative Format Thesis)

Abstract

Suitably cost-effective energy storage will aid in the migration to a low carbon global energy economy. In an attempt to access such technology, recent years have seen a great deal of research into enabling a step change in the performance of lithium-ion batteries. One perceived route to achieving this is an all-solid-state battery, in which the flammable liquid electrolyte used in commercial battery technologies is replaced with a solid that has a high lithium-ion conductivity. Such batteries may have improved energy density and reliability as compared to the current state-of-the-art. Optimising the properties of these solid-electrolytes for successful incorporation into a battery cell has proved challenging, and a commercially viable all-solid cell has yet to be produced. Within this thesis, we argue for a full understanding of point defect chemistry to learn more about the performance and properties of crystalline solid electrolytes. In doing so we discuss these solid-state ionic materials through the lens of semiconductor defect chemistry, highlighting the importance of considering the coupling between defect populations via the electronic chemical potential, or rather the Fermi energy. We show how perturbations to the Fermi energy (facilitated by aliovalent doping) affect the concentrations of all charged defects in the system, questioning the validity of the commonly held assumption that the charge imbalance introduced by aliovalent dopants is compensated by changing concentrations of mobile-ion defects. This leads to potential difficulties in assigning a simple composition-property relationships. This approach reveals that the lithium-ion conductor Li3OCl is, from a defect chemistry perspective, a somewhat poor candidate solid electrolyte and that the garnet ion conductor Li7La3Zr2O12 (LLZO) has a more complex than previously considered doping response. We then build on this model to examine electronic carrier populations in LLZO, discussing possible non-negligible electronic conductivity, which has been linked to battery failure owing to a bulk lithium reduction process. Finally, we explicitly consider dopant-driven changes in defect–host-framework interactions which result in modulations to the potential energy surface for lithium ion transport in the superionic conductor Li10GeP2S12. To close, we briefly discuss the practical and theoretical challenges posed by modelling the point defect chemistry of solid electrolytes more fully. It is hoped that this work will inspire more considerations of the equivalence between a picture of defect chemistry in functional materials that invokes charge-neutral defect reactions and one which considers a system where the concentrations of all charged defects are coupled.

Date of Award	27 Apr 2022
Original language	English
Awarding Institution	University of Bath
Supervisor	Benjamin Morgan (Supervisor) & Alison Walker (Supervisor)