Multi-Scale Modelling of the Li3OCl Solid Electrolyte for Next-Generation Lithium-Ion Batteries

  • Matt Clarke

Student thesis: Doctoral ThesisPhD

Abstract

Sustainable energy storage has attracted considerable research interest in recent years, with Li-ion batteries being amongst the most widely studied technologies. Both the decarbonisation of power generation and vehicle electrification urgently require electrochemical energy storage with enhanced capacities and safety relative to the current state-of-the-art. The replacement of conventional liquid electrolytes with solid-state materials can enhance the safety and energy density of Li-ion batteries, but there remain a number of technical barriers to their use. In particular, the identification of solid electrolytes with high ionic conductivity, soft mechanical properties, and wide electrochemical windows is key. The Li-rich anti-perovskite Li3OCl is a promising material which exhibits all of these properties. However, a failure to characterise the effects of ion doping, Li-ion transport mechanisms, and grain boundary effects across multiple length scales represent significant gaps in understanding. To address this, energy minimization and molecular dynamics (MD) studies have been performed to provide atomistic insights into a promising next-generation battery material. Firstly, an energy minimisation study of Mg- and F-doping in bulk Li3OCl identified strong dopant–defect binding effects, which inhibited long-range Li-ion migration. An MD study of the same system yielded migration barriers of 0.41 – 0.43 eV and 0.84 – 0.89 eV for Mg- and F-doping respectively. The slight increase in migration energies in the Mg-doped system as compared to undoped Li3OCl (LiCl Schottky, 0.29 eV) is offset by the capacity for elevated mobile lithium vacancy concentrations, whilst F-doping is found to be a non-viable doping strategy. Secondly, a systematic search of symmetric grain boundaries identified several grain boundary structures with low formation energies, namely the $\Sigma3[111](111)$, $\Sigma3[211](211)$, $\Sigma3[110](110)$, $\Sigma3[311](0\bar11)$, and $\Sigma5[210](210)$ grain boundary structures. A novel analysis technique was utilised to derive locally resolved Li-ion diffusion coefficients and migration barriers for in both undoped and Mg-doped variants of these structures. Two distinct types of grain boundary structure were identified, with some grain boundaries exhibiting low barriers to lithium ion conduction along the grain boundary interface, and others exhibiting high resistance. Notably, dopant–defect binding effect between Li-vacancies and Mg-dopants did not inhibit Li-ion conduction along conductive grain boundary interfaces as in the bulk system. Finally, a novel micro-scale simulation approach based on a large-scale MD search is performed to identify and characterise the Li-ion transport properties of low energy polycrystal structures in Li3OCl. In contrast to analytical models in the literature, lithium ion conductivity is found to increase with decreasing grain size, attributable to the previously unidentified conductive paths along Li3OCl grain boundaries.
Date of Award13 Dec 2021
Original languageEnglish
Awarding Institution
  • University of Bath
SponsorsCFH Docmail Ltd & UK Research & Innovation
SupervisorMuhammed Islam (Supervisor) & Tim Mays (Supervisor)

Keywords

  • Battery material
  • Battery
  • Solid electrolytes
  • Li-ion
  • computational chemistry
  • Molecular Dynamics
  • atomistic simulations
  • Sustainable material
  • anti-perovskite
  • Li-rich
  • LiRAP
  • Li3OCl
  • Doping

Cite this

'