Materials Modelling Centre, University of Limpopo, Private Bag x1106 Sovenga, 0727, Polokwane, South Africa.

All-solid-state (ASS) lithium-ion batteries (LIBs) are currently being explored as an alternative to the traditional lithium-ion batteries plagued with issues of safety due to the utilization of the highly flammable organic liquid electrolytes. ASS LIBs offer higher energy densities due to efficient packaging and the use of high-potential cathode and solid electrolyte materials with excellent mechanical stability. Nonetheless, solid electrolyte (SE) materials with adequate electrochemical stability, mechanical stability, and transport properties compatible with high-potential cathode materials such as nickel-cobalt-manganese oxides (NMC), spinel-type, and Li-rich layered cathode materials are yet to be found. In this work, a few promising SE and cathode materials for LIBs are studied.  VASP was used to determine the electronic stability, structural properties, and mechanical properties of LiTi2(PO4)3(LTP). LTP has a NASICON-type structure, which is crucial for the transport of lithium ions, and is among other cost-effective and thermally stable SEs. A band gap of ~2.49 eV was obtained, which is comparable to a value of 2.5 eV reported by Fami et al. Furthermore, its mechanical properties were also reproduced to a percentage difference of less than 6% from those obtained from experiments. Nonetheless, its practical utilization in the commercial space is currently limited by its inherent low ionic conductivity (<10-6S/cm). The ionic conductivity of the LTP material can be effectively modified through partial or full substitution of Ti or O. As such, the implementation of VASP in the MedeA environment was utilized to train machine learning force fields (MLFFs) of LTP on the fly from ab initio molecular dynamics (MD) trajectories. The current MLFF was able to reproduce the LTP structure (lattice parameters, volume, and density) within a percentage difference of less than 1% in comparison to DFT data. The current potentials will be further expanded to accurately predict the transport and mechanical properties of LTP and to include the partial substitution of Ti and O with dopants identified through high-throughput screening. Li9Al3(PO4)2(P2O7)3is also among other promising SEs for LIBs consisting of corner-sharing AlO6octahedra and PO4 tetrahedra, which facilitates facile movement of lithium ions in the structure. However, its fundamental structural (electronic, mechanical, and transport) properties, significant for application in ASS LIBs, have not been critically explored.  Hence such properties were calculated and it was found to be electronically stable and brittle. In addition, MLFFs for the Li9Al3(PO4)2(P2O7)3 were trained for the investigation of Li transport properties and the diffusion coefficient was calculated. The Li+ mobility will be further improved using high-throughput screening of promising multivalent cations in the Al site. The generation of on-the-fly MLFFs was also extended to LiMn2O4 spinel, which is a cost-effective, high-potential, and high-rate-capable cathode material for LIBs. The MLFFs will be expanded and trained on LiMn2O4 surfaces to study surface reactions linked to loss of manganese, which causes capacity attenuation, which critically affects the performance of LIBs. The study aims to contribute significant knowledge in the design of durable, long-lasting, safe, and high-energy-density ASS for LIBs.