The rapid development of renewable energy resources in response to growing concerns about fossil fuels dependence and greenhouse gas emissions has raised challenges to developing large-scale energy storage systems for the smooth integration of intermittent energy resources into the grid. Increasing research efforts have been devoted to the study of the sodium-ion battery technology as a promising solution to these challenges, especially in the search for better electrode materials. In this thesis, we focused on the investigation of Na2x[NixTi1-x]O2, a promising bi-functional electrode material that can be used as either positive or negative electrode material in sodium-ion batteries. This thesis aims to gain fundamental understandings of this material using a combination of experimental and computational techniques and to provide insights into the exploration and design of electrode materials for sodium-ion batteries. Firstly, the average and local structural properties and energetics of atomic distribution in P2-Na2/3[Ni1/3Ti2/3]O2 were investigated using Rietveld refinement on neutron diffraction datasets and atomistic simulations based on the Buckingham and Morse interatomic potential models. Both computational and experimental results showed similar nuclear density maps and higher occupancies of Na at edge-sharing sites than face-share sites. The simulations based on both potential models suggested that it is energetically favorable to have an equal amount of Na and transition metal in each of the two layers. The atomic and electronic structure changes during cycling were studied based on density functional theory (DFT) calculations. DFT-based molecular dynamics (MD) simulations showed an expansion of ab plane and contraction of c axis upon Na insertion and a small change of lattice parameters upon Na extraction. DFT calculations revealed that Ni and Ti play a dominant role in the redox reactions upon Na extraction and insertion respectively, along with the participation of O. A higher in-plane electronic conductivity was observed compared to the through-plane one, with both increasing when Na ions were inserted or extracted. The quasi-elastic neutron scattering (QENS) experiments showed that Na ion diffusion can be well described by the Singwi-Sjolander jump diffusion model, where the obtained mean jump length matched the distances between the neighboring edge-share and face-share Na sites. The MD simulations based on DFT calculations showed a better consistency with experimental results than MD based on interatomic potential (IP) models in terms of diffusivity and activation energy. Faster diffusion was observed for compositions with less sodium, i.e., more vacancies. Both Na-deficient phase Na5/9[Ni1/3Ti2/3]O2 and Na-rich phase Na7/9[Ni1/3Ti2/3]O2 showed higher ionic conductivity compared to the pristine phase. Finally, the diffusion and ionic conduction for P2 and O3 Na2x[NixTi1-x]O2 are calculated with machine learning based interatomic potential models. Higher diffusivity and ionic conductivity were observed for the P2 structure.
Read
