Atomistic simulation studies of Li+, Na+ and Mg2+ inserted nano- architectured Ti02 for metal ion batteries

Abstract

Since the rechargeable battery performance involving TiO2 anode strongly depends on its structural, transport and electrochemical properties, it is important to explore different TiO2 nanostructures in order to improve its operation through its ability of hosting mono and multivalent ions. Consequently, the current study involves synthesizing MxTiO2 (M = Li+, Na+, and Mg2+) (x= 0.11, 0.15, 0.19 and 0.23) nanospheres, nanoporous and nanosheets using the molecular dynamics based amorphization and recrystallization technique, employing a DL_POLY code. This is followed by predicting structural and mobile ion transport properties for future rechargeable batteries. Li+, Na+ and Mg2+ were inserted into the amorphous TiO2 nanosphere, nanoporous and nanosheet architectures in order to produce LixTiO2, NaxTiO2 and MgxTiO2 (x= 0.11, 0.15 ,0.19 and 0.23) for each nano-architecture. The amorphous nano-architectures were recrystallized by simulations at 2000 K and resultant structures were cooled from 1500 K to 0 K at intervals of 500 K. Finally all cooled nano-architectures at 0 K were analysed and subsequently heated to 2000 K at 100 K intervals. The TiO2 nanospheres with 0.11, 0.15, 0.23 Li+ and Na+ concentrations had similar microstructural defects, present in pure TiO2 and before Li insertion, characterised by dominant zigzag (brookite) and straight (rutile) tunnels with empty vacancies and Li+ and Na+ filled vacancies with few Li+ and Na+ located on the surface. On the other hand, the TiO2 nanospheres with 0,19 Li+ and Na+ concentrations remained amorphous after recrystallization, cooling and heating; hence their structural and microstructural features were not constructed. Surprisingly, the Mg0.19TiO2 nanosphere recrystallised successfully together with those of 0.11, 0.15, and 0.23 Mg2+ concentrations and their microstructures had zigzag (brookite), straight (rutile), tunnels with empty vacancies, Mg2+ filled vacancies and few Mg2+ were situated on the surfaces. The microstructures of the TiO2 nanoporous structures, with 0.11, 0.15, 0.19 and 0.23 Li+, Na+ and Mg2+ concentrations, showed crystalline patterns of zigzag tunnels (brookite), straight (rutile) tunnels with empty vacancies and Li+, Na+ and Mg2+ filled vacancies after being recrystallised and cooled, except for the 0.23 Mg2+ concentration which showed highly disordered patterns. The microstructures of TiO2 nanosheets, with 0.11, 0.15, 0.19 and 0.23 Li+, Na+ and Mg2+ concentrations, had a mixture of disordered and crystalline (ordered) patterns of zigzag (brookite) and straight (rutile) tunnels with few empty vacancies, and some were filled with Li+, Na+ and Mg2+ where most Li+ and Na+ ions were situated around the surface of the microstructure except for TiO2 nanosheets with 0.23 Mg2+ concentration which was amorphous, similar to that of Mg0.23TiO2 nanoporous architecture. All resulting TiO2 nanostructures and microstructures were further characterised and analyzed by their respective radial distribution functions (RDFs), simulated X-ray diffraction (XRD) patterns, diffusion coefficients and activation energies to study the effects of increased lithiation, sodiation, and magnesiation at various temperatures. The simulated Ti-O radial distribution functions were utilised to confirm the extent of crystallinity of nano-architectures after recrystallisation, cooling, and heating. In addition, simulated X-ray diffraction (XRD) patterns were employed to determine and compare crystallinity at low and elevated temperatures. Furthermore, they depicted rutile and brookite polymorphs in LixTiO2, NaxTiO2 and MgxTiO2, shown by our simulated microstructures, and observed in previous experimental and simulated studies of LixTiO2; which suggests possibility of easy and fast Li+, Na+ and Mg2+ ion passage in the current study. In the case of Li transport, for different nano-architectures, diffusion coefficients (DCs) for most Li concentrations were predominantly near zero at low temperatures and increased gradually above 500 K and significantly at higher temperatures beyond 1000 K. DCs of higher Li concentration nano-architectures tend to be highest. On the contrary, the diffusion coefficient of the amorphous Li0.19TiO2 nanosphere is elevated at low temperatures, however, it is exceeded by those of other concentrations at higher temperatures. On the whole the activation energies of Li in TiO2 nanosphere, nanoporous and nanosheet structures, in appropriate temperature range, were consistent with observed transport properties. In particular, the lowest activation energy for the amorphous Li0.19TiO2 nanosphere concurs with non-zero DC at lower temperatures. On Na transport in various TiO2 nano-architectures, DCs for most Na concentrations were predominantly near zero at low temperatures, except for the Na0.23TiO2 (associated explanation for such deviant behaviour is not yet available). Diffusion coefficients subsequently increased gradually above 500 K to substantial at higher temperatures, with DCs of higher Na concentrations in different nano-architectures tending to be highest; in particular the DC of the amorphous Na0.19TiO2 which was slightly higher than other crystalline NaxTiO2. The activation energy of the nanosphere with the highest Na concentration (Na0.23TiO2) was low, consistent with the trend of non-zero diffusion coefficients at low temperatures. The AEs of nanoporous structures at different Na contents were equivalent and related DCs almost overlap in the 500 to 1000 K range. In the case of nanosheets AEs are also almost equal for all NaxTiO2 with that of the Na0.23TiO2 (highest Na concentration) being highest resulting in a slightly lower diffusivity, except at very high temperatures. These results provide insights and an understanding on how much such TiO2 nano-architectures can enable operations of sodium ion batteries. Non-zero diffusion coefficients of Mg2+ commenced at lower temperatures in nanospheres, especially those with lower Mg concentrations. Related activation energies were also relatively low and increased with Mg2+ concentrations. However, in the nanoporous and nanosheet architectures, the Mg2+ DCs were almost zero below 1000 K and increased significantly above 1200 K with DC being highest in the heavily intercalated nanostructures. The activation energies for the latter two nano-architectures were in the range of 0.400 eV which is higher than in nanospheres, but comparable to those of the nanoporous Li-MnO2. Generally, on matching maximum diffusion coefficients of Li, Na and Mg mobile ions in TiO2 nano-architectures, it is apparent that they are all highest in nanospheres with DCLi > DCNa > DCMg. On comparing magnitudes of DCs in different nano-architectures the following trend emerges; for the monovalent Li and Na mobile ions DCSphere ~ 5DCPorous ~ 3DCSheet and for the divalent Mg mobile ion DCSphere ~ 2.5DCPorous ~ 2.5DCSheet. Such comparisons provide insights and an understanding on how much such TiO2 nano-architectures can enable performance of various mobile ion rechargeable batteries