Sodium ion batteries (NIB) remain attractive to many battery researchers because the earth-abundance of Na opens up a pathway for significant cost reductions in the fabrication of batteries that maintain a high energy density. The cost advantage of Na over Li will only translate into low cost batteries if the concept of earth-abundance is applied consistently. This implies moving away from electrode materials analogous to those in Li-ion batteries which typically contain costly transition metals such as Co. Materials designs starting from LIB analogues also appears to be problematic due to the differences in reaction mechanisms of the Na analogues to the Li transition metal oxides and different structural requirements for Na⁺ mobility. This approach also overlooks a particular strength of sodium compounds compared to LIB materials, in structures with suitable ion transport pathways the heavier and more polarisable Na⁺ can move faster than Li+ ions and thus can yield promising candidates for high rate performance batteries. We previously employed the bond valence (BV) method to analyse and predict ionic transport properties in solids from static and dynamic structure models [1]. Here we present an extension of the approach allowing for a rational and quantitative characterization of insertion-type NIB cathode materials in terms of rate performances based on a diffusion relaxation model.

Bond-valence site energy (BVSE) modelling and DFT simulations were used to probe Na⁺ ion migration paths in over 20 sodium-ion battery cathode materials based on earth-abundant transition metals for which experimental rate performance data were available. The migration barriers calculated based on our BVSE approach matches closely those based on DFT studies. Together with the help of the proposed diffusion relaxation model, this allows for a semi-quantitative prediction of the rate performance of half-cells directly from local structure models. The overall approach is computationally cheap and thus allows for fast screening of candidate structure databases as well as for testing of a series of local structure models (e.g. to investigate the role of antisite defects or stoichiometry deviations towards ionic transport properties) for each structure type. Both molecular dynamics and Kinetic Monte Carlo simulations were also employed to test the rate performances predictions. Finally, specific guidelines are established for the design of novel sodium-ion battery cathode materials in terms of pathway dimensionality, migration barriers, the effect of low-lying unoccupied sites, dopants or antisite defects as well as particle sizes.

For one of the candidate materials that showed strong potential for high rate performance in the computational study, alluaudite-type non-stoichiometric Na_2+xFe_2-x/2(SO₄)₃ (wherein Fe vacancies and Na_Fe’ antisite defects cross-link the otherwise independent 1D Na⁺ pathways [2]), we tested our prediction by building a room-temperature all-solid state sodium-ion battery combining the alluaudite-type cathode with Na₃PS₄ as the solid electrolyte and Na₂Ti₃O₇ as the anode. We also report an alternative synthesis of the alluaudites by mechanical milling of Na₂SO₄ and FeSO₄ followed by annealing at 400°C under argon atmosphere. Favourable rate performance of the all-solid state sodium-ion battery was demonstrated.

[1] S. Adams, R. Prasada Rao, “Understanding ionic conduction and energy storage materials with bond valence-based methods”, Bond Valences (Edited by I. D. Brown et al.), Structure and Bonding 158, pp. 129-160, Springer (2014).

[2] L.L. Wong, H. Chen, S. Adams, “Sodium-ion diffusion mechanisms in the low cost high voltage cathode material Na_2+dFe_2-d/2(SO₄)₃” Phys. Chem. Chem. Phys. 17 (2015), 9186-9193.