Computational Prediction of Transport Properties in Battery Materials

Abstract

Electric vehicles would benefit from batteries with higher energy densities, longer cycle lifetimes, and enhanced safety. State-of-art Li-ion batteries exhibit specific energy densities near 0.26 kW kg–1, limiting the range of electric vehicles and slowing their adoption. Among future battery chemistries, metal-oxygen batteries and solid electrolyte/Li-metal batteries have attracted attention due to their potential to achieve these performance gains. The first goal of this thesis is to predict the electronic and ionic transport properties in the discharge products of metal-oxygen batteries based on alkali-metal anodes (i.e., Li-O2, Na-O2, and K-O2 batteries). Peroxides (Li2O2, Na2O2) or superoxides (LiO2, NaO2, KO2) are the primary discharge products in these batteries. Cells that discharge to superoxides exhibit low charging overpotentials, while those that discharge to peroxides do not. These differences could arise from a higher conductivity within the superoxide; however, this explanation remains speculative given that charge transport in superoxides is relatively unexplored. Here, density functional and quasi-particle methods are used to assess the electronic and ionic conductivities of metal-oxygen discharge products by calculating the equilibrium concentrations and mobilities of intrinsic charge carriers in Na2O2, LiO2, NaO2 and KO2. All compounds are predicted to be electrical insulators, with band gaps exceeding 4 eV. Ionic conductivity in Na2O2 is mediated by negative sodium vacancies, while it is governed by positive oxygen dimer vacancies in lithium, sodium, and potassium superoxides. The predicted ionic conductivities of the superoxides range from 9×10−12 to 4×10−9 S/cm. These values are 8 to 10 orders of magnitude larger than those in lithium and sodium peroxide (9×10−19 to 5×10−20 S/cm). Electronic transport in the peroxides and superoxides is mediated by the hopping of polarons localized on O2 dimers. The predicted equilibrium electronic conductivity in LiO2, 9×10-12 S/cm, is 8 orders of magnitude larger than in Li2O2, Na2O2, NaO2, and KO2 (10-19 to 10-20 S/cm). The moderate conductivity predicted for LiO2 may explain the low overpotentials observed in LiO2 cells. However, given that high conductivities are not predicted for NaO2 or KO2, the enhanced efficiency of these systems should not be attributed to enhanced charge transport; other factors, such as a reduced tendency for electrolyte decomposition, likely explain the small overpotentials in these systems. A second goal of this dissertation is to assess the impact of transition metal (TM) impurities on the performance of Li7La3Zr2O12 (LLZO) solid electrolytes. These impurities are formed by crossover from Li-ion cathodes during interface formation. The presence of TMs in LLZO is hypothesized to impede Li-ion migration, however, the mechanisms responsible for this effect are not understood. Molecular dynamics simulations were used to evaluate the transport rates of Co and three other TMs (Mn, Fe, and Ni) in Al-doped LLZO, and to predict how TM impurities impact Li-ion migration. Fe impurities are the most mobile of the TMs investigated; nevertheless, all TMs exhibit lower diffusivities compared to Li. Importantly, the presence of TMs slows Li-ion migration, with the magnitude of the slowing following the same trend as the TM diffusivities. Because the TMs also migrate along the Li-sublattice, slower-moving TMs impede Li-ion migration via a traffic-jam process. Our work highlights a tradeoff associated with the synthesis of LLZO/cathode solid interfaces: although high-temperature processing increases interfacial contact, and lowers impedance, the use of high temperatures also increases TM crossover from cathode to solid electrolyte, reducing Li-ion mobility.