Atomic Scale Simulations of the Solid Electrolyte Li7La3Zr2O12

Abstract

Solid-state electrolytes are attracting increasing attention for applications in high energy density batteries. At present, Li7La3Zr2O12 (LLZO) is one of the most promising Li solid electrolytes due its favorable combination of high conductivity and chemical stability against Li metal. However, there are several challenges that potentially limit the use of LLZO practically. The work presented in this dissertation characterizes several properties of LLZO at the atomic scale using density functional theory (DFT) and molecular dynamics (MD) calculations. Calculations addressing the electrochemical window of LLZO, the impact of exposure to air, elastic properties, grain boundary transport, and potential dendrite formation mechanisms are presented. Firstly, DFT calculations of absolute band edge positions indicate that LLZO is an excellent electronic insulator with an intrinsic electrochemical window of 0 to 4 V vs. Li/Li+. Next, the impact of exposure to humid air is examined. The thermodynamics of Li2CO3 surface layers is characterized, in combination with the bulk protonation of LLZO. The impact on Li ion transport is examined as a function of proton exchange. The formation of surface contamination layers is predicted to reduce the wettability between Li and LLZO, resulting in increased interfacial resistance. Regarding elastic properties, linear elasticity models and the calculated shear modulus suggest that LLZO should be sufficiently stiff to suppress lithium dendrite formation. However, subsequent experimental studies have shown that elastic properties alone are insufficient for achieving dendrite suppression: microstructural features of the solid electrolyte should also be accounted for. Toward this goal, three hypotheses regarding microstructural features are examined. More specifically, we consider the possibility that dendrites can result from focusing of the Li-ion current caused by (i) limited contact caused by surface contamination and poor wetting at the Li/SE interface or (ii) from fast Li-ion migration along GBs; alternatively, (iii) softening in the vicinity of GBs could foster lithium accumulation during plating. Cleaning the surface of LLZO (scenario 1) appears helpful in delaying the onset of Li penetration, but appears to be insufficient on its own, as dendrites are still observed at high current densities. The ‘fast GB diffusion’ hypothesis is tested by calculating the rate of Li-ion migration along three low-energy GBs of LLZO. These calculations reveal that Li transport is generally reduced in the GB region, ruling out the second hypothesis. GB softening could arise from deviations in density and atomic structure near the GB plane. MD calculations indicate that significant softening can occur in the immediate vicinity of GBs. We propose that nanoscale softening attributed to microstructural features such as GB may also contribute to Li penetration of nominally stiff solid electrolytes.