Locating Impurity Phases in the Lithium-Ion Conductor Al-Doped Li7La3Zr2O12 through Dynamic Nuclear Polarization and Nuclear Magnetic Resonance Spectroscopy

Abstract

An understanding of the nature of the grain boundaries and impurity phases contained in complex mixed metal oxide solid electrolytes is key to the development of improved and more stable solid-state batteries with reduced grain boundary resistances and higher ionic conductivities of the bulk sample. The Li-ion solid electrolyte Li₇La₃Zr₂O₁₂ (LLZO) is one of the most researched electrolytes in the field due to its high ionic conductivity, thermal stability, and wide voltage stability window. Despite its potential, the nature of the impurity and surface phases formed during the synthesis of LLZO and their role and influence on LLZO’s performance when used as an electrolyte remain poorly understood and controlled. In addition, there are limited characterization methods available for detailed studies of these impurity phases, particularly if these phases are buried in or close to the grain boundaries of a dense sintered material. Here, we demonstrate a solid-state nuclear magnetic resonance (ssNMR) and dynamic nuclear polarization (DNP) approach that exploits both endogenous and exogenous dopants to select for either specific impurities or separate bulk vs surface/subsurface phases. Specifically, the location of Al-containing phases within an Al doped LLZO and the impurity phases that form during synthesis are mapped: by doping LLZO with trace amounts of paramagnetic metal ions (Fe³⁺ and Gd³⁺), DNP is used to selectively probe Al- and La-containing impurity phases, respectively, allowing us to enhance the signals arising from the LiAlO₂ and LaAlO₃ impurities and to confirm their identity. A ¹⁷O DNP experiment using Gd³⁺ doped LLZO is performed to identify further La³⁺-containing impurities (specifically La₂Zr₂O₇ and La₂O₃). Finally, a ⁷Li DNP irradiated ⁷Li–²⁷Al dipolar-based heteronuclear multiple quantum correlation experiment is performed by using the radical TEKPol as the polarization agent. This experiment demonstrates that the poorly crystalline LiAlO₂ that is found close to the surfaces of the LLZO composite is coated by a thin Li-containing impurity layer and thus not directly present at the surface.