Lithium-stuffed garnets such as Li₇La₃Zr₂O₁₂  (LLZO) show promise as solid state lithium ion conductors for use in high energy density storage applications due to their chemical and thermal stability and electrochemical performance versus competing electrolytes. Much work has focused on improving the conductivity of these systems by optimising the lithium content through the introduction of lattice vacancies created by the addition of cation dopants (such as Al³⁺ and Ga³⁺)¹ in the Li sites. The effect of adding cations on the observed Li conductivity depends strongly on the identity and concentration of cations; this process must be tuned such that an optimum number of Li ions and vacancies are present and can readily form an interconnected network of tetrahedral and distorted octahedral sites in the lattice, allowing long-range diffusion. The importance of obtaining the cubic garnet phase which is known to exhibit higher conductivities as a result of increased lithium disorder versus the tetragonal polymorph has been acknowledged and pursued by many.²

In addition, recent observations on the reactivity of the garnets with water have shown that lithium undergoes ion exchange with H⁺, resulting in a change in conductivity and degradation of the material, accompanied by the formation of passivating Li hydroxide and carbonates at the surface^3–5. These phenomena occur even upon ambient air exposure of the garnet samples. Despite this knowledge, little is known about the pathway for lithium and proton diffusion during exchange, and many studies on LLZO garnet materials do not account for the role of this mechanism on the properties reported, with samples freely exposed to an air environment during synthesis and measurement.

In this work, we use a Ge(IV) cation substituted at the lithium position to follow the role of aliovalent dopants on the lithium diffusion behaviour, building on previous work using Zn²⁺ and Ga³⁺. The high charge Ge⁴⁺ ion is chosen to achieve a greater theoretical Li vacancy concentration for the lowest amount of added dopant as a result of charge compensation effects. We have found the critical Li ion to vacancy ratio for optimum ionic conduction in the Ge-doped garnet, using ICP-OES to verify the composition and have measured ionic conductivities of 6-8 x10^-4 S cm^-1 using AC impedance for the material of nominal composition Li_6.6La₃Ge_0.1Zr₂O₁₂. XRD confirms that the cubic garnet structure is stabilised by the addition of Ge.

Moreover, the role of moisture in the degradation of LLZO is systematically investigated, through the use of controlled-atmosphere processing (glove box coupled to a high temperature furnace) and deliberate forcing of D/Li exchange in the lattice by exposing the otherwise ‘dry’ samples (synthesised and stored in a protective argon atmosphere) to deuterated water. We use secondary ion mass spectrometry (SIMS) to follow the depth profiles of ion species, specifically Li⁺, D⁺ and H⁺in the undoped and Ge-doped LLZO and have observed diffusion profiles indicative of the exchange process (Figure 1). Preliminary novel work on isotopic Li exchange is underway, following a similar protocol, to calculate the diffusivity in our samples, and the corresponding effect of dopants, stoichiometry and degradation on the observed diffusion profile.

Altogether, this work is enabling the fundamental properties of this solid state garnet electrolyte to be better understood, identifying and optimising the factors governing conduction, with a view to its use in high energy density Li-metal batteries. A knowledge of the degradation processes occurring on Li exchange with protons could reveal important conductivity information about these two charge carriers, enabling us to harness this seemingly detrimental behaviour to our advantage.

References

1.           Bernuy-lopez, C. et al. Chem. Mater. 26,3610–3617 (2014).

2.           Awaka, J., Kijima, N., Hayakawa, H. & Akimoto, J. J. Solid State Chem. 182,2046–2052 (2009).

3.           Li, Y., Han, J.-T., Vogel, S. C. & Wang, C.-A. Solid State Ionics 269,57–61 (2015).

4.           Liu, C. et al. J. Power Sources 282,286–293 (2015).

5.           Larraz, G. et al. J. Mater. Chem. A 3,5683–5691 (2015).

Figure 1: FIB-SEM image of a cross section of undoped LLZO pellet, exposed to D₂O at 100°C (left); and (right) corresponding depth profile of collected SIMS data of selected negative ion species of part of exposed sample. Species counts are normalised to matrix element oxygen-16 ion.