Sebastian Dierickx

Garnet-type Li₇La₃Zr₂O₁₂ (LLZO) is a potential electrolyte material for all-solid-state Li-ion batteries mainly because of its reported excellent chemical stability in contact with Li metal. But good wettability of LLZO and 100% surface coverage of lithium are still a challenge. This study elucidated the suitability of magnetron-sputtered indium in Li(In)/LLZO/Li(In) symmetrical model cells as one of the promising interfacial modifications reported in the literature. Importance was given to the impact of preparation parameters on the surface coverage of Li(In)/LLZO interfaces and the consequences of impedance, cycling stability, and critical current density. SEM and EDXS analyses of In layers of thickness 100 nm to 1 μm revealed complete dissolution of indium in the lithium anode after annealing; 300 nm In layers annealed at 220 °C/10 h provided a surface coverage of >80%, best reproducibility, and a supreme interface resistance R_int of 12.4 Ω·cm². Presuming a surface coverage of 100%, an ultimate interface resistance close to 1 Ω·cm² can be expected. The critical current density was determined as 200–500 μA/cm² at a charge of 100–250 μAh, whereas 500 μA/cm² and above affected cell stability. The increasing voltage plateau was assigned to the increase of the interface resistance R_int and the electrolyte resistance R_G+GB. SEM, EDXS, and X-ray microtomography analyses after voltage breakdown confirmed Li-dendrite growth along grain boundaries into LLZO, often curved parallel to the interface, indicating short-circuiting of the solid electrolyte. Grain boundary characteristics are supposed to be decisive for lithium deposition in and failure of garnet-type solid electrolytes after cycling.

All-solid-state lithium ion batteries may overcome safety issues in future energy storage systems, when a non-flammable solid electrolyte instead of a liquid electrolyte is applied. Garnet-type Li₇La₃Zr₂O₁₂ (LLZO) is an encouraging choice due to its high Li⁺ conductivity and its chemical stability against lithium metal [1]. However, low-resistance LLZO/Li interfaces are deteriorated by incomplete adhesion of lithium metal onto LLZO and by surface contamination [2]. The resulting inhomogeneous distribution of contact areas may be more prone to dendrite formation. Previous studies [3,4] indicate that LLZO/Li interface quality is significantly improved by metallic buffer layers. In this study, we aim for a reliable preparation of low-resistance interfaces between solid electrolyte and lithium metal. At first, magnetron sputtering is applied to deposit an indium buffer layer of 50 to 500 nm thickness onto the polished and cleaned electrolyte surface. For cell assembly, lithium foil is put on both sides of an In-coated LLZO pellet and the test cell is heated to 220°C in argon atmosphere. After that, microstructure and electrochemical properties of the symmetric cells are analyzed. We improved cross-section sample preparation to inspect the quality of LLZO/In/Li interfaces over hundreds of µm. Scanning electron microscopy (SEM, cf. figure 1) confirms excellent adhesion between lithium and LLZO. We use a dry inert atmosphere during preparation, handling and transfer of our samples to avoid artifacts and sample degradation. The FEI Helios G4 FX microscope enables correlative SEM imaging by simultaneous acquisition of up to four different detector signals for interface topography and material contrast. Local chemical composition at the interface is analyzed by energy dispersive X-ray spectroscopy and electron-beam damage is minimized for SEM imaging by selecting optimum electron energies. Electrochemical impedance spectroscopy (EIS) followed by data evaluation via the distribution of relaxation times (DRT) and equivalent circuit modelling [5] deliver the proportional contributions of grain, grain boundaries and interfaces to the total impedance of symmetric cells (Li/In/LLZO/In/Li). Interface microstructures and EIS measurements are correlated for samples with different buffer layer characteristics to allow targeted optimization. We compare the pristine state of the test cells to samples that have been cycled with different current densities. References [1] S. Ramakumar et al., Progress in Materials Science 88 (2017), 325-411 [2] L. Cheng et al., Physical Chemistry Chemical Physics 16 (2014), 18294 [3] W. Luo et al., Advanced Materials 29 (2017), 1606042 [4] C. Tsai et al., ACS Applied Materials & Interfaces 8 (2016), 10617−10626 [5] P. Braun et al., Journal of Electroceramics 38 (2017), 157-167 Figure 1

In most anode-supported SOFCs the fuel electrode is realized by two porous Ni/YSZ layers, the supporting substrate and the anode functional layer (AFL) with specified microstructure properties. These electrodes, especially the AFL, provide a large number of electrochemically active triple phase boundaries, where (i) the electrochemical oxidation of hydrogen couples (ii) the electronic conduction in the nickel matrix, (iii) the ionic conduction in the YSZ-matrix and (iv) the gas diffusion in the pores. In this study a physical meaningful modelling approach based on a three channel transmission line model was developed in order to describe the coupling of all abovementioned processes. The extension of the electrochemical active zone into the substrate is considered for the first time by investigating AFLs varying from 3 to 22 µm. The model is capable of correlating the microstructure and thickness of the AFL with the performance of Ni/YSZ cermet anodes. Anodes with thicker functional layer revealed a poorer initial performance due to enhanced gas diffusion, whereas the tolerance towards sulfur is increased.