Although garnet-type LLZO has been considered one of the most promising solid electrolyte for solid-state batteries (SSB), the instability of electrode/LLZO interface has been obstacle for the practical application of LLZO into the batteries.[1] Thus, a fundamental understanding of the lithium deposition behavior in the interface would aid in elucidating the underlying mechanisms of the short circuit failure due to unstable interface and designing the interface in SSBs. Herein, we successfully used an in operando microscopy technique to probe Li deposition through the LLZO electrolyte in an anode-free solid-state battery setup. More importantly, we carefully examined the interface with artificial interlayers, which revealed that Li plating is strongly dominated by the kinetics of alloying and precipitation through the metal interlayer. In addition, we confirmed that the interlayer also affects the sequential stripping process, influencing the electrochemical performance of the cell. Supported by these intriguing observations, we propose the dynamic roles of the interlayer during battery operation: as a buffer layer and a seed layer. The in-house cell and microscope system shown in Figure 1a were used for the in operando observations of Li deposition on LLZO. Li foil was attached to the bottom of the LLZO pellet and the top surface was pre-coated with the selected interlayer metal. We investigated electrochemical Li deposition behavior in the absence of an interlayer metal. Li metal begins to appear in island shapes under an applied galvanostatic current (0.1 mA cm-2) and continues to grow over time as shown in Figure 1b. It is worth noting that all of the small Li-metal islands first form at pre-existing defects on the pristine LLZO surface during initial lithiation and Li grows preferably at these islands during the subsequent plating. When the metal interlayer was introduced, the deposition behavior was significantly altered. With 30-nm-thick Au layer, we observed that Au interlayer changed color under electrochemical bias, which is indicative of the formation of a Li-Au alloy. These color changes are followed by the formation of island-type small precipitations. From this observation, we propose that Au interlayer plays the unexpected role of a “buffer layer”, which dynamically functions as a medium for Li redistribution by propagating alloying reactions. Given the proposed new role of the interlayer, we expected the Li-metal deposition behavior to critically depend on the thermodynamic and kinetic properties of the interlayer metal and its alloying nature with Li. Si and Ag were chosen in this study, considering the availability of various alloys and the appreciable Li diffusivities in their alloys. When Si layer is applied (Figure 2a), the color changed during the early stages as gold, but subsequent behavior was noticeably different from that of the Au; Li metal preferentially precipitates at only a few sites, and whisker-shaped Li metal rapidly grows at these sites. On the other hand, in Ag interlayer (Figure 2b), the alloying reaction occurs first, followed by the uniform formation of numerous small nuclei. Interestingly, the two reactions occur nearly simultaneously, implying that both the alloying reaction and precipitation in the alloy are so fast that the sequential processes are unable to be distinguished. Regarding on this, the metal interlayers can be regarded as seed matrix for lithium precipitation. Further, Ag interlayer is more reversible upon Li stripping than Au or Si. The better efficacy of the lithium deposition/stripping process with Ag interlayer was further validated by comparative electrochemical testing of two electrode cells constructed with ‘100-nm-thick metal layer│LLZO │Li’ configuration. The results show that the coulombic efficiencies of cells using the various interlayers are closely related to the Li deposition and stripping behavior. The cell with Ag layer, which induces the most-uniform and reversible Li deposition among the tested metals, exhibits the highest efficiency of 71%, while that with Si layer shows 28%. Thus, it indicates that the interlayer can significantly affect the electrochemical performance of anode-free SSBs that employ solid electrolytes by regulating lithiation and de-lithiation behavior at the interface. Considering the ease of interlayer deposition on the LLZO surface and its wide applicability, we expect that our findings will provide useful guidelines for securing optimal interfaces for SSBs. REFERENCES [1] Aguesse, F.; Manalastas, W.; Buannic, L.; Lopez del Amo, J. M.; Singh, G.; Llordés, A.; Kilner, J., ACS Applied Materials & Interfaces 9, 3808-3816. (2017) Figure 1

Recently there has been great interest in porous lithium metal electrodes for rechargeable batteries, which have demonstrated remarkable improvements in cycle life and charge-rate capability compared to planar lithium metal [1]. In this type of cell design, the porous electrode consists of a porous solid electrolyte with void space in the discharged state. During charge, the metal plates onto the surface of the solid electrolyte, filling the pores. During discharge, the pores empty as lithium reacts at the surface of the electrolyte to produce lithium ions which travel through the porous electrolyte. The electrolyte is rigid, so its shape remains fixed over the life of the cell. This design is advantageous because the solid-electrolyte interphase (SEI) is stationary on the surface of the electrolyte, and therefore there is little capacity fade over cycling, because there is no volume change to cause cracks in the SEI. The design also allows space for the metal to deposit into, thereby reducing the cyclic stress of depositing metal pushing against the separator. By reducing the stress of the lithium pushing against the separator, the risk of a dendrite growing through a defect in the separator is reduced. Here we present an electrochemical model for a cell with a porous anolyte based on porous-electrode theory [2], and apply the model to optimize the design of the porous anolyte to maximize cycle life and safety. For example, we find that, because the electronic conductivity of lithium metal is over 7 orders of magnitude higher than the ionic conductivity of present state-of-the-art solid electrolytes, electronically conductive additives provide no benefit for the porous electrode. As the lithium plates onto the electrolyte surface (initially at the interface with the current collector), the plated lithium itself provides a sufficiently electronically conductive network. Furthermore, if the area-specific resistance of the porous electrode is too low, the current distribution can be highly nonuniform, which can have deleterious consequences for cycle life and safety. Finally, we discuss the implications of the model for the necessary mechanical properties of the solid electrolyte to prevent dendrites even in the presence of manufacturing tolerances. References [1] C. Wang, Y. Gong, B. Liu, K. Fu, Y. Yao, E. Hitz, Y. Li, J. Dai, S. Xu, W. Luo, E. D. Wachsman, and L. Hu, NanoLetters 2017, 17(1) p. 565. [2] J. Newman and K. E. Thomas-Alyea, Electrochemical Systems, 3rd edition, Hoboken: John Wiley & Sons, Inc., 2004.

Lithium forms ordered stages when it reacts with graphite. These stages have distinct colors; therefore, optical microscopy gives direct information about the lithium concentration in the graphite. Here we present in situ optical images during charging and discharging of a graphite electrode. Stages are observed to coexist with each other even after extended rest. There is considerable spatial nonuniformity on the microscale. To predict this concentration distribution, we employ a model which combines porous-electrode theory and Cahn-Hilliard phase-field theory to describe the flux of lithium within the graphite. The model closely matches the experimental voltage and concentration distribution. The spatial nonuniformity can be approximated with a relatively simple model of distributed resistances. Finally, we discuss the implications of using the phase-field model instead of a solid-solution model for prediction of lithium plating. The two models give similar predictions of cell voltage and risk of lithium plating under many operating conditions, with the main difference being the relaxation of concentration gradients within particles during rest. The distributed-resistance model shows a higher risk of lithium plating because well-connected particles are overworked as their more-resistive neighbors require a higher driving force for passage of current. (C) The Author(s) 2017. Published by ECS. All rights reserved.

A two-dimensional, transient model of the full-cell sandwich has been developed in order to predict water redistribution during the freezing process. Critical factors affecting the magnitude and direction of water movement are found to be the location of the cell within the stack, the permeability of the porous media of the cell, and the high thermal conductivity of the rib relative to the gas channel.

2 ) 1 ( - n n independent transport properties are necessary to characterize transport completely in a system of n different species. Each necessary transport property measurement corresponds with a binary interaction parameter, Dij between any two species i and j. The table below indicates the interaction parameters that exist between a cation, anion and three solvents (indicated by the subscripts 0, 1 and 2). The three measurable transport properties: the transference number, conductivity, and diffusion coefficient, are functions of the binary interaction parameters of the system, as well as concentration and temperature. Only for a three-species system do these three transport properties fully characterize system transport.