Symposium F.EN03—Overcoming the Challenges with Metal Anodes for High-Energy Batteries

Our team uses an architectural design metaphor to redesign materials and electrodes relevant to electrochemical energy science. Because batteries must balance multiple functions (molecular mass transport, ionic/electronic/thermal conductivity, and electron-transfer kinetics), wiring these functions in three dimensions can move established battery chemistries in new directions. By designing metallic zinc into a sponge form-factor, with co-continuous void and solid, zinc-based alkaline batteries are now rechargeable at high rate and to deep utilization of zinc — without forming separator-piercing dendrites. We solved the historic lack of rechargeability of the zinc anode in alkaline electrolytes because the local current density throughout the sponge never broaches that necessary to launch dendritic growth. All thanks to an inner core of metallic zinc that electronically wires the 4 m² g^–1 of zinc surface area while the pore structure retains the zincate/ZnO discharge product nearby for recovery on the charge step. The nonperiodic, metallically interconnected “sponge” form factor distributes current more uniformly than ad hoc powder composites and ensures extensive charge-discharge cycling without deleterious dendrite formation.

In past decade, because of ultrahigh capacity and lowest reduction potential, Lithium (Li) metal anode has regained great interest aiming to meet the urgent demand for high energy density rechargeable batteries. However, the dendritic growth of Li during cycling causes serious safety concerns and fast performance decay, which primarily restricts its application. Among various strategies about dendrite suppression, the design that introducing Li in a scaffold to fabricate a 3D Li structure is widely accepted. Therefore, the structural features of the scaffold dominate the behavior of Li metal anode. Herein, low-tortuosity, vertically aligned, and well-spaced TiO₂ nanotube arrays with ultrauniform curvature are selected as the host for Li metal anode, and the effect of the surface curvature and nanotube spacing on Li deposition are studied from experiment to theory. The Li metal behaves a unique electrodeposition manner that Li metal homogeneously deposits on the nanotubular wall, causing each Li nanotube to grow in circumference without obvious sign of dendritic formation. The uniform curvature of the anode structure, as well as the space confinement, are the keys to achieve highly stable, dendrite-free Li metal anodes. When this novel Li metal anode was coupled with LiFePO₄ cathode in a full cell, an excellent specific capacity of 132 mAh g^-1 and coulombic efficiency of ~99.85% over 400 cycles are achieved under a high charging rate at 1C. This work sheds light on the effect of structure feature on the stabilization of Li metal anode and takes an inspiring step toward reliable Li metal anode.

The key bottlenecks hindering the practical implementations of metal anodes in high-energy-density rechargeable batteries are the uncontrolled dendrite growth and infinite volume changes during charging and discharging, which lead to short lifespan and catastrophic safety hazard. In principle, these problems can be mitigated or even solved by loading active metals, such as Li, Na and Zn, into a high-surface-area, conductive and porous scaffold. However, a suitable material that can synchronously host a large loading amount of these active metals and endure large current density has not been achieved. Here we report a three-dimensional bicontinuous nanoporous graphene, prepared by nanoporous metal based CVD, as the sought-after scaffold material for metal anodes. The high surface area, large porosity and high conductivity of the nanoporous graphene concede not only dendrite-free stripping/plating but also abundant open space accommodating volume fluctuation of active metals. This ingenious scaffold endows the composite anodes with a long-term cycling stability and ultrahigh rate capability, significantly improving the charge storage performance of high-energy-density rechargeable metal batteries.

Conventional graphite anodes for Li-ion batteries are the most demanding targets for further improvement in order to increase their energy densities. Lithium metal anodes have a much higher energy density but suffer from dendrite formation and volume expansion. To alleviate these problems, three-dimensional (3D) porous carbon scaffolds could be attractive alternatives if lithium metal is deposited in the inside of the porous structures. Nevertheless lithium insertion into porous structures have not been well understood because of difficulties in direct observation during the process. Here we present in-operando study coupled with simultaneously recorded voltage/current profiles for clarifying lithium insertion into porous carbon nanotube (CNT) structures depending on surface characteristics, lithiophobic, lithiophilic, and hybridized lithiophobic/philic surface properties. It was found that lithiophobic graphitic surfaces interfered lithium insertion into the porous media despite the high conductivity along the CNT scaffold. Lithiophilic 3D scaffolds rendered lithium deposited on the outer mouth of the porous electrode, clogging pores and engendering dendrites. This can be attributed to preferential deposition of lithium on lithium metal due to sluggish charge transfer through insulating carboxyl functional groups on graphitic layers. On the other hand, when lithiophilic trenches were created on the lithiophobic graphitic wall of CNT, lithium was uniformly deposited on the surface and then filled up the pores, reaching a high areal capacity of 16 mAh cm^-2 even with a high current density of 8 mA cm^-2 without noticeable dendrite appearance and volume expansion. The functional group near the trench attracts lithium ions into trenches/pores while the preserved graphitic surface allows for charge transfer, giving rise to promising high energy density anode with lithium-impregnated scaffolds by both lithium intercalation and deposition.

The diverse electronegativity among metals provides a rich toolkit to modulate the reactivity of metals via doping different metals while still keeping their metallic behavior for electrochemical purposes. Here we describe a general approach for modulating Li metal properties and enabling stable Li metal batteries by doping with ~0.1 at% Ag or Al. Unlike alloys, our dopant metal (Ag or Al) are atomically dispersed in Li via a non-equilibrium ultrasonication and quenching method during the synthesis and the atomic level dispersion of Ag is evidience by synchrotron X-ray absorption method. The dopants were atomically dispersed in the vacant face-centered sites of the body-centered cubic Li crystals and pull electrons strongly due to higher electronegativity (Ag: 1.98; Al: 1.61 vs. Li: 0.98). As a result, the doped Li anodes have increased work function with reduced chemical reactivity and remained shiny in dry air for months. They also exhibited enhanced Li+/Li redox kinetics and generated thinner but stronger solid-electrolyte interphases in carbonate electrolytes. The dopant atoms are lithophilic and have stronger binding with Li adatoms,
which guide uniform Li deposition and ensures dendrite-free Li interface during battery cycling. Overall, the doped anodes enabled stable operations of not only high current symmetric cells but also practical full cells in which Ni-rich layered cathodes were paired with 30 mm anodes and 7 mL electrolytes. The doping approach is facile and scalable, and opens up new and promising opportunities for designing
practical high energy density metal batteries.

The formation of lithium dendrites has long stood in the way of accessing lithium’s unmatched combination of low reduction potential and high specific capacity for next generation lithium ion batteries. While there has been a body of research directed towards the use of interfacial coatings to improve upon lithium deposition morphologies, one interesting and often overlooked approach to circumvent dendritic growth is the possibility of plating lithium underneath the coating layers. The mechanism of lithium plating underneath offers a refreshing perspective on reframing the question of how to best suppress dendrites and furthermore presents a direction that could greatly benefit from more fundamental understanding.

In this study, we report on a novel mechanism that leads to of plating of lithium underneath a coating layer. A lithium-tin system was chosen as a model coating system due to its ability to form lithium-rich intermetallic which exhibit high Li diffusivity and plays a critical role when considering lithium transport across microns lengths scale. To assess the role of the Lithium-Tin coating on the kinetics of the lithium electrode, Galvanostatic Linear Polarization(GLP) was used to extrapolate the exchange current density using a linear approximation of the Butler-Volmer equation. For the lithium electrode with a Li-Sn coating, the exchange current density of 0.522 ± 0.011 mA cm^-2 was twice as large as that of a bare lithium control indicating that charge transfer kinetics for plating on the lithium tin coated electrodes are more favorable relative to bare lithium. This relationship can also be seen in the large differences in charge transfer resistance observed in impedance measurements.

To study how the enhanced kinetics of the coating layer influences lithium plating, Li-Sn coated and bare lithium electrodes were plated at various current densities for a substantial plating capacity of 4mAh cm^-2 corresponding to ~20 μm of lithium assuming uniform deposition. Unlike bare lithium which always plates on the lithium surface, lithium-tin composite coatings can lead to lithium plating underneath the Li-Sn layer given that there is favorable Li transport kinetics across it. From our studies, there exists a current density ‘transition point’ from which plating occurs below the surface of the coating layer. By combining planar SEM of the surface and cross section of the electrode using Cryo-FIB SEM, we show that current densities below the exchange current density lead to lithium deposition underneath the Li-Sn surface with no evidence of lithium deposits on the surface of the coating. The extension of this model to other systems is presented.

Lithium metal is one of the most promising anode materials for energy dense electrochemical energy storage. However, non-uniform electrodeposition and dendritic growth of Li metal result in low Coulombic efficiency and potential short circuits, which are critical challenges that impede development of practical Li metal batteries. To suppress dendrite growth of Li metal, one strategy is to encapsulate the Li metal with an artificial solid-electrolyte interface (ASEI) layer. The ideal ASEI layer should possess high mechanical strength to prevent the intrusion of Li dendrite, high Li transference number, low bulk/interfacial resistance, and high ionic yet negligible electronic conductivity. Moreover, recent studies show that the morphology of the ASEI layer is critical and that a smooth homogeneous surface is beneficial. A thin film solid electrolyte, lithium phosphorous oxynitride (LiPON), satisfies these requirements. It has a relatively high shear modulus of 7.7 GPa, unity Li⁺ transference number, wide electrochemical stability window (0 - 5.5V vs. Li/Li⁺), ionic conductivity of ~2 × 10^-6 S cm^-1, electronic conductivity of ~8 × 10^-13 S cm^-1, and can fabricated as a smooth isotropic, homogenous thin film via physical vapor deposition. However, it remains a challenge to observe uniform, mirror-like electrodeposition of Li metal under the sole protection of LiPON layers.
Herein we report a multifunctional ASEI design that combines the benefits of a LiPON coating with the enhanced Li morphology conferred by Li_xAu alloys. The 100 nm-thick Au interlayer alloy with Li into a solid buffer layer, which reduces the nucleation energy for Li electrodeposition, resulting in improved Li deposition uniformity. The LiPON layer prevents reactions between the deposited Li and liquid electrolyte. The integrity of the LiPON protective layer was assessed using in operando impedance spectroscopy and ex situ SEM and EDS measurements. At a current density of 0.1 mA cm^-2, it was shown that the LiPON layer remained intact for Li capacities of at least 3 mAh cm^-2. The deposited Li was confined under the LiPON layer creating a smooth, uniform interface, resulting in a shiny Li surface that was “mirror-like.” It was shown that the Au layer is essential to achieving this morphology; discrete, heterogenous Li deposits were observed using Cu current collectors. At higher current densities (up to 1 mA cm^-2) and capacities (6 mAh cm^-2), the LiPON protective coating fractured and nonuniform Li morphologies were observed. Efforts to further optimize this multifunctional ASEI design and engineering application towards long-cycling Li metal batteries will be discussed in the future.

A light-weighting, stable Li metal is critical to high energy-density lithium ion batteries (LIBs). In this presentation, we introduce an inter-layer-calated thin Li metal (ILC-Li) electrode using non-delaminated 2D Ti₃C₂T_x MXene stacks coated on an ultra-thin Li host (Nano Lett. 2020, 20, 2639–2646). The excellent electrical conductivity (higher than graphite) and expanded inter-layer space of about 1.34 nm of the MXene provide a fast e^-/Li⁺ transport to allow Li homogeneous deposition while the layer limits the Li growth along perpendicular direction, thus largely mitigating the dendrite growth. Also, the large surface area of the multi-layer MXene stacks with high mechanical flexibility ensures a high Li capacity and reliable Li nucleation/growth upon plating/stripping. The highly reversible Li deposition/extraction through the confined 2D configuration greatly reduces the dead lithium and electrolyte consumption by forming a thin solid-electrolyte-interphase (SEI) layer. As the result, a small overpotential of less than 135 mV in symmetric cells was achieved after >1050 cycles (>2100 hours) with a very high current density of 10 mA cm^-2 and an areal capacity of 10 mAh cm^-2 while the cell with Li foil showed a larger potential over 400 mV in only 50 cycles with a low current density of 3 mA cm^-2. In a full cell paired with nickel-manganese-cobalt oxide (NMC) or nickel-cobalt-aluminum oxide (NCA) as cathode, the battery with ILC-Li electrode exhibited a clearly improved capacity retention when compared with Li foil, particularly with lean electrolyte conditions of as low as 2.5 μl mAh^-1. The less inorganic-rich SEI layer shown by X-ray photoelectron spectroscopy depth profiling indicates a reduced electrolyte consumption, thus leading to a high energy density up to 366.6 Wh/Kg. In contrast to other coating technologies, the current approach is manufacture scalable by using large sizes of MXene stacks via a facile process, which displays a promising potential in higher energy-density LIBs.

Lithium (Li) metal is an ideal candidate as the anode of Li batteries due to its potential to provide high specific energy. However, Li metal anode is challenging to implement due to several drawbacks. First, Li easily reacts with electrolytes to form a solid electrolyte interface (SEI). The heterogeneous nature of SEI results in local fluctuation of Li⁺ flux and current density, which lead to the formation of dendrites. Second, large volume change during Li stripping/plating creates cracks in the brittle SEI, forms dead Li, and causes further electrolyte consumption. These effects lower the Coulombic efficiency (CE) and devastate the cycle life of Li metal anodes. In order to mitigate the abovementioned degradation pathways, strategies pursued include modifying electrolyte components, utilizing solid electrolytes, employing a shielding cation layer, or chemically pretreating Li metal. Nevertheless, it remains challenging to achieve a stable SEI on Li metal; and therefore, the artificial SEI attracts more attention as an alternative strategy to replace the native SEI on Li.

An ideal artificial SEI must have several key properties. First, previous work proposed benefits of having high modulus coatings on Li; however, our recent work suggested flowability and dynamic property allow SEI to adapt to large volume change during Li stripping/deposition, and result in uniform Li deposition macroscopically. Second, uniform and fast Li⁺ single-ion conduction in the artificial SEI is found to be beneficial to reduce “hot spots”, increase critical Li deposit size, and stabilize the Li metal anode. Finally, the SEI needs to be both chemically and electrochemically inert itself and mitigate electrolyte penetration to minimize deleterious Li-coating and Li-electrolyte side reactions. Nevertheless, few artificial SEIs possess all the desirable properties such as dynamic adaptivity, high ion conductivity, and chemical passivation. Additionally, the majority of the reported Li-metal artificial SEIs are only compatible with ether-based electrolytes, making them incompatible with high-voltage, high-energy-density lithium nickel manganese cobalt oxide (NMC) cathodes that are used in today’s commercial Li-ion batteries.

Herein, we propose an artificial SEI design based on a dynamic polymeric network with high Li⁺ single-ion conductivity. In this network, tetrahedral Al(OR)₄^- (R = soft organic linker) anions are used for the first time as a dynamic motif while providing counter anions for Li⁺. With this dynamic single-ion-conductive network (DSN) as an artificial SEI, we demonstrate over 400 stable plating/stripping cycles in Li||Cu cell using commercial carbonate-based electrolyte. Greater than 85% capacity retention for over 160 cycles in a Li||NMC full battery was achieved using directly coated Li foils and commercial, high active material proportion NMC cathode sheets. This work demonstrates for the first time the promise of dynamic single-ion conductor as a stable and scalable artificial SEI for practical Li metal batteries.

Lithium is an ideal battery anode, with theoretical specific capacity of 3860 mAh/g; replacing conventional graphitic anodes with Li in Li-ion batteries can increase energy density by ~50%. A significant drawback of Li anodes is dendrite formation during cycling, which can lead to short circuiting (a safety hazard and cell death) and to “dead Li,” which drastically reduces cycle life.
A complete dearth of experiments that probe mechanical behavior of electrodeposited Li exists. Most experiments on mechanical properties have focused on thin films and on Focused Ion Beam-carved Li. We developed in-situ experimental methodology that allows to electrochemically charge small-scale battery cells and to observe, in real-time, the formation of Li dendrites and to probe their mechanical response. Experiments reveal: (1) Li nano-deposits are single crystalline and (2) strengths of Li nano-pillars are 16.0 ± 6.82 MPa, which is 24x greater than bulk. This strength enhancement can be explained in terms of the ubiquitous “smaller is stronger size effect” nano-sized single crystalline metals. This work expands the existing strength vs. size property space for Li and helps explain why dendrites can penetrate through much stiffer and harder ceramic solid electrolytes, theorized for several years but never unambiguously demonstrated
Development of high energy density solid-state batteries with Li metal anodes has been limited by uncontrollable growth of Li dendrites in liquid and solid electrolytes. This, in part, may be caused by a dearth of information about mechanical properties of Li, especially at the nano- and micro-length scales and microstructures relevant to Li batteries. We investigate formation, microstructure, and mechanical properties of nano-sized Li electrodeposited in a commercial LiCoO2/LiPON/Cu solid-state thin film cell, grown in-situ in a scanning electron microscope (SEM) equipped with nanomechanical capabilities. Experiments reveal that Li was preferentially deposited at the LiPON/Cu interface, along the valleys that mimic the domain boundaries of underlying LiCoO₂. This pattern of Li formation is consistent with the higher driving force for Li nucleation at the valleys uncovered by electrochemical simulations of Li ion transport in this system. Further charging of the cell resulted in Li nuclei piercing through the 30 nm-thick Cu layer and growing into vacuum through the crack openings. Cryogenic electron microscopy analysis of electrodeposited Li revealed its single crystalline microstructure, and in-situ nano-compression experiments on individual Li nano-pillars with diameters of 360-759 nm revealed their average Young’s modulus to be 6.76 ± 2.88 GPa and an average yield stress of 16.0 ± 6.82 MPa, ~24x higher than what has been reported for bulk polycrystalline Li. We discuss mechanical deformation mechanisms and strength of nano-sized electrodeposited Li in the framework of its microstructure and dislocation-governed nano-scale plasticity of crystals and place it in the parameter space of existing knowledge on small-scale Li mechanics. The enhanced strength of Li at small scales may help explain why Li can penetrate through and fracture much stiffer and harder ceramic solid electrolytes.

Under electrochemical cycling, stress intensification and relaxation within small, confined volumes at the Li/solid-state electrolyte (SSE) interface are thought to be critical factors contributing to mechanical failure of the SSE and subsequent short-circuiting of the device. Nanoindenation has been used to examine the diffusion limited pressure Li can support in the absence of active dislocation sources at high homologous temperatures. Based on the underlying physics of this deformation mechanism, a simple perturbation model coupling local current density, elastic stress and diffusional creep relaxation is introduced. Combining this analysis with the indentation results, it is possible to describe a ‘defect danger zone,’ a length scale which is too large for effective diffusional creep relaxation, but too small for efficient dislocation multiplication. In this instance, the properties of the SSE become critical in controlling localized pressure at the interface.

In contrast to the ‘first generation,’ pioneering analysis of the Li/electrolyte interface by Monroe and Newman, more recent analyses of the Li/SSE interface mechanics examine a similar but uniquely different problem in that they consider non-uniform electrochemical properties and/or filling of morphological defects at the interface during Li charging such that the initial amplitude of the Li protrusion is fixed by a pre-existing interface defect or undulation. Meaningful comparisons between multiple ‘second-generation’ approaches can be challenging. Although they generally address the same problem, assumptions about the interface defect geometry and the self-limiting mechanical and electrochemical behavior of Li place significant limitations on the conclusions and the general applicability of the proposed solution. Despite their differences, the second-generation approaches arrive at generally the same conclusion: Within confined or localized volumes of Li at the interface, Li can support pressures significantly higher than its bulk yield strength (stresses as high as 200 MPa). Such stress concentrations at the interface are of critical concern because they facilitate mechanical failure of the SSE by fracture. The key question then becomes, why is the stress concentration so large relative to the bulk yield strength of Li? Two potential explanations are offered by the second-generation approaches: geometry effects and length scale effects. These effects are not unique to Li. Any metallic anode will experience these effects in a manner that depends on its absolute melting temperature, elastic modulus, yield strength, and self-diffusion coefficient.

Although the length scale effects of high purity Li have only been recently documented, the basic concept of “smaller is stronger” is well documented. As described by Brenner in 1956 and later quantified by Phani et al. in 2013, the yield strength of a metal in small volumes can be described from a purely statistical point of view, as the stress required to plastically flow depends on the probability of finding the necessary dislocations within the stressed volume. Load-displacement data from nanoindentation experiments performed in high purity Li at a homologous temperature of 0.67 graphically illustrate this concept. Similar results obtained in high purity indium, a mechanical surrogate to Li, further illustrate the competition for stress relief. Collectively, an important outcome from these ‘second generation’ approaches is that SSEs with nominally uniform electrochemical properties and flat, smooth surfaces will promote planar, nominally stress-free plating and stripping of Li. Unfortunately, brittle SSE materials will likely experience the same statistical failure issues as structural ceramics. The most robust devices will also implement SSEs that are engineered to provide localized stress relief, potentially through mechanisms such as densification, shear banding and time dependent recovery.

We introduce a new concept to suppress penetration of lithium dendrites through solid electrolytes (SE) by putting the surface of the SE into a state of compressive stress. For a sufficiently high residual compressive stress, cracks have difficulty forming at the surface, and cracks that do form tend to close, inhibiting dendrite penetration. Introduction of compressive surface residual stress solves stress corrosion cracking problems in metals and static fatigue problems in ceramics and glasses. An example of the latter is Gorilla Glass, which is strongly resistant to cracking because of surface compressive stresses. However, the technique will not work if Li ion transport through a SE in compression is substantially reduced. Our DFT and MD calculations for LLZO indicate that the introduction of even high residual compressive stresses (up to 10 GPa) will have only a very modest effect on Li ion transport kinetics, suggesting that the approach is viable. In addition to inhibiting penetration and fracture in solid electrolytes, compressive stresses on the order of GPa will tend to suppress lithium plating inside of solid electrolytes because of an increased overpotential for the Li ion reduction reaction. We discuss and compare several experimental methods for introducing compressive residual stresses at the surfaces of solid electrolytes. We believe that the information here can provide a new avenue for developing high-performance and stable solid electrolytes

Poor stability of Li metal / liquid electrolyte or Li metal / solid electrolyte interfaces leads to chronic problems, such as dendrite formation and capacity loss. The impact of electro-chemo mechanical phenomena on interface stability and dendrite formation are believed to be critical, but are challenging to investigate directly. In-situ curvature measurements are an important tool that can be used to monitor these stresses during battery cycling. Using this technique during Li plating and examining film thickness effects, it is possible to separate contributions from the bulk lithium metal and the solid electrolyte interphase (SEI). These investigations show that the stresses created in the SEI films are much larger than those in the Li metal. Furthermore, these stresses in the SEI can destabilize the electrode / electrolyte interface.
Recently, a variety of methods have been employed to improve the performance and safety of Li metal based batteries by using various artificial SEI layers and/or implementing solid electrolytes. Several of these approaches were also investigated with stress evolution measurements during lithium plating and stripping. This work includes the use of both soft (PEO) and hard (LiF) artificial surface layers. These results indicate that stresses in both the surface film and the bulk lithium evolve differently when artificial SEI layers are employed. Initial attempts to monitor stress evolution with oxide solid electrolytes will also be presented.

Pure lithium in lithium metal batteries is often examined in research projects aiming at improving today's battery technology significantly. But the knowledge about the mechanical properties and the microstructural evolution of pure lithium is incomplete with two particular limitations. First, the available data extensively focuses on the elastic-plastic behavior in some uniaxial loading conditions, the multi-axial plasticity and fracture of pure lithium are rarely taken into consideration. Second, existing studies are mostly performed at the two ends of the length scale axes, either at the nano-scale where dendritic initiation and growth happen or at the bulk-scale where lithium foil is used as an electrode. There is a clear gap at the scale of one single grain or microstructural imperfection. This work aims at providing a systematic study of the multi-axial plasticity and microstructural evolution of pure lithium at the scale of 100 um to 1 mm under various stress states. To manufacture the required complex-shaped experimental specimens for five different stress states including uniaxial tension, notch tension with two different radii, central hole tension, and pure shear, a method that allows safe laser cutting of thick lithium foils in argon atmosphere is developed. The tests are conducted in pure argon as well as in air to quantify the oxidation behavior of lithium. Two active slip systems are observed during deformation and cross-slip is regularly present. Lithium fractures in an extremely ductile manner when the thickness is reduced to zero due to necking. Numerical simulations of the five loading conditions were performed using a multi-axial plasticity model recently developed by Narayan and Anand, which shed light on the full-field deformation and stress information of the samples for fracture examinations. Finally, the developed plasticity model is applied to a lithium-solid electrolyte interface model that couples contact condition, mechanical deformation, and electrochemistry to study the effect of external pressure on the cycling performance of lithium-metal all-solid-state batteries.

While room temperature metallic anodes possess the potential to enable batteries with enormous capacities, their mechanical behavior ultimately plays a key role in the failure modes for both liquid and solid electrolyte battery systems. That is, not only do alkali metal deposits threaten to deform and penetrate the battery separator in liquid electrolyte systems, but the degree of interfacial contact at the solid-state electrolyte/ anode interface possesses significant pressure dependence for solid-state systems. In this talk, we present a comprehensive multi-length scale assessment of the mechanical behavior of lithium, sodium, and potassium metal. Beyond the mechanical properties, we also discuss some practical implications of the anode’s mechanical properties in battery systems.

Zinc is abundant, widely used as structural materials, non-toxic, safe and inexpensive. It has the most negative potential among metals that can be recharged in aqueous electrolyte solutions and high capacity density, and hence is an attractive candidate as a negative electrode material. In alkaline media, the thermodynamically stable discharge product is ZnO, a semiconductor (not insulator), and there is acceptable volume expansion upon its formation (ca. +60%). When anion conductors such as alkaline solutions are used as the electrolyte, as is often the case, the formation of ZnO is at the negative electrode side, which causes tolerable volume changes of the negative electrode when compared to the total metal dissolution and/or deposition at the positive electrode side. The corresponding positive electrode materials can be proton-acceptor types such as NiOOH and oxygen in aqueous alkaline electrolytes. Recent trials include zinc ion acceptors such as organic metal frameworks with zinc ion conductive ionic liquid electrolyte. This presentation mainly focuses on alkaline-based aqueous systems.
The main drawback of the zinc electrode is its insufficient durability, suffered from the morphology changes such as dendrite growth, shape change and densification. This is mostly caused by the formation of highly soluble species Zn(OH)₄^2-, a kinetically favorable discharge product, that diffuses away from the original positions and precipitates wherever it favors. Even if zinc oxide is saturated in the alkaline electrolyte, zinc supersaturation occurs; it stays longer than the time scale of the charge-discharge reactions, even for months. There are several ways to restrict the zinc dissolution, such as complex formation (e.g. calcium zincate [1]), protective film formation (e.g. anion exchange membrane [2]) and the use of electrolyte additives (e.g. pH controller, indifferent salts and organic solvents [3-4]). Though the durability depends on the utilization rates and loaded mass (mAh/cm2), cycle life of more than several hundred cycles be achieved in ideal cases. Mechanical pressure on the electrode (e.g. separator wrapping) is a simple but effective way in restraining the morphology changes. In contrast to electrical charging, mechanical charging is a unique method for air battery systems, mostly for electric vehicles, where soluble zinc species are removed from the cell and regenerated by off-site processes [5].
Passivation during high rate discharging is also problematic, but can be mitigated by constructing composite electrodes with high surface areas. Hydrogen evolution as the self-discharge can be restricted by adding inhibitors such as indium and bismuth.
Analytical methods to clarify the zinc electrode reaction processes are of particular interest, since degradation proceeds with nearly no sign in the electrochemical responses. Since zinc is oxidized in air, in situ (in-cells) or operando (in-current-flow) analyses are favorable. Operando X-ray diffraction (XRD) is helpful in detecting the formation of crystalline ZnO during discharging [6]. XRD mapping analysis indicates how the shape change proceeds [7]. Operando X-ray fluorescence imaging, utilizing the zinc visibility in hard X-ray absorption, is effective in observing zinc behavior with space resolutions [8]. Optical analysis is promising to observe macroscopic changes of the zinc species that significantly affects the electrode durability [9].

[1] R. Jain, et al. J. Appl. Electrochem., 22, 1039 (1992).
[2] K. Miyazaki et al., Electrochemistry, 80, 725 (2012).
[3] A. Nakata et al., J. Electrochem. Soc., 163, A50-A56 (2016).
[4] T. Kakeya et al., J. Power Sources, 407, 180 (2018).
[5] J. Goldstein et al., J. Power Sources, 80, 171 (1999).
[6] A. Nakata et al., Electrochim. Acta, 166, 82 (2015).
[7] A. Nakata et al., APL Materials, 6, 047703 (2018).
[8] A. Nakata et al., Electrochemistry, 83, 849 (2015).
[9] M. Horiuchi et al., Abs. Batt. Symp. Japan, 1E27 (2019).

The electrochemical performance and serviceable life-span of lithium metal batteries (LMBs) are largely governed by the Li metal anode. Particularly, low Columbic efficiency, inferior cycling stability, short life time and potential safety concerns of LMBs are acknowledged to be closely linked to the employed Li anode. As a result, tremendous efforts have been devoted to enhance the electrochemical cycling ability of Li anode and the performance of LMBs have been significantly improved during the last decade. Nevertheless, breakthroughs in propelling the LMB technology from laboratory research to practical commercialization are still missing, which can be mostly attributed to the incomplete understanding of the underlying working/decaying mechanisms of Li anode. To fill in this knowledge gap, a variety of probing techniques, e.g., (Cryo-) TEM, SEM, XPS, NMR, FTIR have been employed to study the underlying working/decaying mechanisms of Li anode during electrochemical cycling, yielding significant amount of valuable insights. Unfortunately, linking these novel insights of the Li anode to the LMB’s performance is still challenging, hence, a comprehensive and in-depth understanding of the correlation between the Li anode and the overall LMB performance remains elusive.

In the current talk, the studies of the morphological evolution of Li anode and its correlation to the LMB’s performance studied by Synchrotron X-ray tomography will be shown. In operando monitoring of the Li-Li symmetrical cells reveals the morphological evolution mechanism of the electrochemically generated/dissolved Li, as well as the underlying internal short circuits. In addition, the effect of the electrochemically generated Li on different types of separators has been also disclosed and it was found that both the mechanical and physical properties of the separator play an important role during Li electrodeposition. Moreover, the close correlation between the morphological evolution of Li anode and the LMB’s electrochemical performance has been established: it was found that the irreversible transformation from original Li bulk to the electrochemically inactive Li electrodeposits results in the observed performance decay. It is hoped that these discoveries could open up new design principles and opportunities to accelerate the practical applications of LMBs.

Lithium (Li) metal anodes have experienced a resurgence of research in recent years, which has been fueled by advances in electrolyte chemistry (both solid and liquid), interfacial engineering, and rational design of electrode architectures^1,2. This has enabled Coulombic efficiency values to push above 99.5%, and cycle life to extend into relevant ranges for transportation applications³. However, while performance metrics are beginning to approach relevant values for consideration of their use in electric vehicles, several fundamental questions remain on how Li metal anodes dynamically evolve during cycling, especially at high current densities. Towards this goal, there is a continued need for new methods to quantitatively measure the evolving morphology, mechanical response, and electrochemical overpotentials of Li metal anodes under realistic conditions.

In this talk, I will discuss recent insights that have been gained on how morphological evolution of non-planar geometries couple with mechanical stresses, both during both electrodeposition and dissolution⁴. To contextualize the critical role of coupled electro-chemo-mechanical phenomena, examples of dynamic Li metal deformation will be discussed during cycling in solid and liquid electrolytes. Specifically, a multi-modal suite of operando analytical techniques will be employed to describe the role of mechanical stresses during the formation of “dead” Li, crack propagation in ceramic solid electrolytes during high-rate cycling, and reversible cycling of Li metal to achieve high Coulombic efficiencies^5–8.

To compliment the cross-sectional perspective provided in our previous operando optical microscopy investigations, recent cell designs that enable plan-view operando of Li metal dynamics will be demonstrated for both liquid and solid electrolyte systems. In liquid electrolytes this allows for new insight into the influence of surface microstructure on nucleation of both dendrites and pits, and the critical role of surface heterogeneity on Coulombic efficiency⁹. In solid electrolytes, plan-view imaging enables direct visualization of Li metal filament propagation in garnet LLZO ceramics, and four distinct filament morphologies are identified¹⁰. Equipped with this fundamental understanding, the talk will aim to address the following critical question in the field: how, when, where, and why does “dead” Li form, and can we control this irreversible Li loss during battery operation?

References
(1) Wood, K. N.; Noked, M.; Dasgupta, N. P. ACS Energy Lett. 2017, 2 (3), 664–672.
(2) Hatzell, K. B. et al. ACS Energy Letters. 2020, 5, 922–934.
(3) Chen, K.-H.; Sanchez, A. J.; Kazyak, E.; Davis, A. L.; Dasgupta, N. P. Adv. Energy Mater. 2019, 9 (4), 1802534.
(4) LePage, W. S.; Chen, Y.; Kazyak, E.; Chen, K.-H.; Sanchez, A. J.; Poli, A.; Arruda, E. M.; Thouless, M. D.; Dasgupta, N. P. J. Electrochem. Soc. 2019, 166 (2), A89–A97.
(5) Wood, K. N.; Kazyak, E.; Chadwick, A. F.; Chen, K.-H.; Zhang, J.-G.; Thornton, K.; Dasgupta, N. P. ACS Cent. Sci. 2016, 2 (11).
(6) Chen, K.-H.; Wood, K. N.; Kazyak, E.; LePage, W. S.; Davis, A. L.; Sanchez, A. J.; Dasgupta, N. P. J. Mater. Chem. A 2017, 5 (23), 11671–11681.
(7) Gupta, A.; Kazyak, E.; Craig, N.; Christensen, J.; Dasgupta, N. P.; Sakamoto, J. J. Electrochem. Soc. 2018, 165 (11), A2801–A2806.
(8) Davis, A. L.; Kazyak, E.; Sakamoto, J.; Dasgupta, N. P.; Garcia-Mendez, R.; Chen, K. H.; Sakamoto, J.; Wood, K. N.; Teeter, G.; Wood, K. N. J. Mater. Chem. A 2020, 8 (13), 6291–6302.
(9) Sanchez, A. J.; Kazyak, E.; Chen, Y.; Chen, K. H.; Pattison, E. R.; Dasgupta, N. P. ACS Energy Lett. 2020, 5 (3), 994–1004.
(10) Kazyak, E.; Garcia-Mendez, R.; LePage, W. S.; Sharafi, A.; Davis, A. L.; Sanchez, A. J.; Chen, K. H.; Haslam, C.; Sakamoto, J.; Dasgupta, N. P. Matter 2020, 2 (4), 1025–1048.

Batteries based on calcium (Ca) metal anodes offer a possibility of improved sustainability and safety with attractive electrochemical metrics. However, Ca battery development is still in its infancy due to basic chemical limitations. When Ca metal comes in contact with an organic electrolyte, a native solid electrolyte interphase (SEI) forms which passivates, and in some instances fully deactivates, the electrode surface. Considering that Ca deposition and dissolution is widely believed to be a surface-film-controlled process, understanding Ca²⁺ transport through this native SEI is paramount. In this talk, characteristic features occurring in cyclic voltammetry (CV) with Ca foils will be discussed as a function of interface composition in a benchmark electrolyte system, Ca(BH₄)₂ in tetrahydrofuran (THF). This electrolyte was found recently by others to support reversible plating/stripping under continuous CV conditions (at 25 mV/s) on planar metal (Au or Pt) working electrodes. As opposed to plating/stripping onto a gold substrate, the first CV cycle on Ca foil reveals a persistent native passivation layer — reflective of the foil preparation process — which manifests as significant reduction and oxidation overpotentials. This native interface can be degraded through the initial electrochemical steps, by processes that depend on whether plating or stripping is first conducted, along with the associated electrochemical protocol. The effect of this first activation step on the resulting electrochemistry will be discussed. Subsequently, the interface evolves to reflect one predominantly governed by CaH₂, which enables further study of the dynamics of this interface under different cycling conditions (scan rate and the imposition of various resting times) on Ca foil. Upon continued cycling, access into the underlying bulk Ca is possible. While this process remains limited by the interface and yields lower current densities than with fresh Ca plating/stripping on Au electrodes, the underlying Ca is not entirely deactivated as concluded by earlier studies, but rather depends on the surface conditioning process and the electrolyte. To further investigate the role of the interface, I will also describe experiments in which nanoscale CaF₂ (~100 nm), an inorganic component found to be present in other Ca SEI, was intentionally formed on Ca foils prior to immersion in an attempt to decouple the foil from the surrounding electrolyte and modulate interface behaviors. Critical differences in interface chemical and electrochemical stability, along with electrochemical performance under CV conditions, will be discussed, and prospects for improving the cyclability of practical Ca foil electrodes will be explored.

Lithium metal electrodes in the presence of liquid electrolytes suffer from unmitigated growth of dendrites that pose safety concerns and lead to electrochemical performance limitation. It necessitates fundamental understanding of the mechanistic origins of interface instability at the mesoscale. This work relies on a detailed mesoscale analysis of the role of disparate mechanisms including electrolyte transport, self-diffusion and electrochemical reaction on the nucleation and growth of electrodeposition. The contrasting influence of temperature on the electrodeposition stability is examined. Finally, the role of mesoscale interactions in devising potential pathways toward electrodeposition stability is elucidated.

Magnesium metal anode has a high volumetric capacity of 3833 mAhcm^-3, rechargeable magnesium battery is expected as an advanced battery system alternate to Li-ion batteries. The SEI-free surface of the magnesium metal leads to the uniform nucleation process resulting in the dendrite free growth during the electrodeposition process. On the other hand, the passivation of the magnesium metal surface initiates the locally-focused current. As a consequence, a risk of dendritic growth of the magnesium metal remains in the electrodeposition process. Here we investigated the nucleation process of magnesium metal on magnesium metal substrates. The initial anodic polarization forms several corrosion pits at the surface of the magnesium substrate. We found that these pits become the preferable sites for the following magnesium deposition. The results suggest that the initial electrochemical activation process is crucial to obtain the uniform plating of magnesium metal.

The challenges with metal anodes for high-energy batteries will likely be overcome as a result of advancements in three research areas: materials structure-property relationships, electrode architectural design and informed operational controls. We have been focusing our efforts on the operational considerations, specifically temperature control, to promote reversible Li plating and stripping. Modulating temperature, both through thermal gradients and transients, has a direct effect on the performance and stability of the electrochemical plating and stripping processes of Li metal. Applying a modest thermal gradient impacts the lithium plating morphology and dictates whether unstable, one-dimensional Li electroplating or stable, more homogeneous electroplating occurs (1). This talk will expand upon our understanding of gradient-influenced plating/strippping behavior and present operational temperature controls necessary for prolonged battery performance and safety. Additionally, thermal regimes applied to improve electrode kinetics and diffusion will show the opportunity to enable fast charging of Li metal and metal-free eletrodes coupled with a LiFePO₄ positive electrode. We use a suite of materials and electrochemical characterization tools (XPS, XRD, optical microscopy, computed tomography and EIS) to guide our understanding of Li metal behaviors in the battery environment.

(1) R. W. Atkinson III, R. Carter and C. T. Love, Energy Storage Materials, 22, 18 (2019).

Increase in the energy density of batteries has been a pressing issue, and metallic lithium is a key electrode material for establishment of such high-energy battery systems. Long history of research has focused on the “dendritic morphology”, but we would still need to continue the basic research for full-scale practical use. Suppressing the dendritic morphology of electrodeposited lithium is so important and surface modification study has been intensively conducted.
In the basic theory of electrodeposition, balance of Li⁺ diffusion and charge transfer kinetics control the morphology and authors confirmed this by lithium plating experiments in ionic liquids [1-3]. Surface modification may adjust this balance or mechanically avoid the growth of dendritic lithium.
We have to consider another point, so called “cross-talk” in the cell. Lithium has the most negative potential among all elements, and electrolytes should be decomposed partly. This is strongly connected to the film formation and then to the cycling efficiency of lithium electrode. Furthermore, the side reaction products may affect the total performance of the cell. In the metal polysulfide/Li cell, a sign of the cross-talk was found and it would lead to the life of the cell [4, 5]. This is quite different from the shuttle mechanism of Li polysulfides in the conventional Li-S cell.

Acknowledgement
A part of this work was conducted by the financial support of JST-CREST and NEDO RISING2 project (JPNP16001).

References
[1] H. Sano, H. Sakaebe, et al., J. Power Sources, 196, 6663-6669 (2011).
[2] H. Sano, H. Sakaebe, et al., ELECTROCHEMISTRY, 80, 777-779 (2012).
[3] H. Sano, H. Sakaebe, et al., J. Electrochem. Soc., 161, A1236-A1240 (2014).
[4] H. Mokudai, H. Sakaebe, et al., to be submitted for publication.
[5] A. Yano, H. Sakaebe, et.al., abstract 3G02, The 87^th ECSJ Spring Meeting, March 2020.

Dendritic growth inside lithium metal batteries is the major issue that prevents the utilization of Li metal anode despite its high energy density. Due to the heterogeneous nature of dendritic growth, the penetrations inside practical batteries occur through few isolated channels, as analysis on multiple postmortem images of the penetrated separators or solid electrolytes advocates. Utilizing the pulled glass capillary cells and COMSOL simulations, we assessed the penetration dynamics focusing on the concentration polarization within the isolated channels with different geometries. We discovered that the varying cross-sectional area has a major impact on the characteristic time for ion depletion on the advancing electrode front, which was previously discovered as the initiation time of dendritic growth. While the ion depletion times still showed a power-law relationship with current densities as the classical Sand equation predicts, the scaling exponent varied, showing a monotonic relationship with the wall curvature near the electrode front. Our findings explain the presence of dendritic growth observed during the system-defined non-transport limited states, as well as deviations of the dendrite initiation time from the classical electro-diffusion equation behaviors. We also propose a new strategic direction for suppressing the dendritic growth and realizing stable lithium metal batteries based on our discovery.

The development of affordable energy harvesting and storage technology has in recent times emerged as a Grand challenge in addressing the irreversible depletion of fossil fuel supplies. It has also emerged as a crucial barrier that must be overcome to lower humanity’s carbon footprint by electrifying transportation and by utilizing renewables-sourced power on the electric grid. One promising solution is to build integrated Photovoltaics(PV)-Battery systems that are high-performing and cost-effective. Currently, the levelized cost of PV energy generation has dropped to $0.05/kWh, which is comparable with gasoline, while the cost of energy storage (i.e. the battery) is over one order of magnitude higher, making the integrated system economically uncompetitive. Zn-ion battery has a low levelized cost of $0.3/kWh relative to vanadium flow battery, Li-ion battery, lead-acid battery and sodium-ion battery($0.7, $0.5, $0.8, $0.7/kWh, respectively).¹ It is obvious that one can further reduce the levelized cost of a battery by prolonging the cycle life in order to maximize its accumulated energy throughput.² Analyses show that, to attain this goal, an at least >99% reversibility much be achieved at the Zn metal anode; however this value is usually <80% in conventional Zn batteries.

In this talk, I will report a type of textured metal anodes that exhibit unprecedented levels of reversibility, > 99.9%. As a first demonstration, I will discuss my recent progress in creating Zn metal batteries that overcome the inherent propensity of the metal to form highly porous microstructures composed of randomly oriented plates. The porous, loosely-connected metal electrodeposits can be only partially reused in the next discharge, leading to a considerable permanent materials lost and the <80% reversibility. Therefore, a compact, planar Zn metal electrodeposition morphology is necessary in order to meet the reversibility requirement. We then show that this characteristic morphology is attributable to the hexagonal close-pack crystal structure of Zn; it tends to expose its close-packed (002) crystal facet, which is parallel to the micro-plates observed. On this basis, we hypothesize that the desirable compact Zn morphology can be obtained by aligning the (002) Zn plates horizontally with respect to the electrode surface. As I will present in the talk, this goal can be fulfilled either by (a) designing an epitaxial substrate that has low lattice misfit with (002) Zn facet,³ or by (b) introducing a hydrodynamic flow normal to the electrode surface.⁴ The resultant (002)-textured Zn electrodes claim >99.9% reversibilities over 10,000 cycles. As a final note, we demonstrate that this concept of textured electrodeposition can be readily generalized to other metals of interest in batteries, e.g. Al.

References
1 Lazard, L. Levelized Cost of Storage Analysis v.1.0 (2015).
2 Zheng, J. et al. Regulating electrodeposition morphology of lithium: towards commercially relevant secondary Li metal batteries. Chemical Society Reviews (2020).
3 Zheng, J. et al. Reversible epitaxial electrodeposition of metals in battery anodes. Science 366, 645, doi:10.1126/science.aax6873 (2019).
4 Zheng, J. et al. Spontaneous and field-induced crystallographic reorientation of metal electrodeposits at battery anodes. Science Advances (in press) (2020).

It is well-known that mass production of aluminum metal is conducted by Hall-Héroult process using high temperature Na₃AlF₆-AlF₃-NaF (or LiF) molten salt system with Al₂O₃. Aluminum metal is one of the most widely used metallic materials for supporting our modern life. However, it is not easy to obtain Al metal without Hall-Héroult process, because Al₂O₃ is thermodynamically very stable as shown in Ellingham diagram and handling temperature for the molten salt is ca. 1273 K. Interestingly, the use of organoaluminum compounds and aluminum halides make it an open possibility to electrochemically deposit Al metal at lower temperature, even below 298 K. Now, chloroaluminate melts consisting of aluminum halides and chemical compounds have attracted attention as Al electroplating baths. Among them, Lewis acidic AlCl₃–1-ethyl-3-methylimidazolium chloride ionic liquid (IL) is mostly employed because of its favorable physicochemical properties including negligible vapor pressure, flame resistance, and high ionic conductivity.¹ Note that Lewis acidic means the AlCl₃ molar fraction is over 50 mol%. In this organic chloroaluminate IL electrolyte, Al metal deposition/stripping reversibly proceeds by the following reaction:

4[Al₂Cl₇]^- + 3e^- ↔ Al + 7[AlCl₄]^-

Similar electrochemical reaction is recognized in a Lewis acidic inorganic 61.0-26.0-13.0 mol% AlCl₃–NaCl–KCl IL (eutectic point: 366 K).² The electrochemical behavior on the Al metal deposition/stripping in these ILs strongly suggests that we can use it for creating Al metal anode rechargeable batteries. Besides, Al metal has several advantages as an anode material, such as very high theoretical capacity (2980 mAh g^-1, 8046 mAh cm^-3), less expensive and mass consumable material, and moderate reactivity in air. Therefore, the number of publications on Al metal anode rechargeable battery (AARB) is increasing for this half a decade.^3,4 AARBs commonly use chloroaluminate anions to control both cathodic and anodic reactions. Cathode is a key for designing good AARBs. Currently layered sp² carbon materials, which can reversibly cause the insertion/deinsertion of chloroaluminate anions, [AlCl₄]^- and/or [Al₂Cl₇]^-, are often employed as cathode active materials.¹

nC + [chloroaluminate anion]^- ↔ C_n[chloroaluminate anion] + e^-

Such carbon materials show desirable cathodic behaviors, for example, high coulomb efficiency and rapid charge-discharge. But, their charge-discharge capacities seem to be insufficient (~ 90 mAh g^-1 in Lewis acidic organic chloroaluminate ILs at 298 K;^1,4 ~ 130 mAh g^-1 in Lewis acidic inorganic ones at 393 K²). In this context, novel high capacity cathode active materials for AARB are being developed. Recently we found out that the sulfur-carbon composite material (SPEG) synthesized from the mixture of sulfur and polyethylene glycol⁵ show appealing electrochemical behaviors, especially high rate capability and excellent cycling durability, in an inorganic 61.0-26.0-13.0 mol% AlCl₃–NaCl–KCl IL. In this presentation, features of the promising cathodes for the AARBs and impact of the ionic liquid electrolytes on the cathodes will be discussed.

Our research on AARBs is being supported by the MIRAI program (grant number JPMJMI17E9), Japan Science and Technology Agency (JST).

References
1. T. Tsuda, G. R. Stafford, and C. L. Hussey, J. Electrochem. Soc., 164, H5007 (2017) and references therein.
2. C.-Y. Chen, T. Tsuda, S. Kuwabata, and C. L. Hussey, Chem. Commun., 54, 4164 (2018).
3. T. Tsuda, I. Kokubo, M. Kawabata, M. Yamagata, M. Ishikawa, S. Kusumoto, A. Imanishi, and S. Kuwabata, J. Electrochem. Soc., 161, A908 (2014).
4. M.-C. Lin, M. Gong, B. Lu, Y. Wu, D.-Y. Wang, M. Guan, M. Angell, C. Chen, J. Yang, B.-J. Hwang, and H. Dai, Nature, 520, 324 (2015).
5. T. Kojima, H. Ando, N. Takeichi, and H. Senoh, ECS Trans., 75, 201 (2017).

Low-temperature (LT) batteries using sodium (Na) metal anodes are highly promising for diverse applications under extreme climate conditions owing to the high specific capacity and low cost of the electrode materials. Unfortunately, the insufficient ionic conductivity of the electrolyte together with the Na dendrite growth at LT significantly influences the performance of Na metal anodes. Herein, we present a novel electrolyte, which not only possesses a superior ionic conductivity at −20 °C but also can derive a robust SEI to effectively suppress the dendrite growth and electrolyte decomposition over cycling, contributing to the realization of battery operation under a lean electrolyte condition. Accordingly, the Na metal anode exhibits an ultrahigh reversible capacity of 50 mAh cm⁻² at 2 mA cm⁻² rate at −20 °C. Furthermore, in the Na metal-Na₃V₂(PO₄)₃ full-cells, a relatively high capacity of 98 mAh g⁻¹ after 1000 cycles and stable Coulombic efficiency over 99.6% at 2C rate are achieved at −20 °C.

The commercial viability of lithium-sulfur batteries is mainly caused by the poor plating and stripping efficiency of the lithium-metal anode. Lithium metal tends to deposit in mossy/dendritic morphology with high surface area. The liquid electrolyte irreversibly decomposes on the reducing lithium surface, consuming both the lithium and electrolyte supply in the cell. Nevertheless, the mechanisms underlying loss of lithium inventory and the role of the lithium-electrolyte interfacial layer (SEI) remain unclear. A clear mechanistic understanding of lithium degradation with cycling in Li-S batteries is necessary to develop strategies towards stabilizing lithium deposition in this unique system.

Here, the anode-free full cell configuration, which combines a fully lithiated Li₂S cathode with a bare nickel foil as the current collector, is used to study the dynamics of lithium deposition. The cycle life of anode-free full cells is dependent on the active lithium inventory in the system, as there is no excess lithium in the system (N/P ratio = 1). By deconvoluting the cycling regimes limited by lithium and sulfur, the lithium inventory loss rate per cycle can be quantitatively estimated. This parameter provides a useful and reliable descriptor of lithium degradation with cycling in cells with a limited lithium inventory.

Analysis using time-of-flight secondary ion mass spectrometry (ToF-SIMS) of the deposited lithium in the anode-free full cells showed that the integrated intensity for Li₂^-, which is indicative of metallic lithium, displayed almost no change with cycling. Thus, a poor correlation between available lithium inventory and amount of metallic lithium detected is discovered. This indicates that the loss of lithium inventory is primarily due to the formation of electrochemically inactive “dead” metallic lithium (Li⁰), and not due to undesirable side reactions with the electrolyte forming SEI components (Li⁺). ToF-SIMS also revealed the growth of a thick layer of electrolyte decomposition products on the surface of the deposited lithium. Since the decomposition products are not ionically conducting, the metallic lithium underneath is rendered electrochemically inactive or “dead”.

It is also found that various hydrogen-containing decomposition products in the bulk of the deposited lithium, such as LiH and LiOH, show a sharp decrease in concentration with cycling. This may be ascribed to the considerable gas evolution that occurs during cycling. These gases include H₂, CH₄, and C₂H₄. This could play an important role in lithium degradation by blocking off ionic and electronic conduction to the trapped metallic lithium, especially if the gases themselves are trapped in void spaces of the porous lithium deposit.

Thus, the mechanisms that underlie the loss of lithium inventory in Li-S batteries are illuminated by a novel application of the anode-free full cell configuration and analytical tools such as ToF-SIMS. The evolution of the lithium-electrolyte interfacial layer (SEI) and its role in causing lithium degradation is demonstrated. It is hoped that the insights generated in this work can motivate efforts towards enabling stable operation of lithium-metal anodes and long cycle life for energy-dense Li-S batteries.

Na metal is regarded as a promising anode for Na batteries owing to its high specific capacity and natural abundance. However, rapid dendrite growth and low reversibility hinder its practical applications. Building an artificial solid electrolyte interphase (SEI) is an effective strategy to stabilize a Na metal anode, yet it is still deficient at high current density and cycling capacity. Here, two artificial SEI layers were constructed through the analogous reactions between Na metal and pure SnCl₄ liquid (SnCl₄-Na electrode) or diglyme solvent with SnCl₂ additive (SnCl₂-Na electrode) to investigate the key parameters for an excellent artificial SEI. Both simulations and experiments demonstrate that high-uniformity and high Na⁺ diffusion properties of SnCl₄-derived protective layer are significant for stabilizing Na metal anodes under rigorous test conditions. Therefore, the SnCl₄-Na electrode protected by homogenous and high Na⁺ diffusion Na-Sn alloy and NaCl achieves an ultralong cycle life for 4000 h and small voltage hysteresis (~100 mV) with a cycling capacity of 3 mAh cm^-2 at 2 mA cm^-2, which is much better than that of the SnCl₂-Na electrode. Moreover, even with a cycling capacity of 5 mAh cm^-2 at 5 mA cm^-2, the SnCl₄-Na electrode still can stably run for ~1500 h. Benefiting from the durable SEI layer, a stable capacity of ~350 mAh g^-1 is observed on a SnCl₄-Na|FeS₂ full cell for 380 cycles.

Li metal batteries have the opportunity to enable high energy density technologies, including Li-air, Li-S, and all-solid-state batteries. However, Li metal anodes suffer from microstructural growth that leads to a loss of Coulombic efficiency and, in some cases, short circuiting and cell death. The composition and structure of the solid electrolyte interphase (SEI) that forms on the surface of Li metal is not well understood, but is believed to play a key role in controlling, and possibly initiating, Li microstructural growth. Here, we use solid-state NMR spectroscopy to provide a detailed characterization of discrete organic and inorganic species that comprise the SEI on Li metal anodes. We examine the SEI and Li ion dynamics through the interphase on Li metal after galvanostatic cycling/polarization in various electrolyte formulations and correlate these changes with differences in Li microstructural growth patterns. Insight into molecular-level changes in SEI composition and dynamics for distinct electrolyte mixtures allows us to rationalize how specific electrolyte components may mediate Li microstructural growth.

Understanding the battery interface is a key-point for improving the lifetime of Li-ion batteries (LIB) as well as design novel electrode materials, beyond LIB. Here, we investigate the SEI formation on the single-crystal metal facets in an organic aprotic electrolyte, by combining idealized experiments with realistic quantum mechanical simulations of the interface (ab-initio molecular dynamics, AIMD). We discover that the LiF formation, accompanied by H₂ evolution, is caused by a reduction of HF impurities and requires the presence of Li at the interface, which catalyzes the HF dissociation. The measured potential of the electrochemical response for these reduction reactions correlates with the work function of the electrode surfaces and that the work function determines the potential for Li⁺ adsorption. The reaction path is investigated further by electrochemical simulations suggesting that the overpotential of the reaction is related to stabilizing the active structure of the interface having adsorbed Li⁺. Li⁺ is needed to facilitate the dissociation of HF which is the source of protons. Further experiments on other proton sources, water, and methanesulfonic acid show that if the hydrogen evolution involves negatively charged intermediates, F^- or HO^-, a cation at the interface can stabilize them and facilitate the reaction kinetics. When the proton source is already significantly dissociated (in the case of a strong acid), there is no negatively charged intermediate and thus the hydrogen evolution can proceed at much lower overpotentials. This reveals a situation where the overpotential for electrocatalysis is related to stabilizing the active structure of the interface, facilitating the reaction rather than providing the reaction energy. Using these data, we generate neural network potentials to extent the length and time scale of AIMD and to fully understand the implications that these findings have for the SEI formation.

Rechargeable lithium sulfur battery has attracted much attention as a next generation battery due to a high theoretical energy density and low cost. In the discharge process, sulfur is reduced to Li₂S with formation of polysulfides as intermediate species. In the case of conventional organic electrolytes used for commercial lithium-ion batteries, polysulfides dissolve in the electrolytes, leading to the self-discharge by the diffusion of polysulfides from the cathode to anode, namely a shuttle effect. The self-discharge has been reported to be prevented by forming a solid electrolyte interphase (SEI) based on functional additives like lithium nitrate. However, the consumption of the additives spoils the prevention of the shuttle effect for long-term cycling. It has been reported that the solubility of polysulfides is very low in the electrolytes composed of lithium bis(trifluoromethylsulfonly)amide (LiTFSA) and such organic compounds as tetraglyme (G4) and sulfolane (SL). Although the superior performance of sulfur cathodes has been reported in these electrolytes, the applicability of a lithium anode in the electrolytes has not been investigated from a fundamental point of view. In the present study, SEIs formed in LiTFSA-G4 and LiTFSA-SL were characterized by electrochemical impedance spectroscopy (EIS), Fourier transform infrared spectroscopy (FT-IR), X-ray photoelectron spectroscopy (XPS), and transmission electron microscopy (TEM).
The formation of a SEI on a Cu substrate was suggested by holding the potential at 0 V vs. Li|Li(I) in LiTFSA-G4 by EIS. The SEI is considered to grow with the elapse of time for holding the potential at 0 V, indicating the decomposition of the electrolyte was sluggish. The SEI with the thickness of about 100 nm was observed on the Cu substrate after the SEI formation. The SEI was suggested to be composed of the organic components derived from the reductive decomposition of the electrolyte. FT-IR and XPS indicated that TFSA^– was included in the SEI. On the other hand, the formation of LiF was not confirmed by XPS, indicating the electrochemical decomposition of TFSA^– was not predominant while the decomposition of TFSA^– after Ar⁺ etching was found to give a peak assignable to LiF. A similar SEI was also found to be formed on a Cu substrate in LiTFSA-SL with addition of a diluent, 1,1,2,2-tetrafluoroethyl 2,2,3,3-tetrafluoropropyl ether (HFE). The thickness of the SEI in LiTFSA-SL-HFE was thinner than that in LiTFSA-G4, suggesting the poor adhesion and/or partial dissolution of the SEI. The formation of LiF was not confirmed by XPS. These results suggested that the SEIs formed in these electrolytes are composed of the ionic species, which are immobilized in the matrix based on the decomposed products of the organic components of the electrolytes.

Acknowledgment
This study was supported by the Advanced Low Carbon Technology Research and Development of Program (ALCA) of the Japan Science and Technology Agency (JST).

Due to the highly reactive and reductive nature of Li-metal, the surface of Li-metal electrodes is typically covered by a layer of solid electrolyte interphase (SEI), which can be formed by electrolyte decomposition or artificial coating. If the SEI cannot passivate the Li-surface, continuous electron leakage through the SEI will lead to more electrolyte decomposition, SEI growth, metallic Li formation and growth inside of the SEI, all of which contribute to the major durability and safety concerns for Li-metal electrodes. Given the multi-components and mosaic structures of SEI, the extended defects such as grain boundaries and interfaces in SEI are likely to be the electron conduction pathways, as the individual SEI components are wide-bandgap insulators in their single crystalline form. Therefore, in this work, the electronic properties of representative grain boundaries (GBs) of the main SEI components (LiF, Li₂O, and Li₂S) on Li in various electrolytes were investigated via density functional theory (DFT) calculations. The GB energy, GB bandgap, excess electron distribution were computed. The total density of states (DOS) of GBs was then aligned with Li Fermi level and Li⁺/Li⁰ potential level to determine their ability to trap electrons from the Li metal electrode. For a systematic comparison, highly coherent tilt and twist GBs with the same ∑ values (∑3 and ∑5 ) in coincidence site lattices (CSL) model were studied to represent low energy special GBs; and GBs with melted GB region and the same orientation of the special GBs were used to demonstrate the influence of the amorphous GBs.
It was found that all the GB bandgaps were smaller than the corresponding bulk bandgaps, with an order of amorphous GBs < Tilt GBs < Twist GBs < single crystal. Among all the calculated GBs, the coherent Li₂S Tilt ∑3 (-12-1)/[111] GB and the amorphous LiF GB showed empty electronic states lower than the Li⁺/Li⁰ depositing level, suggesting that those GB states can trap electrons from Li metal, leading to nucleation of metallic Li inside of the SEI and contributing to the Li dendrite growth. Structural analysis revealed that more under-coordinated atoms in the GBs caused larger bandgap drops and more excess electron localization in the less dense GB regions. These insights suggested that dense SEI structures such as sharp interfaces and well-ordered GBs are preferred to design a fully electronically passivating SEI. Furthermore, to overcome the size limitation of DFT calculations for complex GB structures and facilitate the multiscale simulations, we proposed an analytical model to extrapolate the bandgaps of special GB structures based on the DFT-computable bulk and surface bandgaps. Our predictions of GB bandgaps can be used as inputs for phase field models to simulate Li dendrite growth morphology in both solid and liquid electrolytes.

The promise of Li metal as one of the best imaginable anodes for the next generation of advanced batteries remains intact. However, this promise becoming a reality is not yet possible because of the various issues associated with its high reactivity. Obviously, the immediate expectation is that a great passivation layer will mitigate the problem. Indeed, the immediate contact of Li metal with a liquid electrolyte leads to the spontaneous formation of a “solid-electrolyte interphase” (SEI). The SEI results from the redox reactions where Li becomes oxidized and the electrolyte components become reduced. In this talk, we will discuss how the various anode and electrolyte properties affect such SEI. We have carried out classical reactive molecular dynamics simulations and density functional theory calculations for many different anode surfaces and various electrolyte compositions. As a result, we have observed nucleation and growth of the SEI under various important conditions. The conclusions are very important to understand the evolution of Li metal anodes during battery cycling.

The spontaneously formed and evolved interphase present on a metal anode dictates the Coulombic efficiency, capacity, rate, and lifetime (cycle and calendar) properties of this electrode. Ideally, we would design electrolytes with a detailed knowledge of interphase function to achieve optimum anode performance. Such knowledge is currently lacking for the more energy dense multivalent metal anodes, such as Mg and Ca, cycled in electrolytes that exhibit promise for viable, full cell chemistries. In this presentation, we describe interphase characterization of Mg metal anodes, where Mg serves as a test case for correlating interfacial phenomena with energy storage performance and morphological evolution. We show that Mg exhibits a range of electrolyte dependent behaviors from classic electrocrystallization, marked by near unity Coulombic efficiency and electrochemical epitaxy, to varying degrees of deposit density and crystallography determined by the interphases present at the anode surface. We employ transmission electron microscopy based on a cell to microscope approach to define the role the interphase plays in controlling metal grain size/shape, crystallography and microstructure with cycling and we relate these attributes to the interphase composition and structure. We demonstrate dramatic changes in deposition and dissolution morphology are possible with subtle changes in electrolyte composition, including controlled levels of trace water, added co-ions, and added co-solvents. Our results suggest a path exists toward directing interphase formation based on enhanced electrolyte design.
This work was supported by the Joint Center for Energy Storage Research, an Energy Innovation Hub funded by the U.S. Department of Energy. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.

Lithium metal anodes are of interest due to the high theoretical capacity of lithium (3800 mAh/g) making the lithium anode ideal for use in high energy density applications. Self-discharge is apparent when an electrochemical cell loses capacity at rest, subsequently lowering the energy density. In Li-ion batteries, this can represent up to a 100 % loss in capacity.¹ New Li-ion chemistries using silicon anodes also show evidence of self-discharge.^2-3 Recent work suggests that lithium metal batteries will experience significant capacity losses when at rest in the charged state.^4-6 One of the primary means of self-discharge identified is lithium metal corrosion. This work evaluates self-discharge phenomena in carbonate and ether-based electrolytes commonly used with lithium metal batteries. It is found that the interplay between the solid electrolyte interphase (SEI) and the chemical passivation layer, both of which are electrolyte dependent, determine the extent of self-discharge observed. Specifically, self-discharge is worsened in some electrolytes by the continual breakdown and repair of the SEI upon lithium electrodeposition and dissolution. The use of inorganic artificial SEIs are also studied to mitigate the capacitive losses.

Supported by the Laboratory Directed Research and Development program at Sandia National Laboratories, a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy's National Nuclear Security Administration under contract DE-NA-0003525.

1. Reddy, T. B., Linden's Handbook of Batteries. Fourth ed.; McGraw-Hill Education: 2011.
2. Kalaga, K.; Rodrigues, M.-T. F.; Trask, S. E.; Shkorb, I. A.; Abraham, D. P., Calendar-life versus cycle-life aging of lithium-ion cells with silicon-graphite composite electrodes. Electrochimica Acta 2018, 280, 221-228.
3. Zilberman, I.; Sturm, J.; Jossen, A., Reversible self-discharge and calendar aging of 18650 nickel-rich, silicon-graphite lithium-ion cells. Journal of Power Sources 2019, 425, 217-226.
4. Lin, D.; Liu, Y.; Li, Y.; Li, Y.; Pei, A.; Xie, J.; Huang, W.; Cui, Y., Fast galvanic lithium corrosion involving a Kirkendall-type mechanism. Nature Chemistry 2019, 11, 382-389.
5. Harrison, K. L.; Zavadil, K. R.; Hahn, N. T.; Meng, X.; Elam, J. W.; Leenheer, A.; Zhang, J.-G.; Jungjohann, K. L., Lithium Self-Discharge and Its Prevention: Direct Visualization through In Situ Electrochemical Scanning Transmission Electron Microscopy. ACS Nano 2017, 11, 11194-11205.
6. Wood, S. M.; Fang, C.; Dufek, E. J.; Nagpure, S. C.; Sazhin, S. V.; Liaw, B.; Meng, Y. S., Predicting Calendar Aging in Lithium Metal Secondary Batteries: The Impacts of Solid Electrolyte Interphase Composition and Stability. Advanced Energy Materials 2018, 8 (26).

Li metal is the ultimate anode of choice to improve the energy and power densities of current lithium ion batteries due to its high theoretical capacity and low electrochemical potential. However, challenges associated with the high reactivity and dendrite growth of metallic Li must be addressed before its practical applications. The development of a solid-state electrolyte is the key to enabling Li anode. In order to achieve this goal, the solid-state electrolyte must have high ionic conductivity, adequate mechanical modulus, good chemical stability and adhesion with metallic lithium, and thin sheet processability.

Our strategy to enable Li metal anode is a composite electrolyte approach. The goal is to develop a composite solid electrolyte consisting of a processible soft polymer phase and a highly conductive hard ceramic phase to meet the challenging combination of physical, chemical and manufacturing requirements.

Here we introduce the fabrication of two types of composite electrolytes. One type is to disperse model Li⁺-conducting ceramic particles (LICGC^TM) within a polymer electrolyte matrix (poly(ethylene oxide)-based). Aqueous spray coating is used to produce thin, large area and dense composite films with high loadings of ceramic. The interfacial resistance for ion transport between the polymer and the ceramic is quantified using a trilayer model and methods to minimize the interfacial resistance is discussed. In this type of composite electrolyte, the main ion conduction pathway is through the polymer matrix. A second type of composite electrolyte involves creating a three-dimensionally interconnected structure of the ceramic with greatly reduced interparticle resistance. The main ion conduction pathway is through the ceramic in this case. Symmetrical cells and full cells using these electrolytes with Li metal will be compared.

Acknowledgements: The development of composite electrolytes was sponsored by the U.S. Department of Energy (DOE), Office of Energy Efficiency and Renewable Energy for the Vehicle Technologies Office’s Advanced Battery Materials Research Program (Tien Duong, Program Manager). Neutron and solid state nuclear magnetic resonance experiments were sponsored by DOE’s Office of Science, Basic Energy Sciences, Materials Science and Engineering Division.

Lithium metal and magnesium metal batteries have potential for high system energy density due to the high gravimetric and volumetric charge capacity of these metal anodes. Magnesium metal offers several advantages and disadvantages with respect to lithium metal as an anode. Magnesium is more abundant and resources are geographically widespread. Magnesium metal is less chemically reactive than lithium metal and electrodissolution/deposition happens at elevated potentials, yet issues with inefficiency and side reactions still remain. While lithium metal polymer batteries have been investigated intently for decades, magnesium metal polymer batteries have received little attention. There exist great difficulties in producing magnesium polymer electrolytes that both effectively transport magnesium and support reversible electrodeposition and dissolution of magnesium metal. In this talk, I will review our recent research on ion transport and electrochemistry in traditional salt-doped magnesium polymer electrolytes and magnesium poly(ionic liquid) electrolytes.

Incorporating lithium metal anodes into solid-state batteries (SSBs) is critical for enabling high energy density and specific energy. However, a variety of chemo-mechanical issues at the solid-state electrochemical interface, including interphase growth, contact loss, and lithium morphology evolution, have hindered development of efficient solid-state lithium anodes. Understanding and controlling the dynamic evolution of lithium interfaces is therefore necessary for further development of SSBs. Here, I discuss my group’s efforts using in situ and operando methods to investigate the dynamic evolution of lithium metal and its interphase within SSBs. Multi-modal in situ investigation of interfacial reactions combined with electrochemical experiments reveal how the formation of the interphase is linked to chemo-mechanical cell failure in NASICON-based lithium metal cells. In situ transmission electron microscopy (TEM) shows that the reaction of L_1+xAl_xGe_2-x(PO₄)₃ (LAGP) with lithium to form an interphase is similar to a conversion reaction, in which lithium insertion causes amorphization and volume expansion of ~130%. In situ X-ray tomography experiments of operating LAGP-based cells reveal that the growth of the interphase causes fracture, which is the underlying cause of impedance increase within the cell. Interestingly, we have found that deposited interfacial layers can alter the growth trajectory of the interphase and extend cycling stability of symmetric cells from ~30 hours with unprotected LAGP SSEs to >1000 hours with protected materials. Additional efforts have been focused on the use of operando synchrotron X-ray imaging to understand interfacial dynamics and overall cell evolution during cycling in SSBs with different chemistries. Finally, I will discuss our results on understanding lithium anode evolution at different temperatures in liquid and solid systems. Overall, these results provide fundamental insight into the evolution of the lithium/electrolyte interface, and they show that control over interfacial transformation processes could enable the development of solid-state lithium metal batteries.

Dendrite formation limits the cycle life of lithium and sodium metal anodes and remains a major challenge for their integration into next-generation batteries, even when replacing the liquid electrolyte by a solid electrolyte. Recent studies have pointed out that lithium and sodium metal creep rather than lithium and sodium diffusion is the primary mechanism for replenishing the voids forming in solid metal anodes at the interface to a solid electrolyte upon stripping. Void formation upon stripping causes current constrictions upon plating, promoting dendrite formation. Here we investigate plating and stripping of liquid sodium metal from a ceramic Na-β"-alumina electrolyte at 250 °C, thereby eliminating mass transport limitations due to creep. Employing a porous carbon electrode coating to (1) prevent dewetting of plated liquid sodium from and (2) to supply liquid sodium upon stripping to the sodiophobic Na-β"-alumina surface, we demonstrate extremely high current densities above 1000 mA/cm² and cumulative plated capacities of 10 Ah/cm² without dendrite formation [1]. These values are two orders of magnitude larger than the corresponding values measured at room temperature [2]. We further show that liquid sodium occupies 60% of the porosity of the carbon coating, while excess sodium accumulates on top of the coating. Increasing the thickness of the carbon coating from 50 μm to 200 μm does not affect the Coulombic efficiency, but reduces the constant current rate capability due to the increased flow resistance for liquid sodium in the porous coating. Our results confirm that eliminating void formation is effective in suppressing dendrite formation.

[1] D. Landmann, G. Graeber, M. V. F. Heinz, C. Battaglia, submitted
[2] M.-C. Bay, M. Wang, R. Grissa, M. V. F. Heinz, J. Sakamoto, C. Battaglia, Adv. Energy Mater. 2019, 201902889

Successful prevention of lithium dendrite growth would enable the use of lithium metal as an anode material in next-generation secondary batteries. Mechanically stiff solid polymer electrolytes have been shown to prolong the life of lithium metal cells by partially suppressing lithium dendrite growth. However, we lack fundamental knowledge about the nature of lithium electrodeposition under different electrochemical and electrolyte conditions, as direct observation of this phenomenon is non-trivial.

Synchrotron X-ray microtomography was used to observe lithium metal plating through these solid polystyrene-block-poly(ethylene oxide)-LiTFSI electrolyte membranes under a range of electrochemical conditions: regimes of lithium plating behavior as functions of current density and electrolyte thickness through the block copolymer electrolyte were identified. Impurity particles in the lithium metal were observed at the nucleation sites of protrusions, as reported in other work. Next, stable lithium plating at low current density was used to treat the lithium layer at the electrode-electrolyte interface to investigate the effect of impurity particles on protrusion nucleation and lithium deposition. We show that this “lithium filtering” strategy increases the cycle life of lithium symmetric cells in an unexpected way.

Lithium metal electrodes in the presence of liquid electrolytes suffer from unmitigated growth of dendrites that pose safety concerns and lead to electrochemical performance limitation. It necessitates fundamental understanding of the mechanistic origins of interface instability at the mesoscale. This work relies on a detailed mesoscale analysis of the role of disparate mechanisms including electrolyte transport, self-diffusion and electrochemical reaction on the nucleation and growth of electrodeposition. The contrasting influence of temperature on the electrodeposition stability is examined. Finally, the role of mesoscale interactions in devising potential pathways toward electrodeposition stability is elucidated.

Lithium metal and magnesium metal batteries have potential for high system energy density due to the high gravimetric and volumetric charge capacity of these metal anodes. Magnesium metal offers several advantages and disadvantages with respect to lithium metal as an anode. Magnesium is more abundant and resources are geographically widespread. Magnesium metal is less chemically reactive than lithium metal and electrodissolution/deposition happens at elevated potentials, yet issues with inefficiency and side reactions still remain. While lithium metal polymer batteries have been investigated intently for decades, magnesium metal polymer batteries have received little attention. There exist great difficulties in producing magnesium polymer electrolytes that both effectively transport magnesium and support reversible electrodeposition and dissolution of magnesium metal. In this talk, I will review our recent research on ion transport and electrochemistry in traditional salt-doped magnesium polymer electrolytes and magnesium poly(ionic liquid) electrolytes.

Magnesium metal anode has a high volumetric capacity of 3833 mAhcm^-3, rechargeable magnesium battery is expected as an advanced battery system alternate to Li-ion batteries. The SEI-free surface of the magnesium metal leads to the uniform nucleation process resulting in the dendrite free growth during the electrodeposition process. On the other hand, the passivation of the magnesium metal surface initiates the locally-focused current. As a consequence, a risk of dendritic growth of the magnesium metal remains in the electrodeposition process. Here we investigated the nucleation process of magnesium metal on magnesium metal substrates. The initial anodic polarization forms several corrosion pits at the surface of the magnesium substrate. We found that these pits become the preferable sites for the following magnesium deposition. The results suggest that the initial electrochemical activation process is crucial to obtain the uniform plating of magnesium metal.

The challenges with metal anodes for high-energy batteries will likely be overcome as a result of advancements in three research areas: materials structure-property relationships, electrode architectural design and informed operational controls. We have been focusing our efforts on the operational considerations, specifically temperature control, to promote reversible Li plating and stripping. Modulating temperature, both through thermal gradients and transients, has a direct effect on the performance and stability of the electrochemical plating and stripping processes of Li metal. Applying a modest thermal gradient impacts the lithium plating morphology and dictates whether unstable, one-dimensional Li electroplating or stable, more homogeneous electroplating occurs (1). This talk will expand upon our understanding of gradient-influenced plating/strippping behavior and present operational temperature controls necessary for prolonged battery performance and safety. Additionally, thermal regimes applied to improve electrode kinetics and diffusion will show the opportunity to enable fast charging of Li metal and metal-free eletrodes coupled with a LiFePO₄ positive electrode. We use a suite of materials and electrochemical characterization tools (XPS, XRD, optical microscopy, computed tomography and EIS) to guide our understanding of Li metal behaviors in the battery environment.

(1) R. W. Atkinson III, R. Carter and C. T. Love, Energy Storage Materials, 22, 18 (2019).

Zinc is abundant, widely used as structural materials, non-toxic, safe and inexpensive. It has the most negative potential among metals that can be recharged in aqueous electrolyte solutions and high capacity density, and hence is an attractive candidate as a negative electrode material. In alkaline media, the thermodynamically stable discharge product is ZnO, a semiconductor (not insulator), and there is acceptable volume expansion upon its formation (ca. +60%). When anion conductors such as alkaline solutions are used as the electrolyte, as is often the case, the formation of ZnO is at the negative electrode side, which causes tolerable volume changes of the negative electrode when compared to the total metal dissolution and/or deposition at the positive electrode side. The corresponding positive electrode materials can be proton-acceptor types such as NiOOH and oxygen in aqueous alkaline electrolytes. Recent trials include zinc ion acceptors such as organic metal frameworks with zinc ion conductive ionic liquid electrolyte. This presentation mainly focuses on alkaline-based aqueous systems.
The main drawback of the zinc electrode is its insufficient durability, suffered from the morphology changes such as dendrite growth, shape change and densification. This is mostly caused by the formation of highly soluble species Zn(OH)₄^2-, a kinetically favorable discharge product, that diffuses away from the original positions and precipitates wherever it favors. Even if zinc oxide is saturated in the alkaline electrolyte, zinc supersaturation occurs; it stays longer than the time scale of the charge-discharge reactions, even for months. There are several ways to restrict the zinc dissolution, such as complex formation (e.g. calcium zincate [1]), protective film formation (e.g. anion exchange membrane [2]) and the use of electrolyte additives (e.g. pH controller, indifferent salts and organic solvents [3-4]). Though the durability depends on the utilization rates and loaded mass (mAh/cm2), cycle life of more than several hundred cycles be achieved in ideal cases. Mechanical pressure on the electrode (e.g. separator wrapping) is a simple but effective way in restraining the morphology changes. In contrast to electrical charging, mechanical charging is a unique method for air battery systems, mostly for electric vehicles, where soluble zinc species are removed from the cell and regenerated by off-site processes [5].
Passivation during high rate discharging is also problematic, but can be mitigated by constructing composite electrodes with high surface areas. Hydrogen evolution as the self-discharge can be restricted by adding inhibitors such as indium and bismuth.
Analytical methods to clarify the zinc electrode reaction processes are of particular interest, since degradation proceeds with nearly no sign in the electrochemical responses. Since zinc is oxidized in air, in situ (in-cells) or operando (in-current-flow) analyses are favorable. Operando X-ray diffraction (XRD) is helpful in detecting the formation of crystalline ZnO during discharging [6]. XRD mapping analysis indicates how the shape change proceeds [7]. Operando X-ray fluorescence imaging, utilizing the zinc visibility in hard X-ray absorption, is effective in observing zinc behavior with space resolutions [8]. Optical analysis is promising to observe macroscopic changes of the zinc species that significantly affects the electrode durability [9].

[1] R. Jain, et al. J. Appl. Electrochem., 22, 1039 (1992).
[2] K. Miyazaki et al., Electrochemistry, 80, 725 (2012).
[3] A. Nakata et al., J. Electrochem. Soc., 163, A50-A56 (2016).
[4] T. Kakeya et al., J. Power Sources, 407, 180 (2018).
[5] J. Goldstein et al., J. Power Sources, 80, 171 (1999).
[6] A. Nakata et al., Electrochim. Acta, 166, 82 (2015).
[7] A. Nakata et al., APL Materials, 6, 047703 (2018).
[8] A. Nakata et al., Electrochemistry, 83, 849 (2015).
[9] M. Horiuchi et al., Abs. Batt. Symp. Japan, 1E27 (2019).

Batteries based on calcium (Ca) metal anodes offer a possibility of improved sustainability and safety with attractive electrochemical metrics. However, Ca battery development is still in its infancy due to basic chemical limitations. When Ca metal comes in contact with an organic electrolyte, a native solid electrolyte interphase (SEI) forms which passivates, and in some instances fully deactivates, the electrode surface. Considering that Ca deposition and dissolution is widely believed to be a surface-film-controlled process, understanding Ca²⁺ transport through this native SEI is paramount. In this talk, characteristic features occurring in cyclic voltammetry (CV) with Ca foils will be discussed as a function of interface composition in a benchmark electrolyte system, Ca(BH₄)₂ in tetrahydrofuran (THF). This electrolyte was found recently by others to support reversible plating/stripping under continuous CV conditions (at 25 mV/s) on planar metal (Au or Pt) working electrodes. As opposed to plating/stripping onto a gold substrate, the first CV cycle on Ca foil reveals a persistent native passivation layer — reflective of the foil preparation process — which manifests as significant reduction and oxidation overpotentials. This native interface can be degraded through the initial electrochemical steps, by processes that depend on whether plating or stripping is first conducted, along with the associated electrochemical protocol. The effect of this first activation step on the resulting electrochemistry will be discussed. Subsequently, the interface evolves to reflect one predominantly governed by CaH₂, which enables further study of the dynamics of this interface under different cycling conditions (scan rate and the imposition of various resting times) on Ca foil. Upon continued cycling, access into the underlying bulk Ca is possible. While this process remains limited by the interface and yields lower current densities than with fresh Ca plating/stripping on Au electrodes, the underlying Ca is not entirely deactivated as concluded by earlier studies, but rather depends on the surface conditioning process and the electrolyte. To further investigate the role of the interface, I will also describe experiments in which nanoscale CaF₂ (~100 nm), an inorganic component found to be present in other Ca SEI, was intentionally formed on Ca foils prior to immersion in an attempt to decouple the foil from the surrounding electrolyte and modulate interface behaviors. Critical differences in interface chemical and electrochemical stability, along with electrochemical performance under CV conditions, will be discussed, and prospects for improving the cyclability of practical Ca foil electrodes will be explored.

Lithium is an ideal battery anode, with theoretical specific capacity of 3860 mAh/g; replacing conventional graphitic anodes with Li in Li-ion batteries can increase energy density by ~50%. A significant drawback of Li anodes is dendrite formation during cycling, which can lead to short circuiting (a safety hazard and cell death) and to “dead Li,” which drastically reduces cycle life.
A complete dearth of experiments that probe mechanical behavior of electrodeposited Li exists. Most experiments on mechanical properties have focused on thin films and on Focused Ion Beam-carved Li. We developed in-situ experimental methodology that allows to electrochemically charge small-scale battery cells and to observe, in real-time, the formation of Li dendrites and to probe their mechanical response. Experiments reveal: (1) Li nano-deposits are single crystalline and (2) strengths of Li nano-pillars are 16.0 ± 6.82 MPa, which is 24x greater than bulk. This strength enhancement can be explained in terms of the ubiquitous “smaller is stronger size effect” nano-sized single crystalline metals. This work expands the existing strength vs. size property space for Li and helps explain why dendrites can penetrate through much stiffer and harder ceramic solid electrolytes, theorized for several years but never unambiguously demonstrated
Development of high energy density solid-state batteries with Li metal anodes has been limited by uncontrollable growth of Li dendrites in liquid and solid electrolytes. This, in part, may be caused by a dearth of information about mechanical properties of Li, especially at the nano- and micro-length scales and microstructures relevant to Li batteries. We investigate formation, microstructure, and mechanical properties of nano-sized Li electrodeposited in a commercial LiCoO2/LiPON/Cu solid-state thin film cell, grown in-situ in a scanning electron microscope (SEM) equipped with nanomechanical capabilities. Experiments reveal that Li was preferentially deposited at the LiPON/Cu interface, along the valleys that mimic the domain boundaries of underlying LiCoO₂. This pattern of Li formation is consistent with the higher driving force for Li nucleation at the valleys uncovered by electrochemical simulations of Li ion transport in this system. Further charging of the cell resulted in Li nuclei piercing through the 30 nm-thick Cu layer and growing into vacuum through the crack openings. Cryogenic electron microscopy analysis of electrodeposited Li revealed its single crystalline microstructure, and in-situ nano-compression experiments on individual Li nano-pillars with diameters of 360-759 nm revealed their average Young’s modulus to be 6.76 ± 2.88 GPa and an average yield stress of 16.0 ± 6.82 MPa, ~24x higher than what has been reported for bulk polycrystalline Li. We discuss mechanical deformation mechanisms and strength of nano-sized electrodeposited Li in the framework of its microstructure and dislocation-governed nano-scale plasticity of crystals and place it in the parameter space of existing knowledge on small-scale Li mechanics. The enhanced strength of Li at small scales may help explain why Li can penetrate through and fracture much stiffer and harder ceramic solid electrolytes.

Under electrochemical cycling, stress intensification and relaxation within small, confined volumes at the Li/solid-state electrolyte (SSE) interface are thought to be critical factors contributing to mechanical failure of the SSE and subsequent short-circuiting of the device. Nanoindenation has been used to examine the diffusion limited pressure Li can support in the absence of active dislocation sources at high homologous temperatures. Based on the underlying physics of this deformation mechanism, a simple perturbation model coupling local current density, elastic stress and diffusional creep relaxation is introduced. Combining this analysis with the indentation results, it is possible to describe a ‘defect danger zone,’ a length scale which is too large for effective diffusional creep relaxation, but too small for efficient dislocation multiplication. In this instance, the properties of the SSE become critical in controlling localized pressure at the interface.

In contrast to the ‘first generation,’ pioneering analysis of the Li/electrolyte interface by Monroe and Newman, more recent analyses of the Li/SSE interface mechanics examine a similar but uniquely different problem in that they consider non-uniform electrochemical properties and/or filling of morphological defects at the interface during Li charging such that the initial amplitude of the Li protrusion is fixed by a pre-existing interface defect or undulation. Meaningful comparisons between multiple ‘second-generation’ approaches can be challenging. Although they generally address the same problem, assumptions about the interface defect geometry and the self-limiting mechanical and electrochemical behavior of Li place significant limitations on the conclusions and the general applicability of the proposed solution. Despite their differences, the second-generation approaches arrive at generally the same conclusion: Within confined or localized volumes of Li at the interface, Li can support pressures significantly higher than its bulk yield strength (stresses as high as 200 MPa). Such stress concentrations at the interface are of critical concern because they facilitate mechanical failure of the SSE by fracture. The key question then becomes, why is the stress concentration so large relative to the bulk yield strength of Li? Two potential explanations are offered by the second-generation approaches: geometry effects and length scale effects. These effects are not unique to Li. Any metallic anode will experience these effects in a manner that depends on its absolute melting temperature, elastic modulus, yield strength, and self-diffusion coefficient.

Although the length scale effects of high purity Li have only been recently documented, the basic concept of “smaller is stronger” is well documented. As described by Brenner in 1956 and later quantified by Phani et al. in 2013, the yield strength of a metal in small volumes can be described from a purely statistical point of view, as the stress required to plastically flow depends on the probability of finding the necessary dislocations within the stressed volume. Load-displacement data from nanoindentation experiments performed in high purity Li at a homologous temperature of 0.67 graphically illustrate this concept. Similar results obtained in high purity indium, a mechanical surrogate to Li, further illustrate the competition for stress relief. Collectively, an important outcome from these ‘second generation’ approaches is that SSEs with nominally uniform electrochemical properties and flat, smooth surfaces will promote planar, nominally stress-free plating and stripping of Li. Unfortunately, brittle SSE materials will likely experience the same statistical failure issues as structural ceramics. The most robust devices will also implement SSEs that are engineered to provide localized stress relief, potentially through mechanisms such as densification, shear banding and time dependent recovery.

We introduce a new concept to suppress penetration of lithium dendrites through solid electrolytes (SE) by putting the surface of the SE into a state of compressive stress. For a sufficiently high residual compressive stress, cracks have difficulty forming at the surface, and cracks that do form tend to close, inhibiting dendrite penetration. Introduction of compressive surface residual stress solves stress corrosion cracking problems in metals and static fatigue problems in ceramics and glasses. An example of the latter is Gorilla Glass, which is strongly resistant to cracking because of surface compressive stresses. However, the technique will not work if Li ion transport through a SE in compression is substantially reduced. Our DFT and MD calculations for LLZO indicate that the introduction of even high residual compressive stresses (up to 10 GPa) will have only a very modest effect on Li ion transport kinetics, suggesting that the approach is viable. In addition to inhibiting penetration and fracture in solid electrolytes, compressive stresses on the order of GPa will tend to suppress lithium plating inside of solid electrolytes because of an increased overpotential for the Li ion reduction reaction. We discuss and compare several experimental methods for introducing compressive residual stresses at the surfaces of solid electrolytes. We believe that the information here can provide a new avenue for developing high-performance and stable solid electrolytes

Poor stability of Li metal / liquid electrolyte or Li metal / solid electrolyte interfaces leads to chronic problems, such as dendrite formation and capacity loss. The impact of electro-chemo mechanical phenomena on interface stability and dendrite formation are believed to be critical, but are challenging to investigate directly. In-situ curvature measurements are an important tool that can be used to monitor these stresses during battery cycling. Using this technique during Li plating and examining film thickness effects, it is possible to separate contributions from the bulk lithium metal and the solid electrolyte interphase (SEI). These investigations show that the stresses created in the SEI films are much larger than those in the Li metal. Furthermore, these stresses in the SEI can destabilize the electrode / electrolyte interface.
Recently, a variety of methods have been employed to improve the performance and safety of Li metal based batteries by using various artificial SEI layers and/or implementing solid electrolytes. Several of these approaches were also investigated with stress evolution measurements during lithium plating and stripping. This work includes the use of both soft (PEO) and hard (LiF) artificial surface layers. These results indicate that stresses in both the surface film and the bulk lithium evolve differently when artificial SEI layers are employed. Initial attempts to monitor stress evolution with oxide solid electrolytes will also be presented.

Pure lithium in lithium metal batteries is often examined in research projects aiming at improving today's battery technology significantly. But the knowledge about the mechanical properties and the microstructural evolution of pure lithium is incomplete with two particular limitations. First, the available data extensively focuses on the elastic-plastic behavior in some uniaxial loading conditions, the multi-axial plasticity and fracture of pure lithium are rarely taken into consideration. Second, existing studies are mostly performed at the two ends of the length scale axes, either at the nano-scale where dendritic initiation and growth happen or at the bulk-scale where lithium foil is used as an electrode. There is a clear gap at the scale of one single grain or microstructural imperfection. This work aims at providing a systematic study of the multi-axial plasticity and microstructural evolution of pure lithium at the scale of 100 um to 1 mm under various stress states. To manufacture the required complex-shaped experimental specimens for five different stress states including uniaxial tension, notch tension with two different radii, central hole tension, and pure shear, a method that allows safe laser cutting of thick lithium foils in argon atmosphere is developed. The tests are conducted in pure argon as well as in air to quantify the oxidation behavior of lithium. Two active slip systems are observed during deformation and cross-slip is regularly present. Lithium fractures in an extremely ductile manner when the thickness is reduced to zero due to necking. Numerical simulations of the five loading conditions were performed using a multi-axial plasticity model recently developed by Narayan and Anand, which shed light on the full-field deformation and stress information of the samples for fracture examinations. Finally, the developed plasticity model is applied to a lithium-solid electrolyte interface model that couples contact condition, mechanical deformation, and electrochemistry to study the effect of external pressure on the cycling performance of lithium-metal all-solid-state batteries.

While room temperature metallic anodes possess the potential to enable batteries with enormous capacities, their mechanical behavior ultimately plays a key role in the failure modes for both liquid and solid electrolyte battery systems. That is, not only do alkali metal deposits threaten to deform and penetrate the battery separator in liquid electrolyte systems, but the degree of interfacial contact at the solid-state electrolyte/ anode interface possesses significant pressure dependence for solid-state systems. In this talk, we present a comprehensive multi-length scale assessment of the mechanical behavior of lithium, sodium, and potassium metal. Beyond the mechanical properties, we also discuss some practical implications of the anode’s mechanical properties in battery systems.