Anode-less all-solid-state batteries: recent advances and future outlook

The burgeoning electric vehicle (EV) industry has intensified the demand for secondary batteries with higher energy densities and enhanced safety [1, 2]. In conventional lithium-ion batteries (LIBs), graphite has long been counted as an auspicious anode material owing to its low reduction potential, excellent reversibility, and high electronic/ionic conductivity [3–5]. Nevertheless, extending the driving distance of EVs per charge would require the energy density to be increased beyond the range of commercial LIBs. Along this direction, the exploitation of new anode materials and structures has drawn much attention from the community [6–9]. From the viewpoint of the cell configuration, in particular, the anode-less structure has been considered to be the most appropriate with regard to the energy density because the non-necessity for an active material lessens the electrode volume to the greatest extent. Note that the anode-less system has been investigated in LIBs by modifying the current collector or engineering the electrolyte [10–13].

With respect to safety, the fire hazard posed by an EV battery pack in which series of cells are densely packed has been known to be more difficult to address compared to conventional combustion engines. Upon ignition, an EV battery pack is susceptible to flames rapidly spreading to surrounding packs and other accessories [14], as the cells in the neighboring packs readily satisfy the three conditions for ignition: the presence of oxygen, heat, and fuel. Because it is mostly infeasible to exclude oxygen and heat from the battery system, the attention has naturally focused on the flammable battery components in commercial LIBs, that is, carbonate-based liquid electrolytes. This has raised the question [15]: can these electrolytes be replaced by non-flammable ones without sacrificing the key electrochemical properties? The rising demand for a solution to this problem has resulted in the emergence of all-solid-state batteries (ASSBs) that adopt intrinsically incombustible solid electrolytes (SEs) as alternatives to liquid-based LIBs.

In the field of ASSBs, although the battery community has identified various classes of highly functional SEs [16], numerous researchers have also paid attention to the cell configuration, which closely determines the electrochemical performance and cost. In particular, the anode-less concept has been integrated with various SEs [17, 18] because the anode-less structure would be the most suitable approach to address the shortcoming of ASSBs with regard to the energy density; the implementation of an SE layer would inevitably occupy more volume than a separator membrane whose thickness could be aggressively decreased to the level unattainable by an SE layer [19]. From the perspective of the energy density, the anode-less electrode structure does not require any dead volume in the crystal lattice and inter-particle space in which to store lithium (Li) ions in contrast with its conventional LIB electrode counterpart (figure 1). Additionally, anode-less cells simplify the manufacturing process and lower the production cost. For reference, the use of Li metal foil as an anode could also largely address the energy density issue of ASSBs, but a low cost, scalable production scheme for Li metal foil has remained elusive [20].

Zoom In Zoom Out Reset image size

This perspective discusses research trends and the future outlook for anode-less all-solid-state batteries (ALASSBs). We organized this perspective according to the type of SE, and hope that our contribution will inspire members of the research community to enhance their understanding of the operation and degradation mechanisms of this emerging technology to develop innovative solutions.

SSEs have emerged as the most promising SE options owing to their superior ionic conductivities. SSEs, including glass-ceramics Li₇P₃S₁₁ (LPS) [21], argyrodites Li₆PS₅X (LPSX, X = Cl, Br) [22], and Li₁₀GeP₂S₁₂ (LGPS) [23], were demonstrated to maintain high ionic conductivities and consequently deliver decent cell performance when integrated in a bulk-type cell. Besides the high ionic conductivity, SSEs have ductile mechanical properties [19, 24], enabling them to sustain relatively low interfacial resistance during cycling, a key requirement for most ASSBs. In addition, SSEs showed high performance even when Li metal was integrated, offering high potential in terms of energy density [25–28].

Lee et al reported an ALASSB with an argyrodite LPSCl SSE that exhibited high cyclability over 1000 cycles [29]. By coating the anode current collector with a thin film of silver-carbon (Ag-C) composite, the loss of energy density was minimized and the typical drawback of the anode-less system such as uneven Li plating/stripping was overcome to a great extent based on the mechanism described below.

An anode-less system neither contains an active material to act as a reservoir nor excess Li such as metallic Li foil or powder. This structural feature can impair the Coulombic efficiency (CE) of each cycle due to various undesired interfacial and internal events, which collectively cause the chronic problem of poor cycling stability [30]. These events include the growth of dendritic Li, parasitic side reactions of the electrolyte, and the entrapment of Li at defective sites and reactive functional groups. Many researchers have made diverse efforts to mitigate these drawbacks, of which the integration of a thin protective layer on the anode current collector has had the most remarkable effect [31, 32]. For example, the protective layer comprising the Ag-C composite induces uniform Li plating and stripping for prolonged cycling by taking advantage of the lithiophilicity of Ag [33, 34]; during charging, the Ag nanoparticles act as seeds for Li nucleation, forming a solid solution with Li, which leads to stable and homogeneous Li deposition. In addition, the carbon layer serves as a protective film that limits the physical contact between Li and the SSE layer, which suppresses the formation of uncontrolled Li dendrites (figure 2(a)).

Zoom In Zoom Out Reset image size

Specifically, Lee et al incorporated a Ag-C protective layer in a pouch cell and cycled it at a pressure of 2 MPa (figure 2(b)), and the cell delivered excellent performance with an average CE of 99.8% while maintaining capacity retention of 89% for 1000 cycles (figure 2(c)) [29]. This discovery received extensive attention as the authors demonstrated the practical viability of ASSBs with commercially meaningful cycling performance and energy density. Suzuki et al [35] subsequently reported strategies that entail the use of metal-carbon composite coatings for sulfide-based ASSBs. They found sufficient charging and discharging to be possible even when the protective layer consists only of carbon without the presence of a metal (figure 2(d)). Interestingly, the effect of a carbon-only protective layer also varied depending on the type of carbon that was used. In the case of amorphous carbon such as carbon black, Li repeatedly underwent reversible plating and stripping below the carbon protective layer. However, when graphite was used, only the Li corresponding to the reversible capacity of the graphite was intercalated. Continuation of the plating beyond the reversible capacity of graphite resulted in the deposition of Li on top of the protective layer (figure 2(e)), accompanied by immediate internal short-circuiting. These results clarified the importance of depositing Li underneath the protective layer during charging. At the same time, Li diffusion through the carbon interface is the key driving principle of the carbon-based protective layer in securing stable Li plating underneath. Proceeding beyond the exclusive use of carbon, the combination of various metals such as silver (Ag), zinc (Zn), aluminum (Al), and tin (Sn), with carbon (C) confirmed that Ag-C exhibited the highest performance. The highest performance of Ag is attributed to its lithiophilicity, resulting in the lowest overpotential during alloy formation compared to other alloy metal counterparts [34].

As such, a number of recent studies on sulfide-based ALASSBs adopted a carbon-metal layer to prevent direct physical contact with the SE, and metals capable of forming an alloy with Li were mostly chosen. One notable feature that needs to be taken into account when designing the protective layer is that the formation of the Li-metal alloy is accompanied by an immense volumetric change. With this issue in mind, Oh et al attempted to use a highly elastic binder, namely 'Spandex' (figure 2(f)) [36], to replace the conventional poly(vinylidene fluoride) (PVDF) binder. PVDF does not alleviate the volume change Ag undergoes during the alloying reaction with Li, thereby leading to the formation of additional empty space underneath the protective layer and, eventually, the accumulation of dead Li. This means that the Li trapped beneath the protective layer does not return to the cathode during discharge and, instead, contributes to irreversible capacity (figure 2(g)). The replacement of PVDF with the highly elastic polymeric binder that can strongly interact with the polar surface of the metal overcame this problem by preventing the formation of empty space and thus improving the reversibility of the electrode.

Park et al adopted nickel powder coated with carbon and silver for the protective layer [37]. At the same time, they regulated the internal voids in terms of size by packing the nickel particles with controlled sizes and distributions, which led the authors to learn that the smaller voids result in more uniform penetration of Li metal via the coble creep mechanism. This nickel-based electrode was able to operate successfully even at 30 °C. Furthermore, Lee et al demonstrated stable cycling of SSE-based ALASSBs even at room temperature (RT) using the conversion reaction of silver fluoride (AgF) [38]. During Li plating, the phase of the active material becomes separated into silver-lithium (Ag-Li) alloy and lithium fluoride (LiF). The LiF phase protects the electrode from indiscriminate Li dendritic growth.

The good electrochemical stability and relatively high ionic conductivity of oxide solid electrolytes (OSEs) along with their air stability have motivated a number of studies that focused on developing OSE-based ASSB cells. Li phosphorus oxynitride (LiPON), one of the first OSEs reported by Bates et al [39], has relatively low ionic conductivity >10⁻⁶ S cm⁻¹ at RT and therefore its application has been limited to thin-film batteries. In the following decades, various types of OSEs have been identified, including perovskite-type Li_x La_(2−x)/3TiO₃ [40], NASICON-type LiGe₂(PO₄)₃ [41], Li_1+x Al_x Ti_2−x (PO₄)₃ [42], and garnet-type Li₇La₃Zr₂O₁₂ (LLZO) [43].

An early form of anode-less design with OSEs started with thin-film batteries. Neudecker et al were the first to report a thin-film battery based on a Li-free design [44]. In this system, a LiCoO₂ cathode was deposited on the cathode current collector and the LiPON solid electrolyte was deposited on the copper (Cu) anode current collector. They furthermore indicated that the protective overlayer is imperative for improving the cycle retention. This Li-free design was successfully operated for more than 1000 cycles at 1 mA cm⁻² and more than 500 cycles at 5 mA cm⁻². Several similar studies followed in the course of a few decades; however, the practical application of thin-film batteries remains unattainable owing to their low energy density.

To achieve the practical implementation of ASSBs, bulk-type batteries are being presented as indispensable. Wang et al demonstrated the potential of the 'Li-free' concept for OSE-based ASSBs, without adding liquid electrolyte [45]. They suggested three possible scenarios of Li growth (figure 3(a)): (a) Li deposition on the nucleates resulting in vertical growth and the separation of CC/LLZO; (b) vertical Li plating on the nucleates giving rise to horizontal plastic deformation; (c) horizontal Li plating at the boundary of the nucleates developed from deposited Li. They also proposed that the application of pressure would further increase the potential driving force for the above-mentioned nucleation and growth process. To study the performance of the in-situ formed Li metal anode, they fabricated a CC (current collector)/LLZO/NCA (lithium nickel-cobalt-aluminum oxide) all-solid-state cell. They showed that the CE remains near 100% for 50 cycles at the C/10 rate when an LLZO-laminated-current collector was only adopted on the anodic side. Despite their successful implementation of an anode-less system without the addition of a liquid to the OSEs, the system only operated at elevated temperatures. Chen et al reported ASSBs with an ultra-low N/P (the negative to positive electrode capacity ratio) ratio by adopting a thin film of gold as an anode-LLZTO interfacial layer [18]. Because of the lithiophilic nature of the gold film, the wettability of LLZTO toward Li metal could be realized. They also studied batteries with a zero N/P ratio, which are representative of anode-less systems and noted that unfortunately, the anode-less type cell had a low initial capacity of 76 mAh g⁻¹ and poor cycle retention. Recently, Kravchyk et al suggested porous LLZO scaffolds, either as a host for Li metal or an anode-less layer [46]. Their calculations showed that the LLZO scaffolds not only function well as a Li metal reservoir, but also lower the effective current density during Li plating.

Zoom In Zoom Out Reset image size

On the other hand, as opposed to OSEs, polymer solid electrolytes (PSEs) are flexible and easily deformable, leading to low interfacial resistance, which is beneficial for ASSBs. These favorable mechanical properties make PSEs suitable coating materials. However, due to their intrinsically low ionic conductivity and weak oxidation tolerance, their practical utilization has been limited. Even polyethylene oxide (PEO), the most representative SPE family with high ionic conductivity, is vulnerable to decomposition at oxidative potentials by destroying its ether chain responsible for ionic conduction [47–49]. Consequently, at RT, PSEs were typically used as an artificial protective layer for the Li metal in LIBs. Assegie et al reported anode-less-type LIBs of which the current collector (Cu foil) was modified by coating it with a thin PEO film [50]. The battery cell with the modified current collector demonstrated high CE of approximately 100% over 200 cycles in the half-cell test and capacity retention higher than that of the bare current collector in the full-cell test. Tamwattana et al introduced a current collector coated with a high-dielectric 'LiF@PVDF' polymeric layer for anode-less batteries [31]. Interestingly, by controlling the drying temperature, the crystallization structure of PVDF was varied such that β-phase PVDF, a high-dielectric phase, was formed at a higher temperature than α-phase PVDF. The subsequent addition of high-dielectric LiF nanoparticles resulted in the high dielectricity becoming uniformly effective over the entire layer. The LiF@PVDF-based anode-less half-cell had higher CE and cycle retention than the bare current collector. He et al reported a Li⁺-conducting polyacrylonitrile-coated PEO electrolyte with a sub-5 µm thickness [51]. They proposed an ultrathin SSE with a 3D ceramic framework to provide mechanical strength. Hence, their design of the PSE layer can be interpreted to indicate that the intrinsically low ionic conductivity of the PSE was surmounted by decreasing its layer thickness. Additional polyacrylonitrile layer casting was applied on the cathodic side to overcome the weak oxidation stability of the PSE layer. (figure 3(b)) The bulk resistance of the composite layer was less than 5 Ω and the ionic conductivity was 6.8 × 10⁻⁵ S cm⁻¹ at 25 °C. By combining the composite layer with a NCM811 cathode, the cell had decent cycle retention of 67% over 150 cycles even at a low N/P ratio of 1.1, without any liquid additives (figure 3(c)).

Efforts to maximize the advantages of OSEs and PSEs gave rise to intensive studies of composite polymer electrolytes (CPEs) using both of these solid electrolytes. Zegeye et al reported a garnet LLZTO-polymer composite electrolyte layer, which enables dendrite-free and well-regulated Li deposition on Cu foil at a current density of 0.2 mA cm⁻² at 55 °C [52]. An anode-less layer was formed by spin coating a LLZTO/PEO solution onto the current collector (figure 3(d)). They noted that the CPE exhibited high ionic conductivity and reduced interfacial resistance compared to the use of PEO alone. By adopting a CPE layer on both the anodic and cathodic sides, a full cell containing 15 µL of 1.0 M LiPF₆/EC (ethylene carbonate)/DEC (diethyl carbonate) electrolyte achieved CE of 98.8% and cycle retention of 41.2%, after 65 cycles at 0.2 mA cm⁻².

3.1. General considerations

3.2. ALASSBs using SSEs

3.3. ALASSBs using OSEs, PSEs and CSEs

The authors acknowledge support from the Swiss National Science Foundation (SNF) (Grant No. Sinergia CRSII5_202296), the National Research Foundation of Korea (NRF) (Grant No. NRF-2022M3J1A1054151) and generous support from the Institute of Engineering Research (IOER) and Research Institute of Advanced Materials (RIAM) at Seoul National University.