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

While all-solid-state batteries have built global consensus with regard to their impact in safety and energy density, their anode-less versions have attracted appreciable attention because of the possibility of further lowering the cell volume and cost. This perspective article summarizes recent research trends in anode-less all-solid-state batteries (ALASSBs) based on different types of solid electrolytes and anticipates future directions these batteries may take. We particularly aim to motivate researchers in the field to challenge remaining issues in ALASSBs by employing advanced materials and cell designs.

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].
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.

Sulfide solid electrolytes (SSEs)
SSEs have emerged as the most promising SE options owing to their superior ionic conductivities. SSEs, including glass-ceramics Li 7 P 3 S 11 (LPS) [21], argyrodites Li 6 PS 5 X (LPSX, X = Cl, Br) [22], and Li 10 GeP 2 S 12 (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][26][27][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)).
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 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.

Oxide, polymer, and composite solid electrolytes
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)/3 TiO 3 [40], NASICON-type LiGe 2 (PO 4 ) 3 [41], Li 1+x Al x Ti 2−x (PO 4 ) 3 [42], and garnet-type Li 7 La 3 Zr 2 O 12 (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 LiCoO 2 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.
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][48][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 LiPF 6 /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 .

General considerations
This perspective highlighted recent progress with ALASSBs based on different types of SE. Despite the considerable effort, several limitations hinder the practical implementation of these batteries, calling for further research to clarify various aspects. In this section, issues to be addressed regardless of the type of solid electrolyte for the development of ALASSBs are discussed.
Primarily, the success of ALASSBs would largely depend on effective reversibility during Li plating and stripping [53,54]. The absence of a host with crystallographic sites suitable for Li storage complicates the challenge of attaining reversibility, and Li entrapment and parasitic side reactions are the main obstacles to overcome. Another major difficulty is associated with the operating conditions of a cell; most of the promising results thus far were achieved with cells cycled at high temperatures or under high (stack) pressure. Bridging this gap would be necessary to align the cell operation with the practical conditions in a vehicle or to ensure widespread adoption in other applications. The requirement for high-temperature operation is not only attributed to the limited ionic conductivity of the SE or protective layer at the interface, but also to the temperature dependence of the Li nucleation barrier. Yan et al pointed out that the Li deposition temperature significantly affects the nucleation behavior, which has an impact on the electrochemical performance [55]. They elucidated that the lithiophilicity and Li ion diffusion were enhanced at elevated temperatures, resulting in the formation of large nuclei with low nucleation density. As for the critical nature of the stack pressure, several studies have investigated the correlation between the stack pressure and electrochemical performance. Wang et al reported the importance of carefully considering the stack pressure to minimize pore formation, which is significantly correlated with the polarization and optimal battery performance [56]. According to these researchers, pores are formed when the flux of Li resulting from diffusion and creep in Li metal is insufficient to replenish the flux of Li in the SE. As the stack pressure requires additional accessories in a vehicle, which would sacrifice the energy density (thus, driving mileage) of a battery pack, regulation of the Li deposition without the application of high pressure would be highly desired.
As such, we foresee that ensuring high energy density while enabling operation at low temperature and stack pressure is of the utmost importance for advancement along the technological road of ALASSBs ahead. The introduction of nucleation seeds and lithiophilic interphases or the provision of mechanical supports with high elasticity would be possible ways to circumvent the above-mentioned trade-off.

ALASSBs using SSEs
Even if SSEs have attractive ionic conductivity and ductility, dendritic penetration of Li through the SE layer remains unaddressed especially for the anode-less system. The narrow potential windows of SSEs are another hurdle to overcome, and minimizing the direct contact between Li metal and SSE has served as a main strategy thus far. The Ag-C system mentioned above can be understood from this perspective; Li nucleation seeds (Ag) accelerate Li plating whereas a protective layer (C) mitigates Li dendritic growth. However, this blend imposes an energy barrier for Li to diffuse at the interface through cobble creep mode, requiring high temperature for operation at reasonable C-rates. Hence, the main challenge in the next phase of SSE-based ALASSBs would be to overcome this trade-off between suppression of Li dendrite growth and operation at RT or below, demanding for advanced structural designs and materials selection.

ALASSBs using OSEs, PSEs and CSEs
In the case of ALASSBs using OSEs, the poor contact between the Li metal and OSE constitutes the origin of their insufficient electrochemical performance. This interfacial drawback causes dendritic growth of Li and accelerates the performance degradation over cycling [57]. Thus, the issues to be overcome in OSE-based ALASSBs are in the same line as SSE-based ALASSBs. Introducing a lithiophilic (i.e. LiF, Li 3 N) and flexible interlayer at the interface between the CC and OSE could greatly silence the problem. In the case of ALASSBs with PSEs, the poor ionic conductivity and mechanical properties of PSEs make it very difficult to cycle a cell under practical current density and capacity conditions. Revolutionizing the polymer design or the blending strategy with other liquids and solids might be a viable direction to pursue in getting PSEs competitive over SSEs and OSEs. Perhaps, the combination of these two types of electrolytes (OSEs and PSEs) seemingly leads to outperforming their individuals as they are complementary to each other in physical and electrochemical properties.