Reducing the thickness of solid-state electrolyte membranes for high-energy lithium batteries

Abstract

Rechargeable batteries with lithium metal anodes exhibit higher energy densities than conventional lithium-ion batteries. Solid-state electrolytes (SSEs) provide the opportunity to unlock the full potential of lithium metal anodes and fundamentally eliminate safety concerns caused by flammable liquid electrolytes. Up to now, most studies on SSEs have been focused on enhancing the ionic conductivity and improving the interfacial stability. However, the electrolyte thickness, which has received less attention, also plays an important role in determining the energy density and electrochemical performance of all-solid-state lithium batteries (ASSLBs). Recognizing this, our review evaluates SSE studies beyond traditional factors and focuses on a thickness perspective. We systematically analyze the influence of the electrolyte thickness on the energy densities of ASSLB pouch cells, and highlight the strategies that dramatically reduce the thickness of SSE membranes without sacrificing their mechanical properties. Then, we discuss recent advances and challenges of ASSLBs based on high-voltage and high-capacity cathodes, as well as novel configurations such as bipolar and flexible ASSLBs. Finally, we provide perspectives and suggestions towards high energy-density ASSLBs for future commercialization.

---

Jingyi Wu

Dr Jingyi Wu is currently a post-doctoral fellow in the Department of Materials Science and Engineering, Huazhong University of Science and Technology (HUST), China. She received her PhD degree from School of Chemistry and Chemical Engineering at HUST. Her research is mainly focused on lithium-metal batteries.

Lixia Yuan

Prof. Lixia Yuan received her BS, MS and PhD degrees from Wuhan University. She is now a professor at Huazhong University of Science and Technology (HUST). Her research interest mainly focuses on rechargeable lithium–sulfur batteries.

Wuxing Zhang

Dr Wuxing Zhang is currently an associate Professor at School of Materials Science and Engineering, Huazhong University of Science and Technology (HUST). He received his PhD from HUST in 2002. During 2002–2004, he worked as a postdoctoral researcher at HUST. From 2004 to 2007, he worked as a visiting scholar at Kochi University in Japan. His current research interests focus on next-generation lithium ion batteries and sodium ion batteries.

Zhen Li

Prof. Zhen Li received his PhD from Huazhong University of Science and Technology (HUST) in 2014. He worked as a R&D engineer at Amperex Technology Limited (ATL) from 2009 to 2011, and worked as a postdoctoral researcher in Nanyang Technological University (NTU) from 2015 to 2018. In 2018, he became a professor of materials science in HUST. His research interests are lithium-ion batteries, lithium–sulfur batteries, and solid-state batteries.

Xiaolin Xie

Prof. Xiaolin Xie is a full professor at Huazhong University of Science and Technology (HUST). He received his PhD degree from Sichuan University in 1995, which was followed by 2 year postdoctoral research at Zhejiang University. In December 1997, he joined HUST. Recently, his major interests focus on polymer modification, functional polymer materials, polymer composites and processing rheology.

Yunhui Huang

Prof. Yunhui Huang received his BS, MS and PhD from Peking University. In 2000, he worked as a postdoctoral researcher in Peking University. From 2002 to 2004, he worked as an associate professor in Fudan University and a JSPS fellow at Tokyo Institute of Technology, Japan. He then worked with Prof. John B. Goodenough at the University of Texas at Austin for more than three years. In 2008, he became a chair professor of materials science in Huazhong University of Science and Technology. His research group works on rechargeable batteries and electrode materials.

---

Broader context

All-solid-state batteries (ASSLBs) are promising candidates for next generation safer Li metal batteries, and they have attracted broad interest in energy storage and conversion. In the past decades, numerous investigations have been exploited to improve the ionic conductivity and interfacial stability of solid-state electrolytes (SSEs) and have achieved some significant breakthroughs. Other than the ionic conductivity and interphase, the thickness of SSEs also plays an important role in determining the performance and energy density of ASSLBSs. The excessively large amount of SSE that is commonly used in most studies will inevitably increase the internal resistance and offset the advantages of high-energy electrodes. Thus, urgent attention needs to be paid to evaluating SSEs beyond conventional factors and offering a comprehensive analysis of the effect of the SSE thickness. In this review, we provide a realistic assessment of the effect of the SSE membrane thickness on cell-level energy densities, evaluate the strategies employed to thin SSEs, and offer perspectives towards high energy-density ASSLBs for future commercialization.

1. Introduction

Over the last decades, lithium-ion batteries (LIBs) have undergone rapid development with applications in portable devices, electric vehicles and large-scale energy storage. However, the commercially available intercalation-type LIBs are almost reaching their energy density limit and cannot meet the increasing demand for higher energy density in large-scale storage devices.^1,2 Meanwhile, safety concerns such as leakage, fire and explosion originating from the organic liquid electrolytes have become more critical in the pursuit of higher energy density batteries.^3,4 Metallic lithium (Li) with an ultrahigh theoretical capacity (3860 mA h g⁻¹), the lowest redox potential (−3.04 V vs. the standard hydrogen electrode) and a low weight (0.534 g cm⁻³) has been considered as the ultimate anode material.⁵ Unfortunately, the notorious Li dendrite growth in the liquid electrolyte during Li plating/stripping further aggravates the safety issues.^6,7 The dendrites could potentially penetrate the porous separator, causing internal short-circuit and even fire hazards.^8–10

To fundamentally address the safety concerns, replacing the flammable liquid electrolytes with solid-state electrolytes (SSEs) is an efficient way.^11,12 The inherently low permeability of redox byproducts in SSEs enables the utilization of high-capacity cathodes such as sulfur (S), and the broad electrochemical stability widow makes SSEs suitable for high-voltage cathodes such as LiNi_xMn_yCo_zO₂ (NMC).^13,14 Furthermore, SSEs allow the implementation of a bipolar design, where multiple solid-state cells are alternately stacked in a single package, leading to a higher output voltage, less packaging, higher energy density and reduced cost.¹⁵ The possibilities of integrating SSEs into flexible batteries also shed light on applications in wearable devices.¹⁶

Ideally, SSEs should have high ionic conductivity close to that of a liquid electrolyte (above 10⁻⁴ S cm⁻¹ at room temperature) and stable interphases with the electrodes, both of which are crucial for ion conductance, a wide electrochemical window to pair with high-voltage cathodes, and strong mechanical properties to accommodate electrode volume changes during charge/discharge and to suppress Li dendrite growth.³ For practical applications in energy storage, processing convenience and low cost are two major aspects.¹⁷ Last but not least, the thickness of the SSE is also a critical parameter which influences the internal resistance and the energy-density of all-solid-state Li batteries (ASSLBs).^18–20 The internal resistance depends on the ionic conductance of the SSE and the charge transfer resistance of the electrodes. The ionic conductance (G = σA/l, where G, σ, A, and l represent the ionic conductance, ionic conductivity, surface area and thickness of the SSE, respectively) is inversely proportional to the thickness of the SSE membrane. Higher area-normalized conductance is observed in SSEs with reduced thickness, which originates from the shorter time it takes for ions to transport across SSEs (t = l²/D, where t, l, and D represent the Li⁺ diffusion time, thickness of the SSE and Li⁺ diffusion constant, respectively). As shown in Fig. 1a, to achieve the same area-normalized conductance as a liquid electrolyte using a 25 μm separator, the reduction of the thickness is equally important to the increase of the ionic conductivity of SSEs. SSEs with ionic conductivity such as 0.4 mS cm⁻¹ can obtain the same areal conductance as a liquid electrolyte if the thickness of the SSE is reduced to 10 μm. Ideally, if a certain SSE is reduced to 1/4 of its original thickness, the Li⁺ diffusion time across the SSE will be shortened to 1/16, and the areal conductance will quadruple.

Fig. 1 (a) Area-normalized ionic conductance as a function of electrolyte thickness for various ionic conductivities of SSEs. (b) Gravimetric and volumetric energy densities as the thickness of the SSE is reduced from 500 to 150 and further to 25 μm in Li/NMC811, Li/S and Li/LFP pouch cells (see Section 3 for calculation parameters).

Besides the ionic conductance, the cell-level energy densities of ASSLBs are also inversely related to the thickness of SSEs. The energy densities of Li/LiNi_0.8Mn_0.1Co_0.1O₂ (NMC811), Li/S and Li/LiFePO₄ (LFP) ASSLBs are calculated based on the pouch cell settings. As illustrated in Fig. 1b, the thickness of the SSE plays an enormous role in determining the cell-level energy densities, and thinner SSEs are essential to achieve higher gravimetric and volumetric energy densities. Unfortunately, conventional SSEs are usually over 100 μm thick, and the excessively large amount of SSE will inevitably offset the advantages of high-energy electrodes. By a rough calculation, if the thickness of the SSE is reduced from 150 to 25 μm (assuming a sulfide SSE with a density of 1.96 g cm⁻³, and a cathode film containing 80 wt% active materials and 15 wt% SSE), the component ratio of the SSE will decrease from 56 to 25%, which is comparable to that of the liquid electrolyte system (electrolyte amount of 2.5 g A h⁻¹ and cathode active materials of 95 wt%²¹) (Fig. 2). This corresponds to an over 30% increase in active materials, which contributes to higher energy output. However, the SSE also functions as a separator, and the thinning of the membrane will inevitably deteriorate its mechanical strength and increase the risk of membrane fracture or Li dendrite penetration, both of which lead to internal short circuit, cell failure and even safety hazards. Thus, the challenges associated with the thin SSE design mainly lie in the dilemma between minimizing the thickness and sustaining the mechanical strength.

Fig. 2 Battery configurations and components’ weight ratios based on lithium batteries using (a) a liquid electrolyte (electrolyte amount of 2.5 g A h⁻¹ and cathode active materials of 95 wt%), (b) conventional SSEs and (c) thin SSEs (a sulfide SSE with a density of 1.96 g cm⁻³, and a cathode film containing 80 wt% active materials and 15 wt% SSE).

In this review, we start by briefly introducing the main challenges and recent developments of SSEs. Then we analyze the relationships between the electrolyte thickness and energy density (gravimetric and volumetric energy densities) for various systems. Following that, we focus on the strategies to minimize the thickness of SSE membranes, and highlight the solutions suitable for mass production. In addition, we discuss the recent progress of ASSLBs paired with high-voltage and high-capacity cathodes, and present the potential and challenges of bipolar and flexible ASSLBs. Finally, we conclude by providing perspectives and suggestions towards high energy-density ASSLBs for future commercialization.

2. Fundamental of SSEs

SSEs generally can be classified into 3 categories: solid polymer electrolytes (SPEs), inorganic solid electrolytes (ISEs) and composite polymer electrolytes (CPEs).^22,23 SPEs, which consist of a polymer matrix and dissolved lithium salts, have favorable interfacial stability with electrodes, excellent processability, low cost and flexibility, but suffer from low room-temperature (RT) ionic conductivity, inferior electrochemistry stability and unsatisfactory mechanical strength.¹⁶ The thickness of most SPEs is in the range of 80 to 200 μm. ISEs are crystalline or glassy inorganics where ion transport relies on their defect sites.²⁴ ISEs possess high ionic conductivity, a broad electrochemical window, and sufficient mechanical strength, but exhibit poor interfacial contacts with electrodes, costly production, low throughput and a lack of flexibility. The widely used powder pressing method produces ISEs up to 1 mm thick. CPEs, which are constructed from an SPE matrix and inorganic fillers, could potentially inherit the advantages of both SPEs and ISEs while overcoming their shortcomings.²⁵ However, most CPEs are over 100 μm thick, and it still remains a huge challenge to greatly reduce the thickness of CPEs without compromising the mechanical properties.

2.1 SPEs

SPEs are composed of Li salts dissolved in a polymer matrix, where the optimal polymer presents a high dielectric constant to facilitate the dissociation and transport of ions. Among various polymer matrices such as poly(ethylene oxide) (PEO),²⁶ polyacrylonitrile (PAN),²⁷ poly(vinylidene fluoride) (PVDF),²⁸ poly(propylene carbonate) (PPC)²⁹ and poly(methyl methacrylate) (PMMA),³⁰ PEO is the most researched and widely used one due to its superior solubility for Li salts.³¹ Under an electrical field, the solvated Li⁺ hops from one coordination site to another, associated with breaking/forming of Li–O bonds, and long-range Li⁺ transport is achieved by local relaxation and continuous segmentation rearrangement of the polymer chains (Fig. 3a).³² It is generally believed that ion transport mainly occurs in the amorphous phase of PEO chains, whereas the crystalline phase presents an obstacle for ion transfer. Thus, the ionic conductivity of a PEO-based SPE highly depends on the ratio of the amorphous to crystalline phases. Since PEO is mainly crystalline below the glass transition temperature (t_g, ∼65 °C), the ionic conductivity of a PEO-based SPE is quite low at RT (10⁻⁷–10⁻⁵ S cm⁻¹), restricting it in application in RT batteries. Many strategies have been explored to increase the RT ionic conductivity of PEO-based SPEs, such as polymer architecture optimization,^33,34 salt design,^35,36 and organic plasticizer^37,38 or inorganic filler addition.^39,40 Among them, the introduction of a plasticizer or inorganic filler is a facile yet effective approach to increase the motion ability of PEO chains.

Fig. 3 (a) Li⁺ migration and hopping in the amorphous region of SPEs by polymer segmental motion. Reproduced with permission.³² Copyright 2015, Royal Society of Chemistry. (b) Typical Li⁺ migration mechanisms in ISEs: vacancy and interstitial. (c) Diagram showing the interaction between fillers, LiTFSI and PEO that leads to decreased PEO crystallinity and ion decoupling. Reproduced with permission.⁷⁹ Copyright 2017, Elsevier. (d) Possible Li⁺ conducting pathways in CPEs with aligned nanowires.

Plastic crystals such as succinonitrile (SN) with a dielectric constant as high as 55 exhibit high solubility of Li salts at RT, thus showing considerably higher ionic conductivity than PEO. In 2004, Alarco et al. first discovered that SN with dissolved 5 mol% lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) reaches a RT ionic conductivity as high as 3 × 10⁻³ S cm⁻¹, comparable to that of the liquid electrolyte.⁴¹ The electrolyte also shows a broad electrochemical window up to 6 V. Later in 2007, Fan et al. demonstrated SN as a universal plasticizer in SPEs.³⁷ Due to the existence of a highly conductive SN–LiTFSI phase, 50% (PEO–LiTFSI)–50% SN exhibits a conductivity of 5 × 10⁻⁴ S cm⁻¹, 80 times higher than that of the PEO–LiTFSI film at the same temperature. Echeverri et al. systematically and theoretically studied the relationship between ionic conductivity and the coexistence regions of the ternary phase diagram of PEO/LiTFSI/SN.⁴² SN is found to show dual-functionality as both a good plasticizer and an effective ionizer to dissociate Li⁺ from the LiTFSI salt. With the optimized design, a high ionic conductivity of 2.9 × 10⁻³ S cm⁻¹ can be achieved at RT. Despite the fact that the introduction of SN brings a significant increase in the RT ionic conductivity, the possible polymerization of nitriles in the presence of metallic Li limits the application of SN-based SPEs in ASSLBs.^41,43

Liquid electrolytes and conventional SPEs are dual-ion conductors, where both Li⁺ cations and the counter anions are mobile. During discharge, the ambipolar conductivity leads to the accumulation of anions on the anode side since electrodes only exchange cations (Li⁺) with the electrolyte.⁴⁴ This results in a concentration gradient and electrode polarization, and triggers the growth of dendrites in LMBs. Single Li⁺ conducting SPEs (SLC-SPEs) have been designed to overcome the challenges of the dual-ion conductors, where anions are immobilized by a polymer matrix, inorganic fillers or anion acceptors, and only Li⁺ contributes to the flow of charge inside the battery.^44–47 Bouchet et al. designed a polyanionic block copolymer electrolyte P(STFSILi)-b-PEO-b-P(STFSILi), which consists of a linear PEO block for ionic conductivity and a graft-diblock P(STFSILi) for mechanical strength and anion immobilization.⁴⁵ The weak interaction between Li⁺ and the grafted TFSI anions with delocalized negative charge enables a high dissolution level of Li⁺. The obtained SLC-SPE exhibits a high t_Li⁺ of 0.85 and ionic conductivity of 1.3 × 10⁻⁵ S cm⁻¹ at 60 °C, enhanced mechanical strength (10 MPa) and a broadened electrochemical stability window up to 5 V, all of which contribute to an impressive improvement in the performance of Li/SLC-SPE/LFP ASSLBs at 60 °C. Lago et al. reported a –SO₂N(−)SO₂CF₃ moiety grafted nano-Al₂O₃ filler to enable unipolar conductance in a PEO/PEGDME polymer matrix.⁴⁶ Long-term stability of 130 cycles with a specific capacity of 125 mA h g⁻¹ was achieved in Li/LFP ASSLBs at 70 °C. The unipolar conductivity of the SLC-SPE effectively enhances the ion selectivity and the electrochemical performance of the ASSLBs, however, the unsatisfactory RT ionic conductivity remains a bottleneck for their application.

2.2 ISEs

ISEs can be mainly classified into oxides, including NASICON-type, perovskite-type, garnet-type, etc., and sulfides. Fast ionic conductance can be achieved by ion hopping through vacancies or interstitials (Fig. 3b).^3,24,48

Na superionic conductor (NASICON)-type ISEs are crystalline phosphates generally with the AM₂(PO₄)₃ formula (A = Li, Na or K; M = Ge, Zr or Ti). Among them, Li_1+xAl_xTi_2−x(PO₄)₃ (LATP) and Li_1+xAl_xGe_2−x(PO₄)₃ (LAGP) are widely studied due to their high ionic conductivity at RT.⁴⁹ The reported ionic conductivity of LATP could reach as high as 1.09 × 10⁻³ S cm⁻¹.^50,51 However, NASISON-type ISEs are not stable with Li metal, which hampers their application.⁵² Li can be introduced into perovskite-type ABO₃ (A = La, Sr, or Ca; B = Al or Ti) on the A site, creating a perovskite-type ISE with a general formula of Li_3xLa_2/3−xTiO₃ (LLTO, 0 < x < 0.16).⁵³ The conductivity of LLTO heavily depends on the concentration of both Li⁺ and vacancies, and with the optimized design, a high RT conductivity of 10⁻³ S cm⁻¹ can be found in Li_0.34La_0.56TiO₃.⁵⁴ Similar to NASICON-type ISEs, the reduction from Ti⁴⁺ to Ti³⁺ when in contact with Li metal limits the application of LLTO in ASSLBs.^54,55 Unlike NASICON or perovskite-type ISEs, garnet-type ISEs exhibit high chemical stability with Li metal, which makes them an attractive candidate for ASSLBs. The general chemical formula of garnet-type ISEs is A₃B₂(XO₄)₃ (A = Ca, Mg, Y, La or rare-earth; B = Al, Fe, Ga, Ge, Mn, Ni, V; X = Si, Ge, Al), where the A, B and X cations are in 8, 6, and 4 coordinated sites, respectively.⁵⁶ One representative garnet-type ISE is a Li₇La₃Zr₂O₁₂ (LLZO)-based Li-rich material, whose ionic conductivity can reach 3 × 10⁻⁴ S cm⁻¹ at RT.⁵⁷ A small amount of Ta, Al, Ga, Nb, or Te is doped to further improve the ionic conductivity to ∼10⁻³ S cm⁻¹. Despite the fact that garnet-type ISEs exhibit high ionic conductivity and a wide electrochemical window, the poor interfacial contact with the Li anode and the sensitivity towards H₂O and CO₂ significantly decrease their performance.^58,59

Unlike the crystalline oxides that are fabricated by high-temperature sintering, glass-ceramic sulfides are usually prepared by low-temperature heat-treatment or liquid phase reactions. The glassy Li₂S–SiS₂ system was first studied in 1986 and later achieved a RT ionic conductivity of 6.9 × 10⁻⁴ S cm⁻¹ with Li₃PO₄ doping.^60,61 The glass-ceramic type electrolyte in the form of xLi₂S–(1 − x)P₂S₅ has been widely reported to exhibit high ionic conductivity reaching an exceptionally high 10⁻² S cm⁻¹ at RT.⁶² So far, sulfide electrolytes possess the highest ionic conductivity among all SSEs, together with a relatively “soft” contact with electrodes compared to that of the oxides. However, sulfides are extremely sensitive to moisture and generate hazardous H₂S gas, which is a major drawback to application in practical batteries.⁶³

2.3 CPEs

Inorganic fillers can be incorporated into SPEs to develop composite polymer electrolytes with the purpose of increasing the ionic conductivity, strengthening the mechanical properties, and improving the interfacial contact with the electrodes. CPEs, which combine the advantages of both SPEs and ISEs, are considered promising to construct high-performance ASSLBs.^64,65 The inorganic fillers can be classified as inert (non-ionic conductive) fillers and active fillers (ionic conductive). The enhanced ionic conductivity from the inert fillers mainly comes from the reduced polymer crystallinity to accelerate segmental motion.^66,67 The surface groups on the ceramic nanoparticles may also engage in the Lewis acid–base interaction with Li salts or polymer chains to facilitate Li⁺ dissociation (Fig. 3c).^68,69 The active fillers directly participate in the ion transport, leading to additional ionic pathways.^70–72

Ceramic nanoparticles (5.8–13 nm) were first introduced into a PEO SPE by Croce et al. in 1998.⁷³ The ionic conductivity increases by over 3 orders of magnitude and reaches 2 × 10⁻⁵ S cm⁻¹ at 31 °C, benefiting from the combined effect of large surface area and Lewis-acid characteristics, which prevents the recrystallization of PEO chains during the cooling process (from above 60 °C to ambient temperature). A t_Li⁺ of 0.6 is also achieved, much higher than that of 0.2–0.3 in the ceramic-free PEO electrolyte. Although the addition of nanoparticles is advantageous, nanomaterials with high surface energy tend to aggregate, leading to reduced surface area and deteriorated Li⁺ pathways. To overcome the agglomeration of nanoparticles, Pan et al. proposed the use of inorganic polyhedral oligomeric silsesquioxane (POSS) to cross-link poly(ethylene glycol) (PEG) and produce a CPE with a controllable network structure.⁶⁷ Enhancement in both ionic conductivity (∼1 × 10⁻⁴ S cm⁻¹ at RT) and mechanical strength (33.6 MPa) is realized, which is attributed to the suppressed polymer crystallization and the uniformly dispersed POSS nanoparticles in the cross-linked network.

Constructing interconnected filler networks can further improve the ionic conductivity and mechanical strength of CPEs. Lin et al. reported an in situ synthesis of a SiO₂-aerogel reinforced CPE with an improved RT ionic conductivity of 6 × 10⁻⁴ S cm⁻¹ and high modulus of 0.43 GPa.⁷⁴ The ultrafine and continuous SiO₂ network with large surface area boosts the Lewis acid–base interaction with anions and enables more efficient Li⁺ conductance. The strong SiO₂ backbone also endows the CPE with excellent mechanical strength, which is beneficial for suppressing Li dendrite growth. A more powerful strategy for fabricating continuous Li⁺ pathways in CPEs is to adopt 3D fast ion conductor networks. Bae et al. synthesized a 3D nanostructured hydrogel derived pre-percolated Li_0.35La_0.55TiO₃ (LLTO) framework, serving as a 3D nanofiller to prevent agglomeration and provide continuous Li⁺ conducting pathways.⁷⁵ The LLTO/PEO CPE reaches a RT ionic conductivity close to 10⁻⁴ S cm⁻¹ with improved thermal/electrochemical stability.

The tortuous filler/polymer interface junctions in CPEs present the main hindrances for realizing optimized conductivity. Recently, the concept of a CPE with aligned nanowires has been introduced by Liu et al.⁷⁶ The vertically aligned Li_0.33La_0.557TiO₃ nanowires (LLTO NWs) significantly enhance the ionic conductivity of the CPE due to its long-term ionic conducting pathway without obstructive cross-junctions (Fig. 3d). The RT ionic conductivity of the CPE with vertically aligned LLTO NWs is ten times higher than the CPE with randomly distributed NWs. Later, Zhai et al. used an ice-templating process to produce a CPE with vertically aligned and continuous Li_1+xAl_xTi_2−x(PO₄)₃ (LATP) nanoparticles.⁷⁷ The aligned fast ionic conductors provide direct channels for Li⁺ transport and improve the RT ionic conductivity to 5.2 × 10⁻⁵ S cm⁻¹, 3.6 times that of the CPE with random LATP nanoparticles. Zhang et al. used a melt infiltration method to introduce a PEO polymer electrolyte into anodized aluminum oxide (AAO) discs and produce vertical, continuous and non-tortuous interfaces.⁷⁸ The interfacial effects decrease the crystallization temperatures and reduce the crystallinity of PEO to 14.7%, lower than pure SPE (22.7%) and the CPE with NWs (19.7%) at temperatures below crystallization. The CPE exhibits a RT ionic conductivity of 5.82 × 10⁻⁴ S cm⁻¹ and excellent Li dendrite suppression ability.

3. Effects of the SSE thickness on the energy density

ASSLBs that are expected to power electric vehicles and portable devices require both high gravimetric and volumetric energy densities. To quantify the influence of the SSE thickness on the energy densities, we provide simulations of gravimetric and volumetric energy densities of three types of ASSLBs (Li/NMC811, Li/S and Li/LFP) paired with various SSEs in a practical pouch cell model. A cathode capacity of 3 mA h cm⁻² and N/P ratio of 2 are set for all ASSLBs.^9,80,81 In practical ASSLBs, it is extremely challenging to reach the theoretical capacities, and thus we reasonably assume stable specific capacities of 170, 1000 and 150 mA h g⁻¹ for the NMC811, S and LFP cathode, respectively.^82–85 Furthermore, the cathode active material accounts for 80%, 50% and 80% with a cathode density of 3.0, 0.4 and 2.0 g cm⁻³ for the NMC811, S and LFP cathode, respectively.⁸⁶ The cathode active material ratios in solid-state batteries are lower than their counterparts using liquid electrolytes because a non-negligible amount of SSE needs to be incorporated in the cathode film for sufficient ion conduction. A cathode current collector (Al foil, 12 μm, double-sided coated) and other components in pouch cells (taps, laminate film, etc., accounting for 5 wt% of the whole pouch cell) are also included in the calculation.⁸⁷ PEO, LLZO, double-layered LPS-LGPS and two LLZO/PEO hybrids are chosen as representatives of SPEs, oxide ISEs, sulfide ISEs and CPEs, with the density calculated to be 1.26 (EO : Li ratio of 10 : 1), 5.1, 1.96, 1.48 (20 wt% LLZO in PEO SPE, denoted as CPE20) and 3.17 g cm⁻³ (80 wt% LLZO in PEO SPE, denoted as CPE80), respectively.⁸⁷

As shown in Fig. 4a–c, both the gravimetric and volumetric energy densities increase upon a reducing thickness of the SSE. The light-weight SPE and CPE20 are more promising in delivering high gravimetric energy densities compared to the heavy oxide ISE and CPE80. On the other hand, the volumetric energy density is not affected by the different types of SSEs for a given system, since the thicknesses of the electrodes are fixed and the total thickness of a pouch cell is only determined by the thickness of the SSE. Compared to the Li/NMC811 and Li/LFP ASSLBs, the increase in the gravimetric energy density of the Li/S ASSLB is more significant with a reduced SSE thickness. This is because the weight and the weight reduction ratio of the SSE account for a larger portion of the whole pouch cell using the high-capacity and low-density sulfur cathode. For instance, when reducing the thickness of the SSE from 500 to 10 μm in a Li/S pouch cell, weight reductions of 85.5%, 94.6%, 89.6%, 87.2% and 92.6% of the whole pouch cell using the SPE, oxide ISE, sulfide ISE, CPE20 and CPE80 are observed, respectively, which are higher than the cases in Li/NMC811 (70.0%, 89.2%, 78.0%, 73.1% and 84.5% in order) and Li/LFP pouch cells (67.7%, 88.3%, 76.1%, 71.0% and 83.2% in order).

Fig. 4 Estimating the practical gravimetric and volumetric energy densities of (a) Li/NMC811, (b) Li/S and (c) Li/LFP ASSLBs. The calculated gravimetric and volumetric energy densities of (d) Li/NMC811, (e) Li/S, and (f) Li/LFP ASSLBs. (1) A baseline cell with 3 mA h cm⁻² areal capacity, 170 (NMC811), 1000 (S), 150 (LFP) mA h g⁻¹ specific capacity, 80% (NMC811), 50% (S) and 80% (LFP) active material, an N/P ratio of 2 and an SSE thickness of 150 μm. (2) Increasing the areal capacity to 4 mA h cm⁻². (3) Increasing the active materials to 90% (NMC811), 60% (S) and 90% (LFP). (4) Increasing the specific capacity to 200 (NMC811), 1200 (S) and 165 (LFP) mA h g⁻¹. (5) Reducing the N/P ratio to 1. (6) Reducing the SSE to 45 μm. (7) Reducing the SSE to 25 μm.

In the Li/NMC811 ASSLB, the gravimetric energy densities using a 10 μm SPE, oxide ISE, sulfide ISE, CPE20 and CPE80 are 410, 358, 399, 407 and 382 W h kg⁻¹, respectively, much higher than using 500 μm SSEs (123, 39, 88, 109 and 59 W h kg⁻¹ in order) (Fig. 4a). However, the gravimetric energy density still falls short of the 500 W h kg⁻¹ target. Fig. 4d illustrates the feasible pathways under different scenarios to achieve the targets of 500 W h kg⁻¹ in gravimetric energy density and 1000 W h L⁻¹ in volumetric energy density. We start with the Li/NMC811 ASSLB employing a conventional 150 μm thick CPE20, and the baseline cell delivers energy densities of 229 W h kg⁻¹ and 441 W h L⁻¹. With the optimized cathode design and the reduced N/P ratio, the energy densities reach 308 W h kg⁻¹ and 622 W h L⁻¹. The gravimetric energy density further increases to 459 and 507 W h kg⁻¹ by reducing the thickness of the SSE to 45 and 25 μm, respectively. At the same time, high volumetric energy densities of 1089 and 1271 W h L⁻¹ are achieved using 45 and 25 μm CPE20, respectively. Similar calculations can be made for the other types of SSEs. Generally speaking, to achieve a gravimetric energy density above 500 W h kg⁻¹ in Li/NMC811 ASSLBs, the thickness of the SSEs should be controlled below 30 μm for the SPE, 25 μm for CPE20, 20 μm for the sulfide ISE, and 10 μm for the oxide ISE and CPE80.

As for Li/S ASSLBs, it is possible to realize a gravimetric energy density >500 W h kg⁻¹ with a 10 μm SPE or CPE20 (Fig. 4b). After electrode optimization, 45 μm CPE20 would be sufficient for an output of 512 W h kg⁻¹ (Fig. 4e). Similarly, a 55 μm SPE, 35 μm sulfide ISE, 25 μm CPE80 and 15 μm oxide ISE are necessary to reach the target of 500 W h kg⁻¹. On the other hand, the volumetric energy density of Li/S ASSLBs is relatively low and a 25 μm SSE only leads to an output of 473 W h L⁻¹, which originates from the high porosity and low density of the S cathode. Finally, for Li/LFP ASSLBs, it is unlikely to deliver a high gravimetric energy density of 500 W h kg⁻¹ due to the relatively low capacity of the LFP cathode (Fig. 4f). However, a high volumetric energy density of 755 W h L⁻¹ is obtained with a 25 μm thick SSE, and it can further increase to above 800 W h L⁻¹ as the thickness is reduced to 15 μm. It should be noted that some parameters in the calculation are quite challenging to reach in practice, so the requirement on the SSE thickness should be more stringent for practical ASSLBs. Nevertheless, the development of ultrathin SSEs is essential for the realization of high-energy-density ASSLBs.

4. Strategies to reduce the thickness of SSEs

As discussed previously, the internal resistance and gravimetric/volumetric energy densities of ASSLBs are greatly dependent on the thickness of the SSE. The internal resistance is determined by the ionic conductance of the SSE and the charge transfer resistance of the electrodes. The ionic conductance, by definition, is inversely proportional to the thickness of the SSE membrane. Fig. 5a shows the conductance and thickness of SSEs in recent publications. Although the reported SSEs have a wide range of ionic conductivities (10⁻⁶–10⁻⁴ S cm⁻¹ at RT), the conductance in general becomes lower as the thickness increases. Similar trends can be observed in the cell-level energy density vs. SSE thickness data (Fig. 5b), which highlights the importance of minimizing the thickness of the SSE. Unfortunately, the thickness of ISEs produced by the widely used powder-pressing method is up to 1 mm, while SPEs and CPEs are relatively thinner, with thicknesses generally ranging from 80 to 200 μm, which are still far too thick to achieve the target gravimetric/volumetric energy densities.⁸⁸ On the other hand, SSEs not only function as ionic conductors but also serve as separators to avoid contact between the cathode and anode. Reducing the thickness of the SSE is accomplished by increasing the risks of internal short-circuit due to the mechanical failure of SSEs. Thus, a balance needs to be found between minimizing the thickness and sustaining the mechanical integrity of SSEs. In recent years, a number of strategies have been employed to successfully produce thin SSE membranes with sufficient mechanical strength, such as supporting SSEs with scaffolds, developing tape casting techniques, and utilizing emerging strategies such as 3D printing. The traditional thin-film deposition-based fabrication techniques are too expensive and time-consuming to be scaled-up in manufacturing, and are not in the scope of this review.

Fig. 5 (a) The thickness of SSEs and the corresponding area-normalized ionic conductance in the literature: SPEs,^{20,78,89–96} ISEs,^97–105 and CPEs.^{67,69,83,106–115} (b) The thickness of SSEs and the corresponding gravimetric energy densities in reported Li/LFP ASSLBs.^{20,39,90,100,102,108,110,116}

4.1 Scaffold supports

Infiltrating solid electrolytes into porous scaffolds is an effective strategy to fabricate thin SSEs with enhanced mechanical stability. The robust and flexible scaffold provides mechanical support for the soft polymer electrolyte or brittle ceramics and suppresses Li dendrite growth during cycling even with reduced SSE thickness. The thickness of the resulting SSE is governed by the thickness of the scaffold (Table 1).

4.1.1 Polymer scaffolds. Mechanically reliable polymer scaffolds are common supporting materials to build thin SSEs, and the structure of the scaffold plays a vital role in determining the ionic conductivity. Since a polymer scaffold is usually a poor or non-ion conductor and ion transport only occurs in the impregnated electrolyte region, large porosity and interconnectivity of pores are important in achieving high ionic conductivity.

In 2014, Choi et al. reported a new class of thin and deformable plastic crystal SPEs by combining PEO/SN with a porous polyethylene terephthalate (PET) nonwoven.¹⁸ The 3D plastic crystal CPE is formed by in situ UV polymerization of ethoxylated trimethylolpropane triacrylate (ETPTA) monomer inside a PET scaffold in the presence of plastic crystal SN (N-PCPE). The highly porous (∼70% porosity) and thin (thickness ∼13 μm) PET is composed of multi-fibrous layers with interconnected and large-sized pores (more than 5 μm), which contributes to the well-developed 3D continuous ionic pathway. The ionic conductance of N-PCPE is 5.7 × 10⁻⁴ S cm⁻¹ at 30 °C, slightly lower than that of the control CPE containing the UV-crosslinked ETPTA/SN SPE without a PET scaffold (X-PCPE) due to the presence of the non-ionic conductive PET matrix. However, the thin N-PCPE (25 μm) is advantageous in delivering an ionic conductance (0.38 S at 30 °C) more than twice that of the thick X-PCPE (0.18 S at 30 °C, thickness 200 μm), owning to a much shorter Li⁺ diffusion distance. The resulting Li₄Ti₅O₁₂ (LTO)/LiCoO₂ (LCO) solid-state battery achieves stable cycling at a high current density of 2C at 30 °C. Moreover, N-PCPE shows great tolerance to bending stress and the pouch cell employing N-PCPE can light a LED bulb and maintain stable cycling under serious deformation. Later, an even thinner plastic crystal SPE of 17 μm was fabricated by in situ polymerization on an electrospun PAN fiber membrane.¹⁹ The molten precursor of PVA–CN, SN and Li salts were injected into the PAN mat and filled into the cells, and then the as-prepared cells were heated at 70 °C, which prepares the cross-linked SPE (SEN) and forms ASSLBs in one step (Fig. 6a). The PAN mat ensures sufficient mechanical strength and the obtained SEN shows enhanced stress and strain compared to the counterpart PVA–CN/SN without PAN backbones (thickness of 150 μm) (Fig. 6b). The 17 μm SEN exhibits a RT conductance of 0.30 S, which is even higher than that of the liquid electrolyte (0.25 S) because of the remarkably reduced membrane thickness (Fig. 6c). The RT electrochemical performance of Li/LFP using SEN is comparable to the cell using a liquid electrolyte. Recently, Jiang et al. pressed Li_6.75La₃Zr_1.75Ta_0.25O₁₂ (LLZTO) powders connected by fibrous polytetrafluoroethylene (PTFE) into a nylon mesh and then immersed the SSE into a melted SN/LiTFSI mixture to obtain a thin (100 μm), robust and flexible CPE with a high RT ionic conductivity of 1.2 × 10⁻⁴ S cm⁻¹ and stable cycling in both Li/LFP and Li/NMC ASSLBs.¹¹⁷

Fig. 6 (a) Schematic illustration of the in situ synthesis of SEN. (b) Stress–strain curves of SEN, the PAN membrane and the commercial separator. (c) Ionic conductance comparison of SEN with a PVA–CN/SN solid electrolyte and liquid electrolyte. Reproduced with permission.¹⁹ Copyright 2015, Wiley-VCH Verlag GmbH & Co. KGaA. (d) Schematic illustration of the roll-to-roll process that can be adopted to assemble ASSLBs with PPL electrolyte. (e) Rate performance of the PPL electrolyte at 60 °C. Reproduced with permission.⁹⁰ Copyright 2019, Wiley-VCH Verlag GmbH & Co. KGaA.

Commercially available polyolefin separators, which are highly porous, mechanically strong and flexible, chemically stable and cost-efficient, are a promising candidate as the scaffold for thin SSEs. A thin (36 μm) and asymmetric organic–inorganic hybrid solid electrolyte (ASE) was developed by Duan et al. using a separator as the scaffold.¹⁰⁷In situ polymerization of poly(ethylene glycol) methyl ether acrylate (PEGMEA) in the interior and on the cathode side of the separator enables good interfacial contact, whereas an Li₇La₃Zr₂O₁₂ (LLZO) layer topped with a thin polymer electrolyte layer facing the anode side prevents Li dendrite penetration. The total thickness of the soft polymer electrolyte is about 30.4 μm, including 5.4 μm on the exterior and 25 μm impregnated in the pores of the separator, and the thickness of the dense LLZO layer is about 5.7 μm. The ASE exhibits an ionic conductivity of 1 × 10⁻⁴ S cm⁻¹ at 55 °C and a broad electrochemical window up to 4.8 V. Wu et al. fabricated an ultrathin polymer electrolyte (7.5 μm) by integrating a polyethylene (PE) separator with a PEO polymer electrolyte (PPL) (Fig. 6d).⁹⁰ The robust yet flexible polyethylene separator offers mechanical support for the soft PEO electrolyte, effectively preventing short-circuits even under mechanical deformation. PPL exhibits a Young's modulus of 445 MPa, more than three orders higher than that of the thick PEO SPE without separator support, and the ultrathin membrane maintains structural integrity after a bending test of 10 000 cycles. Due to the reduced internal resistance, the Li/LFP ASSLB with the ultrathin PPL can successfully cycle at a current density as high as 10C at 60 °C (Fig. 6e), outperforming that of the thicker CPE (∼27 μm) with a similar structure. The results demonstrate that the battery performance can be significantly improved by minimizing the SSE thickness without changing the ionic conductivity. Adopting battery separators as the scaffold promotes the possibility of roll-to-roll assembly, which is a big step forwards in industrial fabrication of ASSLBs. However, the thermal stability of polyolefin separators needs further improvement to enable safe ASSLBs in a wider temperature window.

Porous polyimide (PI) films with excellent heat resistance can be prepared by methods such as electrospinning, track-etching, phase inversion and so on.^20,91,106 Hu et al. infiltrated Li_6.75La₃Zr_1.75Ta_0.25O₁₂ (LLTZO)/PVDF slurry into an electrospun PI fiber network and prepared a 20 μm-thick LLTZO/PVDF/PI solid electrolyte.¹⁰⁶ The thin SSE presents improved electrochemical and thermal stabilities, and can be stably cycled with a Li/NMC battery in a voltage window of 2.5–4.3 V. Wan et al. prepared an 8.6 μm-thick SSE made of a PI host with vertically aligned nanopores and PEO polymer electrolyte fillers.²⁰ Molecular simulation reveals that Li⁺ diffuses faster along the aligned direction, with a calculated diffusion coefficient in the aligned direction of 1.3 × 10⁻⁸ cm² s⁻¹, high than that of 5.7× 10⁻⁹ cm² s⁻¹ in the random direction. The increased ionic conductivity in aligned nanopores is probably due to the nanoconfinement of channels and/or polymer-chain alignment at the PI/PEO interfaces. Li/LFP ASSLBs show stable cycling for 200 cycles at 60 °C, and over 120 mA h g⁻¹ capacity at 40 and 30 °C. This work demonstrates that ultrathin SSEs with vertically aligned ion channels are promising for achieving high-performance ASSLBs. Unfortunately, the porosity of the PI film made by the track-etching method is only 11%. Strategies to fabricate polymer hosts with a large number of vertically aligned channels need to be explored.

Other than polymer-based electrolytes, ductile sulfides can also be assembled into thin films with mechanical support of a scaffold. Nam et al. prepared a thin (∼70 μm) and bendable sulfide ISE reinforced with a mechanically compliant poly(paraphenylene terephthalamide) (PPTA) nonwoven (NW) scaffold (Fig. 7a).⁹⁸ Sulfide slurry [Li₃PS₄ (LPS) or Li₁₀GeP₂S₁₂ (LGPS) in anhydrous toluene] was coated on Ni foil using a doctor-blade, followed by solvent evaporation. The sulfide layer was then cold-pressed into a 16 μm-NW substrate with a high porosity of 70% and large opening sizes (∼40 μm). Although the resulting NW-SSE film exhibits lower RT ionic conductivity (2 × 10⁻⁵ S cm⁻¹) compared to the conventional sulfides (7 × 10⁻⁵ S cm⁻¹, 700 μm thick), due to the presence of nonionic-conducting NWs, it demonstrates a three-fold improvement in conductance, i.e. 37 vs. 14 mS. The flexible sulfide electrolyte can be both free-standing and bendable, enabling a bipolar cell architecture, which further improves the energy densities of ASSLBs. Later, a 40 μm thick sulfide ISE supported by electrospun PI nonwovens was reported by Kim et al. (Fig. 7b).¹¹⁸ Solution processable Li₆PS₅[Cl,Br] was dropped onto a PI nonwoven, followed by drying under a vacuum and further heat treatment at 400 °C to enhance the crystallinity. The considerably thinner and lighter PI-LPSClBr (40–70 μm) exhibits an ionic conductance of 15–29 mS, similar to that of the conventional Li₆PS₅Cl_0.5Br_0.5 pellet (700 μm thick, 29 mS) (Fig. 7c), which gives rise to a significant advantage in the cell-based energy density of ASSLBs (Fig. 7d).

Fig. 7 (a) Schematic illustrating the synthesis of the bendable sulfide NW-SE films with two different structures. Reproduced with permission.⁹⁸ Copyright 2015, American Chemical Society. (b) Schematic showing the fabrication of the PI-supported Li₆PS₅[Cl,Br] membranes. (c) Conductance of PI-Li₆PS₅Cl_0.5Br_0.5 as a function of membrane thickness. (d) Comparison of the SSE membrane thickness and areal capacity for NMC/graphite all-solid-state batteries. Reproduced with permission.¹¹⁸ Copyright 2020, American Chemical Society.

4.1.2 Inorganic scaffolds. Polymer scaffolds are usually poor or non-ion conductors, and Li⁺ can only transport in the open pores that are impregnated with electrolyte, leading to reduced ionic conducting pathways. Therefore, utilizing ionic conducting inorganic networks is beneficial for improving the ionic conductivity. An interconnected Li_6.4La₃Zr₂Al_0.2O₁₂ (LLZO) network was synthesized by electrospinning and subsequent calcination, and then a PEO polymer electrolyte was infiltrated into the LLZO network to produce a free-standing CPE (40–50 μm).¹¹⁰ While the 3D LLZO framework is not as robust as the polymer scaffold, it does provide better mechanical reinforcement for the soft PEO electrolyte than other nanofillers, and the CPE membrane shows some flexibility. The CPE exhibits good RT ionic conductivity of 2.5 × 10⁻⁴ S cm⁻¹ due to the continuous and long-range interconnected ion transport with minimum filler aggregation. Later, the same group used bacteria cellulose (BC) as a sacrificial template and fabricated a hybrid electrolyte with a cubic Li₇La₃Zr₂O₁₂ (c-LLZO) 3D network and PEO polymer electrolyte.¹¹² Highly porous BC with great water absorption capacity can absorb the precursor salt solution and construct the interconnected c-LLZO nanofiber network after calcination. The large aspect ratio and the intertwined structure, together with the addition of adhesive PEO electrolyte, effectively address the brittleness of the inorganic electrolyte and the as-prepared 70–100 μm CPE can withstand bending. The RT ionic conductivity reaches 1.12 × 10⁻⁴ S cm⁻¹.

Minimized thickness, mechanical reliability, and good adhesion with the solid electrolyte are crucial characteristics of the scaffold to achieve thin SSEs with favorable ionic conductance and ASSLBs with high energy density. The chemical, electrochemical and thermal stabilities of the scaffold are also important parameters for SSEs. Porous polymers with remarkable mechanical strength are a popular matrix to construct SSEs with thickness even lower than 10 μm. High porosity and interconnected pore structure are required due to the poor ionic conducting nature of the polymer matrix. An ionic conducting inorganic network provides additional ion pathways with a sacrifice in mechanical strength, which leads to relatively thicker SSEs above 40 μm. For practical application, the possibility of roll-to-roll manufacture is desirable to allow for the implementation of sheet-like thin SSEs in the existing LIB production lines. In this regard, adopting commercially available porous polymer scaffolds offers opportunities for fast and scalable manufacturing of thin SSEs at a low-cost. Although large-area ceramic precursor films can be obtained by electrospinning or employing templates, the ceramic network after sintering is not robust enough to stand roll-pressing, which leads to limited scalability and a major challenge for practical application of this method.

Table 1 Thin SSEs with scaffold supports

SSE composition	Fabrication method	SSE thickness (μm)	RT ionic conductivity (S cm^−1)	Active materials	Electrochemical performance	Ref.
“—” means data not available.
ETPTA/SN	In situ polymerization on PET fibers	25	5.7 × 10^−4	LTO/LCO	128 mA h g^−1 (0.2C, RT)	18
PVA–CN/SN	In situ polymerization on PAN fibers	17	4.5 × 10^−4	Li/LFP	155 mA h g^−1 (0.1C, RT)	19
C-PEGDE	In situ polymerization on cellulose fibers	30	8.9 × 10^−5	Li/LFP	135 mA h g^−1 (0.05C, RT)	89
PVDF/HEC/EC-DEC	Dip coating onto a PP separator	16	7.8 × 10^−4	Li/LNMO	133 mA h g^−1 (0.2C, RT)	119
PEO	Solution infiltration into a PE separator	7.5	3.7 × 10^−5	Li/LFP	160 mA h g^−1 (0.1C, 60 °C) 135 mA h g^−1 (0.1C, 30 °C)	90
PEO	Melt infiltration to a PI film	8.5	2.3 × 10^−4	Li/LFP	176 mA h g^−1 (0.1C, 60 °C) 150 mA h g^−1 (0.1C, 30 °C)	20
PEO	Melt infiltration to PI fibers	10	6.7 × 10^−6	Li/LFP	163 mA h g^−1 (0.1C, 60 °C)	91
PEO/SN	Solution infiltration to cellulose	55	5 × 10^−4	Li/LCO	152 mA h g^−1 (0.1C, 60 °C)	120
PEO	Melt infiltration to AAO	60	5.82 × 10^−4	—	—	78
LLZTO/PVDF	Infiltration to PI fibers	20	1.23 × 10^−4	Li/NMC	160 mA h g^−1 (0.1C, RT)	106
LLZTO/PTFE/SN	Hot pressing into Nylon mesh	100	1.2 × 10^−4	Li/LFP	157 mA h g^−1 (0.1C, RT)	117
LLZO/PEO (asymmetric)	In situ polymerization/tape casting on a PP separator	36	1 × 10^−4	Li/LFP	161 mA h g^−1 (0.2C, 55 °C)	107
LPS	Infiltration to Kevlar fibers	100	3 × 10^−4	Li/Li2S	800 mA h g^−1 (0.05C, RT)	97
LPS	Cold pressing into PPTA NWs	70	2 × 10^−4	LTO/LCO	140 mA h g^−1 (0.4C, RT)	98
LPS[Cl,Br]	Infiltration to PI fibers	70	2 × 10^−4	Graphite/NMC	146 mA h g^−1 (0.1C, RT)	118
β-Li3PS4-S	Infiltration to Kevlar fibers	60	2 × 10^−5	Li–In/LNO-NCA	156 mA h g^−1 (18 mA g^−1, 60 °C)	121
LLZO/PEO	Infiltration to a LLZO network	40–50	2.5 × 10^−4	—	—	110
c-LLZO/PEO	Infiltration to a LLZO network	70–100	1.12 × 10^−4	—	—	112

---

4.2 Tape casting

Tape casting is a well-established technique to produce films with tailored thickness on a large scale.^122–124 Ceramic or polymer powder and necessary additives are dispersed in a solvent to form a pseudoplastic slurry which is then cast onto a release film or directly on the electrode by a doctor blade.^125,126 By adjusting the slurry composition and height of the scraper, the thickness of the SSE can be precisely controlled from a few micrometers to 1 mm (Table 2).

4.2.1 Cathode-supported SSEs. A SSE slurry can be directly coated on the electrode, forming a cathode-supported SSE with good interfacial contacts between layers after solvent evaporation. Yan et al. tape cast a slurry composed of Li₇La₃Zr₂O₁₂ (LLZO), a dispersant, adhesives and a surfactant directly on the LFP cathode, and produced cathode-supported LLZO as thin as 3–5 μm without further cold or hot pressing (Fig. 8a).¹²⁷ The RT ionic conductivity of the LLZO thin film is 2.4 × 10⁻³ S cm⁻¹, slightly lower than that of the bulk LLZO, which is probably due to the existence of more grain boundaries between nanoparticles. Later, a Li₃PS₄ (LPS) slurry (LPS : poly(propylene carbonate) (PPC) = 100 : 3 w/w in anisole) was directly coated on the NMC cathode by Yamamoto et al. (Fig. 8b and c).⁹⁹ After heat treatment at 225 °C for 30 min under a vacuum to remove the ion-blocking PPC binder, the ionic conductivity of the LPS ISE reaches 5.2 × 10⁻⁴ S cm⁻¹. The sheet-like NMC battery with a 59 μm thick ISE layer can stably cycle for 350 cycles and post-mortem SEM analysis shows that the dense ISE layer maintains intimate contact with both electrodes (Fig. 8d and e). In 2019, Chen et al. fabricated a cathode-supported PEO/Al₂O₃ CPE with a thickness of 19 μm by bilayer tape casting (Fig. 8f).¹⁰⁸ The cathode layer and the electrolyte layer are well integrated with tight contact, which reduces the interfacial resistance and enhances Li⁺ transport through the solid-state battery. The Li/LFP ASSLB exhibits an initial capacity of 125 mA h g⁻¹ and 78% capacity retention after 410 cycles at 30 °C and 0.1C. As the temperature increases to 50 °C, a discharge capacity of 167 mA h g⁻¹ can be achieved at 0.1C.

Fig. 8 (a) Schematic illustrating the synthesis of the cathode-supported LLZO ultrathin film. Reproduced with permission.¹²⁷ Copyright 2017, American Chemical Society. Photographs of (b) a LPS-PPC slurry and (c) a cathode-supported LPS electrolyte. (d) Cross-sectional SEM image of the NMC–graphite full cell employing the LPS electrolyte after 350 cycles, and (e) the magnified figure of the LPS electrolyte layer. Reproduced with permission.⁹⁹ Copyright 2017, Nature Publishing Group. (f) Schematic illustrating the preparation of the cathode-supported PEO/Al₂O₃ CPE. Reproduced with permission.¹⁰⁸ Royal Society of Chemistry. (g) Schematic showing the fabrication of the LLZO film using the tape-casting method. (h) Breaking force of oxide films with different thicknesses. Reproduced with permission.¹⁰⁰ Copyright 2019, Wiley-VCH Verlag GmbH & Co. KGaA.

4.2.2 Freestanding SSEs. Although cathode-supported SSEs ensure favorable contact between the electrode and electrolyte, intermixing between the two layers may occur during the coating of the SSE layer. Additionally, the mechanical strength of the SSE layer may not be high enough, especially when the thickness of the SSE is reduced. Therefore, free-standing thin SSE membranes by tape casting were investigated. In 2019, a free-standing perovskite-type Li_0.34La_0.56TiO₃ (LLTO) ISE was successfully synthesized by Jiang et al. via tape casting a slurry composed of LLTO powders, a dispersant, a binder, a plasticizer and a solvent on a release film, followed by sintering at a high temperature of 1260 °C (Fig. 8g).¹⁰⁰ As expected, the mechanical strength of the free-standing ceramic film is heavily dependent on the film thickness. The breaking force decreases dramatically from 1959 to 109 mN as the thickness is reduced from 115 to 25 μm, and only pallets with a thickness of 41 μm or higher are robust enough to be used in battery assembly (Fig. 8h). As the pallets are thinned from 160 to 41 μm, the area specific resistance is reduced by 90% (2153 vs. 207 Ω cm⁻²), and the ionic conductivity is tripled (2 × 10⁻⁵vs. 6.7 × 10⁻⁶ S cm⁻¹), probably attributed to the decreased grain boundary resistance. The Li/LFP ASSLB using the 41 μm-thick pallet exhibits an initial capacity of 145 mA h g⁻¹ and capacity retention of 86.2% after 50 cycles at 0.1C and 65 °C.

4.2.3 Multilayer design. The frangibility of free-standing thin ceramic films still poses a major challenge for battery assembly. Thus, smart electrolyte configurations need to be investigated. In 2015, Zhu et al. constructed a porous-dense bilayer structure of the Li_1.3Al_0.3Ti_1.7(PO₄)₃ (LATP) electrolyte.¹⁰¹ A highly porous layer provides storage and loading space for cathode impregnation and the ultrathin and ultra-dense electrolyte layer gives high Li⁺ conductivity. The integrated structure alleviates the demands for mechanical strength, allowing the thickness of the electrolyte to be reduced from 600 to 36 μm. Later in 2017, a bilayer ISE was prepared by laminating porous and dense tapes followed by co-sintering to remove polymer pore formers and binders. Garnet-type Li₇La_2.75Ca_0.25Zr_1.75Nb_0.25O₁₂ (LLCZNO) with a thin dense layer (∼20 μm) provides continuous Li⁺ conduction and a thick porous layer (50–100 μm) mechanically supports the thin layer and confines the cathode materials.¹²⁸ The integrated sulfur cathode with a loading of 7.5 mg cm⁻² delivers an initial discharge capacity of 645 mA h g⁻¹ and a high initial Coulombic efficiency of 99.8%. Recently, Yang and Hitz et al. took one step further and constructed porous–dense–porous trilayer architectures by tape casting and co-sintering (Fig. 9a).^129,130 The dense and exceptionally thin (∼14 μm) Li₇La₃Zr₂O₁₂ (LLZ) center layer is strengthened by the porous outer layers (∼70 μm each side), making the cell strong enough to be handled with ease (Fig. 9b). The dense layer that is free of structural defects can simultaneously block Li dendrite penetration and diffusion of electrode components such as polysulfides, while the highly porous layers on the opposite sides of the dense layer form a 3D ion conductive framework for accommodation of electrode materials with high loadings. The Li filled trilayer symmetric cell has an area specific resistance of 2–10 Ω cm⁻², significantly lower than garnets with thickness of 100–150 μm (19–40 Ω cm⁻²). The trilayer Li–S cells with 5.4 mg cm⁻² Li and 5.3 mg cm⁻² sulfur reached a sulfur-specific capacity of 1244 mA h g⁻¹ and could potentially achieve a cell-level energy density as high as 450 W h kg⁻¹ (Fig. 9c). Thin oxide ISEs with similar bilayer or trilayer structure and high active material loadings have been further reported with projected energy density as high as 900 W h kg⁻¹.^131–133 Despite advantages in energy density, the fabrication of ASSLBs using multilayered ISEs is still challenging. For the porous layer, the trade-off between sufficient mechanical strength and large porosity needs careful consideration. Pore engineering is necessary to provide interconnected channels while avoiding close pores. Additionally, other than highly flowable molten sulfur and its discharge products, infiltrating electrode slurries such as LFP and NMC into the porous framework remains questionable.

Fig. 9 (a) Schematic of the garnet trilayer framework. (b) Cross-sectional SEM image of the garnet trilayer electrolyte. (c) The voltage profiles of a Li–S pouch cell with a sulfur loading of 5.3 mg cm⁻² and a lithium loading of 5.4 mg cm⁻² at 0.22 mA cm⁻². Reproduced with permission.¹²⁹ Copyright 2019, Elsevier.

Tape casting is a facile and industry-compatible technique to produce large-area solid-state sheets. The thickness of the electrolyte layer can be fine-tuned by adjusting the slurry composition and scraper height. Cathode-supported SSE layers as thin as a few micrometers can be achieved by casting the electrolyte slurry on top of the cathode layer. Cathode-supported SSEs are usually blessed with good electrode/electrolyte contact and sufficient flexibility, whereas they are at risk of interlayer-mixing during slurry casting of the top layer. Freestanding SSE layers prevent possible interlayer-mixing and possess stronger mechanical properties. However, since the mechanical strength deteriorates quickly with reduced thickness, freestanding SSEs are usually thicker than cathode-supported ones. Smart electrolyte configurations such as a bilayer or trilayer architecture are developed to provide mechanical support for the ultrathin electrolyte layer with the help of a porous top-layer which also serves as a host for high electrode material loading. The combination of a thin SSE and thick electrode leads to a significant improvement in energy density.

Table 2 Thin SSEs prepared by tape casting

SSE composition	SSE architecture	SSE thickness (μm)	RT ionic conductivity (S cm^−1)	Active materials	Electrochemical performance	Ref.
“—” means data not available.
PEO/PVDF/Al2O3	Cathode-supported	19	∼9 × 10^−5	Li/LFP	125 mA h g^−1 (0.1C, RT)	108
LLTO/PEO	Cathode-supported	15	2.3 × 10^−4	Li/S	415 mA h g^−1 (0.05C, RT)	134
LLZO/PEO	Cathode-supported	20–75	1.6 × 10^−4 (60 °C)	Li/S	900 mA h g^−1 (0.05 mA cm^−2, 37 °C)	84
PEO/ZrO2-IL	Cathode-supported	20–30	8.5 × 10^−5	Li/S	600 mA h g^−1 (37 °C)	69
LLZ/PAN/PC-DEC	Cathode-supported	3	6.0 × 10^−4	LTO/LMFP	14 mA h (1C, RT)	135
LPS/PEG–Ti	Cathode-supported	∼20	1.6 × 10^−4	Li/LFP	125 mA h g^−1 (0.05C, RT)	109
LLZO	Cathode-supported	3–5	2.4 × 10^−3	Li/LFP	160 mA h g^−1 (0.1C, RT)	127
LPS	Cathode-supported	59	5.2 × 10^−4	Li/NMC	155 mA h g^−1 (0.17C, RT)	99
LLZTO/PEO	Freestanding	30	1.12 × 10^−5	Li/LFP	155 mA h g^−1 (0.1C, 60 °C)	39
LAGP-PAN	Freestanding	25	3.7 × 10^−4	Li/NMC622	175 mA h g^−1 (0.1C. RT); 175 mA h g^−1 (0.5C, RT)	83
				Li/NMC811
LLZO	Freestanding	41	2 × 10^−5	Li/LFP	150 mA h g^−1 (0.1C, RT)	100
LAGP	Freestanding	70	2 × 10^−5	—	—	136
LATP	Bilayer	36	1.2 × 10^−4	Li/O2	14 200 mA h g^−1 carbon (0.15 mA cm^−2, RT)	101
LLCZNO	Bilayer	20	—	Li/S	645 mA h g^−1 (0.2 mA cm^−2, RT)	128
LLCZN	Trilayer	28	—	—	—	130
LLZ	Trilayer	14	—	Li/S	1244 mA h g^−1 (0.22 mA cm^−2, RT)	129

---

4.3 Other strategies

4.3.1 3D printing. Additive manufacturing, also known as 3D printing, is emerging as a revolutionary process for the construction of electrochemical energy storage devices with accurately controlled spatial geometries and architectures.¹³⁷ Different from the conventional methods that produce 2D planar structures, 3D novel structures such as pores, grids and columns can be easily designed by the 3D printing technique. In 2018, Zokoll et al. synthesized a hybrid SSE composed of a 3D ordered bicontinuous conducting ceramic and insulating polymer microchannels by the 3D printing technique.¹³⁸ Polymer templates with various architectures are printed by stereolithography, and after impregnation with Li_1.4Al_0.4Ge_1.6(PO₄)₃ (LAGP) powders and removal of the polymer template by combustion, structured LAGP scaffolds with empty channels are achieved, which are then filled with a non-conductive polymer such as polypropylene or epoxy to obtain the hybrid electrolyte. The bicontinuous microarchitecture enhances both the mechanical strength (especially the brittleness of the ceramics) and ionic conductivity (1.6 × 10⁻⁴ S cm⁻¹). More recently, McOwen et al. developed “conformal” and “self-supporting” solid electrolyte inks based on Li₇La₃Zr₂O₁₂ (LLZ).¹³⁹ Through optimization of the ink composition, LLZ inks with desired rheological or structural properties can be achieved, which are used to print ordered SSEs with various microstructures, such as ultrathin (5–10 μm) conformal film, stacked-array, or log-cabin (Fig. 10a and b). The cycling of Li/Li symmetric cells demonstrates that 3D-printed LLZ with an ordered and continuous structure effectively blocks Li dendrite penetration and dramatically lowers the battery resistance.

Fig. 10 (a) Schematic showing the fabrication process of 3D-printed SSEs. (b) Photograph of a 3D-printed SSE thin film and the cross-sectional SEM image of the 3D-printed thin film with 5–10 μm thickness after sintering. Reproduced with permission.¹³⁹ Copyright 2018, WILEY-VCH Verlag GmbH & Co. KGaA. (c) Schematic illustrating the formation of L₂S–P₂S₅ in a polymer membrane. (d) Cycling performance of FeS₂/Li–In alloy batteries employing the L₂S–P₂S₅ in polymer SSE. Reproduced with permission.¹⁴⁰ Copyright 2015, WILEY-VCH Verlag GmbH & Co. KGaA.

4.3.2 Hot pressing. Mechanical pressing of dry electrolyte powders can eliminate the use of solvent and the possible side reactions caused by the solvent. This method usually starts with a ball-milling process, where the local high temperature and pressure promote the uniform mixing of powders, followed by hot-pressing to ensure a dense structure and uniform thickness. While pallets of rigid oxides produced by mechanical pressing are usually very thick, ductile sulfide glass can be pressed into thin films with the help of deliberately chosen adhesives. Whiteley et al. successfully fabricated a 64 μm-thick L₂S–P₂S₅ film paired with a polyimine binder by ball-milling and a hot isostatic press at 100 °C under 100 MPa (Fig. 10c).¹⁴⁰ During hot pressing, the malleable thermoset polyimines flow through the pores, fill the voids of the densified sulfides and form a continuous cross-linked network, providing excellent adhesion for glassy sulfides to maintain the structural integrity at a thickness as thin as 64 μm. Moreover, unlike the solution process where polymer binders wrap around the ISE particles, polyimines in the solvent-free hot pressing process don’t interfere with the interparticle contact of the sulfide particles, so a high ionic conductivity of 1 × 10⁻⁴ S cm⁻¹ is realized with 80 wt% sulfides. Li/FeS₂ ASSLBs employing the L₂S–P₂S₅ thin film can be stably cycled for more than 200 cycles (Fig. 10d).

4.3.3 In situ reaction. Taking advantage of the high reactivity of Li metal, some SSEs can be in situ formed on the surface of Li by the spontaneous reaction with the precursors. Xie et al. fabricated a Li–Al–O based SSE by the chemical reaction between LiOH on the Li surface and triethylaluminum (TEAL) in the presence of a LiTFSI containing electrolyte.¹⁴¹ The thickness of the Li–Al–O SSE is determined by the thickness of the LiOH surface layer, which can be controlled by tuning the dipping time of Li metal in 4000 ppm H₂O/dimethyl ether. A 40 μm-thick SSE is obtained by dipping for 3 min, and the RT ionic conductivity reaches 1.42 × 10⁻⁴ S cm⁻¹.

SSEs with minimized thickness not only alleviate the battery resistance, but also boost the energy density of ASSLBs. However, the dilemma between the thickness and the mechanical properties of SSEs has not been fully solved. It is still challenging to synthesize thin SSE membranes with sufficient mechanical strength and favorable ionic conductivity. Moreover, the scale-up from laboratory research to industrial mass production of thin SSEs requires convenient fabrication at a cost-effective price. In this regard, a solution-based casting process which possesses a similar fashion to the conventional LIB manufacturing is most promising for large-scale production. Continuous processes that are adoptable in current LIB fabrication, such as the roll-to-roll process, are highly desirable.

5. Applications of thin SSEs in ASSLBs

5.1 High-voltage lithium-metal batteries

Coupling the Li metal anode with high-voltage cathodes such as NMC is a promising approach to maximize the energy density of LMBs.^142–145 Due to the low oxidative stability of PEO, PEO-based electrolytes decompose at voltages higher than 3.9 V (vs. Li/Li⁺), which limits their use to cathodes with lower-voltages such as LFP.^14,120 Crosslinking PEO with inorganic or organic components can effectively broaden the electrochemical window of PEO-based electrolytes, however, the majority of the reported studies are still applied in LFP batteries. An in situ hydrolysis of tetraethyl orthosilicate (TEOS) in PEO solution was proposed by Lin et al. to synthesize a PEO/monodisperse ultrafine SiO₂ CPE, which has outstanding 5.5 V vs. Li⁺/Li voltage stability.¹¹⁶ Xu et al. prepared a PEO/perovskite-type Li_3/8Sr_7/16Ta_3/4Zr_1/4O₃ CPE with strong bonds between TFSI anions and ceramic fillers and increased the electrochemical window of the SSE to a higher voltage over 5 V.⁷⁰ Besides fillers, electrolyte additives such as lithium borates (typically lithium bis(oxalato)borate (LiBOB), and lithium difluoro(oxalate)borate (LiDFOB)) that increase the oxidation resistance of liquid electrolytes are readily transferable to SSEs by forming a more robust cathode–electrolyte interphase.^120,146,147

Polymers such as PAN, PVDF, PVDF–HFP, and poly(ionic liquid) (PIL) have higher antioxidation ability, and are compatible with high-voltage cathodes. Ma et al. prepared a polyethylene supported ultrathin PVDF/hydroxyethyl cellulose (PVDF/HEC) polymer electrolyte by dipping a PE separator in a coating solution.¹¹⁹ The resulting polymer electrolyte (16 μm) was then soaked in liquid electrolyte to obtain a gel polymer electrolyte with oxidation stability up to 5.25 V, far higher than the 4.3 V of the liquid electrolyte. The Li/LiNi_0.5Mn_1.5O₄ cell cycled between 3.5 and 5 V exhibits a high initial discharge capacity of 133 mA h g⁻¹ and stable cycling with 88% retention after 200 cycles at 0.2C. Dong et al. prepared an ultra-strong bacterial cellulose supported poly(methyl vinyl ether-alt-maleic anhydride) plus propylene carbonate/lithium difluoro(oxalato) borate multifunctional gel polymer electrolyte (∼80 μm) which demonstrates superior electro-oxidative resistance (5.2 V vs. Li⁺/Li) and ultra-long cycle life in a 4.45 V-class Li/LCO battery.¹⁴⁶ Although the aforementioned polymers are oxidation-resistant, the ionic conductivities in those dry polymers are too limited to compete with PEO, so a large amount of liquid additives needs to be included, which contradicts the safety benefits brought by SSEs. Moreover, some antioxidative polymers such as PAN are not compatible with the Li metal anode due to severe side reactions.

The asymmetrical design of SSEs that optimizes the compatibility with both the high-voltage cathode and Li anode and provides sufficient ionic conductivity is investigated by researchers. Zhou et al. first investigated a bilayer polymer electrolyte with a high-voltage stable layer of poly(N-methyl-malonic amide) (PMA) contacting the cathode and a low-voltage stable layer of PEO contacting the anode side (Fig. 11a).¹⁴⁸ In this bilayer structure, the PMA layer is isolated by the PEO layer to avoid reduction by metallic Li, and the PEO layer is protected by the PMA layer from high-voltage oxidation. The 120–130 μm-thick bilayer SPE has a wide redox window of 4.75 V (Fig. 11b), a good ionic conductivity of 2.05 × 10⁻⁴ S cm⁻¹ at 65 °C and high flexibility. The Li/LCO ASSLB cycled in a voltage window of 2.5–4.25 V exhibits a capacity retention of 108.5 mA h g⁻¹ at 0.2C after 100 cycles and 65 °C, corresponding to a retention rate of 91.2% (Fig. 11c). Duan et al. fabricated an asymmetric PAN@LAGP CPE with PAN enriched on one side while LAGP on the other side, due to the different solvent evaporation rate on the two sides of the electrolyte (Fig. 11d).⁸³ Oxidation-resistant PAN and reduction-tolerant PEGDA are then coated on the opposite sides of the CPE, forming a heterogeneous multilayered composite solid electrolyte (HMSE) as thin as 25 μm. The PAN layer on the PAN-rich side provides good electrolyte/cathode compatibility and the PEGDA layer contacting the LAGP-rich side enables stability with the Li anode. The HMSE has a broad electrochemical window of 0–5 V (Fig. 11e), and high RT ionic conductivity of 3.7 × 10⁻⁴ S cm⁻¹, and can be stably cycled with both NMC811 and NMC622 cathodes in a voltage window of 2.8 to 4.3 V (Fig. 11f). Later, the same group constructed Janus interfaces of a ceramic electrolyte by coating PAN and PEO polymer electrolytes on opposite sides of a Li_1.4Al_0.4Ti_1.6(PO₄)₃ (LATP) pallet.¹⁴⁹ The upper PAN (15 μm) facing the NMC622 cathode guarantees high-voltage tolerance, whereas the lower PEO (25 μm) facing Li metal provides improved interfacial stability. The construction of Janus interfaces endows the ceramic pallets with excellent stabilities towards both the high-voltage cathode and Li anode.

Fig. 11 (a) Stacking model of the PMA–PEO bilayer electrolyte. (b) CV curve of a Li/PEO/PMA/Fe (stainless steel) battery (red line) and a Li/PEO/PMA/PEO/Fe battery (black line) from 2.5 to 5.05 V at 65 °C. (c) Voltage profiles of the Li/PEO/PMA/LCO ASSLB at 65 °C and 0.2C. Reproduced with permission.¹⁴⁸ Copyright 2018, WILEY-VCH Verlag GmbH & Co. KGaA. (d) Schematic diagram of the HMSE. (e) Electrochemical window of the HMSE. (f) Voltage profiles of the Li/HMSE/NMC811 ASSLB at 0.5C. Reproduced with permission.⁸³ Copyright 2019, WILEY-VCH Verlag GmbH & Co. KGaA.

High voltage ASSLBs with high operating voltage and electrode density are promising to reach the target energy density of 500 W h kg⁻¹ and 1000 W h L⁻¹ at the cell level (see Section 3 for simulations). ISEs generally possess sufficient electrochemical oxidation stability and can be paired with high-voltage cathodes. SPEs with a narrower electrochemical window cannot simultaneously ensure stable anode and cathode (>4 V vs. Li/Li⁺) interphases. Cross-linking or introducing electrolyte additives proves effective in enhancing the oxidation stability of PEO-based SPEs. Multilayer asymmetric design that can render stable cathode and anode interphases as well as remarkable ionic conductivity is considered an attractive direction for constructing thin SSEs in high-voltage ASSLBs. It is worth noting that although many studies reported approaching 5 V oxidation stability of SSEs by CV or LSV scans, during full cell testing, a lower voltage cut-off of 4.2–4.3 V was applied, mainly due to electrolyte decomposition and interphase instability at higher voltages. Further extending the voltage window of SSEs for larger capacities is beneficial for improving the energy density of the cathode.

5.2 High-capacity lithium–sulfur batteries

The low cost, non-toxic, and naturally abundant sulfur cathode with a high specific capacity of 1675 mA h g⁻¹ endows Li–S batteries with a superior theoretical energy density of 2600 W h kg⁻¹.^150–152 However, the soluble polysulfide shuttle between electrodes causes severe problems such as poor sulfur utilization and unsatisfactory cycle life.^153–155 Compared to liquid electrolytes, SSEs are believed to suppress the shuttle effect due to the slower dissolution and diffusion of polysulfides in the solid-state.^156,157 DeGott first developed all-solid-state Li–S batteries (ASSLSBs) using a PEO electrolyte in 1986, and then the research on SPEs for ASSLSBs mainly focused on PEO-based electrolytes.¹⁵⁸ However, the low ionic conductivity of PEO-based electrolytes at RT further slows down the sluggish sulfur chemistry and leads to lower sulfur utilization and poorer cycling performance compared to the liquid electrolyte counterpart. Elevating the temperature above the melting point (65 °C) promotes the ionic conductivity of PEO-based SPEs but at the expense of exacerbating the polysulfide shuttle in the semi-molten polymer. Therefore, enabling RT or near RT operation of PEO-based ASSLSBs is considered as a promising direction. The RT ionic conductivity of PEO-based SPEs is generally boosted by introducing various fillers. Tao et al. added LLZO powders to PEO solution and then cast the dispersion on the sulfur cathode to fabricate a 20–70 μm thick CPE.⁸⁴ The CPE exhibits an ionic conductivity of 9.5 × 10⁻⁶ and 1.1 × 10⁻⁴ S cm⁻¹ at 20 and 40 °C, respectively, and the ASSLSB delivered a specific capacity of more than 900 mA h g⁻¹ at 37 °C. Sheng et al. prepared a PEO-based electrolyte (20–30 μm) containing ionic liquid grafted oxide nanoparticles (IL@NP) with a high ionic conductivity of 2.32 × 10⁻⁴ S cm⁻¹ at 37 °C.⁶⁹ The reduced crystallinity of the polymer electrolyte and the weakened interaction between Li⁺ and PEO contribute to the enhancement in the ionic conductivity. The ASSLSB delivers a reversible specific capacity of ∼600 mA h g⁻¹ at 37 °C. Zhu et al. designed a bilayer structure by infiltrating PEO/LLZO electrolyte on the surface and in the pores of a carbon nanotube/sulfur cathode membrane.¹³⁴ The integrated framework of the SSE and cathode reduces the interfacial resistance and provides continuous Li⁺ pathways. ASSLSBs with a 15 μm thick SSE deliver a steady specific capacity of 415 mA h g⁻¹ at RT.

Compared with polymer-based electrolytes, ISEs with high RT ionic conductivity and the ability to physically block the polysulfide shuttle have flourishing prospects in Li–S batteries. Among them, sulfides that are relatively soft and have similar chemical components to sulfur or sulfur-based cathode materials are the most promising SSEs for ASSLSBs with good interfacial contact. A Li₂S–P₂S₅ glass-type electrolyte was reported by Scrosati et al. and a high capacity of 1600 mA h g⁻¹ was achieved at 25 °C.¹⁵⁹ Xu et al. prepared a thin (100 μm) Li₃PS₄ (LPS) solid electrolyte by dropping an LPS suspension on a Kevlar nonwoven, followed by pressing onto a stainless steel supported Li₂S cathode.⁹⁷ The Li/Li₂S cell with 2.54 mg cm⁻² Li₂S exhibits a high reversible discharge capacity of 949.9 mA h g⁻¹ at 0.05C and stable cycling for 100 cycles at 0.2C. A superior cell-level energy density of 370.6 W h kg⁻¹ can be realized using the thin SSE paired with a high Li₂S loading cathode of 7.64 mg cm⁻². The thick porous layer in bilayer or trilayer garnet type SSEs can also be used to hold sulfur, providing continuous ionic pathways while accommodating volume changes of sulfur during discharge/charge.^128,129 The thin and dense SSE layer suppresses the polysulfide shuttle and allows fast ion transport.

The polysulfide shuttle effect can be reduced to some degree in SPEs and CPEs due to lower dissolution and diffusion of polysulfides in the polymer matrix. However, the capacity and rate performance at RT or near-RT are still poor because of the slow Li⁺ transport and large interfacial resistance. Integrating the S cathode with SSEs can lower the interfacial resistance and provide an opportunity to form thin SSEs, which shows great potential for fabrication of ASSLBs with enhanced performance. Additionally, similar to that in high-voltage ASSLBs, bilayer or asymmetric design of SSEs integrating different functions should be explored to tackle the challenges of the polysulfide shuttle, ionic conductivity and mechanical strength. Sulfide ISEs with no polysulfides generated in the charge/discharge process are considered the most promising type of SSEs for ASSLSBs. Thin membranes can be made by infiltrating ductile sulfides into the polymer scaffold. However, the chemical and electrochemical instability remain a major challenge for their practical application. Oxides with dense structure show remarkable performance in suppressing the polysulfide shuttle, but the large interfacial resistance with the electrodes needs to be addressed. Thin SSE design is particularly important for heavy oxides to develop high-energy-density ASSLSBs. Thin oxides with thick electrodes produced by a cost-effective and scalable process are the goal.

5.3 Bipolar solid-state batteries

In contrast to the conventional LIBs, where liquid electrolytes immerse all components, the confined SSE enables the implementation of bipolar stacking, where mono cells are alternately stacked in a single package without an external electrical connection.^15,160 The bipolar design allows for building up of the total output voltage and reducing the packaging materials, which leads to the improvement of the energy density and reduction of the manufacturing cost for large-scale battery systems (Fig. 12a).¹⁷ The stacking strategies for bipolar design are generally classified into 2 categories, the lamination of free-standing SSEs and the layer-by-layer coating of SSEs on the electrode.

Fig. 12 (a) Comparison at the pack level of conventional liquid LIBs and ASSLBs. (b) Discharge curves of the 12 V bipolar cell at various temperatures. Reproduced with permission.¹³⁵ Copyright 2016 Elsevier. (c) CV curves of the bipolar cells at the 1st cycle at 0.1C. (d) SEM image of the bipolar cell with 3-stacked mono cells. Reproduced with permission.¹⁶² Copyright 2019, WILEY-VCH Verlag GmbH & Co. KGaA.

Nam et al. synthesized a freestanding sulfide SSE reinforced by a PPTA scaffold via a cold-press “transfer” process, and the thin (70 μm) and flexible SSE enables facile stacking of mono-cells.⁹⁸ The voltage of the bipolar LTO/LCO cell consisting of two mono-cells is doubled and exhibits a plateau at 4.5 V. Shin et al. reported a free-standing and flexible biphasic solid electrolyte (30 μm thick) composed of perovskite-type Li_0.29La_0.57TiO₃ (LLTO) and PEO by a tape casting method.¹⁶¹ The Li/NMC bipolar cell based on 3 mono-cells is successfully operated in a voltage range of 9.2–12.0 V, demonstrating successful stacking without short circuit or current leakage. The bipolar cell delivers a specific capacity of 125 mA h g⁻¹ with a capacity retention of 83% after 50 cycles.

Layer-by-layer coating of electrode and SSE slurries on a current collector is simple and scalable. Yoshima et al. demonstrated a 12 V-class bipolar battery composed of a thin SSE (3 μm) of Li₇La₃Zr₂O₁₂ (LLZ) particles coated with 4 wt% PAN-based gel polymer electrolyte.¹³⁵ LiMn_0.8Fe_0.2PO₄ (LMFP, cathodes) and LTO (anodes) are slurry coated on the opposite sides of an Al current collector, and then the SSE slurry was coated on both sides of the bipolar electrode, arranged in a face-to-face configuration, before the heating polymerization process. The 12 V pouch cell (20 mA h) constructed by stacking 5 bipolar electrodes shows stable cycling at 60 °C (85% capacity retention after 200 cycles), good rate capacity from 0.2 to 20C at 25 °C, and sufficient capacity retention at −40 °C (50% capacity retention of that at 25 °C), demonstrating successful operation in a wide temperature range (Fig. 12b). Recently, Kim et al. reported a bipolar ASSLSB based on two thermodynamically immiscible gel electrolytes of ethyl methyl sulfone (EMS) and tetraethylene glycol dimethyl ether (TEGDME).¹⁶² The EMS gel electrolyte embedded in the sulfur cathode shows the dual-functionality of a binder and sulfur utilization promoter, whereas the TEGDME composite gel electrolyte functions as a polysulfide-repelling separator membrane (∼50 μm). Bipolar cells were connected by a UV curing-assisted stepwise printing process, showing good electrolyte/electrode interfacial contact (Fig. 12d). The open-circuit voltage of the bipolar cell is proportional to the number of mono-cells, which is 5.6 V for a 2-stacked cell and 8.4 V for a 3-stacked cell (Fig. 12c). Bipolar cells with 2-stacked and 3-stacked mono-cells deliver initial discharge capacities of ∼1000 mA h g⁻¹ and ∼900 mA h g⁻¹, respectively. The bipolar cells also show exceptional advances in mechanical foldability and safety, which is promising in realization of scalable fabrication of high-energy-density ASSLBs.

5.4 Flexible solid-state batteries

With the rapid development of flexible electronics, flexible batteries have attracted increasing attention as power sources for flexible and wearable devices.^163,164 SPEs and CPEs are preferable for flexible ASSLBs due to their flexibility and the facile lamination processing. SSEs with scaffold support usually show great bending-tolerance, and are successfully applied in flexible ASSLBs. A 25 μm-thick SPE composed of in situ cross-linked ETPTA on a PET nonwoven in the presence of plastic crystal SN shows strong resistance against mechanical deformations, which enables a stable electrochemical performance even in a severely wrinkled state.¹⁸ PEO-based CPEs supported by porous scaffolds such as PI and PE all show great flexibility and high tolerance to mechanical abuse, and demonstrate steady performance in flexible pouch cells under deformation.^{20,90,106,117} With deliberate design, ISEs supported by a flexible polymer matrix can also exhibit sufficient bendability to enable flexible ASSLBs. A 100 μm bendable ISE was obtained by pressing LLZTO interconnected by a fibrous PTFE binder into nylon mesh.¹¹⁷ The Li/NMC532 pouch cell using this flexible ISE can light up a diode array even under harsh conditions such as bending and cutting. Flexible Li₆PS₅Cl_0.5Br_0.5 infiltrated PI (PI-LPSClBr) was prepared by dropping sulfide solution onto electrospun PI fibers followed by heat treatment.¹¹⁸ The thin (40–70 μm) PI-LPSClBr film shows great flexibility and can wrap around a 3 cm-diameter rod. Hence, the development of ultrathin SSEs can also extend the application in flexible ASSLBs.

6. Summary and perspectives

The replacement of liquid electrolytes with SSEs makes LMBs promising for next-generation energy storage systems. The thickness of the SSE membrane needs to be reduced below 25 μm to potentially realize high gravimetric/volumetric densities of 500 W h kg⁻¹/1000 W h L⁻¹. However, the reported thicknesses of SPEs and CPEs are generally in the range of 80–200 μm, and the dense ISE pallets prepared by the widely used powder pressing method are up to 1 mm thick. Those thick SSEs inevitably offset the advantages of high-energy electrodes by the excessively large amount of inactive components. Therefore, more attention has been paid to reducing the thickness of SSEs in recent years. The thickness, together with the ionic conductivity and mechanical strength, of SSEs proves crucial in meeting the criteria of high-performance and safe ASSLBs (Fig. 13). High ionic conductivity is the essential prerequisite of a functional SSE. Reducing the electrolyte thickness facilitates ion transport and improves the performance of the resultant ASSLBs. Meanwhile, the mechanical reliability of SSEs is a key aspect related to safety. Thin SSEs must have sufficient strength to resist fracture or Li dendrite penetration over long-term cycling. Employing robust scaffolds and developing advanced casting techniques are two main strategies to balance the minimized thickness against reliable mechanical properties.

Fig. 13 Proposed criteria for practical SSEs.

Fig. 14 shows the variety of scaffold-support SSEs. Great efforts have been dedicated to reducing the thickness of SSEs with enhanced mechanical strength by introducing flexible and robust polymer scaffolds. The facile infiltration process can produce thin SSEs below 10 μm. However, polymer scaffolds are usually poor ion conductors, and Li⁺ can only diffuse in the impregnated electrolyte region, which leads to the inevitable reduction of the ionic conductivity. High porosity and interconnected pores of the scaffold are crucial in facilitating ion conduction in the SSE. Vertically aligned pores provide low-tortuosity ion diffusion, but the fabrication of polymers with vertically aligned channels is usually costly, and the porosity falls behind compared to scaffolds with randomly distributed pores. Future development of polymer scaffolds needs to focus on pore structure engineering and creating cost-effective polymers with abundant vertically aligned ion channels. Biomass with well-aligned channel structure may be a promising candidate as electrolyte supports due to its low cost and natural abundance.

Fig. 14 Illustration of the applicable electrolytes, minimum thicknesses, and advantages and challenges of various scaffold supported SSEs.

Inorganic scaffolds can function as interconnected filler networks of polymer electrolytes to increase the ion conduction either by reducing the polymer crystallinity or by directly participating in ion transport. However, an inorganic network is not as robust as a polymer scaffold, and thereby the resultant SSEs are usually thicker (>40 μm). For greater robustness and flexibility, a large aspect ratio and highly entangled structure are necessary. Furthermore, minimizing the thickness of the inorganic network with ordered structure for optimized ion transport is worth future exploration.

The scalable and industry-compatible tape casting technique is considered an attractive direction for commercialization of thin SSEs. The thickness of the electrolyte layer is controlled by the slurry composition and scraper height. Fig. 15 illustrates the advantages and challenges of thin SSEs produced by various tape casting methods. With the help of the cathode layer to provide mechanical support, SSEs even thinner than 5 μm can be obtained with good interfacial contact. During fabrication, careful consideration must be taken in choosing the solvents and binders of the cathode and the SSE slurries to avoid layer mixing-up or dissolution of the cathode layer when applying the SSE slurry. Tape casting an oxide SSE slurry on a release film allows the following high-temperature sintering (usually >800 °C) to achieve improved crystallinity and higher ionic conductivity. However, combustion of polymer binders makes the freestanding membrane brittle, resulting in processing difficulties. A balance between the ionic conductivity and processability of thin oxide membranes should be considered. Laminating thick porous and thin dense layers provides an opportunity to fabricate defect-free oxide SSEs thinner than 20 μm. The thick porous layer ensures mechanical strength and high electrode material loading. Although tape casting enables scalable production of layered ISEs, the electrode materials must be infiltrated into the porous layer, which requires a multistep fabrication process. Strategies to simplify the cell fabrication process are necessary. Continuous processes that are adoptable in current LIB fabrication, such as the roll-to-roll process, are the ultimate goal for commercially-viable production of ASSLBs.

Fig. 15 Illustration of the applicable electrolytes, minimum thicknesses, and advantages and challenges of various tape cast SSEs.

The environmental sensitivity of different electrolyte systems is an important factor influencing the manufacture, performance, cost and safety of solid-state batteries. Sulfides exhibit the highest ionic conductivity among all SSEs, approaching that of liquid electrolytes. However, sulfides and their precursors are extremely sensitive to moisture and produce toxic H₂S gas. Therefore, sulfide SSEs must be handled under an inert atmosphere, which is a main drawback in practical applications. Ductile sulfides can be pressed into porous scaffolds, solution cast on top of the cathode film, or hot pressed into thin films. Oxides are generally considered stable in the atmosphere, although garnet-type oxides do react with H₂O and CO₂ in air and form LiOH and Li₂CO₃ on the surface. Thin oxide films can be fabricated by infiltrating into a porous polymer network, and solution casting onto a cathode or release film. Multilayer designs that integrate the dense electrolyte layer and the porous supporting layer are advantageous in producing ultrathin oxide SSEs. Polymers are generally stable in air but the dissolved salts easily adsorb H₂O, so the preparation of SPEs is usually carried out in a dry-room or Ar-filled glovebox. The most effective strategy to produce thin SPEs is to support them by robust polymer or inorganic networks. The mechanical strength of SPEs is too weak to be cast into free-standing thin films without the risk of short-circuiting during cycling. Incorporating inorganic fillers into the polymer matrix improves the mechanical properties of the resulting CPEs and provides opportunities for thinning films by tape casting.

For thin SPEs, the design should still focus on enhancing the ionic conductivity to allow steady operation at RT, increasing the mechanical strength to suppress Li dendrites, and broadening the electrochemical window to pair with high-voltage cathodes. Composite or asymmetric systems seem a promising direction to achieve the goals, and exploration of new polymer or salt design is also very important. For ISEs, the interfacial instability between the electrolyte and Li metal anode remains a critical problem affecting the performance of ASSLBs. As ISE membranes get thinner, such problems are becoming more critical, since thinner ISEs would be penetrated more easily by the interfacial side-reactions and dendrite growth. Tremendous efforts have been made to address the interface issue by coating ISEs with various interlayer materials. However, operation on thin ISEs is challenging due to the reduced mechanical properties. One promising approach is to construct Li composites or Li alloys by introducing a second phase to guarantee close contact between the Li anode and thin ISE and to suppress Li dendrite growth.^58,165–168

Fabrication convenience and cost are critical parameters for the scale-up from lab research to industrial mass production. From this perspective, oxides that need high-energy ball-milling and high temperature/pressure post-sintering or sulfides that require a stringent inert atmosphere seem to be difficult to be implemented industrially. Solution-based fabrication followed by a roll-to-roll process is more promising for the short-term application of high energy-density ASSLBs. Synthesizing thin SSEs using commercially available materials or biomass is also rewarding. At the same time, more efforts need to be put into bipolar designs and flexible batteries to enable large-scale energy storage systems. It also should be noted that the majority of thin SSEs are tested in coin-cell geometries. Therefore, although extensive academic research has been carried out, there is still a big gap between lab-scale scientific findings and industrial-level applications. Testing thin solid-state electrolytes in large pouch-cell settings close to commercial implementation is necessary to obtain realistic assessments of electrochemical performances and energy densities.

There is no doubt that SSEs will play an important role in the development of safe and high-energy-density Li metal batteries. The design and development of thin and robust SSE membranes have received growing academic and industrial interest, and more breakthroughs are expected in the near future.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors thank the financial supports of the National Key R&D Program of China (2016YFB0100302), Natural Science Foundation of China (No. 51972131, 51632001 and 52002138) and China Postdoctoral Science Foundation (2019M662613).

References

1. D. Larcher and J. M. Tarascon, Nat. Chem., 2015, 7, 19–29.
2. M. Li, J. Lu, Z. Chen and K. Amine, Adv. Mater., 2018, 30, 1800561.
3. A. Manthiram, X. W. Yu and S. F. Wang, Nat. Rev. Mater., 2017, 2, 16103.
4. J. Duan, X. Tang, H. Dai, Y. Yang, W. Wu, X. Wei and Y. Huang, Electrochem. Energy Rev., 2019, 3, 1–42.
5. J. Wu, L. Yuan, Z. Li, X. Xie and Y. Huang, Mater. Horiz., 2020, 7, 2619–2634.
6. J. Wu, X. Li, Z. Rao, X. Xu, Z. Cheng, Y. Liao, L. Yuan, X. Xie, Z. Li and Y. Huang, Nano Energy, 2020, 72, 104725.
7. J. Duan, Y. Zheng, W. Luo, W. Wu, T. Wang, Y. Xie, S. Li, J. Li and Y. Huang, Natl. Sci. Rev., 2020, 7, 1208–1217.
8. P. Albertus, S. Babinec, S. Litzelman and A. Newman, Nat. Energy, 2018, 3, 16–21.
9. J. Xiang, L. Yang, L. Yuan, K. Yuan, Y. Zhang, Y. Huang, J. Lin, F. Pan and Y. Huang, Joule, 2019, 3, 2334–2363.
10. X. Zheng, Z. Gu, X. Liu, Z. Wang, J. Wen, X. Wu, W. Luo and Y. Huang, Energy Environ. Sci., 2020, 13, 1788–1798.
11. C. Chen, D. Liu and X. Xiong, Acta Phys.-Chim. Sin., 2020, 202008078,  DOI:10.3866/pku.whxb202008078.
12. J. Wu, H. Zeng, Q. Shi, X. Li, Q. Xia, Z. Xue, Y. Ye and X. Xie, J. Power Sources, 2018, 405, 7–17.
13. S.-J. Tan, X.-X. Zeng, Q. Ma, X.-W. Wu and Y.-G. Guo, Electrochem. Energy Rev., 2018, 1, 113–138.
14. H. T. Xu, H. R. Zhang, J. Ma, G. J. Xu, T. T. Dong, J. C. Chen and G. L. Cui, ACS Energy Lett., 2019, 4, 2871–2886.
15. K. N. Jung, H. S. Shin, M. S. Park and J. W. Lee, ChemElectroChem, 2019, 6, 3842–3859.
16. D. Zhou, D. Shanmukaraj, A. Tkacheva, M. Armand and G. Wang, Chem, 2019, 5, 2326–2352.
17. J. Schnell, T. Günther, T. Knoche, C. Vieider, L. Köhler, A. Just, M. Keller, S. Passerini and G. Reinhart, J. Power Sources, 2018, 382, 160–175.
18. K.-H. Choi, S.-J. Cho, S.-H. Kim, Y. H. Kwon, J. Y. Kim and S.-Y. Lee, Adv. Funct. Mater., 2014, 24, 44–52.
19. D. Zhou, Y.-B. He, R. Liu, M. Liu, H. Du, B. Li, Q. Cai, Q.-H. Yang and F. Kang, Adv. Energy Mater., 2015, 5, 1500353.
20. J. Wan, J. Xie, X. Kong, Z. Liu, K. Liu, F. Shi, A. Pei, H. Chen, W. Chen, J. Chen, X. Zhang, L. Zong, J. Wang, L. Q. Chen, J. Qin and Y. Cui, Nat. Nanotechnol., 2019, 14, 705–711.
21. M. Ue, K. Sakaushi and K. Uosaki, Mater. Horiz., 2020, 7, 1937–1954.
22. J. Wan, J. Xie, D. G. Mackanic, W. Burke, Z. Bao and Y. Cui, Mater. Today Nano, 2018, 4, 1–16.
23. R. Chen, W. Qu, X. Guo, L. Li and F. Wu, Mater. Horiz., 2016, 3, 487–516.
24. Z. Gao, H. Sun, L. Fu, F. Ye, Y. Zhang, W. Luo and Y. Huang, Adv. Mater., 2018, 30, 1705702.
25. S. Li, S. Q. Zhang, L. Shen, Q. Liu, J. B. Ma, W. Lv, Y. B. He and Q. H. Yang, Adv. Sci., 2020, 7, 1903088.
26. G. S. MacGlashan, Y. G. Andreev and P. G. Bruce, Nature, 1999, 398, 792–794.
27. Q. Wang, W.-L. Song, L.-Z. Fan and Y. Song, J. Membr. Sci., 2015, 486, 21–28.
28. X. Zhang, S. Wang, C. Xue, C. Xin, Y. Lin, Y. Shen, L. Li and C. W. Nan, Adv. Mater., 2019, 31, e1806082.
29. J. Zhang, J. Zhao, L. Yue, Q. Wang, J. Chai, Z. Liu, X. Zhou, H. Li, Y. Guo, G. Cui and L. Chen, Adv. Energy Mater., 2015, 5, 1501082.
30. H.-W. Chen, T.-P. Lin and F.-C. Chang, Polymer, 2002, 43, 5281–5288.
31. L. Yue, J. Ma, J. Zhang, J. Zhao, S. Dong, Z. Liu, G. Cui and L. Chen, Energy Storage Mater., 2016, 5, 139–164.
32. Z. Xue, D. He and X. Xie, J. Mater. Chem. A, 2015, 3, 19218–19253.
33. P. P. Soo, B. Huang, Y.-I. Jang, Y.-M. Chiang, D. R. Sadoway and A. M. Mayes, J. Electrochem. Soc., 1999, 146, 32–37.
34. D. R. Sadoway, J. Power Sources, 2004, 129, 1–3.
35. H. Zhang, X. Judez, A. Santiago, M. Martinez-Ibañez, M. Á. Muñoz-Márquez, J. Carrasco, C. Li, G. G. Eshetu and M. Armand, Adv. Energy Mater., 2019, 9, 1900763.
36. G. G. Eshetu, X. Judez, C. Li, O. Bondarchuk, L. M. Rodriguez-Martinez, H. Zhang and M. Armand, Angew. Chem., Int. Ed., 2017, 56, 15368–15372.
37. L. Z. Fan, Y. S. Hu, A. J. Bhattacharyya and J. Maier, Adv. Funct. Mater., 2007, 17, 2800–2807.
38. M. Forsyth, L. Porcarelli, X. Wang, N. Goujon and D. Mecerreyes, Acc. Chem. Res., 2019, 52, 686–694.
39. C. Z. Zhao, X. Q. Zhang, X. B. Cheng, R. Zhang, R. Xu, P. Y. Chen, H. J. Peng, J. Q. Huang and Q. Zhang, Proc. Natl. Acad. Sci. U. S. A., 2017, 114, 11069–11074.
40. L. Chen, Y. Li, S.-P. Li, L.-Z. Fan, C.-W. Nan and J. B. Goodenough, Nano Energy, 2018, 46, 176–184.
41. P. J. Alarco, Y. Abu-Lebdeh, A. Abouimrane and M. Armand, Nat. Mater., 2004, 3, 476–481.
42. M. Echeverri, N. Kim and T. Kyu, Macromolecules, 2012, 45, 6068–6077.
43. C. Wang, K. R. Adair, J. Liang, X. Li, Y. Sun, X. Li, J. Wang, Q. Sun, F. Zhao, X. Lin, R. Li, H. Huang, L. Zhang, R. Yang, S. Lu and X. Sun, Adv. Funct. Mater., 2019, 1900392.
44. H. Zhang, C. Li, M. Piszcz, E. Coya, T. Rojo, L. M. Rodriguez-Martinez, M. Armand and Z. Zhou, Chem. Soc. Rev., 2017, 46, 797–815.
45. R. Bouchet1, S. Maria, R. Meziane, A. Aboulaich, L. Lienafa, J.-P. Bonnet, T. N. T. Phan, D. Bertin, D. Gigmes, D. Devaux, R. Denoye and M. Armand, Nat. Mater., 2013, 12, 452–457.
46. N. Lago, O. Garcia-Calvo, J. M. Lopez del Amo, T. Rojo and M. Armand, ChemSusChem, 2015, 8, 3039–3043.
47. A. Blazejczyk, M. Szczupak, W. Wieczorek, P. Cmoch, G. B. Appetecchi, B. Scrosati, R. Kovarsky, D. Golodnitsky and E. Peled, Chem. Mater., 2005, 17, 1535–1547.
48. T. Famprikis, P. Canepa, J. A. Dawson, M. S. Islam and C. Masquelier, Nat. Mater., 2019, 18, 1278–1291.
49. C. R. Mariappan, C. Yada, F. Rosciano and B. Roling, J. Power Sources, 2011, 196, 6456–6464.
50. H. Aono, E. Sugimoto, Y. Sadaoka, N. Imanaka and G.-Y. Adachi, Solid State Ionics, 1990, 40–41, 38–42.
51. H. Morimoto, H. Awano, J. Terashima, Y. Shindo, S. Nakanishi, N. Ito, K. Ishikawa and S.-I. Tobishima, J. Power Sources, 2013, 240, 636–643.
52. J. K. Feng, L. Lu and M. O. Lai, J. Alloys Compd., 2010, 501, 255–258.
53. O. Bohnke, Solid State Ionics, 2008, 179, 9–15.
54. S. Stramare, V. Thangadurai and W. Weppner, Chem. Mater., 2003, 15, 3974–3990.
55. S. Wenzel, T. Leichtweiss, D. Krüger, J. Sann and J. Janek, Solid State Ionics, 2015, 278, 98–105.
56. V. Thangadurai, S. Narayanan and D. Pinzaru, Chem. Soc. Rev., 2014, 43, 4714–4727.
57. J. Awaka, N. Kijima, H. Hayakawa and J. Akimoto, J. Solid State Chem., 2009, 182, 2046–2052.
58. J. Wen, Y. Huang, J. Duan, Y. Wu, W. Luo, L. Zhou, C. Hu, L. Huang, X. Zheng, W. Yang, Z. Wen and Y. Huang, ACS Nano, 2019, 13, 14549–14556.
59. J. Duan, L. Huang, T. Wang, Y. Huang, H. Fu, W. Wu, W. Luo and Y. Huang, Adv. Funct. Mater., 2020, 30, 1908701.
60. J. H. Kennedy, S. Sahami, S. W. Shea and Z. Zhang, Solid State Ionics, 1986, 18–19, 368–371.
61. A. Hayashi, S. Hama, T. Minami and M. Tatsumisago, Electrochem. Commun., 2003, 5, 111–114.
62. N. Kamaya, K. Homma, Y. Yamakawa, M. Hirayama, R. Kanno, M. Yonemura, T. Kamiyama, Y. Kato, S. Hama, K. Kawamoto and A. Mitsui, Nat. Mater., 2011, 10, 682–686.
63. H. Muramatsu, A. Hayashi, T. Ohtomo, S. Hama and M. Tatsumisago, Solid State Ionics, 2011, 182, 116–119.
64. F. Croce, G. B. Appetecchi, L. Persi and B. Scrosati, Nature, 1998, 394, 456–458.
65. T. Wang, R. Zhang, Y. Wu, G. Zhu, C. Hu, J. Wen and W. Luo, J. Energy Chem., 2020, 46, 187.
66. S. Choudhury, R. Mangal, A. Agrawal and L. A. Archer, Nat. Commun., 2015, 6, 10101.
67. Q. Pan, D. M. Smith, H. Qi, S. Wang and C. Y. Li, Adv. Mater., 2015, 27, 5995–6001.
68. J. Shim, H. J. Kim, B. G. Kim, Y. S. Kim, D.-G. Kim and J.-C. Lee, Energy Environ. Sci., 2017, 10, 1911–1916.
69. O. Sheng, C. Jin, J. Luo, H. Yuan, C. Fang, H. Huang, Y. Gan, J. Zhang, Y. Xia, C. Liang, W. Zhang and X. Tao, J. Mater. Chem. A, 2017, 5, 12934–12942.
70. H. Xu, P. H. Chien, J. Shi, Y. Li, N. Wu, Y. Liu, Y. Y. Hu and J. B. Goodenough, Proc. Natl. Acad. Sci. U. S. A., 2019, 116, 18815–18821.
71. W. Liu, N. Liu, J. Sun, P. C. Hsu, Y. Li, H. W. Lee and Y. Cui, Nano Lett., 2015, 15, 2740–2745.
72. J. Gao, Q. Shao and J. Chen, J. Energy Chem., 2020, 46, 237.
73. F. Croce, G. B. Appetecchi, L. Persi and B. Scrosati, Nature, 1998, 394, 456–458.
74. D. Lin, P. Y. Yuen, Y. Liu, W. Liu, N. Liu, R. H. Dauskardt and Y. Cui, Adv. Mater., 2018, 30, e1802661.
75. J. Bae, Y. Li, J. Zhang, X. Zhou, F. Zhao, Y. Shi, J. B. Goodenough and G. Yu, Angew. Chem., Int. Ed., 2018, 57, 2096–2100.
76. W. Liu, S. W. Lee, D. Lin, F. Shi, S. Wang, A. D. Sendek and Y. Cui, Nat. Energy, 2017, 2, 17035.
77. H. Zhai, P. Xu, M. Ning, Q. Cheng, J. Mandal and Y. Yang, Nano Lett., 2017, 17, 3182–3187.
78. X. Zhang, J. Xie, F. Shi, D. Lin, Y. Liu, W. Liu, A. Pei, Y. Gong, H. Wang, K. Liu, Y. Xiang and Y. Cui, Nano Lett., 2018, 18, 3829–3838.
79. Y. Lin, X. M. Wang, J. Liu and J. D. Miller, Nano Energy, 2017, 31, 478–485.
80. J. Liu, Z. Bao, Y. Cui, E. J. Dufek, J. B. Goodenough, P. Khalifah, Q. Li, B. Y. Liaw, P. Liu, A. Manthiram, Y. S. Meng, V. R. Subramanian, M. F. Toney, V. V. Viswanathan, M. S. Whittingham, J. Xiao, W. Xu, J. Yang, X.-Q. Yang and J.-G. Zhang, Nat. Energy, 2019, 4, 180–186.
81. B. Liu, J.-G. Zhang and W. Xu, Joule, 2018, 2, 833–845.
82. Z. Zhang, S. Chen, X. Yao, P. Cui, J. Duan, W. Luo, Y. Huang and X. Xu, Energy Storage Mater., 2020, 24, 714–718.
83. H. Duan, M. Fan, W. P. Chen, J. Y. Li, P. F. Wang, W. P. Wang, J. L. Shi, Y. X. Yin, L. J. Wan and Y. G. Guo, Adv. Mater., 2019, 31, e1807789.
84. X. Tao, Y. Liu, W. Liu, G. Zhou, J. Zhao, D. Lin, C. Zu, O. Sheng, W. Zhang, H. W. Lee and Y. Cui, Nano Lett., 2017, 17, 2967–2972.
85. K. Pan, L. Zhang, W. Qian, X. Wu, K. Dong, H. Zhang and S. Zhang, Adv. Mater., 2020, 32, e2000399.
86. J. W. Choi and D. Aurbach, Nat. Rev. Mater., 2016, 1, 16013.
87. X. Yang, J. Luo and X. Sun, Chem. Soc. Rev., 2020, 49, 2140–2195.
88. Y. J. Nam, D. Y. Oh, S. H. Jung and Y. S. Jung, J. Power Sources, 2018, 375, 93–101.
89. Y. Cui, X. Liang, J. Chai, Z. Cui, Q. Wang, W. He, X. Liu, Z. Liu, G. Cui and J. Feng, Adv. Sci., 2017, 4, 1700174.
90. J. Wu, Z. Rao, Z. Cheng, L. Yuan, Z. Li and Y. Huang, Adv. Energy Mater., 2019, 9, 1902767.
91. Y. Cui, J. Wan, Y. Ye, K. Liu, L. Y. Chou and Y. Cui, Nano Lett., 2020, 20, 1686–1692.
92. R. Khurana, J. L. Schaefer, L. A. Archer and G. W. Coates, J. Am. Chem. Soc., 2014, 136, 7395–7402.
93. B. Zhang, Y. Zhang, N. Zhang, J. Liu, L. Cong, J. Liu, L. Sun, A. Mauger, C. M. Julien, H. Xie and X. Pan, J. Power Sources, 2019, 428, 93–104.
94. X. Yang, Q. Sun, C. Zhao, X. Gao, K. R. Adair, Y. Liu, J. Luo, X. Lin, J. Liang, H. Huang, L. Zhang, R. Yang, S. Lu, R. Li and X. Sun, Nano Energy, 2019, 61, 567–575.
95. P. Walke, K. M. Freitag, H. Kirchhain, M. Kaiser, L. van Wüllen and T. Nilges, Z. Anorg. Allg. Chem., 2018, 644, 1863–1874.
96. H. Wang, Q. Wang, X. Cao, Y. He, K. Wu, J. Yang, H. Zhou, W. Liu and X. Sun, Adv. Mater., 2020, 32, e2001259.
97. R. Xu, J. Yue, S. Liu, J. Tu, F. Han, P. Liu and C. Wang, ACS Energy Lett., 2019, 4, 1073–1079.
98. Y. J. Nam, S. J. Cho, D. Y. Oh, J. M. Lim, S. Y. Kim, J. H. Song, Y. G. Lee, S. Y. Lee and Y. S. Jung, Nano Lett., 2015, 15, 3317–3323.
99. M. Yamamoto, Y. Terauchi, A. Sakuda and M. Takahashi, Sci. Rep., 2018, 8, 1212.
100. Z. Jiang, S. Wang, X. Chen, W. Yang, X. Yao, X. Hu, Q. Han and H. Wang, Adv. Mater., 2020, 32, 1906221.
101. X. B. Zhu, T. S. Zhao, Z. H. Wei, P. Tan and G. Zhao, Energy Environ. Sci., 2015, 8, 2782–2790.
102. X. G. Hao, Q. Zhao, S. M. Su, S. Q. Zhang, J. B. Ma, L. Shen, Q. P. Yu, L. Zhao, Y. Liu, F. Y. Kang and Y. B. He, Adv. Energy Mater., 2019, 9, 1901604.
103. S. Yu, A. Mertens, H. Tempel, R. Schierholz, H. Kungl and R.-A. Eichel, ACS Appl. Mater. Interfaces, 2018, 10, 22264–22277.
104. M. Finsterbusch, T. Danner, C.-L. Tsai, S. Uhlenbruck, A. Latz and O. Guillon, ACS Appl. Mater. Interfaces, 2018, 10, 22329–22339.
105. D. Xie, S. Chen, Z. Zhang, J. Ren, L. Yao, L. Wu, X. Yao and X. Xu, J. Power Sources, 2018, 389, 140–147.
106. J. Hu, P. He, B. Zhang, B. Wang and L.-Z. Fan, Energy Storage Mater., 2020, 26, 283–289.
107. H. Duan, Y. X. Yin, Y. Shi, P. F. Wang, X. D. Zhang, C. P. Yang, J. L. Shi, R. Wen, Y. G. Guo and L. J. Wan, J. Am. Chem. Soc., 2018, 140, 82–85.
108. X. Chen, W. He, L.-X. Ding, S. Wang and H. Wang, Energy Environ. Sci., 2019, 12, 938–944.
109. T.-J. Cai, Y.-H. Lo and J.-J. Wu, Mater. Today Energy, 2019, 13, 119–124.
110. K. K. Fu, Y. Gong, J. Dai, A. Gong, X. Han, Y. Yao, C. Wang, Y. Wang, Y. Chen, C. Yan, Y. Li, E. D. Wachsman and L. Hu, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 7094–7099.
111. Z. Wan, D. Lei, W. Yang, C. Liu, K. Shi, X. Hao, L. Shen, W. Lv, B. Li, Q.-H. Yang, F. Kang and Y.-B. He, Adv. Funct. Mater., 2019, 29, 1805301.
112. H. Xie, C. Yang, K. K. Fu, Y. Yao, F. Jiang, E. Hitz, B. Liu, S. Wang and L. Hu, Adv. Energy Mater., 2018, 8, 1703474.
113. P. Zhu, C. Yan, M. Dirican, J. Zhu, J. Zang, R. K. Selvan, C.-C. Chung, H. Jia, Y. Li, Y. Kiyak, N. Wu and X. Zhang, J. Mater. Chem. A, 2018, 6, 4279–4285.
114. X. Wang, H. Zhai, B. Qie, Q. Cheng, A. Li, J. Borovilas, B. Xu, C. Shi, T. Jin, X. Liao, Y. Li, X. He, S. Du, Y. Fu, M. Dontigny, K. Zaghib and Y. Yang, Nano Energy, 2019, 60, 205–212.
115. Y. Zhao, C. Wu, G. Peng, X. Chen, X. Yao, Y. Bai, F. Wu, S. Chen and X. Xu, J. Power Sources, 2016, 301, 47–53.
116. D. Lin, W. Liu, Y. Liu, H. R. Lee, P. C. Hsu, K. Liu and Y. Cui, Nano Lett., 2016, 16, 459–465.
117. T. Jiang, P. He, G. Wang, Y. Shen, C. W. Nan and L. Z. Fan, Adv. Energy Mater., 2020, 10, 1903376.
118. D. H. Kim, Y. H. Lee, Y. B. Song, H. Kwak, S. Y. Lee and Y. S. Jung, ACS Energy Lett., 2020, 5, 718–727.
119. X. Ma, X. Zuo, J. Wu, X. Deng, X. Xiao, J. Liu and J. Nan, J. Mater. Chem. A, 2018, 6, 1496–1503.
120. C. Wang, T. Wang, L. Wang, Z. Hu, Z. Cui, J. Li, S. Dong, X. Zhou and G. Cui, Adv. Sci., 2019, 6, 1901036.
121. Y. Li, X. Wang, H. Zhou, X. Xing, A. Banerjee, J. Holoubek, H. Liu, Y. S. Meng and P. Liu, ACS Energy Lett., 2020, 5, 955–961.
122. B. Timurkutluk and S. Dokuyucu, Ceram. Int., 2018, 44, 17399–17406.
123. X. Chen, W. Ni, X. Du, Z. Sun, T. Zhu, Q. Zhong and M. Han, J. Mater. Sci. Technol., 2019, 35, 695–701.
124. D. H. S. Tan, A. Banerjee, Z. Chen and Y. S. Meng, Nat. Nanotechnol., 2020, 15, 170–180.
125. A. Li, X. Liao, H. Zhang, L. Shi, P. Wang, Q. Cheng, J. Borovilas, Z. Li, W. Huang, Z. Fu, M. Dontigny, K. Zaghib, K. Myers, X. Chuan, X. Chen and Y. Yang, Adv. Mater., 2020, 32, e1905517.
126. R. K. Nishihora, P. L. Rachadel, M. G. N. Quadri and D. Hotza, J. Eur. Ceram. Soc., 2018, 38, 988–1001.
127. X. Yan, Z. Li, Z. Wen and W. Han, J. Phys. Chem. C, 2017, 121, 1431–1435.
128. K. Fu, Y. Gong, G. T. Hitz, D. W. McOwen, Y. Li, S. Xu, Y. Wen, L. Zhang, C. Wang, G. Pastel, J. Dai, B. Liu, H. Xie, Y. Yao, E. D. Wachsman and L. Hu, Energy Environ. Sci., 2017, 10, 1568–1575.
129. G. T. Hitz, D. W. McOwen, L. Zhang, Z. Ma, Z. Fu, Y. Wen, Y. Gong, J. Dai, T. R. Hamann, L. Hu and E. D. Wachsman, Mater. Today, 2019, 22, 50–57.
130. C. Yang, L. Zhang, B. Liu, S. Xu, T. Hamann, D. McOwen, J. Dai, W. Luo, Y. Gong, E. D. Wachsman and L. Hu, Proc. Natl. Acad. Sci. U. S. A., 2018, 115, 3770–3775.
131. S. Xu, D. W. McOwen, C. Wang, L. Zhang, W. Luo, C. Chen, Y. Li, Y. Gong, J. Dai, Y. Kuang, C. Yang, T. R. Hamann, E. D. Wachsman and L. Hu, Nano Lett., 2018, 18, 3926–3933.
132. B. Liu, L. Zhang, S. Xu, D. W. McOwen, Y. Gong, C. Yang, G. R. Pastel, H. Xie, K. Fu, J. Dai, C. Chen, E. D. Wachsman and L. Hu, Energy Storage Mater., 2018, 14, 376–382.
133. S. Xu, D. W. McOwen, L. Zhang, G. T. Hitz, C. Wang, Z. Ma, C. Chen, W. Luo, J. Dai, Y. Kuang, E. M. Hitz, K. Fu, Y. Gong, E. D. Wachsman and L. Hu, Energy Storage Mater., 2018, 15, 458–464.
134. P. Zhu, C. Y. Yan, J. D. Zhu, J. Zang, H. Jia, X. Dong, Z. Du, C. M. Zhang, N. Q. Wu, M. Dirican, X. W. Zhang and Y. Li, Energy Storage Mater., 2019, 17, 220–225.
135. K. Yoshima, Y. Harada and N. Takami, J. Power Sources, 2016, 302, 283–290.
136. A. Paolella, W. Zhu, G. L. Xu, A. La Monaca, S. Savoie, G. Girard, A. Vijh, H. Demers, A. Perea, N. Delaporte, A. Guerfi, X. Liu, Y. Ren, C. J. Sun, J. Lu, K. Amine and K. Zaghib, Adv. Energy Mater., 2020, 10, 2001497.
137. P. Chang, H. Mei, S. Zhou, K. G. Dassios and L. Cheng, J. Mater. Chem. A, 2019, 7, 4230–4258.
138. S. Zekoll, C. Marriner-Edwards, A. K. O. Hekselman, J. Kasemchainan, C. Kuss, D. E. J. Armstrong, D. Cai, R. J. Wallace, F. H. Richter, J. H. J. Thijssen and P. G. Bruce, Energy Environ. Sci., 2018, 11, 185–201.
139. D. W. McOwen, S. Xu, Y. Gong, Y. Wen, G. L. Godbey, J. E. Gritton, T. R. Hamann, J. Dai, G. T. Hitz, L. Hu and E. D. Wachsman, Adv. Mater., 2018, 30, e1707132.
140. J. M. Whiteley, P. Taynton, W. Zhang and S.-H. Lee, Adv. Mater., 2015, 27, 6922–6927.
141. M. Xie, X. Lin, Z. Huang, Y. Li, Y. Zhong, Z. Cheng, L. Yuan, Y. Shen, X. Lu, T. Zhai and Y. Huang, Adv. Funct. Mater., 2019, 30, 1905949.
142. X. Ren, L. Zou, X. Cao, M. H. Engelhard, W. Liu, S. D. Burton, H. Lee, C. Niu, B. E. Matthews, Z. Zhu, C. Wang, B. W. Arey, J. Xiao, J. Liu, J.-G. Zhang and W. Xu, Joule, 2019, 3, 1662–1676.
143. X. Wang, Y. L. Ding, Y. P. Deng and Z. Chen, Adv. Energy Mater., 2020, 10, 1903864.
144. H. Maleki Kheimeh Sari and X. Li, Adv. Energy Mater., 2019, 9, 1901597.
145. W. Li, B. Song and A. Manthiram, Chem. Soc. Rev., 2017, 46, 3006–3059.
146. T. T. Dong, J. J. Zhang, G. J. Xu, J. C. Chai, H. P. Du, L. L. Wang, H. J. Wen, X. Zang, A. B. Du, Q. M. Jia, X. H. Zhou and G. L. Cui, Energy Environ. Sci., 2018, 11, 1197–1203.
147. J. Li, W. Li, Y. You and A. Manthiram, Adv. Energy Mater., 2018, 8, 1801957.
148. W. Zhou, Z. Wang, Y. Pu, Y. Li, S. Xin, X. Li, J. Chen and J. B. Goodenough, Adv. Mater., 2019, 31, e1805574.
149. J.-Y. Liang, X.-X. Zeng, X.-D. Zhang, T.-T. Zuo, M. Yan, Y.-X. Yin, J.-L. Shi, X.-W. Wu, Y.-G. Guo and L.-J. Wan, J. Am. Chem. Soc., 2019, 141, 9165–9169.
150. A. Manthiram, Y. Fu, S. H. Chung, C. Zu and Y. S. Su, Chem. Rev., 2014, 114, 11751–11787.
151. D. Q. He, J. T. Meng, X. Y. Chen, Y. Q. Liao, Z. X. Cheng, L. X. Yuan, Z. Li and Y. H. Huang, Adv. Funct. Mater., 2020, 2001201.
152. J. Wu, X. Li, H. Zeng, Y. Xue, F. Chen, Z. Xue, Y. Ye and X. Xie, J. Mater. Chem. A, 2019, 7, 7897–7906.
153. J. Wu, H. Zeng, X. Li, X. Xiang, Y. Liao, Z. Xue, Y. Ye and X. Xie, Adv. Energy Mater., 2018, 8, 1802430.
154. D. He, Y. Liao, Z. Cheng, X. Sang, L. Yuan, Z. Li and Y. Huang, Sci. China Mater., 2020, 63, 1663.
155. J. Wu, N. You, X. Li, H. Zeng, S. Li, Z. Xue, Y. Ye and X. Xie, J. Mater. Chem. A, 2019, 7, 7644–7653.
156. E. Umeshbabu, B. Zheng and Y. Yang, Electrochem. Energy Rev., 2019, 2, 199–230.
157. P. G. Bruce, S. A. Freunberger, L. J. Hardwick and J.-M. Tarascon, Nat. Mater., 2012, 11, 19–29.
158. D. Marmorstein, T. H. Yu, K. A. Striebel, F. R. McLarnon, J. Hou and E. J. Cairns, J. Power Sources, 2000, 89, 219–226.
159. T. Yamada, S. Ito, R. Omoda, T. Watanabe, Y. Aihara, M. Agostini, U. Ulissi, J. Hassoun and B. Scrosati, J. Electrochem. Soc., 2015, 162, A646–A651.
160. S.-H. Kim, K.-H. Choi, S.-J. Cho, J. Yoo, S.-S. Lee and S.-Y. Lee, Energy Environ. Sci., 2018, 11, 321–330.
161. H. S. Shin, W. G. Ryu, M. S. Park, K. N. Jung, H. Kim and J. W. Lee, ChemSusChem, 2018, 11, 3184–3190.
162. S. H. Kim, J. H. Kim, S. J. Cho and S. Y. Lee, Adv. Energy Mater., 2019, 9, 1901841.
163. W. Liu, M.-S. Song, B. Kong and Y. Cui, Adv. Mater., 2017, 29, 1603436.
164. H. J. Peng, J. Q. Huang and Q. Zhang, Chem. Soc. Rev., 2017, 46, 5237–5288.
165. J. Duan, W. Wu, A. M. Nolan, T. Wang, J. Wen, C. Hu, Y. Mo, W. Luo and Y. Huang, Adv. Mater., 2019, 31, 1807243.
166. H. Xu, S. Li, C. Zhang, X. L. Chen, W. J. Liu, Y. H. Zheng, Y. Xie, Y. H. Huang and J. Li, Energy Environ. Sci., 2019, 12, 2991–3000.
167. H. Xu, S. Li, X. Chen, C. Zhang, W. Liu, H. Fan, Y. Yu, Y. Huang and J. Li, Adv. Energy Mater., 2019, 9, 1902150.
168. Y. Huang, B. Chen, J. Duan, F. Yang, T. Wang, Z. Wang, W. Yang, C. Hu, W. Luo and Y. Huang, Angew. Chem., Int. Ed., 2020, 59, 3699–3704.

---