Recent advances in lithium metal protective strategies with a stable interface

Abstract

Lithium metal batteries (LMBs) have regained attention for the next-generation storage system due to their exceptional energy densities. However, Li metal anodes suffer from serious dendrite growth, unstable solid electrolyte interface (SEI), and enormous volume fluctuation, resulting in low cycling stability and compromised safety. This review systematically summarizes the recent advances in lithium metal protective strategies toward high-performance LMBs. First, 3D lithium metal host designs with homogeneous and gradient structures are developed with the aim of optimizing the structure and nucleation kinetics to guide Li plating behavior and achieve uniform lithium deposition. Second, interfacial engineering involves the construction of an artificial solid electrolyte interface (SEI) and the improvement of the innate SEI through interfacial modulation to protect lithium electrodes. Third, electrolyte additives and solid-state electrolytes are developed to form a stable SEI and suppress the dendrite formation. Finally, this review outlines the current challenges and future rational designs for the protection of LMBs to promote further development of high-energy-density LMBs.

---

1. Introduction

The growing demand for high-energy-density battery systems, driven particularly by portable electronics and electric vehicles, has spurred the development of sustainable energy and its related industries, as well as research into energy storage equipment. Although research on lithium-ion batteries has advanced significantly over the past three years, the current market-standard graphite anode batteries (with an energy density of around 300 Wh kg⁻¹) still fall short of the requirements for next-generation high-energy-density storage systems.¹

Lithium metal possesses an ultra-high theoretical specific capacity (3860 mAh g⁻¹) and the lowest redox potential (−3.04 V vs. SHE). Consequently, lithium metal batteries (LMBs) can achieve a high energy density of over 500 Wh kg⁻¹ and are thus regarded as a leading candidate for the next-generation high-energy-density energy storage devices.² Despite the above-mentioned merits, the practical applications of LMBs have been restricted by the intrinsic notorious problems of the lithium metal anode (LMA). In contrast to the intercalation mechanism of graphite-based anodes, LMAs conduct a simple conversion between the Li metal and the Li ion.^3,4 This repeated plating/stripping process leads to the growth of disordered Li dendrites and serious volume changes. The fragile dendritic structures are prone to detachment from the LMA, forming electrochemically inactive “dead Li”.⁵ Continuous volume fluctuations cause severe mechanical damage and necessitate constant SEI repair, while the accumulation of “dead Li” might lead to low coulombic efficiency (CE), poor cycling stability, and reduced Li utilization rate. More seriously, the uncontrolled growth of Li dendrites may puncture the separator, posing serious safety hazards such as battery short circuits.⁶

Due to the low redox potential (−3.04 V vs. SHE), metal Li spontaneously reacts with a liquid electrolyte, which decomposes at the electrode–electrolyte interfaces to form an SEI layer.⁶ Ideally, the SEI layer is electronically insulating and ionically conductive, serving as a protective barrier that isolates metal Li from the electrolyte.⁷ However, during repeated Li plating/stripping, inhomogeneous Li deposition and incomplete dissolution induce a large volume change, causing fracture of this inherently fragile SEI. The freshly exposed Li metal then reacts with the electrolyte, triggering the repeated interface reaction. Concurrently, lithium ions preferentially deposit at the SEI cracks, initiating the initial nucleation of Li dendrites.⁸ In traditional liquid electrolytes, the high reactivity of Li metal can easily provoke side reactions and rapid electrolyte consumption. The electrode corrosion and continuous depletion of electrolytes will affect the irreversible capacity decay of the battery.⁹ Therefore, surface engineering techniques that establish a chemical stable protective barrier have been extensively explored to mitigate these issues and enhance the battery performance.

In lithium metal batteries, interfacial stability is fundamentally governed by the dynamic competition among thermodynamic spontaneous reactions, kinetic mass transfer processes, and structural–mechanical factors. The extremely high reducing nature of lithium metal inevitably leads to interfacial reactions, and the key to achieving interfacial stability lies in constructing and maintaining a dynamically balanced interface capable of regulating ion transport, mechanical stress, and chemical reaction kinetics.

Dendritic lithium, lithium corrosion, and the consumption of liquid electrolytes are the causes of the failure of LMBs. Along with intensive research efforts for LMBs, typical strategies such as composite anode designs,^10–13 artificial SEI (ASEI),^14,15 aqueous electrolyte engineering,^16–18 and solid-state electrolytes^3,19–21 all contribute to stable Li electrodes. Hence, among them, structural designs mainly involve nucleation site modifications, and 3D structural hosts are aimed at promoting uniform and compact deposition of lithium.²² On the other hand, homogenizing the Li⁺ flux on the surface of the lithium anode endows the anode with a tunable plating/stripping behavior. The artificial SEI with certain mechanical strength regulates the Li⁺ flux and inhibits the uncontrollable growth of lithium dendrites.²³ In addition, configuring liquid electrolytes with additives and replacing flammable organic liquid electrolytes with solid electrolytes are efficient strategies to generate a stable SEI and alleviate metal anode safety issues.^24,25 Solid-state electrolytes (SSEs) with a robust polymer network can impede the lithium dendrite growth.²⁶ Different from combustible organic electrolytes, the inherent security of solid electrolytes endows LMBs with prolonged cycling stability and overall safety.

The vast majority of strategies converge on the ultimate goal of “constructing a stable interface.” For instance, both 3D hosts and flexible polymer interlayers aim to promote uniform lithium nucleation and manage interfacial stress, thus contributing to the formation of a mechanically stable SEI. Additives and artificial SEI coatings are designed to directly modify interfacial chemistry and enhance ion transport. In contract, 3D porous hosts and solid-state electrolytes work to suppress dendrite growth, thereby establishing a stable interface layer. Indeed, trade-offs often exist among different strategy objectives. For example, while a high-modulus inorganic SEI or solid-state electrolytes (SSEs) help mechanically block dendrite growth, they are typically accompanied by poor solid–solid contact and high interfacial impedance, which hinder ion transport. Similarly, while 3D porous hosts homogenize deposition by increasing the surface area, this enlarged interface may exacerbate side reactions, leading to accelerated electrolyte depletion. Thermodynamically stable SEI components (such as LiF) often exhibit relatively low ionic conductivity. This inherent trade-off highlights a central challenge in the interface design: achieving both interfacial stability and efficient ion transport necessitates careful and balanced design.

Lithium deposition/stripping involves significant volume changes and localized stress concentration. A static interfacial structure cannot withstand the dynamic lithium plating/stripping process, making mechanical adaptability a universal requirement for achieving long-term stable cycling. In addition, an ideal lithium metal anode design comprises three key elements: uniform lithium-ion flux, robust interfacial mechanics, and minimal kinetic barriers. Furthermore, unlike liquid electrolyte systems, solid-state electrolytes face the challenge of balancing solid–solid contact impedance against high-modulus suppression of dendrite growth, with primary solutions centered on interface fusion/welding techniques.

This review comprehensively examines the obstructions and protective strategies for lithium metal anodes reported in recent years, with emphasis on three primary approaches (Fig. 1). First, a rational design of three-dimensional host structures with homogeneous or gradient architectures is developed to regulate the lithium deposition behavior. Second, interfacial engineering focuses on artificial SEI construction and innate SEI modification. Third, electrolyte optimization using electrolyte additives and solid-state electrolytes is related to stable SEI formation. We discuss these strategies in detail and present the further development prospects of high-performance LMBs.

Fig. 1 Schematic of the design principles and modification strategies to improve the stability of LMA, including 3D lithium metal host design, artificial solid electrolyte interface construction, use of electrolyte additives and solid-state electrolyte development.

2. Lithium metal host design

Bare Li foil undergoes significant deformation and volume expansion during cycling due to its intrinsic malleability and insufficient mechanical strength, leading to low coulombic efficiency (CE), dendrite growth, and rapid capacity decay. Moreover, the use of excess Li (>50 mAh cm⁻²) and high negative/positive (N/P) ratios (>50) remain critical challenges in high-energy-density battery systems.²⁷ The composite host design has emerged as a promising strategy, offering high specific surface areas and tunable mechanical properties. The 3D porous architecture combined with tailored surface chemistry helps mitigate volume variation and homogenize local current density, thereby significantly suppressing the overall macroscopic deformation of the electrode and maintaining long-term interfacial stability.

Common techniques for fabricating composite Li metal anodes based on 3D hosts include mechanical rolling, molten Li infusion, and electrodeposition. Among various host materials, 3D carbon-based frameworks—such as graphenes, carbon nanofibers (CNFs), and carbon nanotubes (CNTs)—and 3D metal-based scaffolds (e.g., Cu and Ni foams) have been widely recognized as effective architectures for stabilizing Li metal anodes.^28–32

2.1 Carbon-based hosts

The 3D carbon skeleton possesses high electrical conductivity and stable electrochemical stability. Most carbon materials are nonpolar substrates, and they exhibit poor wettability of Li. However, their surface chemistry can be tailored and loaded with lithiophilicity to improve the wettability. Huang et al. introduced high-activity silver nanoparticles into a CNF surface to enhance the interfacial dynamics.³³ Cao et al. fabricated a metal anode by the electrochemical deposition of Li on a 3D CNF matrix with SiO_x, and the composite anode can be operated over 1350 h in a symmetrical cell.³⁴ Importantly, the preparation techniques of the carbon host are easily operated for structural properties (such as surface area, pore structure, electrical conductivity, and lithiophilic sites) and large-scale production.

First, the inherent mechanical flexibility will contribute to construct flexible electrodes and eliminate the volume stress during the electrochemical process. Moreover, modified carbon-based hosts with lithiophilic sites can make the Li-ion flux uniform, thereby facilitating the uniform deposition of Li and mitigating dendrite growth. For instance, a graphene host modified with pyrrolic and pyridinic nitrogen, which can induce a low nucleation overpotential of 22 mV, exhibited a dendrite-free morphology.³⁵ Zhang et al. constructed a flexible Cu₃P/CoP@C/CNT host to plate metallic Li, where the confinement ability of the interlaced hierarchical porous structure can relieve volume expansion.³⁶ An oxygen-functionalized CNF host (Ni@PCNF-O) decorated with Ni nanoparticles was designed to electrodeposit Li (Fig. 2a).³⁷ The mesoporous structure and surface oxygen groups of the CNF electric field could regulate the whole electric field to achieve uniform nucleation and growth. The Ni@PCNF-O symmetrical cell exhibited stable cycling performance (1200 cycles) at 0.5 mA cm⁻² to 1 mAh cm⁻². Deng et al. utilized atomic layer deposition to cover ZnO on a 50 µm-thick GO film (0VCCS-rGO&ZnO), showing a near-zero volume change upon cycling (Fig. 2b and c).³⁸ After the infusion of molten Li, lithium metal was confined within the micro-scale cavities of the multilayer rGO structure. The continuous multilayer architecture of the 0VCCS-rGO&ZnO host effectively prevented electrolyte penetration and subsequent reaction with the metallic lithium. After stripping/plating cycles, the composite anode exhibited negligible volume changes, as shown in Fig. 2c. This volumetric stability contributed to the formation of a mechanically stable, inorganic-rich SEI and provided effective protection for the lithium metal.

Fig. 2 (a) Schematic of the Li plating behavior on the Ni@PCNF-O host. Reproduced with permission.³⁷ Copyright 2021, Elsevier Ltd. (b) Schematic of the fabrication process of the porous rGO&ZnO host and Li@rGO&ZnO. (c) Cross-sectional SEM images and 3D schematic of the pre-cycling 50-µm-thick Li@rGO&ZnO electrode, Li@rGO&ZnO electrode after Li stripping, Li@rGO&ZnO electrode after Li plating, and Li@rGO&ZnO electrode after 400 cycles. Reproduced with permission.³⁸ Copyright 2021, Nature Research. (d) Li deposition mechanism on ISHCP. (e) Li deposition mechanism on Ni₂P@ISHCP. Reproduced with permission.³⁹ Copyright 2022, Elsevier Ltd.

The asymmetric cell composed of 0VCCS@rGO&ZnO and Li foil delivered an average CE of 99.99% over 2000 cycles. More interestingly, an interconnected graphitized hollow carbon sphere with Ni₂P nanoparticles was constructed as the Li deposition host (Ni₂P@ISHCP).³⁹ The carbon architecture with a large specific surface area of 2.66 cm³ g⁻¹ effectively reduced the local current density, but the poor lithiophilicity of ISHCP and the high Li⁺ concentration resulted in serious Li metal dendrite growth (Fig. 2e). The selective Li deposition mechanism of Ni₂P@ISHCP is shown in Fig. 2e; Li metal is guided to selectively nucleate and uniformly deposit within the carbon cavity, benefiting from the low nucleation energy barrier of Ni₂P. The Ni₂P@ISHCP symmetrical battery achieved low polarization (12 mV) and stable cycling for 1000 h at 1 mA cm⁻² to 1 mAh cm⁻². Hao et al. proposed a high-performance Li anode host that integrated a CNF scaffold with CoO_x.⁴⁰ The embedded CoO_x favored the formation of a Li₂O-rich SEI, significantly improving the Li⁺ transport kinetics and achieving uniform Li deposition.

In brief, the surface chemistry of carbon-based hosts can be artificially modified to introduce lithiophilic sites (e.g., metallic active sites or functional groups), while their three-dimensional (3D) porous structure helps lower the local current density, thereby promoting uniform Li deposition. However, certain carbon-based hosts suffer from insufficient mechanical stability, and the additional mass of the carbon scaffold reduces the overall energy density of lithium metal batteries (LMBs). Notably, the high specific surface area of such hosts inevitably increases the electrode–electrolyte interface area, leading to substantial liquid–electrolyte consumption. Therefore, guided by computational simulations, the rational design of lightweight, mechanically robust carbon hosts with an optimized specific surface area will be crucial for developing high-energy-density LMBs.

2.2 Metal-based hosts

In studies of LMAs, commonly used metallic scaffolds include Ti-mesh, Ni mesh, Ni foam and Cu foam.^41–43 Metal current collectors can homogenize the electric field; however, most of them are inactive toward Li nucleation. Introducing lithiophilic materials (CuO, ZnO, Ni₂P, Si, and nitrogen-doped carbon) into a metal scaffold can effectively modify its lithiophobic nature, thereby greatly reducing the energy barrier for Li nucleation.^44–47 A 3D Ni foam (NSNF) modified with a lithiophilic NiSe coating was reported to regulate the uniform Li nucleation and growth (Fig. 3a).⁴³ The composite anode was constructed by mechanical rolling, and the surface of the NSNF shows good compatibility with the Li metal. Consequently, the composite anode with a robust Li₂Se-rich SEI demonstrated high rate capacity (10 mA cm⁻²) and stable cycling performance (6600 cycles at 5 mA cm⁻² to 1 mAh cm⁻²) in a symmetric cell. Wang et al. employed a uniform ultrathin Co₃O₄ nanoarray layer to decorate the 3D Ni host, which could effectively homogenize the Li⁺ flux for regulating Li plating/stripping behaviors (Fig. 3b).⁴⁸ Well-arranged Co₃O₄ nanosheets, despite their low intrinsic electronic conductivity, exhibit high Li⁺ adsorption energy. The adsorbed Li⁺ ions subsequently nucleate on the highly conductive NF substrate, guiding metallic Li growth along the Co₃O₄ nanoarrays, as evidenced in Fig. 3c and d. The highly conductive network ensures an extremely low charge transfer resistance, thereby facilitating rapid Li⁺ reduction. Uniform Co₃O₄ nanoarrays with abundant pores effectively mitigate the volume expansion and reduce local current density. The synergistic combination of lithiophilic Co₃O₄ and the designed conductivity gradient promotes a “bottom-up” Li deposition mechanism. This structural advantage endows the Li-uCo₃O₄@NF composite anode with a high coulombic efficiency (96.5%) maintained over 260 cycles and a long lifespan of over 1100 h (1 mA cm⁻²/1 mAh cm⁻²). Significantly, COMSOL simulation and operando optical microscopy were performed to verify this “bottom-up” deposition mode and Li plating morphologies. Yu et al. designed a 3D Ni mesh with lithiophilic Ni₃Sn₂ and NiF₂ coating (NSF).⁴² In the subsequent molten Li infusion and electrodeposition process, the in situ configured Li–Sn alloy and LiF-rich SEI synergistically suppressed dendrite growth. The LNSF symmetric cell cycled for 1400 h at 1 mA cm⁻², while the LNSF‖LiFePO₄ full cell exhibited a capacity retention of 92.8% after 200 cycles at 0.5C.

Fig. 3 (a) Schematic of the fabrication process of the NANF-hosted Li metal. Reproduced with permission.⁴³ Copyright 2024, Wiley Online Library. (b) Illustration of the synthetic procedure and Li plating behavior of uCo₃O₄@NF. (c) Uniform Li plating behavior in the “bottom-up” mode in the uCo₃O₄@NF array. (d) SEM image of uCo₃O₄@NF. Reproduced with permission.⁴⁸ Copyright 2024, Elsevier Ltd. (e) Schematic of the multifunctional layer evolution in the LNSF mesh. Reproduced with permission.⁴² Copyright 2025, Elsevier Ltd.

Metal scaffolds exhibit a stable structure and superior mechanical strength, but the mass densities are ten or even dozens of times that of lithium metal. The metal scaffold may occupy the major mass and volume of the LMBs, which does not conform to the concept of high energy density.

Metal hosts typically rely on their high modulus and rigidity to physically constrain the volume expansion during lithium deposition. However, within rigid pores, the mechanical constraints on lithium deposition and dissolution are often spatially uneven. Lithium tends to preferentially deposit in areas with weaker constraints and easier stress relaxation, such as pore openings or large cavities, rather than filling the entire internal space of the host uniformly. This phenomenon can instead lead to localized lithium protrusions and accelerate early battery failure. During long-term cycling, the host skeleton may gradually undergo micro-plastic deformation or even fatigue cracks under the cyclic stress induced by the repeated intercalation/deintercalation. This effect is particularly detrimental to brittle intermetallic compound coatings or surface modification layers. Therefore, the future design philosophy for hosts should shift from “pursuing absolute rigid constraints” to “achieving adaptive mechanical guidance.” Ideal host materials should possess a gradient modulus or controllable elasticity and be capable of guiding uniform lithium nucleation and growth through surface chemistry modulation.

During lithium deposition, if the deposited lithium metal fails to form and maintain intimate physical contact with the inner surface of the host material, interfacial micro-voids are generated. These micro-voids cause the lithium trapped within them to lose electrical contact with the current collector, thereby transforming into electrochemically inert “dead lithium”. This process directly results in irreversible loss of active lithium, manifesting as capacity decay and a decline in coulombic efficiency. Simultaneously, during the repeated deposition and stripping of lithium, the periodic volume expansion and contraction induce significant cyclic stress and strain within the rigid host framework. This long-term, repeated mechanical stress can lead to irreversible plastic deformation of the host structure or initiate micro-cracks and even fractures in brittle regions, further compromising structural integrity and accelerating battery failure.

Table 1 summarizes the host materials, fabrication methods, and electrochemical performance of the composite Li anode, providing guidance for the development of LMBs.

Table 1 Cycling performance of symmetric cells and full cells based on composite anodes

Host materials	Fabrication method	Areal capacity of composite anode	Cycle life of symmetric cells (current density–capacity, cycle time)	Cathode (mg cm^−2)	Cycle life of full batteries (cycle number-retention, rate, N/P ratio)	Ref.
CNF@ZnO@Sn	Electrodeposition	3 mAh cm^−2	1 mA cm^−2–1 mAh cm^−2, 600 h	LFP, 10	300–92.1%, 0.5C, 3	55
CNF with gradient pore	Electrodeposition	3 mAh cm^−2	0.1 mA cm^−2–0.1 mAh cm^−2, 600 h	LFP	370–70.0%, 0.5C	56
Carbon cloth	Molten infusion	—	3 mA cm^−2–3 mAh cm^−2, 1000 h	—	—	57
CNF-CoPx	Electrodeposition	7 mAh cm^−2	3 mA cm^−2–1 mAh cm^−2, 600 h	LFP, 10.5	500–89.5%, 1C	58
CNF-Fe3C/Fe2O3	Electrodeposition	3 mAh cm^−2	3 mA cm^−2–1 mAh cm^−2, 400 h	LFP, 1.9	500–66.0%, 1C, 2	59
Ni foam-Co3O4	Electrodeposition	5 mAh cm^−2	1 mA cm^−2–1 mAh cm^−2, 1600 h	LFP	300–69.4%, 1C	60
CNF-Sn	Electrodeposition	15 mAh cm^−2	2 mA cm^−2–1 mAh cm^−2, 1600 h	NCM, 1.2	300–48.0%, 10C	61
CNF-Co/N	Electrodeposition	10 mAh cm^−2	1 mA cm^−2–1 mAh cm^−2, 1100 h	LFP, 5	500–89.0%, 1C	62
Ni foam–MnO2	Electrodeposition	4 mAh cm^−2	3 mA cm^−2–1 mAh cm^−2, 900 h	LFP, 1.5	800–81.2%, 2C	63
Ni foam–ZnO	Molten infusion	—	5 mA cm^−2–1 mAh cm^−2, 400 h	LFP, 4.24	450, 1C	64

---

2.3 Gradient hosts

While 3D porous hosts can alleviate dendrite growth to some extent, the Li metal tends to gradually deposit and aggregate on the top surface of the host due to the inherently high local Li⁺ concentration in the upper region, especially under high current density or large areal capacity.⁴⁹ Moreover, as the deposition capacity increases, active sites may be covered, losing their regulatory function. Ultimately, this phenomenon leads to obvious dendrite growth and low space utilization rate. To address this, the gradient-structured architecture involves spatial design and physicochemical gradient characteristics such as gradient porosity, gradient electronic conductivity, and lithiophilic–lithiophobic transitions, which are conducive to Li⁺ flux guidance, electric field adjustment, and local charge transport optimization within the electrode.^50,51 It has been demonstrated that variations in porosity or pore size gradients along the longitudinal direction can optimize the Li⁺ transport and induce a bottom-up deposition behavior.⁵² As illustrated in Fig. 4a–c, the normal 3D carbon cloth (CC) host presented high Li⁺ concentration on the top of the electrode, leading to the dendrite formation and growth.⁵³ Although the ZnO/CC host with uniformly distributed ZnO facilitated more uniform Li growth, the “top-growth” mode was still observed under high capacity or current density. The ZnO-gradient design introduced a higher Li⁺ flux into the electrode and promoted the Li⁺ migration from the top to the bottom, ultimately realizing a “bottom-up” deposition mode. Yu and co-workers constructed a bilayer host with an electronic conductivity gradient, where the insulative Si₃N₄ top layer improved the electrolyte wettability and the carbon black/LiNO₃-mixed bottom layer effectively enhanced charge transfer (Fig. 4d).¹⁰ Especially, the electronic–ionic conductive network of the bottom layer could reduce the Li nucleation barrier, enabling the asymmetric cell to work for over 600 cycles at 0.5 mA cm⁻² with a coulombic efficiency of 98.65%. As shown in Fig. 4e–h, the cross-sectional and top-view SEM images confirmed the dendrite-free growth. The full cell paired with a LiFePO₄ cathode maintained stable operation for 200 cycles. Song et al. proposed a dual-gradient pore-size/lithiophilicity matrix composed of a macroporous carbon fiber (CF) layer and a microporous carbon black (CB) particle layer.⁵⁴ The macroporous CF on the upper side ensured Li⁺ transport to the bottom of the host, and the microporous structure at the bottom could maintain the geometric integrity of the matrix. Gradient porosity creates a gradient in ionic transport resistance, where the large-pore region at the top facilitates the infiltration of liquid electrolytes and macroscale ion transport. The dual-gradient electrode induced preferential Li⁺ deposition at the bottom and ensured the long-term cycling stability performance in symmetric cells. Furthermore, the ex situ SEM images and in situ optical images confirmed the bottom-up Li deposition behavior. The macropore region at the top serves as a “buffer layer” or “reserved space,” accommodating subsequently deposited lithium and alleviating stress caused by volume expansion. The gradient structure guides a more uniform distribution of ion flow and electric field, avoiding excessive localized lithium accumulation and resulting in a denser, flatter deposition of lithium metal. This “bottom-up” deposition mechanism prevents the formation of mossy or dendritic lithium protrusions on the top surface near the separator.

Fig. 4 (a)–(c) Schematic of the deposition principle of CC, ZnO/CC, and GZn/CC hosts. Reproduced with permission.⁵³ Copyright 2025, Elsevier Ltd. (d) Structure and Li deposition mechanism of Si₃N₄-C/LiNO₃. (e)–(g) Cross-sectional SEM images of Si₃N₄-C/LiNO₃ with 0.5 mAh cm⁻², 2 mAh cm⁻², and 5 mAh cm⁻² Li deposition. (h) Top-view SEM image of Si₃N₄-C/LiNO₃ with 2 mAh cm⁻². Reproduced with permission.¹⁰ Copyright 2025, Elsevier Ltd.

Therefore, gradient hosts can eliminate the adverse impact of inhomogeneous Li⁺ flux, concentration polarization, and electrical field under high current densities, endowing LMBs with enhanced energy density in terms of practical application. As for high-energy-density batteries, the low cost, light, and simple preparation of lithiophilic hosts have to be taken into account. Rational gradient design adjusts the spatial electric field and Li⁺ flux to regulate Li growth behavior, reducing the possibility of short circuit and improving the safety of LMBs.

The above statements afford the composite Li anode features limited Li capacity, low N/P ratio (<3), and high current density (>3 mA cm⁻²).³² The composite lithium anode exerts a physical confinement effect that suppresses lithium dendrite growth and mitigates volume changes. Significantly, the composite Li avoids vast excess of Li metal in practical application. The fabrication of composite Li anode involves electrodeposition, molten Li infusion, and roll-in methods. Electrodeposition can precisely control the capacity loading of Li metal and adjust the N/P ratio in full cells, while the complex procedures cause sacrificial electrodes and incur extra high costs. The molten Li infusion method requires high-temperature and strict conditions. The roll-in method is a flexible and operable manufacturing technique for industrial and large-scale applications. In practical applications, the appropriate hosts, cost, and fabrication process have to be taken into account. Density functional theory (DFT) should be further employed to simulate the electric field and current density distribution in the electrode, in order to further elaborate the deposition mode of Li.

3. Artificial SEI construction

Research has demonstrated that the native SEI fails to maintain stability during Li plating and stripping. Interfacial engineering aims to construct artificial protective layers to suppress Li dendrite growth and mitigate side reactions. The core concept is to pre-establish a stable and functionalized interface—known as an ideal artificial SEI—before the LMA contacts the electrolyte.⁶⁵ An ideal artificial SEI requires high ionic conductivity, sufficient mechanical strength and chemical stability to homogenize the Li⁺ flux and withstand stress change, thereby maintaining a stable protective interface. Various artificial SEI layers can be classified into inorganics, organics, and organic–inorganic hybrid materials.

3.1 Inorganic SEI

Inorganic compounds such as rGO, SiO_x, Al₂O₃, and Li₃N exhibit high Young's modulus, dielectric constant, and ionic conductivity, and can act as physical barriers to inhibit dendrite penetration.^66–69 It has been reported that the ionic conductivity and Young's modulus of SEI should be above 1.5 × 10⁻³ S cm⁻¹ and 4.0 GPa to withstand stress concentrations and homogenize Li deposition.⁷⁰ The reported synthesis processes involve drop coating, scraping, spin coating, and atomic layer deposition (ALD). ALD is able to accurately control the thickness and consistency of atomic metal oxide coatings on the lithium metal surface. For instance, a consistent 14 nm-thick Al₂O₃ layer was fabricated by ALD to enhance the stability of LMA.⁷¹ Hu et al. protected Li metal with a hybrid SEI layer of lithium-antimony (Li₃Sb) alloy and lithium fluoride (LiF), in which Li₃Sb was confirmed as a superionic conductor for Li⁺ transport.⁷² A gallium–lithium (GaLi) protecting layer was fabricated via the ion-exchange reaction of gallium chloride and metal Li (Fig. 5a).⁷³ Compared to bare Li, the modified Li anode enabled favorable Li⁺ transfer kinetics and achieved uniform Li deposition even at high area capacity (Fig. 5b and c). Inorganic components (Li₂S, Li₃N, and Li₂O) possess high ionic conductivity. Thus, the rapid Li⁺ diffusion rate can be regulated to induce Li dense deposition. Chen et al. constructed a flower-shaped Li₃N layer on the surface of the Li anode using the N₂ plasma.⁷⁴ This artificial SEI layer with high Young's modulus (48 GPa) and ionic conductivity (0.502 mS cm⁻¹) not only physically suppressed the dendrite growth, but also isolated Li metal and organic electrolyte. The assembled Li/LiCoO₂ full cell exhibited a capacity retention of 96% at 5C. This plasma activation technique provides a scalable and efficient strategy to construct stable LMBs with improved electrochemical performance. The graphitic material works well in the protective layer due to its stronger mechanical stability. Zhu et al. proposed a bi-layer artificial SEI layer via comprehensive first-principles calculations.⁷⁵ This artificial SEI consisted of inorganic components (e.g., LiF, Li₂O, Li₃N, and Li₂CO₃) and graphitic layer (graphene and h-BN), where the defects of graphitic layer acted as Li⁺ diffusion channels. Tang et al. designed an air-stable inorganic-rich artificial SEI (LiF, LiOH, Li₂O and Li_xSiF_y) on the surface of metal Li using the spontaneous conversion and alloying reaction of metal Li with MgSiF₆·6H₂O in a hot rolling process (Fig. 5d).⁷⁶ The artificial layer formed in situ exhibited superior adhesion characteristics. Scanning electron microscopic (SEM) images showed a dense artificial SEI, which endowed fast Li⁺ transport, isolated both the air and the electrolyte, and significantly inhibited Li dendrite growth, in contrast to bare Li (Fig. 5e and f). In addition, Mg and Li_xMg_y alloy phase under the SEI layer act as lithiophilic sites to induce low overpotential and Li⁺ uniform plating. As a result, the symmetric Li‖Li cells demonstrated stable cycling for over 1500 h at 1 mA cm⁻² with 1 mAh cm⁻². Han's team proposed a self-regulated lithium-rich antiperovskite (LiRAP) layer as an adaptive SEI (LiRAP-ASEI) to suppress Li dendrite growth.⁷⁷ As shown in Fig. 5g and h, inorganic LiRAP-ASEI underwent repeated cracking and self-regulation due to the volumetric stress generated during the plating/stripping process. A Li‖Cu half-cell with a LiRAP interfacial layer had 99.5% coulombic efficiency over 1500 cycles at 0.5 mA cm⁻².

Fig. 5 (a) Schematic of the fabrication of GaLi–Li. (b) and (c) Cross-section SEM images of the bare Li and GaLi–Li after 10 cycles at 1 mA cm⁻² with 1 mAh cm⁻². Reproduced with permission.⁷³ Copyright 2021, Elsevier Ltd. (d) Schematic of the fabrication of the MSF@Li via the hot-rolling process. (e) and (f) SEM images of bare Li and MSF@Li after 10 cycles at 1 mA cm⁻² with 1 mAh cm⁻².⁷⁶ (g) Schematic of the cycling stability of the self-regulated LiRAP-ASEI layer. (h) SEM images of the self-regulated LiRAP SEI layer with different cycles at 2 mA cm⁻² with 2 mAh cm⁻². Reproduced with permission⁷⁷ Copyright 2020, Elsevier Ltd.

Inorganic SEI components (e.g., LiF) exhibit high Young's modulus (∼65 GPa) and high surface energy, which can withstand the mechanical stress generated during lithium deposition/stripping and enhance SEI stability. Their moderate Li⁺ diffusion barrier and low Li⁺ diffusion anisotropy help homogenize the local current density on the electrode surface. For instance, LiF exhibits exceptional chemical and electrochemical stability, which effectively suppresses the reductive decomposition side reactions of the electrolyte at the electrode interface, thereby establishing itself as a critical inert barrier for the stable operation of batteries. The high surface energy of LiF reduces the Li⁺ nucleation energy barrier, guiding uniform lithium deposition and suppressing dendrite growth.

3.2 Organic SEI

Although the inorganic artificial SEI facilitates fast Li⁺ diffusion, their rigid nature often prevents them from accommodating the volume changes of metal anode, leading to the worse conformality and fracture. Conventional coating strategies often employ elastic organic polymers with ionic conductivity such as poly(vinyl alcohol) (PVA), poly(vinylidene fluoride) (PVDF), polyurethane (PU), polydimethylsiloxane (PDMS), and poly(acrylamide) (PAM) on the Li surface.^78–80 These organic polymeric coatings provide sufficient elasticity and mechanical flexibility to adapt to volume change during cycling. As shown in Fig. 6a, a physical artificial SEI network (C-Li@P) was constructed by electrospinning of a mixed solution of carboxymethyl guar gum (CMGG) and polyacrylamide (PAM). Lithiophilic CMGG-Li and the amide functionality of PAM created a chemically and physically stable cation/anion-derived SEI layer. The hollow nanofiber guided Li⁺ transport and fostered Li nuclei and plating. As shown in Fig. 6b, the morphological evolution of C-Li@P revealed its smooth nanofiber structure and showed few apparent substances on the exterior nanofibers after 50 cycles, confirming the buffering effect of artificial SEI. The C-Li@P membrane-covered Li symmetric cell maintained a stable plating/stripping cycling for 1500 h, surpassing 750% of bare Li. Wu et al. built a PVA-based SEI (DPN) with rich lithiophilic/anionphilic groups on Li metal.⁸⁰ Fig. 6c displays the interpenetrating structure of PVA chains and γ-CD rings, where H bonds provide flexible mechanical robustness. The SEM images of the cycled DPN@Li showed good harmony between the Li metal and the DPN layer, while the stable lithiophilic ability and considerable resilience are conducive to achieve long-term cycling stability (Fig. 6d and e).

Fig. 6 (a) Schematic of the fabrication process of the C-Li@P membrane. (b) Schematic and cross-sectional and top views of the C-Li@P membrane. Reproduced with permission.⁷⁹ Copyright 2024, Wiley Online Library. (c) Schematic of the plating process. (d) Cross-sectional view of DPN@Li after Li plating. (e) Cross-sectional view of DPN@Li after Li stripping. Reproduced with permission.⁸⁰ Copyright 2024, Elsevier Ltd.

Flexible organic coating can form intimate interfacial contact with Li to adapt to volume stress. In situ organic coatings are formed via a spontaneous reaction between the Li metal and organic inducers to create strong binding. Importantly, modifying organic molecules with fluorine-base groups or lithium-salt-like moieties can adjust the stability or ionic conductivity of artificial SEI. Li et al. created a negatively charged artificial SEI by the in situ polymerization of 2,3,7,8-tetrakis((trimethylsilyl)ethynyl)pyrazino[2,3-g]quinoxaline-5,10-dione (PPQ) on Li foil (PPQ-Li), which exhibited strong mechanical strength (Young's modulus of 7.39 GPa), as shown in Fig. 7a.⁸¹ The negatively charged polymer layer performed a repulsion of anions to achieve a preferential Li⁺ transport (t_Li⁺ = 0.74). This artificial SEI served as effective physical barriers to inhibit dendritic penetration in terms of its mechanical stability. The NCM‖PPQ-Li cell maintained a capacity retention of 76% after 350 cycles. A protection layer (LiOPs) formed by octaphenylsiloxane (OPS) and LiFSI salt was proposed by Fang et al.⁸² Insulated OPS layer not only exhibited strong Li⁺ adsorption to regulate the Li⁺ flux, but also induced Li deposition under an artificial LiOP layer to limit dendrite development (Fig. 7b). Compared to bare Li, the LiOPs reduced dendritic formation and stabilized the LMA interface, forming dense and uniform lithium particles on the LiOP electrode (Fig. 7c and d). Wang et al. utilized N-vinylcarbazole (NVK) to polymerize an artificial SEI on the Li anode surface, yielding highly Li⁺ conductive Li₃N that enhances uniform Li⁺ flux and promotes homogeneous Li deposition (Fig. 7e).⁸³ The corresponding Li‖Li symmetric cell exhibited an excellent stability with 350 cycles at 1 mA cm⁻² to 0.5 mAh cm⁻² (Fig. 7f).

Fig. 7 (a) Schematic of the in situ process of PPQ-Li. Reproduced with permission.⁸¹ Copyright 2021, Elsevier Ltd. (b) Schematic of bare Li and LiOPs. (c) SEM images of the Li electrode after Li plating. (d) SEM images of the LiOP electrode after Li plating. Reproduced with permission.⁸² Copyright 2023, Wiley Online Library. (e) Schematic of the artificial SEI with PVK. (f) Cycling performance of Li‖Li symmetric cells with blank, 0.25 wt% VK-containing, and 0.25 wt% PVK-containing electrolytes at 1.0 mA cm⁻² with 0.5 mAh cm⁻². Reproduced with permission.⁸³ Copyright 2025, AAAs.

Some organic SEI layers exhibit insufficient elastic strength and difficulty in resisting lithium dendrites during long-term cycling. The modifiable structure of organic molecules brings infinite possibilities for the construction of organic artificial SEI. This target is to improve the mechanical strength and ionic conductivity. More, more research has focused their attention on organic inorganic composite ASEIs, taking into account the integrated merit of both components.

3.3 Organic/inorganic composited SEI

Inorganic–organic hybrid coatings that integrate the advantages of both flexible organic components and rigid inorganic components with uniform Li⁺ diffusion and dendrite suppression have attracted much attention in the protection of LMAs. The synergistic interaction significantly reduces the interfacial side reaction and improves the cycling stability of LMBs, while its inherent chemical and mechanical compatibility with the Li anode further ensures robust interfacial stability. For example, Wu et al. proposed a double-layer artificial layer on the Li surface, where robust outer PVDF-hexafluoropropylene (PVDF-HFP) prevented the dendrite growth, and the inner lithiophilic Li–Mg facilitated the uniform Li deposition under a PVDF-HFP layer, as presented in Fig. 8a.⁸⁴ Han et al. developed an in situ organic/inorganic hybrid SEI layer (LiBr-HBU) by immersing the Li foil into 1-(6-bromohexanoyl)-3-butylurea (BUBU) (Fig. 8b).⁸⁵ The LiBr-HBU hybrid coating exhibited strong attachment to the Li surface and a synergistic “soft-rigid” property. As shown in Fig. 8c–f, the morphology evolution of the LiBr-HBU SEI layer revealed that the hybrid SEI layer maintained the intact structure and tight binding with Li foil after the repeated plating/stripping cycles. As a result, the symmetric cell with the hybrid SEI layer cycled steadily for over 1333 h at an ultrahigh current density of 15 mA cm⁻². A dual-protective MXene/COF interface with superior mechanical flexibility was reported by Li et al. The artificial SEI consists of soft covalent-organic framework (COF) spheres embedded in rigid MXene sheets, which helps to alleviate SEI cracking and prevent dendrite puncture. Organic–inorganic composite SEIs can combine the advantage of both components, with high ionic conductivity and robust mechanical strength.

Fig. 8 (a) Schematic of the fabrication process of the Li@LMP anode. Reproduced with permission.⁸⁴ Copyright 2025, Elsevier Ltd. (b) Schematic of the in situ formation of a LiBr-HBU SEI on Li surface and the deposition mechanism. (c)–(f) Morphology evolution of LiBr-HBU during Li plating/stripping cycling. Reproduced with permission.⁸⁵ Copyright 2024, Wiley Online Library.

The artificial SEI layer mainly involves the chemical treatment and physical coating of the metal anode surface, forming a functionalized interface before adding the electrolyte. The anode protection strategy is widely used in metal anode protection.⁸⁶ As discussed, the artificial SEI is expected to exhibit high ionic conductivity, mechanical strength, elasticity, and electrical insulation. Inorganic components have high ionic conductivity and mechanical rigidity, facilitating Li⁺ transport for long cycling stability. Flexible organic coating can adapt to mechanical stress, homogenize the Li⁺ flux, and form tight interaction with LMA. The artificial SEI strategy aims to accelerate the charge transfer kinetics and construct a robust interface. The organic–inorganic artificial SEI will be the main development direction, and the balance between organic and inorganic components is essential in terms of their own advantages.

There exists a close coupling and trade-off relationship between the mechanical properties and the ionic conductivity of the SEI, as shown in Fig. 9. For instance, high-modulus inorganic SEI components (e.g., Li₂O and LiF) typically exhibit high grain-boundary resistance and large ion-migration barriers, which result in relatively low ionic conductivity.

Fig. 9 Mechanical–ionic property diagram for SEI materials.

During cycling, electrode volume changes and persistent electrolyte side reactions induce mechanical fatigue in the SEI, such as cracking or interfacial detachment. This leads to localized ion transport resistance and uneven current distribution, accelerating lithium dendrite formation that further imposes greater mechanical stress on the SEI layer. By constructing gradient structures or incorporating self-healing designs, the mechanical strength and ionic conductivity of the SEI can be synergistically enhanced, thereby effectively delaying or even blocking its degradation spiral and ultimately achieving long-cycle-life battery performance.

4. Electrolyte additives and solid-state electrolyte development

The SEI derived from the reduction of electrolyte on the Li metal surface forms a protective membrane that isolates the Li metal from the electrolyte. The most common electrolyte in commercialization is the carbonate electrolyte, while the intrinsic SEI exhibit limited stability. In liquid electrolyte, Li⁺ coordinates with 4–6 solvent molecules to form a solvation structure.⁸⁷ Moreover, the solvation structure must undergo desolvation before Li⁺ migrates through the SEI and a lower desolvation energy facilitates Li⁺ transport. Electrolyte modification is a straightforward strategy to regulate Li deposition morphology by forming an SEI with specific components and adjusting the solvation structure. Numerous studies have been conducted to improve the electrolyte compositions. For instance, an inorganic-rich SEI is regarded as a superior interfacial layer, where the anions of inorganic salts alter the solvation structure of Li⁺. Due to their low lowest unoccupied molecular orbital (LUMO) energy, these anions are preferentially reduced on the anode surface, resulting in a native SEI enriched with prominent inorganic components (such as LiF and Li₃N).⁸⁸ An optimized SEI layer effectively inhibits unwanted side reactions and renders controlled lithium deposition and stripping. Common additives include inorganic salts (such as LiNO₃,⁷³ LiI,⁸⁹ and LiPF₆,⁹⁰) organic compounds (such as diethyl dibromomalonate,⁹¹ 2-(triphenylphosphoranylidene)succinic anhydride,⁹² thiourea,⁹³ polydimethylsiloxane,⁹⁴ and tris(pentafluorophenyl)borane⁹⁵). Fluorinated additives generally contribute to a LiF-rich SEI with high mechanical strength, helping to suppress electrolyte decomposition and lithium dendrite growth. Liu's group demonstrated that hexafluoroglutaric anhydride (F6-0) induces a dense LiF-rich SEI with LiNO₃ (LNO) as an adjuvant in regular carbonate electrolytes (Fig. 10a).⁹⁶ The reduction efficiency of F6-0 increased to 91% under the assistance of LNO, promoting uniform and dense SEI dominated by LiF form on the anode (Fig. 10b). As evidenced in Fig. 10c, a flat Li deposition layer was favorable for the highly reversible Li cycling. In F6-0-LNO coordination, the Li symmetric cell operated stably for 1000 h with low voltage hysteresis. Unlike single SEI component, multicomponent effects caused by electrolyte additives could explore the potential of the complex SEI layer. Fang et al. confirmed that symmetric Li cells were more stable in an ether-based electrolyte system (LiTFSI/DME-DOL-LiNO₃), which was due to the bifunctional effects of methylmagnesium chloride (CH₃MgCl).⁹⁷ As presented in Fig. 10d, CH₃Mg⁺ preferentially reduced and formed an inorganic-rich SEI and a Li–Mg alloy layer, while Cl⁻ generated LiCl species on Li metal anodes. Li–Mg alloy exhibited a much low overpotential for Li nucleation, and high-quality inorganic SEI that contained LiF and LiCl contributed to suppress Li corrosion. This synergistic interaction demonstrated excellent performance in practical Li metal batteries. As Fig. 10e, the morphology of deposited Li featured thick, columnar-like Li, confirming the optimized Li deposition behavior in a Lewis basic CH₃MgCl system. Zhang et al. synthesized a perfluoroalkylsulfonyl quaternary ammonium nitrate (PQA-NO₃) additive to utilize the synergistic effects of cationic (PQA⁺) and anionic (NO³⁻) components.¹⁷ The PQA⁺ cation contributed to the formation of inorganic-rich SEI with fast Li⁺ transport, while the NO₃⁻ anion regulated the solvation structure and oxidation stability at the cathode/electrolyte interface. Liu et al. proposed a dimethyl sulfoxide (DMSO) with highly concentrated LiNO₃ (4.0 M) for fluoroethylene-carbonate (FEC)-based electrolyte to form a LiNO₃ saturated electrolyte (LiNO₃-S) (Fig. 10f).⁹⁸ In the primary Li⁺ solvation sheath, the aggregate solvation structure of NO³⁻ was preferentially reduced to inorganic Li₂O, Li₃N, LiN_xO_y, and LiF in the SEI layer, which could effectively facilitate ionic diffusion and mechanical strength. Fig. 10g and h show the top and side views of the deposited Li in the LiNO₃-S, and a dense and compact structure verifies the capacity of inorganic-rich SEI to regulate Li deposition. Appropriate additives are introduced into the electrode to form an adaptable SEI at the electrolyte/electrode interface towards protecting the metal anode.⁹⁹

Fig. 10 (a) Schematic of the construction of the LiF-rich SEI on the Li anode with F6L. (b) TEM image of the SEI formed in F6L. (c) SEM images of the Li metal in F6L Li plating/stripping cycling. Reproduced with permission.⁹⁶ Copyright 2024, NPG. (d) Schematic of the Lewis-basic electrolyte additive-mediated interfacial chemistry mechanism. (e) SEM image of deposited Li on Cu after 15 cycles in the LN-20 electrolyte. Reproduced with permission.⁹⁷ Copyright 2025, Wiley Online Library. (f) Schematic of the plated Li on bare Cu in the LiNO₃-S electrolyte. (g) and (h) SEM images of the plated Li (3 mAh cm⁻³) in the LiNO₃-S electrolyte. Reproduced with permission.⁹⁸ Copyright 2020, Wiley Online Library.

The above introduction pertains to SEI film-forming additives used for protecting LMA. Various additives with different functions, such as conductive additives, flame retardant additives, and overcharge prevention additives have been developed.

Most notably, the reactions between certain additives (such as film-forming and flame-retardant additives) and the lithium metal surface lack selectivity and may be accompanied by the generation of small-molecule gases (e.g., HF, C₂H₄, CO, and NH₃).^100,101 This leads to increased internal battery pressure, electrode interface delamination, and ultimately accelerated capacity decay.¹⁰² This issue is particularly prominent in lithium metal batteries, the high reactivity of lithium metal tends to trigger excessive additive decomposition, and dendrite growth on the lithium metal surface further exacerbates gas entrapment and interface failure.

Moreover, many additives effective in stabilizing lithium metal anodes—such as ether-based solvents and LiNO₃—possess limited oxidation stability, and thus, are incompatible with high-voltage cathodes (>4.0 V) including NMC811 and lithium-rich materials.¹⁰³ This can lead to the formation of a thick and highly resistive cathode electrolyte interphase (CEI) layer on the cathode surface, and may even trigger lattice oxygen release and structural collapse. To achieve a high lithium-ion transference number (t₊) and construct an inorganic-rich SEI, the high-concentration electrolyte strategy has been widely adopted.¹⁰⁴ However, a significant increase in salt concentration substantially increases the electrolyte viscosity, impeding the bulk transport kinetics of lithium ions. This results in aggravated battery polarization, diminished power performance, and deteriorated low-temperature performance. Furthermore, the poor wettability toward porous electrodes, especially thick electrodes, further limits the utilization efficiency of active materials.

In conclusion, research on electrolyte additives is receiving increasing attention in academia due to their convenience and efficiency in electrode protection, and more importantly, the ability to precisely control the SEI layer. However, the complex side reactions and detrimental effects on ionic conductivity and wettability contradict the intended purpose of the additives.

A solid-state electrolyte (SSE) with a high modulus can contribute to suppress dendrite growth and is extensively used in all-solid-state LMBs. Their nonflammable, solid-state LMBs offer enhanced safety compared to traditional LMBs that use flammable organic electrolytes. For practical applications, key features such as ionic conductivity, interfacial compatibility, and electrochemical stability impact the battery performance. SSE can be categorized into inorganics, polymer solid electrolytes and composite solid electrolytes.

Inorganic ceramic electrolytes are mainly divided into sulfides and oxides. Oxide-based SSEs such as perovskite-type Li_3xLa_2/3−xTiO₃ (LLTO),¹⁰⁵ NASICON-type Li_1+xAl_xTi_2−x(PO₄)₃,¹⁰⁶ Li_1+xAl_xGe_2−x(PO₄)₃ (LAGP),¹⁰⁷ Li_6.4La₃Zr_1.4Ta_0.6O₁₂ (LLZTO),¹⁰⁸ Li_6.55La_2.95Ca_0.05Zr_1.5Ta_0.5O₁₂ (LLCZTO),¹⁰⁹ and garnet-type Li₇La₃Zr₂O₁₂ (LLZO)¹¹⁰ exhibit high σ, wide electrochemical window, and thermal and electrochemical stability. However, owing to the poor interfacial compatibility between inorganic SSEs and electrodes, the severe interfacial contact resistance had a negative influence on the successful implementation of all-solid-state LMBs.¹¹¹ This resolve approach that has been widely investigated is constructing interlayer among inorganic SSE and electrodes.^112,113 As for polymer solid electrolytes, polymer matrixes such as PEO, PVDF, and polyethylene glycol containing Li salts such as LiTFSI and LiFSI show great interface compatibility.^114–116 The constrained motion of polymer segments at the crystalline state leads to insufficient ion conductivity (10⁻⁷ S cm⁻¹ at room temperature), hence the operating temperature typically exceeds 60 °C.^25,117 After the addition of inorganic filers such as ceramic filler LLZO,¹¹⁸ LLZTO,¹¹⁹ MOF,¹²⁰ and molecular sieve,¹²¹ polymer chains can significantly expedite Li⁺ transport kinetics and improve mechanical property, that is inorganic/polymer hybrid SSE. Integrating the merits of inorganic and polymer electrolytes, composite electrolytes exhibit good ionic conductivity (10⁻⁴–10⁻³), flexibility, and rigidity.^{115,122–124}

Numerous efforts have been devoted to constructing a composite SSE. However, compared to liquid electrolytes, the limited ionic conductivity still falls far short of the requirements of high rate, leading to the undesired dendrite growth.¹²⁵ Wang et al. presented an enthalpy–entropy strategy to introduce poly(ionic liquid) extracted from N-butyl-N-methylpyrrolidinium bis(trifluoromethanesulfonyl)imide (PYR14TFSI) into PEO-based SSEs.¹²⁶ Such strategy aided in mitigating the strong Li⁺–PEO coordination, achieving an ionic conductivity of 0.117 mS cm⁻¹ and a Li⁺ transference number of 0.71. Li's group proposed a molecular level regulation on PEO-based SSE with CaF₂ as fillers.¹²⁷ The PEO–CaF₂ electrolyte demonstrated an ionic conductivity of 0.31 mS cm⁻¹, which was attributed to the strong coordination of Ga²⁺ with a Li salt anion and the ether–oxygen in the PEO chain. Moreover, CaF₂ further physically disrupted the crystallinity of PEO and spontaneously reacted with Li to generate a LiF-rich solid electrolyte interface (Fig. 11a). In contrast to the ordinary PEO electrolyte, the PEO–CaF₂ electrolyte achieved uniform Li deposition (Fig. 11b). The assembled Li‖LEP pouch cell exhibited stable cycling performance at 60 °C at 0.5C, maintaining a capacity retention of 86.81% after 1000 cycles (Fig. 11c). To improve the mechanical strength, a polyacrylonitrile (PAN)-reinforced composite solid electrolyte (PEO/PAN/LLZTO) was fabricated to suppress dendrite penetration.¹²⁸ Besides, inorganic LiF and Li₃N spontaneously formed at the electrolyte/Lithium interface, promoting uniform Li⁺ deposition. Zhang et al. employed an in situ thermal treatment to construct an ultrathin (10 µm) poly(ethylene glycol) diacrylate (PEGDA)-based electrolyte with a PAN fiber (Fig. 11d).¹²⁹ The PAN fiber not only provided mechanical strength but also promoted anion trapping, hence allowing high ionic conductivity (0.88 mS cm⁻¹). The symmetrical Li‖Li with a PAN-fiber-PEGDA electrolyte operated for 600 h at 0.2 mA cm⁻². Table 2 summarizes the electrochemical performance of lithium batteries using solid-state electrolytes.

Fig. 11 (a) Schematic of CaF₂ affecting the PEO electrolyte. (b) SEM images of the Li anode with PEO and PEO–CaF₂ electrolytes after 50 cycles at 0.1 mA cm⁻². (c) Cyclability of the Li‖LEP pouch cell at 0.5C under 60 °C. Reproduced with permission.¹²⁷ Copyright 2025, the American Chemical Society. (d) Schematic of the synthesis process of PAN fiber-PEGDA SSE. (e) Schematic of the effect of PAN fibers. Reproduced with permission.¹²⁹ Copyright 2025, Wiley Online Library.

Table 2 Performance of LMBs using solid-state electrolytes

Electrolyte	Ionic conductivity	Operation temperature	Cathode	Cycle life of full batteries (Cycle number-Retention)	Current	Ref.
Li2S6–PEO	0.17 mS cm^−1	40 °C	LFP, 3–5 mg cm^−2	—	—	130
MOF–PEO	0.691 mS cm^−1	27 °C	LFP, 3.0 mg cm^−2	228–80%	0.2C	131
PEO/CPTP	—	50 °C	LFP	300–81%	0.2C	132
VN-g-C3N4–PEO	—	30 °C	LFP, 2.0 mg cm^−2	300–90%	0.2C	133
PEO/PVDF	0.556 mS cm^−1	60 °C	LFP, 2–2.3 mg cm^−2	100–95.7%	0.5C	134
Ga-CeO2/LiTFSI/PEO	0.13 mS cm^−1	60 °C	LFP	—	—	135
PEO@Li7La3Zr2O12 nanowires	1.53 mS cm^−1	60 °C	LFP, 1.68 mg cm^−2	—	—	136
PEG-COF@Li1.3Al0.3Ti1.7(PO4)3	2.55 mS cm^−1	30 °C	LFP, 2–3 mg cm^−2	200–98.8%	1C	137

---

SSEs are widely regarded as an ideal solution for achieving high-safety lithium metal batteries. However, their practical application faces a fundamental contradiction: it is challenging to simultaneously achieve both high ionic conductivity (σ) and excellent dendrite suppression capability (high mechanical strength) within a single material system. Ionic conduction relies on low-energy-barrier pathways available for ion migration within crystalline or amorphous structures. Achieving high ionic conductivity typically requires meeting the following conditions: (1) a network of energetically similar migration sites within the crystal lattice (e.g., lithium sites in garnet-type LLZO) or continuous diffusion channels in amorphous structures; (2) a sufficiently high carrier concentration (such as vacancies or interstitial ions); and (3) a low migration activation energy (E_a).¹³⁸ These structural characteristics are often associated with relatively open, weakly bonded, or polarizable chemical environments—for instance, S²⁻ is more polarizable than O²⁻. In contrast, the high mechanical strength originates from dense atomic packing and strong chemical bonds. For oxide electrolytes such as LLZO, strong Li–O and M–O bonds impart high hardness (G > 50 GPa) but simultaneously restrict the migration freedom of Li⁺. Additionally, their high lattice energy results in a low intrinsic lithium vacancy concentration, often requiring defect introduction through high-valence cation doping (e.g., Nb⁵⁺ and Al³⁺) to enhance the conductivity.¹³⁹ Therefore, at the atomic scale, an intrinsic conflict exists between structural features conducive to fast ion transport (openness and weak bonding) and those enabling high mechanical strength (density and strong bonding).

The poor interfacial contact between the solid electrolyte and the electrode can bring about dendrite growth and serious interfacial impedance. For instance, by using magnetron sputtering to deposit the ZnO layer on the LLCTO substrate, the solid–solid interface resistance sharply decreases to 1%.¹⁰⁹ A large volume charge will generate dendrites and voids at the interface after long-term cycling. Furthermore, a narrow electrochemical window is also an important obstacle for practical applications.

5. Multi-strategy synergistic protection mechanism

The interfacial instability of lithium metal anodes arises from the coupling of multiple factors, and single-factor protection strategies often come with inherent limitations. While 3D host structures can mitigate volume expansion, their increased specific surface area exacerbates interfacial side reactions. Artificial SEI layers can enhance interfacial stability, yet they are prone to cracking under cycling stress. Electrolyte engineering allows the tuning of SEI composition but cannot directly constrain lithium's volume changes. Hence, future efforts should focus on developing integrated solutions that combine multiple complementary strategies to achieve a stable lithium metal interface.

The core of achieving dynamic stability in lithium deposition and the interface lies in the regulation of deposition morphology, management of cyclic stress, and maintenance of interfacial chemistry.

The synergy between the host structure and the artificial SEI lies in the combination of physical confinement and chemical protection. It involves the host structure providing macroscopic mechanical support and spatial confinement, while the artificial SEI constructs a chemically and electrochemically stable layer at the microscopic interface. Together, they guide and stabilize the deposition behavior of lithium metal. Specifically, the three-dimensional lithiophilic host provides space for lithium deposition and alleviates volume strain, while the artificial SEI layer enhances the mechanical strength of the interface, reduces the diffusion barrier for lithium ions, and effectively blocks continuous side reactions with the electrolyte.¹⁴⁰

Artificial SEI enables the precise pre-design of interfacial composition and structure but lacks dynamic self-healing capability during cycling. While electrolyte additives can repair the SEI in situ, the composition and structure of the resulting layers are difficult to control precisely. The synergy between the two strategies bridges the gap from static interface design to dynamic adaptive maintenance. To elaborate, the dense and robust artificial SEI (e.g., LiF) serves as an initial barrier that effectively isolates direct contact between the lithium metal and the electrolyte, thereby significantly reducing initial side reactions and electrolyte consumption.^141,142 When micro-cracks or local damage occur in the artificial SEI during cycling, functional additives in the electrolyte (FEC and LiNO₃) are capable of preferentially decomposing at damaged sites, forming “patch-like” repair layers that enable dynamic in situ restoration of the interface.

The ternary synergistic system represents the top-level paradigm of coherent “interface–structure–electrolyte” design. Through deep mutual adaptation and functional coupling among the three components, it achieves holistic optimization of lithium metal anode performance, simultaneously addressing the three core challenges of uniform deposition, volume expansion, and interfacial chemical stability.

6. Breakthrough prospects in LMB research

6.1 Biomimetic interface engineering

Traditional interface engineering lacks self-adaptive capability and is prone to cracking or delamination due to stress accumulation. Once damaged, it cannot self-repair, resulting in continuous exposure of the electrode to the electrolyte, which triggers uncontrollable side reactions and irreversible consumption of active materials, severely compromising the battery's cycling stability and lifespan.

Bio-inspired intelligent interface engineering represents a cutting-edge battery material design paradigm. It simulates the dynamic adaptation, self-healing, and environmental response capabilities of biological systems to construct intelligent interface protection layers for highly reactive electrodes such as lithium metal.

Inspired by the environmental response characteristics of biological systems, constructing intelligent interface layers with thermal, mechanical, and other stimulus-responsive capabilities enables dynamic adaptation to battery operating conditions. These layers maintain efficient ion transport during normal operation and rapidly activate protective mechanisms under abnormal conditions (e.g., high temperature and mechanical damage), thereby addressing safety hazards in lithium metal batteries. Li et al. developed a cell membrane-inspired artificial layer (CAL) with biomimetic ion channels. Through in situ conversion during cycling, the negatively charged CAL forms a robust transition layer rich in lithiophilic inorganic components, facilitating lithium-ion diffusion.¹⁴³

6.2 AI-multiscale computational platform for rational electrolyte design

The development of traditional electrolyte systems has long relied on empirical “trial-and-error” methods, making it difficult to accurately predict and regulate the complex synergistic effects among their multiple internal components. In contrast, artificial intelligence technologies based on machine learning can establish a “mechanism-data” program. This program is capable of rapidly integrating the molecular structures, key physicochemical properties (e.g., redox potential, viscosity, and ionic conductivity) of electrolyte components (including lithium salts, solvents and additives) and their corresponding battery performance data, thereby establishing a mapping relationship between molecular characteristics and macroscopic performance. Furthermore, by combining quantum chemical calculations and molecular dynamics simulations, it can deeply analyze the solvation structure and the desolvation dynamics at the interface. This enables the rational design of novel electrolyte systems capable of in situ forming an SEI with ideal mechanical strength, high ionic conductivity, and self-healing functionality at the electrode interface.

6.3 In situ and operational characterization techniques

In situ optical microscopy has been widely employed in the design of lithium metal host materials, enabling real-time observation of the morphological evolution of lithium deposition layers on lithium foil or host surfaces.⁴⁸

In situ transmission electron microscopy (in situ TEM), leveraging its atomic-scale spatial resolution, enables real-time observation of lithium dendrite nucleation and growth, as well as the dynamic evolution of the solid electrolyte interface (SEI) structure during battery charging and discharging. For example, Wang et al. performed in situ TEM for the real-time observation of the cross-sectional interface between the solid-state lithium metal and the LLZO electrolyte, clearly revealing the formation and evolution of pores at the interface.¹⁴⁴ In another study, Li et al. employed in situ TEM to observe the stripping process of lithium whiskers using Li@Li₂O as a model system.¹⁴⁵ This work not only provided in-depth insights into the stripping mechanism of lithium metal but also established a criterion for the mechanical instability of the SEI based on these observations. The study confirmed that the ratio of the thickness (t) of the supporting solid structure (SSS) to the radius (r) of the lithium deposit (t/r) is a key parameter controlling the stripping behavior of lithium metal. When the SSS is mechanically unstable (i.e., t/r < 0.21), it undergoes buckling and contraction, ultimately leading to the formation of “dead lithium.”

Atomic force microscopy (AFM) operates by detecting the interaction forces between the fine probe and the sample surface, thereby enabling the analysis of surface topography and mechanical properties (such as Young's modulus and adhesion). When integrated with an electrochemical environment, in situ AFM can monitor the dynamic evolution of Young's modulus at the electrode interface under operating conditions in real time, providing crucial data support for establishing quantitative “structure–mechanics–performance” correlations at the interface. For instance, to investigate the formation mechanisms of mossy lithium and lithium dendrites, Hausen et al. combined atomic force microscopy with cyclic voltammetry (CV) to perform in situ measurements on lithium metal anodes.¹⁴⁶

In situ X-ray photoelectron spectroscopy (in situ XPS) enables real-time analysis of the chemical composition, elemental oxidation states, and types of chemical bonds within the SEI layer during electrochemical cycling. Wood et al. utilized in situ XPS to investigate the reactions between the Li anode and the LPS solid electrolyte. The measurements indicated that electrochemically driven Li⁺ leads to the formation of Li₂S and Li₃P decomposition phases.¹⁴⁷

7. The pervasive limitation of non-standardized testing conditions

A coherent framework for interface stability is only as reliable as the data underpinning it. A critical, yet often understated, limitation in the existing literature is the substantial heterogeneity under electrochemical testing conditions. This variability hinders reliable direct comparisons between different strategies and may render conclusions about “stability” valid only within a narrow and impractical parameter space. Undoubtedly, assessments of interfacial stability based on such testing parameters exhibit evident limitations.

(1) Low current (<1 mA cm⁻²) and areal capacity (<1 mAh cm⁻²) minimize polarization effects and volumetric strain. Interfacial strategies that exhibit excellent performance in LMBs in benign tests (e.g., 0.5 mA cm⁻², <1 mAh cm⁻²) may rapidly degrade under high current and high areal capacity conditions (>3 mA cm⁻², >3 mAh cm⁻²) required for practical applications, often due to ion transport limitations or mechanical stress accumulation. High current densities exacerbate kinetic limitations, while high areal capacities pose severe challenges to the mechanical integrity of the interface.

(2) A large excess of lithium (N/P ≫ 10)acts as a reservoir to mask continuous Li loss caused by poor CE resulting from chemical instability, and excessive electrolyte (electrolyte-to-capacity, E/C) in flooded cells can continuously replenish the components decomposed at the unstable interface, thereby effectively concealing true interfacial reactivity. High N/P ratios (N/P > 50) and excessive electrolytes (high E/C ratio) can significantly mask the intrinsic instability of the interface. Consequently, the “long cycle life” observed under such conditions holds a little predictive value for evaluating practical battery systems, such as anode-free or lean-Li configurations (N/P < 3) and lean-electrolyte designs (E/C < 3 g Ah⁻¹).

(3) Deep cycle testing (high depth of discharge, DOD) is the ultimate litmus test for determining a battery's full-volume adaptive capability. Cycle life numbers are meaningless without the context of DOD.

Ignoring differences in testing conditions will obscure the root causes of failures, lead to erroneous conclusions, and hinder the establishment of universal theories. To this end, rigorous validation protocols are essential. The stability of the anode interface must be assessed under harsh conditions, such as low N/P ratio and lean electrolyte. A comprehensive evaluation demands definitive verification in a full-cell configuration matched with a high-loading cathode to accurately gauge the overall performance and energy density. In conclusion, to ensure reproducibility and enable fair comparison across studies, it is imperative that research reports explicitly document key parameters, including current density, areal capacity, N/P ratio, E/C ratio, cathode loading, testing temperature, and formation cycle details.

The design of LMBs should involve the practical application, such as large current density, high specific capacity and low N/P ration. In response, an increasing number of studies have emerged, focusing on these key concerns and contributing substantial progress in the field. The full battery with a composite 0VCCS-Li@rGO&ZnO anode and a LiFePO₄ cathode (3 mAh cm⁻²) demonstrated excellent cycle life (>1000 cycles) at the practical N/P ratio (2.3).³⁸ Kong and co-workers developed a random-aligned CNF host for the Li metal anode.¹⁴⁸ With a low N/P ratio (2), the full cell paired with a high-loading LFP electrode (11.26 mg cm⁻²) exhibited long lifespan of over 400 cycles at 3.0C. Moreover, an r/a-HPCNF‖LFP pouch cell with low N/P (2) and E/C (2.8) ratio displayed an exceptional gravimetric energy density (260 Wh kg⁻¹) and robust cycling stability, confirming its practical applications. Each approach offers distinct advantages; however, accompanying challenges necessitate further research to accelerate the practical application of LMBs.

In lithium–metal anode battery research, the relationship between the reported “performance inflation” and the “practical relevance” of a study fundamentally reflects a mapping between laboratory metrics and industrial value. This connection may be closely aligned or, conversely, significantly decoupled. A critical understanding and clear articulation of this relationship is essential for assessing the true merit and impact of research. The reported “performance inflation” generally refers to quantifiable gains achieved under controlled laboratory settings. Key advancements highlighted in papers—such as uniform lithium deposition morphology, a stable SEI, high CE, exceptional cycle life, and superior rate capability—are frequently obtained under simplified, idealized test conditions. These include the use of excess lithium and electrolytes, small-format electrodes, and mild charge–discharge protocols, which pose the risk of performance overestimation and often limit the work to proof-of-concept demonstrations with little direct relevance to industrial scale-up.

The “practical relevance” of research refers to the real-world value of laboratory performance improvements in addressing industrialization challenges, which must be evaluated across four dimensions: feasibility (whether the improvement can be replicated under harsh conditions such as lean electrolyte, limited lithium source, and practical cathode matching); reliability (whether the battery can maintain stability under complex operating conditions such as wide temperature ranges, high-rate fast charging, and dynamic loads); economic viability (whether the materials and processes are cost-controllable and scalable); and safety (whether it can pass rigorous safety tests).

Future research endeavors should (1) standardize reports by specifying test conditions (recommended: N/P ≤ 3 and E/C ≤ 5 g Ah⁻¹); (2) validate findings with pouch-type full cells, calculating practical energy density at the cell level; and (3) incorporate key tests such as fast-charging/-discharging, wide-temperature operation, and post-cycling interface analysis. These efforts aim to narrow the gap between “performance inflation” and “practical relevance”.

8. Conclusion and perspectives

LMBs are considered as next-generation high-energy storage devices. The unstable Li anode is the most critical factor for the commercialization of LMBs, and the main concern is the lithium dendrite growth. Inhomogeneous Li deposition during the cycling process causes continuous dendrite growth and “dead lithium” stacking, which ultimately impale the separator and even lead to internal short circuits. Numerous studies have demonstrated that the Li anode can be protected by constructing hosts and stable barriers, and the battery performance and safe issues have been improved. In this review, the current design, interface engineering, SEI properties and solid-state electrolytes are summarized. All measures are committed to improving CE, which is conducive to retard capacity degradation and prolong the lifespan. Considerable research has achieved remarkable achievements, there are still many spaces worth exploring for the commercialization of LMBs. Theoretical and technical challenges are listed to further explore the commercial development of practical LMBs.

(1) Development of a light host for high energy density. The mass of the host needs to be reduced so that it is the same as or lower than that of Li metal to achieve a specific energy of over 400 Wh kg⁻¹ for pouch batteries.

(2) Development of new techniques for the composite Li anode. Electrodeposition and molten Li infusion methods are high-risk and high-cost operations. The host requires a certain mechanical strength and flexibility to withstand the high pressure of roll-in technology, which is inconsistent with light hosts. Thus, the optimized roll-in method and developed new techniques are required for the scalable production of the composite lithium metal electrode.

(3) Exploring optimization for interfacial protection. The ex situ SEI layer is artificially prepared on the lithium surface before battery assembly. The organic–inorganic artificial SEI exhibits robust mechanical strength and flexibility, and exploring the binding effect between the artificial SEI and metal Li may be the topic of further research in interfacial engineering.

(4) The principal challenge for SEI lies in balancing mechanical integrity with efficient ion conduction. The homogeneous and dense LiF-rich inorganic SEI effectively prolongs the lifespan of LMBs, while the relatively low conductivity has a negative impact on the mass-transfer kinetics. The correlation between the complex SEI components and the LMB performance should be further explored through a systematic theoretical calculation, while the mechanism of electrolyte additives can be further analyzed through systematic simulation.

(5) Exploring the influence of additives on the solvated structure and desolvation mechanism in depth. Currently, relevant research studies are still in computational simulation, and further studies on the desolvation mechanism need to be conducted by combing experiments and in situ/operando characterizations.

(6) The ionic conductivity of the solid electrolyte is far below the liquid electrolyte, especially for the high energy density and long-cycle durability in practical applications. Developing ultrathin SSEs may be conducive to improve the ionic conductivity. The ultra-thin thickness will weaken the inhibitory effect of SSE on dendrites. Addressing this balance is crucial for improving the stability and electrochemical performance of solid-state LMBs.

In conclusion, the research processes of metal anode protection are reviewed in terms of the fabrication, structure, mechanism of different strategies to stimulate the substantial advances of the LMB application. Furthermore, future research studies on LMBs integrate material designs, advanced characterizations, and theoretical calculations, with the hope of overcoming the barriers in the translation from basic research to engineering applications.

Author contributions

S. Li and X. Wang wrote the manuscript. R. Li, M. Zhang, J. Wang, S. Chen, H. Liu, C. Cao, J. Mi, Y. Feng, Q. Zhang and M. Ge revised the manuscript. All authors contributed to the discussions.

Conflicts of interest

The authors declare no conflict of interests.

Data availability

No new data were generated or analysed as part of this review.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (no. 52202256, 22378292, and 22375047), the Natural Science Foundation of Jiangsu Province of China (BK20240956 and BK20220612), and the Shanxi Province Science Foundation (no. 202403021224005 and 202303021211072). The authors also acknowledge the funds from the Young Elite Scientists Sponsorship Program of Jiangsu Association for Science and Technology.

References

1. R. Schmuch, R. Wagner, G. Hörpel, T. Placke and M. Winter, Nat. Energy, 2018, 3, 267.
2. 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.
3. H. Lin, Y. Hu, Y. Zhou, W. Hu, J. Lin, H. Luo, Y. Jin, D. Zhao, J. Chen, M. Tao, P. Shan, J. Liang, Y. Wei and Y. Yang, Adv. Energy Mater., 2025, e02728.
4. Y. Wu, C. Wang, C. Wang, Y. Zhang, J. Liu, Y. Jin, H. Wang and Q. Zhang, Mater. Horiz., 2024, 11, 388.
5. K.-H. Chen, K. N. Wood, E. Kazyak, W. S. LePage, A. L. Davis, A. J. Sanchez and N. P. Dasgupta, J. Mater. Chem. A, 2017, 5, 11671.
6. J.-F. Ding, R. Xu, C. Yan, B.-Q. Li, H. Yuan and J.-Q. Huang, J. Energy Chem., 2021, 59, 306.
7. J. Tan, J. Matz, P. Dong, J. Shen and M. Ye, Adv. Energy Mater., 2021, 11, 2100046.
8. D. Anseán, M. Dubarry, A. Devie, B. Y. Liaw, V. M. García, J. C. Viera and M. González, ECS Meet. Abstr., 2016, 6, 900.
9. X. Li, J. Zheng, X. Ren, M. H. Engelhard, W. Zhao, Q. Li, J. Zhang and W. Xu, Adv. Energy Mater., 2018, 8, 1703022.
10. M. Yu and X. Ning, Chem. Eng. J., 2025, 503, 158554.
11. Z.-H. Zeng, Y.-W. Tian, Q.-Y. Gao, Z.-R. Li, Z. Qian, Z.-S. Xia, Q.-Q. Yang, Z.-Y. Hu, H. S. H. Mohamed, Y. Li and B.-L. Su, J. Colloid Interface Sci., 2025, 699, 138196.
12. K. Chen, H. Liao, X. Yu, L. Li, S. Hu, Y. Zeng, H. Zhang, A. Bahadur, Y. Liu and X. Wang, J. Alloys Compd., 2025, 1018, 179043.
13. C. Shi, M. Su, H. Yu, Q. Zhang, P. Yue, B. Song, D. Cao and K. Zhu, J. Colloid Interface Sci., 2025, 693, 137643.
14. C. Yan, X. Cheng, Y. Yao, X. Shen, B. Li, W. Li, R. Zhang, J. Huang, H. Li and Q. Zhang, Adv. Mater., 2018, 30, 1804461.
15. M. Wang, Z. Peng, W. Luo, F. Ren, Z. Li, Q. Zhang, H. He, C. Ouyang and D. Wang, Adv. Energy Mater., 2019, 9, 1802912.
16. N. Zhang, L. Du, J. Zhang, H. Xu, X. Zhou, L. Mai and L. Xu, Adv. Funct. Mater., 2023, 33, 2210862.
17. W. Zhang, Y. Lu, Q. Feng, H. Wang, G. Cheng, H. Liu, Q. Cao, Z. Luo, P. Zhou, Y. Xia, W. Hou, K. Zhao, C. Du and K. Liu, Nat. Commun., 2025, 16, 3344.
18. J. Zhang, J. Ran, W. Jia, X. Yan, L. Feng, W. Zou and C. Fan, Chem. Commun., 2025, 61, 15417.
19. J. Huang, C. Li, D. Jiang, J. Gao, L. Cheng, G. Li, H. Luo, Z. Xu, D. Shin, Y. Wang, Y. Lu and Y. Kim, Adv. Funct. Mater., 2025, 35, 2411171.
20. Z. Lin, Q. Yao, S. Yang, H. Song, Z. Yu, Z. Li, S. Chen, M. Wang, Z. Wang, G. Zhang, L. Zhang, Z. Yu, X. Song, K. Zhou, W. Li, L. Yu, J. Xu and K. Chen, Adv. Funct. Mater., 2025, 35, 2424110.
21. J. Huang, C. Li, D. Jiang, J. Gao, L. Cheng, G. Li, H. Luo, Z. Xu, D. Shin, Y. Wang, Y. Lu and Y. Kim, Adv. Funct. Mater., 2025, 35, 2411171.
22. J. Zheng, H. Huang, H. Zhou, H. Zhao, H. Li, G. Jiang, W. Han, H. Zhang, L. Li, X. Geng, B. An and C. Sun, Chem. Mater., 2025, 37, 1393.
23. Y. Xu, Y. Sun, Y. Sun, H. Fang, Y. Jiang and B. Zhao, Surf. Interfaces, 2024, 44, 103768.
24. M. Al-Abbasi, Y. Zhao, H. He, H. Liu, H. Xia, T. Zhu, K. Wang, Z. Xu, H. Wang, W. Zhang, Y. Lai and M. Ge, Carbon Neutralization, 2024, 3, 108.
25. H. He, L. Wang, M. Al-Abbasi, C. Cao, H. Li, Z. Xu, S. Chen, W. Zhang, R. Li, Y. Lai, Y. Tang and M. Ge, Energy Environ. Mater., 2024, 7, e12699.
26. A. B. Yaroslavtsev, Russ. Chem. Rev., 2016, 85, 1255.
27. F. Qiu, X. Li, H. Deng, D. Wang, X. Mu, P. He and H. Zhou, Adv. Energy Mater., 2019, 9, 1803372.
28. C. Dong, H. Shi, H. Cui, S. Yu, Y. Li, Y. Ma, Y. Guo, Y. Dong, L. Zhang, C. Li, Y. Yu and Z.-S. Wu, Energy Storage Mater, 2025, 75, 103987.
29. Y. Wang, X. Li, L. Wu, J. Tan, G. Liu, C. Ye, L. Ma, Z. Liu, M. Ye and J. Shen, Energy Storage Mater, 2024, 66, 103247.
30. D. Chen, C. Chen, H. Yu, S. Zheng, T. Jin, N. W. Li and L. Yu, Adv. Funct. Mater., 2024, 34, 2402951.
31. H. Liu, J. Di, P. Wang, R. Gao, H. Tian, P. Ren, Q. Yuan, W. Huang, R. Liu, Q. Liu and M. Feng, Carbon Energy, 2022, 4, 654.
32. X. Wang, Z. Chen, X. Xue, J. Wang, Y. Wang, D. Bresser, X. Liu, M. Chen and S. Passerini, Nano Energy, 2025, 133, 110439.
33. A. Huang, H. Huang, S. Li, X. Pan, S. Sun, X. Su, H. Geng, L. Li, M. Maximov, J. Ren and S. Peng, Adv. Funct. Mater., 2025, 2506258.
34. Y. W. Nahm, J. S. Lee, J. H. Choi, J. S. Cho and Y. C. Kang, Small, 2025, 21, 2504223.
35. R. Zhang, X. Chen, X. Chen, X. Cheng, X. Zhang, C. Yan and Q. Zhang, Angew. Chem., 2017, 129, 7872.
36. X. Zhang, S. Jin, M. H. Seo, C. Shang, G. Zhou, X. Wang and G. Li, Nano Energy, 2022, 93, 106905.
37. Y. Ouyang, W. Zong, J. Wang, Z. Xu, L. Mo, F. Lai, Z.-L. Xu, Y.-E. Miao and T. Liu, Energy Storage Mater, 2021, 42, 68.
38. L. Deng, Y. Liu, H. Qi, Y. Yang, Z. Wang, L.-T. Dong, J. Zhan, K.-P. Song, D. Qi, Y. Xu, Y. Sang, J. Yang, J.-J. Wang, Z. Zheng, S. Wang, C. Gao, H. Liu and H. Chen, Nat. Nanotechnol., 2025, 20, 1439.
39. H. Jiang, Y. Zhou, H. Zhu, F. Qin, Z. Han, M. Bai, J. Yang, J. Li, B. Hong and Y. Lai, Chem. Eng. J., 2022, 428, 132648.
40. Z. Hao, D. Liu, X. Zuo, H. Yu and Y. Zhang, Adv. Mater., 2024, 36, 2404983.
41. K. Huang, Z. Li, Q. Xu, H. Liu, H. Li and Y. Wang, Adv. Energy Mater., 2019, 9, 1900853.
42. B. Yu, L. Fei, M. Li, Z. Ma, B. Guo, J. Chen, X. Wang, Z. Xue, L. Zhou and J. Wang, Energy Storage Mater, 2025, 82, 104579.
43. C. Li, C. Yang, T. Huang, Y. Wang, J. Yang, Y. Jiang, J. Mao, S. Zheng and S. Xia, Adv. Funct. Mater., 2024, 34, 2407149.
44. S. Wu, T. Jiao, S. Yang, B. Liu, W. Zhang and K. Zhang, J. Mater. Chem. A, 2019, 7, 5726.
45. Y. Zhou, K. Zhao, Y. Han, Z. Sun, H. Zhang, L. Xu, Y. Ma and Y. Chen, J. Mater. Chem. A, 2019, 7, 5712.
46. W. Xu, Y. Li, Y. Wang, N. Xin, K. Wang, H. Li and C. Yuan, Int. J. Electrochem. Sci., 2025, 20, 101017.
47. Z. Wang, Z. Li, Y. Wang, C. Zhu and K. Chen, J. Energy Storage, 2024, 101, 113837.
48. Y. Wang, X. Li, L. Wu, J. Tan, G. Liu, C. Ye, L. Ma, Z. Liu, M. Ye and J. Shen, Energy Storage Mater, 2024, 66, 103247.
49. X. Cheng, T. Hou, R. Zhang, H. Peng, C. Zhao, J. Huang and Q. Zhang, Adv. Mater., 2016, 28, 2888.
50. H. Zhang, X. Liao, Y. Guan, Y. Xiang, M. Li, W. Zhang, X. Zhu, H. Ming, L. Lu, J. Qiu, Y. Huang, G. Cao, Y. Yang, L. Mai, Y. Zhao and H. Zhang, Nat. Commun., 2018, 9, 3729.
51. S. Hong, D. Jung, J. Kim, Y. Lee, S. Cho, S. H. Joo, H. Lee, K. Lee and S. Lee, Adv. Funct. Mater., 2020, 30, 1908868.
52. H.-J. Noh, M.-H. Lee, B. G. Kim, J.-H. Park, S.-M. Lee and J.-H. Choi, ACS Appl. Mater. Interfaces, 2021, 13, 55227.
53. X. Jia, M. Wang, S. Chen, C. Cui and J. Zhu, Chem. Eng. J., 2025, 510, 161788.
54. Z. Song, Z. Wang, C. Wei, J. Xing, Q. Zhao, A. Zhou, Q. Deng, H. Li and J. Li, ACS Nano, 2025, 19, 38724.
55. T. Li, S. Gu, L. Chen, L. Zhang, X. Qin, Z. Huang, Y. He, W. Lv and F. Kang, Small, 2022, 18, 2203273.
56. C. Zhang, K. Chen, X. Li, Z. Wang, S. Yang, R. Bagherzadeh, F. Lai, C. Zhang, G. He, I. P. Parkin, Y.-E. Miao and T. Liu, ACS Appl. Energy Mater., 2025, 8, 6203.
57. L. Hong, L. Wu, J. Gong, K. Pan, J. Xu, Y. Wang, G. Shi and H. Pang, Carbon, 2025, 243, 120495.
58. Z. Hao, D. Liu, X. Zuo, H. Yu, T. You and Y. Zhang, Adv. Funct. Mater., 2025, 35, 2415251.
59. A. Huang, H. Huang, S. Li, X. Pan, A. Wang, H. Chen, T. Wang, L. Li, M. Maximov, J. Ren, Y. Wu and S. Peng, Adv. Energy Mater., 2025, 15, 2403576.
60. W. Zhang, J. Wang, H. Zhang, Q. Dong, S. Zhang, B. Sun, Z. Chen, H. Guo, X. Han, Y. Deng and W. Hu, Small Struct., 2024, 5, 2300358.
61. H. Zhuang, T. Zhang, H. Xiao, X. Liang, F. Zhang, J. Deng and Q. Gao, Energy Environ. Mater., 2023, 6, e12470.
62. C. Zhao, S. Xiong, H. Li, Z. Li, C. Qi, H. Yang, L. Wang, Y. Zhao and T. Liu, J. Power Sources, 2021, 483, 229188.
63. Y. Fan, H. Li, X. He, Y. Huang, C. Sun, T. Zhu, H. Liu, E. Huangzhang, F. Sun and J. Nan, ACS Appl. Energy Mater., 2022, 5, 10034.
64. K. Wang, W. Wang, J. Deng, X. Jiang, G. Xu, H. Gong, N. Zhang and D. Li, J. Alloys Compd., 2021, 889, 161597.
65. Y. Yuan, F. Wu, G. Chen, Y. Bai and C. Wu, J. Energy Chem., 2019, 37, 197.
66. C. Li, Y. Cui, S. Lin, P. Ma, Y. Ji, Z. Cen, G. Yu, S. Li, S. Liu and D. Wu, Adv. Mater., 2025, 37, 2416377.
67. T. Zhou, W. Tang, J. Lv, Y. Deng, Q. Liu, L. Zhang and R. Liu, Small, 2023, 19, 2303924.
68. Y. Zhang, J. Lu, Z. Jin, X. Xie, L. Wei, P. Li, Y. Zhao, J. Zhao, C. Xu, W. Wang and Z. Zhang, Adv. Funct. Mater., 2025, 35, 2417160.
69. Y. Liu, D. Lin, P. Y. Yuen, K. Liu, J. Xie, R. H. Dauskardt and Y. Cui, Adv. Mater., 2017, 29, 1605531.
70. J. Wang, Z. Zhang, H. Ying, G. Han and W.-Q. Han, Chem. Eng. J., 2021, 411, 128534.
71. A. C. Kozen, C.-F. Lin, A. J. Pearse, M. A. Schroeder, X. Han, L. Hu, S.-B. Lee, G. W. Rubloff and M. Noked, ACS Nano, 2015, 9, 5884.
72. A. Hu, W. Chen, X. Du, Y. Hu, T. Lei, H. Wang, L. Xue, Y. Li, H. Sun, Y. Yan, J. Long, C. Shu, J. Zhu, B. Li, X. Wang and J. Xiong, Energy Environ. Sci., 2021, 14, 4115.
73. Y. Zhou, J. Zhang, K. Zhao, Y. Ma, H. Zhang, D. Song, X. Shi, L. Zhang and Y. Ding, Energy Storage Mater, 2021, 39, 403.
74. K. Chen, R. Pathak, A. Gurung, E. A. Adhamash, B. Bahrami, Q. He, H. Qiao, A. L. Smirnova, J. J. Wu, Q. Qiao and Y. Zhou, Energy Storage Mater, 2019, 18, 389.
75. J. Zhu, P. Li, X. Chen, D. Legut, Y. Fan, R. Zhang, Y. Lu, X. Cheng and Q. Zhang, Energy Storage Mater, 2019, 16, 426.
76. W. Tang, Y. Zhang, X. Zhang, S. Yao and R. Liu, J. Mater. Chem. A, 2025, 13, 14411.
77. B. Han, D. Feng, S. Li, Z. Zhang, Y. Zou, M. Gu, H. Meng, C. Wang, K. Xu, Y. Zhao, H. Zeng, C. Wang and Y. Deng, Nano Lett., 2020, 20, 4029.
78. R. Lu, A. Shokrieh, C. Li, B. Zhang, K. Amin, L. Mao and Z. Wei, Nano Energy, 2022, 95, 107009.
79. H. Song, J. Lee, M. Sagong, J. Jeon, Y. Han, J. Kim, H. Jung, J. Yu, J. Lee and I. Kim, Adv. Mater., 2024, 36, 2407381.
80. Q. Wu, Y. Mei, H. Huang, F. Zhou, H. Li and H. Chen, Mater. Today, 2024, 75, 112.
81. X. Li, M. Lv, Y. Tian, L. Gao, T. Liu, Q. Zhou, Y. Xu, L. Shen, W. Shi, X. Li, Y. Lu, X. Liu and S. Xiao, Nano Energy, 2021, 87, 106214.
82. S. Fang, F. Wu, S. Zhao, M. Zarrabeitia, G. Kim, J. Kim, N. Zhou and S. Passerini, Adv. Energy Mater., 2023, 13, 2302577.
83. H. Wang, M. Wu, J. He, G. Jiang, S. Qi, J. Huang, D. Wu and J. Ma, Energy Mater. Adv., 2025, 6, 0105.
84. Y. Wu, Y. Wang, C. Wei, Z. Song, H. Liu, Y. Ding, D. Wang, A. Zhou and J. Li, J. Power Sources, 2025, 641, 236857.
85. D. Han, Z. Wang, S. Chen, J. Zhou, S. Chen, M. Wang, D. Wu, X. Meng, C. W. Bielawski and J. Geng, Small, 2024, 20, 2405453.
86. C. Cao, Z. Xu, H. Li, K. Wang, R. Li, H. Wang, H. Dong, W. Zhang, H. Liu, S. Chen, B. Fei and M. Ge, Chem. Eng. J., 2025, 523, 168778.
87. K. Chen, X. Shen, L. Luo, H. Chen, R. Cao, X. Feng, W. Chen, Y. Fang and Y. Cao, Angew. Chem., Int. Ed., 2023, 62, e202312373.
88. H. Yue, Y. Yang, Y. Xiao, Z. Dong, S. Cheng, Y. Yin, C. Ling, W. Yang, Y. Yu and S. Yang, J. Mater. Chem. A, 2019, 7, 594.
89. G. Wang, X. Xiong, D. Xie, X. Fu, X. Ma, Y. Li, Y. Liu, Z. Lin, C. Yang and M. Liu, Energy Storage Mater, 2019, 23, 701.
90. L. Sheng, D. Zhu, K. Yang, Y. Wu, L. Wang, J. Wang, H. Xu and X. He, Nano Lett., 2024, 24, 533.
91. P. Liu, S. Shen, Z. Qiu, T. Yang, Y. Liu, H. Su, Y. Zhang, J. Li, F. Cao, Y. Zhong, X. Liang, M. Chen, X. He, Y. Xia, C. Wang, W. Wan, J. Tu, W. Zhang and X. Xia, Adv. Mater., 2024, 36, 2312812.
92. J. Jeong, J.-N. Lee, J.-K. Park, M.-H. Ryou and Y. M. Lee, Electrochim. Acta, 2015, 170, 353.
93. Q. Wang, C. Yang, J. Yang, K. Wu, C. Hu, J. Lu, W. Liu, X. Sun, J. Qiu and H. Zhou, Adv. Mater., 2019, 31, 1903248.
94. J. Meng, F. Chu, J. Hu and C. Li, Adv. Funct. Mater., 2019, 29, 1902220.
95. L. Li, M. Liu, P. Yang, W. Yuan and J. Chen, J. Colloid Interface Sci., 2024, 676, 613.
96. J. Liu, W. Hao, M. Fang, X. Chen, Y. Dong, Y. Chen, Z. Wang, X. Yue and Z. Liang, Nat. Commun., 2024, 15, 9356.
97. R. Fang, S. Ma, L. Ding, Y. Gu, X. Dong, D. Si, J. Tian, X. Lin, B. Mao and J. Li, Angew. Chem., Int. Ed., 2025, 64, e202502048.
98. S. Liu, X. Ji, N. Piao, J. Chen, N. Eidson, J. Xu, P. Wang, L. Chen, J. Zhang, T. Deng, S. Hou, T. Jin, H. Wan, J. Li, J. Tu and C. Wang, Angew. Chem., Int. Ed., 2021, 60, 3661.
99. Y. Zhao, C. Cao, N. Zhang, F. Liang, H. Dong, H. He, S. Li, Y. Feng, R. Li, W. Gu, B. Fei and M. Ge, Chem. Eng. J., 2025, 504, 158737.
100. Y. Xiang, M. Tao, X. Chen, P. Shan, D. Zhao, J. Wu, M. Lin, X. Liu, H. He, W. Zhao, Y. Hu, J. Chen, Y. Wang and Y. Yang, Nat. Commun., 2023, 14, 177.
101. P. Liu, L. Yang, B. Xiao, H. Wang, L. Li, S. Ye, Y. Li, X. Ren, X. Ouyang, J. Hu, F. Pan, Q. Zhang and J. Liu, Adv. Funct. Mater., 2022, 32, 2208586.
102. S. Y. Han, C. Lee, J. A. Lewis, D. Yeh, Y. Liu, H.-W. Lee and M. T. McDowell, Joule, 2021, 5, 2450.
103. Y. Chen, M. Li, Y. Jie, Y. Liu, Z. Zhang, P. Yu, W. Li, Y. Liu, X. Li, Z. Lei, P. Yan, T. Cheng, M. D. Gu, S. Jiao and R. Cao, Joule, 2025, 9, 101885.
104. L. Chen, Y. Xu, A. Liu, B. Cheng, S. Wang, X. Zhang, Y. Hua, L. Jiang, C. Fang, J. Han and P. K. Chu, Carbon Neutralization, 2025, 4, e70036.
105. A. Lakshmanan, R. Gurusamy, A. Ramani, N. Srinivasan and S. Venkatachalam, Ceram. Int., 2024, 50, 30240.
106. Z. Jian, Y. Hu, X. Ji and W. Chen, Adv. Mater., 2017, 29, 1601925.
107. M. Kotobuki, E. Hanc, B. Yan, J. Molenda and L. Lu, Ceram. Int., 2017, 43, 12616.
108. C. Yildirim, F. Flatscher, S. Ganschow, A. Lassnig, C. Gammer, J. Todt, J. Keckes and D. Rettenwander, Nat. Commun., 2024, 15, 8207.
109. L. Zhang, J. Yang, K. Jing, C. Li, Y. Gao, X. Wang and Q. Fang, ACS Appl. Mater. Interfaces, 2020, 12, 13836.
110. V. Thangadurai, S. Narayanan and D. Pinzaru, Chem. Soc. Rev., 2014, 43, 4714.
111. A. Banerjee, X. Wang, C. Fang, E. A. Wu and Y. S. Meng, Chem. Rev., 2020, 120, 6878.
112. S. A. Pervez, P. Ganjeh-Anzabi, U. Farooq, M. Trifkovic, E. P. L. Roberts and V. Thangadurai, Adv. Mater. Interfaces, 2019, 6, 1900186.
113. 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.
114. D. Zhou, D. Shanmukaraj, A. Tkacheva, M. Armand and G. Wang, Chem, 2019, 5, 2326.
115. K. Pan, L. Zhang, W. Qian, X. Wu, K. Dong, H. Zhang and S. Zhang, Adv. Mater., 2020, 32, 2000399.
116. Y. Wu, Y. Li, Y. Wang, Q. Liu, Q. Chen and M. Chen, J. Energy Chem., 2022, 64, 62.
117. L. Balo, H. Gupta, S. K. Singh, V. K. Singh, S. Kataria and R. K. Singh, J. Ind. Eng. Chem., 2018, 65, 137.
118. Y. Li, W. Zhang, Q. Dou, K. W. Wong and K. M. Ng, J. Mater. Chem. A, 2019, 7, 3391.
119. H. He, J. Shang, S. Li, C. Cao, H. Zhang, W. Zhang, H. Liu, Y. Feng, R. Li, S. Chen, B. Fei and M. Ge, J. Mater. Chem. A, 2024, 12, 22971.
120. S. Duan, L. Qian, Y. Zheng, Y. Zhu, X. Liu, L. Dong, W. Yan and J. Zhang, Adv. Mater., 2024, 36, 2314120.
121. S. Qian, H. Zhu, C. Sun, M. Li, M. Zheng, Z. Wu, Y. Liang, C. Yang, S. Zhang and J. Lu, Adv. Mater., 2024, 36, 2313456.
122. P. Yadav, M. S. Hosen, P. K. Dammala, P. Ivanchenko, J. Van Mierlo and M. Berecibar, Solid-State Ionics, 2023, 399, 116308.
123. L. Gao, J. Li, J. Ju, B. Cheng, W. Kang and N. Deng, Compos. Sci. Technol., 2020, 200, 108408.
124. Y. Yang, S. Yang, X. Xue, X. Zhang, Q. Li, Y. Yao, X. Rui, H. Pan and Y. Yu, Adv. Mater., 2024, 36, 2308332.
125. C.-C. Su, M. He, R. Amine, T. Rojas, L. Cheng, A. T. Ngo and K. Amine, Energy Environ. Sci., 2019, 12, 1249.
126. R. Wang, C. Tian, X. Li, T. Li, Y. Fang, K. Pan and Z. Zhou, Adv. Funct. Mater., 2025, e16074.
127. T. Li, Q. Zheng, J. Li, Z. Zhao, W. Huang, B. Zhang, G. Zhao, T. Wu, D.-L. Peng, Q. Xie, Y. Xu and K. Amine, ACS Energy Lett., 2025, 10, 2228.
128. H. He, R. Wang, Q. Qiu, M. Li, S. Chen, Y. Tang, Y. Feng, H. Shao, R. Li, C. Cao, B. Fei and M. Ge, Colloids Surf., A, 2025, 727, 138292.
129. Y. Zhang, J. Yu, H. Shi, S. Wang, Y. Lv, Y. Zhang, Q. Yuan, J. Liang, T. Gao, R. Wei, X. Chen, L. Wang, Y. Yu and W. Liu, Adv. Funct. Mater., 2025, 35, 2421054.
130. R. Fang, B. Xu, N. S. Grundish, Y. Xia, Y. Li, C. Lu, Y. Liu, N. Wu and J. B. Goodenough, Angew. Chem., 2021, 133, 17842.
131. H. Li, W. Tang, Y. Zou, S. Deng and S. Deng, J. Power Sources, 2025, 631, 236249.
132. Q. Li, F. Tian, L. Lu, Q. Lv, X. Shang, Z. Li and M. Wu, J. Power Sources, 2025, 641, 236900.
133. X. Zhou, G. Xu, K. Huang, Z. Yan, H. Hu, X. Li, G. Wang, W. Jiang, L. Yang, J. Huang, Y.-X. Wang, Y. Liang and W.-H. Lai, Nano Lett., 2025, 25, 12132.
134. J. Li, K. Zhu, J. Wang, K. Yan, J. Liu, Z. Yao and Y. Xu, Mater. Technol., 2022, 37, 240.
135. H. Chen, D. Adekoya, L. Hencz, J. Ma, S. Chen, C. Yan, H. Zhao, G. Cui and S. Zhang, Adv. Energy Mater., 2020, 10, 2000049.
136. Z. Wan, D. Lei, W. Yang, C. Liu, K. Shi, X. Hao, L. Shen, W. Lv, B. Li, Q. Yang, F. Kang and Y. He, Adv. Funct. Mater., 2019, 29, 1805301.
137. B. He, Y. Zhong, Z. Yan, T. Liu, X. Liu, Y. Du, Y. Yang, L. Yang, R. Zhang, Y. Su, Z. Huang, B. Xu and G. Zhang, Adv. Funct. Mater, 2025, 35, 2416779.
138. Y. Han, Y. Chen, Y. Huang, M. Zhang, Z. Li and Y. Wang, Ceram. Int., 2023, 49, 29375.
139. M. J. Counihan, Z. D. Hood, H. Zheng, T. Fuchs, L. Merola, M. Pavan, S. L. Benz, T. Li, A. Baskin, J. Park, J. H. Stenlid, X. Chen, D. P. Phelan, J. W. Lawson, J. G. Connell, J. Janek, F. H. Richter and S. Tepavcevic, Adv. Energy Mater., 2025, 15, 2406020.
140. D. Zhang, B. Li, S. Wang and S. Yang, ACS Appl. Mater. Interfaces, 2017, 9, 40265.
141. Y. Han, R. Fang, C. Lu, K. Wang, J. Zhang, X. Xia, X. He, Y. Gan, H. Huang, W. Zhang and Y. Xia, ACS Appl. Mater. Interfaces, 2023, 15, 31543.
142. H. Zhuang, H. Xiao, T. Zhang, F. Zhang, P. Han, M. Xu, W. Dai, J. Jiao, L. Jiang and Q. Gao, Angew. Chem., Int. Ed., 2024, 63, e202407315.
143. Y. Li, T. Wang, J. Chen, X. Peng, M. Chen, B. Liu, Y. Mu, L. Zeng and T. Zhao, Sci. Bull., 2023, 68, 1379.
144. H. Gao, C. Lin, Y. Liu, J. Shi, B. Zhang, Z. Sun, Z. Li, Y. Wang, M. Yang, Y. Cheng and M.-S. Wang, Sci. Adv., 2025, 11, eadt4666.
145. C. Wang, R. Lin, Y. He, P. Zou, K. Kisslinger, Q. He, J. Li and H. L. Xin, Adv. Mater., 2023, 35, 2209091.
146. S. Benning and F. Hausen, ECS Meet. Abstr., 2018, 4, 256.
147. K. N. Wood, K. X. Steirer, S. E. Hafner, C. Ban, S. Santhanagopalan, S.-H. Lee and G. Teeter, Nat. Commun., 2018, 9, 2490.
148. D. Kong, X. Wang, J. Wei, S. Chu, H. Li, W. Guo, Y. Mu, K. Zhu, Y. Han, W. Li, A. Xie, N. Wang, Y. Zhao and Y. Zhao, ACS Nano, 2026, 20, 1614.

---