Issues Concerning Interfaces with Inorganic Solid Electrolytes in All-Solid-State Lithium Metal Batteries

Abstract

All-solid-state batteries have attracted wide attention for high-performance and safe batteries. The combination of solid electrolytes and lithium metal anodes makes high-energy batteries practical for next-generation high-performance devices. However, when a solid electrolyte replaces the liquid electrolyte, many different interface/interphase issues have arisen from the contact with electrodes. Poor wettability and unstable chemical/electrochemical reaction at the interfaces with lithium metal anodes will lead to poor lithium diffusion kinetics and combustion of fresh lithium and active materials in the electrolyte. Element cross-diffusion and charge layer formation at the interfaces with cathodes also impede the lithium ionic conductivity and increase the charge transfer resistance. The abovementioned interface issues hinder the electrochemical performance of all-solid-state lithium metal batteries. This review demonstrates the formation and mechanism of these interface issues between solid electrolytes and anodes/cathodes. Aiming to address the problems, we review and propose modification strategies to weaken interface resistance and improve the electrochemical performance of all-solid-state lithium metal batteries.

1. Introduction

With the rapid development of energy in electric and hybrid vehicles, there is an increasing demand for high-performance rechargeable batteries [1]. Traditional lithium-ion batteries with the organic liquid electrolyte cannot completely meet the demand: (1) The narrow electrochemical window hinders the application of cathodes with high-voltage and lithium metal anodes [2]; (2) When the current increases, a lithium-ion concentration gradient appears and results in high internal resistance [3]; (3) They can only work safely within the temperature range of 0–40 °C; (4) They react with the anode to produce a solid electrolyte interphase (SEI), resulting in continuous loss of electrolyte and lithium [4,5,6,7]. Especially, liquid electrolyte faces the risk of leakage and burning, which threaten the safety of cells. Due to the inevitable problems of liquid batteries, all-solid-state batteries with lithium metal anodes have received more and more attention (Figure 1). Since 2011, the number of published papers on solid-state lithium metal batteries has increased year by year. The lithium metal anode has an extremely high theoretical capacity (3960 mAh g⁻¹) and the lowest electrochemical potential (−3.04 V vs. SHE) [1]. When used as an anode, they make high-energy batteries feasible. However, inevitable lithium dendrites growth in the organic liquid electrolytes may lead to short-circuit and combustion in practical use [8,9]. Replacement of the liquid electrolyte with the solid electrolyte makes lithium metal anode more practical (Figure 2). There will be no risk of electrolyte leakage, combustion, and explosion in all-solid-state batteries [10,11].

Solid-state batteries show the following advantages: (1) Lithium metal can be used as anode due to the suppression of growth of dendrites, significantly increasing the energy density compared with commercial graphite anode [12]; (2) Due to the chemical and electrochemical stability of solid electrolytes, a series of safety problems such as leakage, corrosion, and even flammability will be avoided, which is particularly important for grid-scale storage and transportation applications including airplanes and automobiles [13,14,15]; (3) The lithium ionic conductivity of some electrolytes even exceed that of liquid electrolytes. Additionally, inorganic solid electrolytes exhibit excellent safety due to their flame retardancy and thermal stability [16,17].

Inorganic solid electrolytes include NASICON-type (Li_1.4Al_0.4Ti_1.6(PO₄)₃, LATP) [18,19], Perovskite-type (Li_3xLa_2/3−xTiO₃, LLTO) [20,21], Sulfide-type (Li₇P₃S₁₁, LPS) [22,23], Garnet-type (Li₇La₃Zr₂O₁₂, LLZO) [24,25], etc. Garnet-type electrolyte (LLZO) is a typical oxide electrolyte, and its ionic conductivity at room temperature can reach 10⁻³ to 10⁻⁴ S cm⁻¹ [26,27]. It has the characteristics of high thermal stability, wide chemical windows, excellent chemical stability against air, and electrochemical stability against lithium metal. The sulfide electrolyte is mainly composed of Li₂S and sulfides (SiS₂ [28], P₂S₅ [29], and GeS₂ [30]). Due to the larger radius of S²⁻, the lithium-ion transmission channel of the sulfide electrolyte is larger, so the ion conductivity can even reach 10⁻² S cm⁻¹ at room temperature, higher than that of the oxide electrolyte [31].

Although solid electrolytes have gained much attention, they must meet a variety of performance requirements, such as high ionic conductivity at room temperature, high migration of Li⁺ ions, chemical stability, sustainability, and economic feasibility [32,33,34]. There is enough indirect evidence that the solid-solid interface between electrodes and electrolytes hinders the migration of lithium ions, leading to high interface impedance and hindering the commercialization of solid-state batteries [35,36,37]. When solid electrolytes replace organic liquid electrolytes, the original solid-liquid interface is replaced by a solid-solid interface, which contacts worse and results in a high interface resistance [38]. It should be noted that the mechanical failure and degradation of the interface caused by electrolyte decomposition, or the formation of a lithium-depleted space charge layer caused by the large chemical potential difference are also reasons for the high interface resistance. Since solid electrolyte processes excellent chemical/electrochemical stability, so the possibility of interface decomposition can be ignored [39].

On the other hand, the solid electrolyte makes the lithium metal anode available in an all-solid-state battery for a high energy density. However, the wettability of lithium metal anode to the solid electrolytes is generally poor [40,41,42]. The resulting high contact resistance is an important reason for the growth of lithium dendrites. Therefore, this review takes Garnet-type and sulfide electrolytes as examples to discuss the issues and improvement strategies of the interface in all-solid-state lithium metal batteries. The interface issues are discussed in two main parts: (1) the interface between solid electrolyte and cathode and (2) the interface between solid electrolyte and lithium metal anode (Figure 3).

2. Interface Issues in Solid-State Lithium Metal Batteries with Electrodes

Commercial lithium-ion batteries with liquid organic electrolytes show a relatively good performance. The liquid electrolyte has good wettability to the active material particles, which promotes the rapid transfer of lithium ions in the interfaces and exhibits excellent interface dynamics [43]. However, during repeated charging and discharging, the deterioration and exhaustion of the liquid electrolyte will reduce the lifespan of the battery. Additionally, when the battery is short-circuited, the organic liquid electrolyte will decompose and cause a combustion explosion [10,44]. The introduction of the solid electrolyte effectively avoids these problems. However, impedance from the interface weakens the performance of solid-state batteries.

2.1. Interface Issues between Cathode and Solid Electrolyte

The solid electrolyte is rigid, and its surface is rough, which significantly reduces the contact area with electrode active particles, hindering the transfer dynamics of lithium ions and leading to a high interface impedance. Although the ionic conductivity of the bulk solid electrolyte reaches 10⁻² to 10⁻³ S cm⁻¹, the diffusion of lithium ions is still significantly weakened by the high interface impedance, with which the solid electrolyte cannot fully take advantage of the high ionic conductivity.

2.1.1. Poor Physical Contact

Most intercalated cathodes will have a volume change of about 5% during the cycle (Figure 4), and the microscopic particles may repeatedly crack [45]. The volume change of conversion electrodes such as sulfur is 80% [46] and 300% for silicon [47]. Because the elastic deformation of solid electrolytes is very limited, their only means to confront electrode size changes is the displacement or irreversible deformation (cracking, etc.) [48]. So, it is generally believed that when the rigid solid electrolyte match electrodes, there will inevitably be nano- or micro-pores in the interface, causing contact loss, blocking the ion conduction path, and triggering high interface impedance [49]. During repeated charging and discharging processes, the physical contact between the cathode and the electrolyte is weakened more, then the ion transport path is further blocked. Weak physical contact and the resulting heterogeneous interface chemical properties may cause an uneven charge state and stress distribution, significantly affecting the interface stability during cycling.

2.1.2. Space Charge Layer

The formation of a lithium-depleted space charge layer due to the large chemical potential difference is an important reason for the high interface resistance [50,51]. When two materials with different chemical potentials are in contact, electrons and atoms cannot establish local charge neutrality and charges will be generated at the interface to form a space charge layer. Ohta et al. believe that when the sulfide solid-state electrolyte is in contact with the oxide cathode [52], the space charge layer on the cathode side will quickly disappear because the electrons can neutralize the excess lithium ions. So, the space charge layer is usually distributed on the solid-state electrolyte side. The poor kinetics caused by the deviation from electrical neutrality greatly hinders the conduction of ions and electrons and leads a poor electrochemical performance. Inserting an additional ion-conductive and electron-insulated oxide layer as a buffer layer is an effective way to suppress the space charge layer (Figure 5).

2.1.3. Element Cross-Diffusion

Due to the narrow electrochemical window, the sulfide electrolyte is easily oxidized by the oxide cathode under a high voltage. The adverse chemical reaction or element diffusion between the oxide cathode and the sulfide electrolyte has been verified by calculation and experimental results [53,54]. Structural disturbances caused by interface chemical reactions can also lead to a higher interface resistance. Element cross-diffusion is also considered to be due to the instability of the thermodynamic interface [55]. Generally, elements P, Co, and S are easily diffused. Experiments using scanning transmission electron microscopy and energy-dispersive X-ray spectroscopy showed that the diffusion of Co element from the LiCoO₂ cathode to the sulfide electrolyte is quite serious [56,57]. A distinct interfacial layer appeared between LiCoO₂ (Figure 6a) and SSE. This diffusion of Co was observed to be 30 nm (Figure 6b) [58]. Sakuda et al. [56] firstly observed the existence of a 10 nm interface layer between LiCoO₂ and Li₂S-P₂S₅ through a transmission electron microscope (TEM). A. Banerjee et al. [59] verified that the interface layer products caused by element-diffusion of high-voltage cathode LiNi_0.85Co_0.1Al_0.05O₂ (NCA) and Li₆PS₅Cl (LPSCl) are Ni₃S₄, LiCl, Li₃PO₄ and oxidized LPSCl (Figure 6c).

2.2. Interface between Lithium Metal Anode and Solid Electrolyte

Lithium has the lowest potential (−3.04 V), the smallest mass density (0.59 g cm⁻³), and an extremely high specific capacity (3860 mAh g⁻¹) [60], which make it possible for high energy density batteries with lithium metal anode. However, in liquid batteries, the application of lithium metal is significantly limited by the growth of lithium dendrites [61]. Due to uneven lithium plating and stripping (Figure 7), lithium dendrites usually form during repeated electrochemical cycles [62,63]. In the liquid electrolyte, the growth of lithium dendrites is unimpeded. Because the tip electric field is stronger, lithium tends to deposit on the protrusions, which manifests as an unlimited growth of dendritic lithium. As the battery life increases, dendrites will continue to grow, and may eventually pierce the separator, causing a short circuit or even burning and explosion. In addition, owing to the high chemical activity of lithium, an electronically insulating solid electrolyte interphase (SEI) will be formed immediately when the lithium metal and the organic liquid electrolyte come into contact [10]. Because of poor elastic properties, SEI break during the repeated cycles, then the newly exposed lithium meets the electrolyte and generates new SEI quickly. The repeated generation and rupture of SEI leads to the formation of dead lithium and the loss of effective reversible capacity [62]. When a solid electrolyte is used, the growth of lithium dendrites can be significantly suppressed. The high shear modulus of the solid electrolyte can inhibit the growth of lithium dendrites. Monroe and Newman et al. believe that when the shear modulus of the electrolyte reaches 6.8 GPa (2 times that of lithium metal) [64], it can prevent the puncture of lithium dendrites. Almost all inorganic solid electrolytes meet this condition. However, some studies have found that lithium dendrites will still grow along grain boundaries or across grains in Garnet-type electrolytes [65].

2.2.1. Weak Chemical/Electrochemical Stability

Lithium metal generally shows poor chemical or electrochemical stability with solid electrolytes [66]. The stability window of the electrolyte is determined by the energy separation (E_g) between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) (Figure 8a). When the electrochemical potential (μ_a) of the anode is higher than LUMO, the electrolyte will be reduced unless a passivation layer is formed to prevent electron transfer. When the electrochemical potential (μ_c) of the cathode is lower than HOMO, the electrolyte will be oxidized unless a passivation layer is generated. Therefore, only when the electrochemical window is larger than the potential difference between μ_a and μ_c, the electrolyte can maintain thermodynamic stability. That is:

$$e{V}_{oc}=e({\mu }_c-{\mu }_a)\le E_g$$

Unfortunately, because lithium metal has the lowest electrochemical potential (−3.04 V), which is easily higher than the LUMO of the solid electrolyte, it is difficult for the electrolyte to maintain thermodynamic stability. According to the intrinsic properties of the lithium metal and the electrolyte, there are generally two situations at the interface [67]. In the first case, the solid electrolyte does not react with lithium metal at all and forms a distinct two-dimensional boundary instead, which appears as a stable interphase (SI) (Figure 8b). The second situation is that the chemical reaction of anode and electrolyte is under the thermodynamic drive. One type is a mixed-conductor interphase (MCI) (Figure 8c), which has partial electronic and ion conductivity. The interphase gradually grows into the electrolyte, and eventually, electrons are transported through the electrolyte to cause self-discharge. If this intermediate phase is electronically non-conducting, it is like SEI (Figure 8d). The ionic conductivity of SEI will significantly affect the electrochemical performance of the battery.

2.2.2. Poor Physical Contact

Physical contact significantly affects lithium-ion transport and dynamic diffusion (Figure 9a) [68]. Due to the high interfacial energy, the weak wettability of lithium metal and solid electrolyte leads to poor contact (Figure 9b). The local microscopic tension and stress on the anode side will cause the deterioration of close contact between the electrode and the electrolyte [69]. Poor physical contact will reduce the reaction kinetics of the electrode and weaken the performance of the electrolyte. At the same time, the limited and uneven contact area brings a high impedance, causing uneven deposition of lithium. Lithium is in contact during plating and detached during peeling, which intensifies the growth of dendrites. Unregularly deposited lithium will fill the pores and cracks at the adjacent interface, eventually causing permanent failure of the electrolyte (Figure 9c) [70].

3. Modification of Interface between Cathode and Solid Electrolyte

This chapter mainly discusses the interface characteristics between sulfide/Garnet-type oxide solid electrolytes and cathode. It is generally believed that the interface stability between the sulfide electrolyte and cathode is worse than that between the oxide electrolyte and cathode [18,71,72].

3.1. Interface between Sulfide Electrolyte and Cathode

All-solid-state batteries make it possible to have a long cycle life [71,73]. However, they face some critical problems in the practical application, such as the low operating current density and low power density. Therefore, sulfide electrolytes began to gain attention. The interface between sulfide electrolytes is less stable than oxides, but they provide better ionic conductivity and better workability [72]. In recent reports, the ionic conductivity of sulfide electrolytes is comparable to that of organic liquid electrolytes [74,75]. The bottleneck in the application of sulfide electrolytes is that considerable electrical resistance is often observed near the cathode/sulfide electrolyte interface [76]. Since the oxygen of the oxide cathode material is more attractive to lithium than the sulfur of the sulfide solid electrolyte, a large amount of lithium in the sulfide electrolyte moves to the oxide cathode material, resulting in a lack of lithium in the electrolyte side and forming a space charge layer. The cathode acts as a mixed conductor so that electrons can also migrate to neutralize the lithium-ion concentration on the cathode side. The result of neutralization is that the lithium ions on the sulfide side will still be attracted by the oxygen of the cathode and continue to migrate, resulting in the gradual depletion of charge carriers in this area and the gradual enlargement of the space charge layer. This is the origin of the cathode/sulfide electrolyte interface resistance [52].

3.1.1. Suppressing Space Charge Layer

A buffer layer with ionic conductivity applied between the solid electrolyte and the cathode can effectively suppress the formation of the space charge layer. The buffer layer commonly needs to meet the following characteristics: (1) High ionic conductivity and no electronic conductivity are required; (2) It should be thin enough to reduce the impact of the buffer layer resistance; (3) The anion of the buffer layer has a strong attraction to lithium ions to avoid a space charge layer, which generally takes oxygen ions. All the above three conditions require precise processes to form a thin and uniform buffer layer.

Jun et al. applied a buffer layer of LiNbO₃ (LNO) between the LiCoO₂ (LCO) cathode and sulfide electrolyte (β-Li₃PS₄, LPS) (Figure 10) [76]. It was found that a space charge layer was formed at the beginning of charging. The intervention of the buffer layer LNO inhibited the space charge layer and provided lithium transport paths. It formed a matching interface without lithium adsorption space and suppressed the uneven distribution of lithium. Ohta et al. used Li₄Ti₅O₁₂ to coat the surface of LiCoO₂ to make a cathode/solid electrolyte interface, and observed by electrochemical impedance spectroscopy (EIS) that the minimum resistance of the sample was 20 times lower than that of the uncoated sample [52]. Xu et al. found that when the surface of LiCoO₂ was covered by TaO₃ nanosheets, the electrode resistance was reduced by two orders of magnitude [77]. The above method confirms that embedding or coating a buffer layer between the cathode and the sulfide solid electrolyte is currently effective to suppress the generation of the space charge layer. Although there have been many discussions on the space charge layer, currently no direct evidence can verify it. Moreover, the space charge layer in all-solid-state batteries explained by electronic potential is difficult to explain the Debye-screening length of ion conductor materials. Therefore, the influence of the space charge layer on the interface impedance needs further verification.

3.1.2. Inhibiting the Element Cross-Diffusion

The structural disorder caused by interface chemical reaction or diffusion leads to poor electrochemical performance. Woo J.H. et al. applied ALD (atomic layer deposition) method to coat Al₂O₃ on LiCoO₂ to inhibit the element diffusion [58]. EIS research showed that the Al₂O₃ layer on LiCoO₂ effectively reduced the interface resistance. The microstructure and elemental analysis using a high-resolution transmission electron microscope (TEM) and energy spectrometer (EDS) showed that the coating layer inhibits the diffusion effect of Co, P, and other elements between the LiCoO₂/Li_3.15Ge_0.15P_0.85S₄. So far, researchers have found that applying SiO₂ [78], TiO₂ [79], ZrO₂ [80], and Li₂SiO₃ [78] coating layers on the surface of the cathode can also have a similar effect on inhibiting the diffusion of the above elements. It is worth noting that the coating thickness must be paid close attention to because the thick coating will make the resistance of the insulated coating itself become the main part of the interface impedance. Sun et al. found that Li₂WO₄ (LWO) can prevent the diffusion of Co and O from LCO into the Li₆PS₅Cl solid electrolyte (Figure 11a), thereby significantly reducing the interfacial impedance [81]. Figure 11b,c shows the SEM images and EDS point spectrum results of the bare LCO battery and the LWO-LCO battery after 100 cycles, respectively. Obviously, after the introduction of LWO, the Co element will be difficult to diffuse into the Li₆PS₅Cl solid electrolyte.

3.1.3. Improving the Weak Physical Contact

Weak physical contact between solid-solid interfaces is also a problem that needs to be paid attention to. To ensure full contact between the electrolyte and the cathode, many strategies have been proposed, including preparing nano-scale active materials, constructing good ionic and electronic pathways with the cathode, and increasing the ionic conductivity of the active material [55].

Ball milling and liquid phase synthesis are common methods for preparing nano-active materials. Han et al. [82] designed a bottom-up preparation method, dissolving Li₂S, polyvinylpyrrolidone, and Li₆PS₅Cl in ethanol, followed by a co-precipitation and carbonization process. Nano-sized Li₂S and Li₆PS₅Cl grew in situ in a soft carbon matrix (Figure 12a). This nanostructure provided a fast ion conduction path and a buffer layer to reduce the stress and strain during charging and discharging. The carbon material with high electronic conductivity was evenly distributed within the cathode and electrolyte, providing a three-phase contact (Figure 12b), effectively increasing the contact area between the active materials. In addition, the coating structure can suppress the volume change of the cathode during charging and discharging. The battery exhibited excellent cycle life and rate characteristics (Figure 12c). Zhang et al. [83] verified that the NCM cathode was disconnected from the solid electrolyte during the stripping process. They monitored internal pressure changes through LiCoO₂/Li₁₀GeP₂S₁₂/In cells. According to Figure 12d, the internal pressure increased almost linearly during charging and decreased during discharge, which is consistent with the charge-discharge curve. At a high discharge rate, the capacity obtained was smaller, so the pressure change was smaller (Figure 12e). The decrease in capacity was consistent with that in internal pressure, indicating that the irreversible process of the electrode/electrolyte interface is related to the particle contact loss. Shin et al. [84] ball-milled TiS₂ and Li₂S-P₂S₅ to obtain nano-level cathode and achieved high electrochemical performance (Figure 12f). The prepared nano-TiS₂-electrolyte composite material effectively contributed to the improvement of the physical contact of interfaces (Figure 12g).

3.2. Interface between LLZO and Cathode

The interface resistance between Garnet-type electrolyte (LLZO) and LiCoO₂ cathode has been a major research focus in recent years for the optimization of solid-state battery performance. The chemical potential of Garnet-type electrolyte and LiCoO₂ cathode is not much different, so there is no significant space charge layer. The volume change, elements diffusion, and poor chemical stability are the main factors to limit the use of Garnet-type electrolytes.

3.2.1. Suppressing the Element Cross-Diffusion

The internal stress of the cathode materials was generated during ion deintercalation. Among the cathode materials, LCO shows a volume change of 2%, LFP of 6.81%, NMC of 3.4%, and NCA of 5.9% [85,86]. The volume change of the cathode causes lithium-induced stress, leading the crystal to crack and pulverize. The solid electrolyte has mechanical rigidity, so the confrontation of the stress will cause mechanical degradation. Bucci et al. [87] calculated that when the volume expansion exceeded 3%, the entire solid electrolyte system will crack and increase the curvature of the electrode, leading to uneven lithiation and local stress, and ultimately causing battery failure. To overcome the volume change caused by weak physical contact and stress, high-temperature sintering technology is generally used to sinter LLZO and the cathode material together. Zhang et al. performed co-sintering (a relatively low temperature at 873 K) in the air after ball milling LLZO with commercial LCO and NCM cathodes (Figure 13a) [88]. Under this process, LLZO can spontaneously cover the surface of the cathode material, then the thickness of the surface layer can be reduced to nanometer level after co-sintering. They calculated the interfacial exchange energies of Ni-La and Ni-Li using first-principles density functional theory (DFT). Figure 13b shows the 3D model of the NCM/LLZO interface, whereas Figure 13c,d shows the exchange energies of Ni-Li and Ni-La, which are −1.20 and −0.85 eV, respectively. They also calculated the exchange energy of the LCO/LLZO interface and found that the Co-Li exchange occurred more easily than that of Ni-Li, and the Mn-Li exchange was less likely to occur. Exchanges of Co-La and Mn-La do not occur at the interface. Therefore, compared with NCM + LLZO, the sample LCO + LLZO exhibits a more stable electrochemical performance.

It is worth noting that Park et al. [89] found that the ionic conductivity is significantly reduced when the Garnet cubic phase structure is transformed into a tetragonal phase (Figure 14a). When LiCoO₂ was directly formed on the Garnet-type electrolyte matrix, the transformation from the cubic phase to the tetragonal phase will be easier. In this case, the cathode can only provide a low capacity of 35 mAh g⁻¹, which was only 25% of the theoretical capacity. Cross-diffusion of elements was still not evitable. The formation of a LiCoO₂ film on a Garnet-type oxide solid electrolyte substrate at 700 °C would result in the formation of LaCoO₃ at the interface [90]. Zarabian et al. found that even at 400 °C, side-reaction products containing Co in various oxidation states existed [91]. Herein, although high-temperature sintering can improve the physical contact, the chemical instability between the cathode and LLZO can also lead to the formation of high resistance. On the other hand, this high-temperature process is also accompanied by the loss of lithium raw materials, which not only leads to high interface resistance but also affects electrochemical performance, especially the initial Coulomb efficiency and cycle life. So, surface modification of the Garnet-type electrolyte (for example, the introduction of a co-diffusion surface layer or Li₃BO₃, etc.) is used to reduce the cross-diffusion of elements and the transformation of the cubic phase to the tetragonal phase. Introducing ionic conductive materials with low melting point materials into composite cathodes is a conventional strategy. Li₃BO₃ (LBO) has a low melting point (~700 °C) and good ionic conductivity. It can be melted as a binder for active materials during the sintering process [92]. After annealing at 700 °C, from the comparison of the backscattered secondary electron (BSE) image and the secondary electron image of the cross-section (Figure 14b), LBO appeared in the pores among LCO particles due to the melting and diffusion during the annealing process. Han et al. [93] further designed Li_2.3C_0.7B_0.3O₃ (LCBO) to improve the performance of all-solid-state batteries (Figure 14c). The ionic conductive eutectic LCBO was directly formed on the surface of LLZO and LCO. This design avoided the direct contact between the cathode and the electrolyte, and effectively inhibited the element diffusion and chemical reaction between LCO and LLZO.

3.2.2. Strengthening the Physical Contact and Resisting the Volume Change

With the co-sintering strategy, the low electronic conductivity of the cathode and the electrolyte may be the reason for the poor ionic and electronic diffusion dynamics. N-type semiconductor indium tin oxide (ITO) with high free carrier concentration (10²⁰–10²¹ cm⁻³) and low resistivity (<10⁻³ Ω cm) is a common additive to improve electronic conductivity. Liu et al. [95] incorporated highly conductive ITO into a composite cathode composed of LCO and LBO, significantly reducing the interface resistance. However, this type of electrode generally must be tested at a higher temperature. Therefore, polymer interlayers are an alternative to operating at room temperature or relatively high temperature (around 65 °C). Wang et al. [94] used poly(acrylonitrile-co-butadiene) (PAB) to coat the NMC cathode. It improved the physical contact between the cathode and the electrolyte and suppressed the side reaction of the interface and the cross-diffusion of elements. Fu et al. [96] built a three-dimensional porous solid electrolyte and a sulfur cathode to form a composite electrode (Figure 15a,b). In this work, there were two LLZO layers with different densities: a thin and dense electrolyte layer was used to isolate the cathode and anode and improved the mechanical strength to prevent lithium dendrites from puncturing. The thick and porous layer was compounded with the sulfur cathode to increase physical contact and provide a lithium-ion conduction channel. Additionally, strong mechanical properties can buffer the volume expansion during charging and discharging. This bi-functional garnet-type electrolyte helped achieve excellent electrochemical performance.

4. Modification of Interface between Lithium Metal Anode and Solid Electrolyte

The lithium metal anode has an extremely high theoretical capacity (3960 mAh g⁻¹) and the lowest electrochemical potential (−3.04 V vs. SHE) [97]. However, the formation of powdered lithium and lithium dendrites in the lithium metal/liquid electrolyte system has restricted the development of lithium metal batteries from the 1990s to the beginning of the 21st century, because they make lithium metal batteries face serious safety concerns. The growth of lithium dendrites is mainly caused by the concentration gradient of lithium ions and the uneven surface of the anode during cycling [98,99]. Additionally, lithium metal and any organic solvent are thermodynamically unstable and will react instantly to form a new passivation layer called SEI. SEI cracks in the repeated depositing and peeling. Exposed fresh lithium will meet the electrolyte and generates new SEI immediately. Along with this process, lithium dendrite grows until the short-circuit [49,100,101]. With the development of solid electrolytes, lithium metal batteries have gradually returned to the field of vision. Compared with porous gel electrolytes and polymer electrolytes, solid electrolytes are denser and have higher strength and hardness, which can effectively prevent the penetration of lithium dendrites and improve battery safety.

However, the high interface impedance of the solid electrolyte/lithium metal interface still hinders the rate performance and power density of solid-state batteries. Li et al. found that the interface resistance between LLZO electrolyte and lithium metal was estimated to be 1700 Ω·cm⁻², which was greater than the total resistance of LLZO particles [102]. There are two main problems in the interface between the lithium metal anode and the solid electrolyte: interface stability and interface impedance [10]. First, due to the strong reducibility of lithium metal, some high-valent cations in the solid electrolyte are easily reduced by electrons, resulting in a high interface resistance phase. At the same time, due to the poor chemical stability of the solid electrolyte, there may be an interface reaction with the electrode. For example, the contact of perovskite Li_3xLa_2/3-xTiO₃ (LLTO) with metallic lithium causes Ti⁴⁺ to be reduced, resulting in a higher electronic conductance. Some solid electrolytes with high oxidation state elements, such as Li_1+xAl_xM_4+2-x(PO₄)₃ (M = Ge⁴⁺ or Ti⁴⁺, namely LAGP or LATP) and Li_4−xGe_1−xP_xS₄ (LGPS, Ge⁴⁺), have been verified to be reduced and form an interface layer that conducts ions and electrons, leading to continuous decomposition of the electrolyte [103,104]. Secondly, due to the small effective contact area of the solid-solid interface, it is difficult to achieve close physical contact, resulting in a higher interface impedance [105]. At the same time, the volume of lithium metal changes repeatedly during cycles, while the electrolyte does not change, which makes the interface stress larger and damages the interface structure. As a result, the physical contact deteriorates further [65,106].

Compared with polymer electrolytes, interface stability and interface impedance are more prominent in inorganic ceramic electrolytes. Improving the interface affinity and stability can be achieved by constructing a lithiophilic buffer structure, modifying the electrolyte surface, and directly modifying the lithium metal surface.

4.1. Constructing Buffer Layer

Construct buffer layer is an important way to improve the interface between lithium metal anode and electrolyte. NASICON electrolyte has relatively high lithium ionic conductivity (10⁻⁴–10⁻³ S cm⁻¹), which is more stable in the air and easier to be prepared. However, NASION-type electrolytes are generally unstable with lithium metal [107]. LAGP (Li_1.5Al_0.5Ge_1.5P₃O₁₂) is a very important NASION-type electrolyte with high ionic conductivity and a wide chemical window (>6 V) [100]. However, Ge⁴⁺ in LAGP will be reduced by lithium metal, because the reaction of Ge⁴⁺ to Ge²⁺ and Ge⁰ has a higher free reaction enthalpy (Figure 16a). Liu et al. [101] sputtered an amorphous Ge film on the surface of LAGP with a thickness of only 60 nm (Figure 16b). Not only can the reduction of Ge⁴⁺ be inhibited, but it can also help the lithium metal to make close contact with the electrolyte (Figure 16c). LLZO electrolytes are also reduced after contact with lithium metal, which leads to high interfacial charge-transfer impedance and lithium dendrites formation. Niu et al. [108] introduced a nano-scale Li phosphorous oxynitride (LiPON) layer on the Nb-LLZO pellets. LiPON reacted with lithium to produce a lithiophilic, electronically insulating, and ionic conductive interphase, which contributed to a stable lifespan of over 2000 h without any short circuit.

Introducing a soft or liquid interlayer with high conductivity and excellent chemical stability between lithium and solid electrolyte is also a major strategy. Compared to inorganic solid electrolytes, gel polymer electrolyte has good flexibility, and viscoelasticity, and is lightweight, so it is often used as an interlayer. Liu et al. used PVDF-HFP polymer and an appropriate amount of liquid electrolyte to form a gel interlayer, which was placed between the electrolyte and lithium metal (Figure 16d) [109]. The gel interlayer showed a high lithium ionic conductivity (5 × 10⁻⁴ S cm⁻¹) and electrochemical stability. Due to its good wettability with lithium metal, its interface impedance dropped significantly from 1.4 × 10³ to 214 Ω cm². Huo et al. [110] introduced an ionic liquid [BMIM]TF₂N to wet the PEO/garnet complex to form a composite electrolyte (SPE). It was stable at room temperature and showed a high ionic conductivity (Figure 16e). Compared to the electrolyte without ionic liquid, the addition of [BMIM]TF₂N increased the conductivity by an order of magnitude at room temperature, reaching 2.2 × 10⁻⁴ S cm⁻¹. Therefore, a high and stable battery capacity was obtained. The ionic conductivity of Li₁₀GeP₂S₁₂ at room temperature can reach 2.2 × 10⁻³ S cm⁻¹, but it was not chemical stable with lithium metal. Yao et al. [111] introduced 75% Li₂S-24% P₂S₅-1% P₂O₅ to improve the interface impedance with lithium metal with Li₁₀GeP₂S₁₂. The results of the cyclic voltammetry of 75% Li₂S-24% P₂S₅-1% P₂O₅ showed that there was no obvious peak in the voltage window of −0.5 V to 10 V (vs. Li⁺/Li), which means it was chemically stable to lithium. The all-solid-state lithium-sulfur battery under this strategy showed a high capacity, good cycle stability, and excellent rate performance.

4.2. Surface Modification of Electrolyte

According to the first-principles calculation results, almost all solid electrolytes are thermodynamically unstable to lithium, except for LLZO due to its lowest reduction potential [112]. Therefore, LLZO is the most suitable solid-state electrolyte for lithium metal anode. However, the poor wettability of lithium is an important challenge to LLZO. Poor physical contact from the poor wettability leads to high interface resistance and uneven current distribution. Due to the poor wettability between garnet and molten lithium, the effect of heating or melting lithium to reduce the interface impedance is very limited [10]. Therefore, surface modification on electrolytes has been performed to enhance the affinity with lithium, thereby reducing the interface impedance.

Luo et al. [113] introduced ultra-thin amorphous silicon of only about 10 nm between the electrolyte/lithium metal interface by plasma chemical vapor deposition (Figure 17a). Silicon formed Li_xSi during the lithiation process. The lithiated silicon specimens are good lithium ionic conductors (D_Li+ ≈ 10⁻¹² m² s⁻¹), effectively improving the lithium-philicity of the electrolyte and reducing the electrochemical interface impedance. Han et al. [114] used atomic layer deposition technology to deposit an Al₂O₃ layer with a thickness of about 5.2 nm on the surface of Li₇La_2.75Ca_0.25Zr_1.75Nb_0.25O₁₂ (LLCZN) (Figure 17b). The deposited layer reduced the interface impedance from 1710 Ω cm² to 34 Ω cm². SAED and EELS results showed that Li₂Al₄O₇ with a high crystal symmetry structure was formed during charging and discharging. The lithiophilic Li₂Al₄O₇ improved the wettability of the interface and greatly enhanced the effective contact area between molten lithium and the electrolyte (Figure 17c). ZnO [115] is also deposited on the surface of the garnet to promote wetting with lithium metal. In this case, ZnO was reduced by metallic lithium and formed a Zn-Li alloy, a stable and continuous conduction layer between the lithium metal and the electrolyte. Ge, Al, Sn, Mg [116,117,118,119] and other metals that are compatible with lithium metal can be also used on the surface of solid electrolytes. Huo et al. [120] used a simple conversion reaction between Cu₃N and molten lithium at 200 °C to form a mixed ion electronic conductive layer (MCL) to modify the LLZO/Li interface (Figure 17d). Li₃N had high Li⁺ conductivity (~10⁻³ S cm⁻¹) and low Li⁺ migration energy barrier (0.007–0.038 eV) at room temperature, which was conducive to the rapid transmission of lithium ions in the interface layer. The interface resistance was significantly reduced from 1138.5 Ω cm² to 83.4 Ω cm². Generally, the lithium alloy layer may fall off from the garnet electrolyte after hundreds of cycles [118], while the MCL was more stable on the garnet. The uniformly distributed Cu nanoparticles inside the MCL can not only disperse uniform electron distribution to reduce the nucleation of lithium dendrites but also buffer the volume change of lithium.

5. Conclusions

Large-scale energy storage systems have become an important part of the future power grid. Energy storage technology with high safety, long life, and high specific capacity has become an urgent need. At present, the research on all-solid-state lithium metal batteries focuses on the development of high lithium-ion conductivity and high stability of electrolytes, but the interface problem is still a core issue. Issues including weak physical contact, chemical/electrochemical unstable interface, space charge layer, element cross-diffusion, and others restrict the development and application of all-solid-state lithium metal batteries. After unremitting efforts, there have been many strategies proposed to improve the interface issues and achieved excellent electrochemical performances.

Despite the fruitful results, all-solid-state lithium metal batteries still face challenges:

(1)

Various in-situ/ex-situ characterization methods are required to understand interface properties. The characterization content on the time scale could include the evolution of charge transfer and element diffusion. The characterization content on the spatial scale could include the distance range of element diffusion, the range of space charge layer, and the size of volume expansion.

(2)

Quantitative estimation of interfacial charge transfer kinetics is required. Poor interfacial dynamics arise from weak physical contact, space charge layers, ion diffusion, and other factors. However, the impact of these parameters is currently in the qualitative description stage, it is still difficult to play a guiding role in the proposal of improvement measures.

(3)

The performance of all-solid-state lithium metal batteries needs to be adjusted according to different needs. The current strategies to solve the interface problem are constantly emerging, but it is difficult to have a battery that can meet all needs. The interface modification strategy needs to be adjusted according to the actual needs, such as high-energy density or high-power density.

Most studies are performed in a small-scale production in the lab. When scaling up to the industrial application, energy density and cost should be considered. It is necessary to optimize and modify the thickness of the solid electrolyte in the battery to meet the total package weight criterion and reduce the practical in-plane resistance of the system. The interphase modification strategies also should turn into a feasible and green way. We believe that all-solid-state lithium metal batteries can still gain high attention due to their high energy density. It is also believed that through unremitting efforts and cooperation, the interface problem can be greatly improved, and the all-solid-state lithium metal battery can take another step toward practical application.