Li‐containing alloys beneficial for stabilizing lithium anode: A review

Due to the soaring growth of electric vehicles and grid‐scale energy storage, high‐safety and high‐energy density battery storage systems are urgently needed. Lithium metal anodes, which possess the highest theoretical specific capacity (3860 mA h g−1) and the lowest electrochemical potential (−3.04 V vs standard hydrogen electrode) among anode materials, are regarded as the ultimate choice for high‐energy density batteries. However, its safety problems as well as the low Coulombic efficiency during the Li plating and stripping process significantly limit the commercialization of lithium metal batteries. Recently, Li‐containing alloys have demonstrated vital roles in inhibiting lithium dendrite growth, controlling interfacial reactions and enhancing the Coulombic efficiency (CE) as well as cycle life. Accordingly, in this perspective, the progresses of lithium alloys for robust, stable, and dendrite free anodes for rechargeable lithium metal batteries are summarized. The challenges and future research focus of lithium‐containing alloys in lithium metal batteries are also discussed.


INTRODUCTION
Metallic lithium as an anode in a rechargeable battery was first explored by Whittingham in 1970s at Exxon, and its commercialization was realized by Moli Energy in the late 1980s. [1][2][3] Nevertheless, frequent accidents, including fires caused by dendrite formation, brought serious safety issues to the public eye, which ultimately lead Moli Energy to recall all the cells. Hence, the first attempt at the commercialization of lithium-metal anode rechargeable batteries ended in failure. Then, Sony developed graphite anodes to replace the metallic Li anode and, when paired with the LiCoO 2 cathode, successfully built reliable Li-ion cells that have been widely used until now. 2,[4][5][6] However, with the booming growth in consumer electronic devices and electric vehicles, lithium ion batteries with higher energy density are urgently required. [7][8][9][10] As a result, both anodes and cathodes with higher energy density would be beneficial. The theoretical specific capacity of graphite anode is only 372 mA h g −1 , which does not match well with the high specific capacity cathode. Therefore, in recent decades, rechargeable lithium metal batteries using a Li anode has attracted unprecedented attention, as it has the highest theoretical capacity (3860 mA h g −1 , or 2061 mA h cm −3 ) and lowest electrochemical potential (−3.04 V vs. the standard hydrogen electrode) among all known anode materials. 2,[11][12][13][14][15] Simultaneously, the rapid development of cathodes alternative to conventional intercalation cathodes used in state-of-the-art Li-ion batteries, that is, the sulfur (S) cathode (with energy density of ≈2600 Wh kg −1 ) for the Li-S battery and the oxygen cathode (with energy density of ≈11,140 Wh kg −1 ) for the Li-air battery, 16,17 has increased the urgency of rechargeable Li metal batteries. However, the safety concerns associated with the lithium metal anode is not the only limitation facing these batteries, which also suffer from poor cycle stability and low Coulombic efficiency. Accordingly, their practical application is hindered. Tremendous efforts have been devoted to solving the notorious lithium dendrite problem by employing advanced electrolytes, separators, and novel electrode materials/structures. With regards to the electrolyte, increasing the concentrations of lithium salts and adding inorganic or organic additives into the electrolyte not only helps to stabilize the spontaneous solid electrolyte interphase (SEI) films, which reduce side reactions, but could also control the nucleation and growth of metallic lithium thus enhancing the stability of lithium anodes during the stripping and plating processes. 15,[18][19][20][21] Recently, our group employed octaphenyl polyoxyethylene as an electrolyte additive to enable a stable complex layer on the surface of the lithium anode. This surface layer not only promoted uniform lithium deposition, but also facilitated the formation of a robust SEI film. 22 Conversely, developing new types of modified separators is expected to physically suppress the growth of lithium dendrites. For instance, fabricating a stable tissue-directed/reinforced bifunctional separator/protection film (TBF) on the surface of the Li anode in situ effectively protected the lithium from corrosion caused by O 2 , discharge intermediates, H 2 O, and the electrolyte, which reduced the morphology change of the surface of lithium anode. 23 In terms of novel electrode materials/structures, infusion of the Li metal into a carbon framework or the synthesis of lithium-based composites [24][25][26] allows the growth of lithium dendrites to be controlled, but also solves the issue of infinite volume change of lithium-based electrodes. 18,27 For instance, Koratkar and co-workers described defect-induced plating of metallic lithium within the interior of a porous graphene network. 26 The network acted as seed points that initiated plating of lithium metal and prevented dendritic growth.
Based on the aforementioned considerations, a stable operation of Li metal anodes achieved through the inhibition of lithium dendrite growth and control of the side reactions/volume change is critical for next-generation battery technologies. Thus, in this short perspective article, we briefly review the Li-containing alloys reported in the literatures that stabilize Li metal anodes and propose a few new suggestions for protecting the Li metal by reasonably combining the lithium-containing alloys with other strategies.

FAILURE MECHANISMS OF METALLIC LITHIUM ANODES
Before introducing the practical applications of lithium alloy in present metal lithium batteries, it is necessary to understand the failure mechanisms of the lithium anode. Actually, during the past 10 years, deep and fundamental understanding on the failure mechanisms of the Li metal anode has been extensively discussed. No matter whether it is in LIBs, Li-S, Li-O 2 , or solid-state electrolyte Li-metal batteries, these issues will generally cause the failure of the Li anode 28-30 : 1. Uncontrolled interfacial reactions: the uncontrolled interfacial reactions, also called the side reactions, happen on the Li-metal surface with the organic electrolytes. As the Li-metal in the organic electrolyte is highly reactive, corrosive reactions easily proceed, which deplete the electrolytes and generate thick SEI layers, leading to increasing resistance and low Coulombic efficiency. 28 2. Uncontrollable Li dendrite growth: the Li dendrites could penetrate through the separator and cause cell short circuit, resulting in a series of safety concerns. 31 Meanwhile, the dendritic Li could also produce so-called "dead Li" via electrical detachment of Li from the current collector; significantly affecting the cycle life of the Li-metal battery. 31 It is now widely accepted that unstable and inhomogeneous SEI films, inhomogeneous electric fields, and inordinate lithium-ion flux are the origination of the dendrite growth. 2,18 3. Infinite relative volume changes: as Li is "hostless" in nature, during repeated stripping/plating processes, the relative volume change of Li anode is virtually infinite, which will initiate cracks on the SEI, leading high structural instability for the Li anode. 2 As the specific example illustrated in the previous report, 2,32 a single-sided commercial electrode needs to reach an areal capacity of 3 mA h cm −2 , indicating a huge change in thickness of ∼14.6 μm for Li. In terms of future applications, this value could be even higher, providing a formidable challenge on the SEI stability.

HOW DO LI-CONTAINING ALLOYS SOLVE THE PRESENT ISSUES?
The first use of lithium alloys as negative electrodes in commercial batteries that operate at ambient temperatures was in Wood's metal alloys in lithium-conducting button type cells by Matsushita in Japan. Development on these alloys started in 1983, 33,34 and became commercially available somewhat later. 34 Since then, researchers not only used lithium alloys as the anodes to reduce the activity of Li anode and stabilize it, 35 but also developed various kinds of other strategies, that is, constructing rational design of anode structure (3D hosts and 3D current collectors), 24,36 suppressing Li dendrite growth via modification of electrolytes, separators and anode interfacial engineering, 2,37 and so on.

Directly using lithium alloys to replace metallic lithium as anodes
In the lithium-free alloy anodes, such as Sn-Sb, Sn-Co, Ni-Sn alloy, and so on, without pre-stored lithium, the overall energy density is limited by the low-capacity lithium metal oxide cathodes, while the pure lithium metal anode faces high reactivity and uncontrolled dendrite growth. 38 Li-containing alloy anodes inherit the desirable properties of both alloy anodes and pure Li metal anodes. Lithium alloy anodes for rechargeable ambient temperature lithium batteries have been studied since the early 1970. 39,40 During the past 40 years, a great deal of literature have been reported using lithium-containing alloys as the anode materials for lithium ion batteries. 18,29,[31][32][33]37,[41][42][43] From the reported literatures, we can conclude that there are several kinds of methods to synthesize these lithium alloys, including; fusion reaction, electrochemical lithiation/deposition, magnetron sputtering, and ball-milling, as shown in Table 1. Each method has their own advantages and disadvantages, for example, the fusion reaction method and ball-milling methods are easy to operate and the equipment is not expensive. Electrochemical lithiation method is operationally more complex but applicable to nearly any lithium alloys. With the magnetron sputtering method, the components and the thickness of alloys is easily adjusted, but not commonly to each alloy and the equipment is expensive. Lithium alloys can effectively reduce Li nucleation overpotential and decrease interfacial resistance, guiding the formation and growth of non-dendritic Li. 24 These lithium alloy anodes can mainly divided into two categories: binary lithium alloys and ternary lithium alloys, with the different lithium alloy anodes having their own advantages and disadvantages.
For example, Li-Si alloy anodes exhibit multiple attractive properties: (i) fully lithiated Li x Si alloy has a sufficiently low potential of around 10 mV versus Li/Li + to prelithiate all types of anodes including graphite, Si, Ge, and Sn 44 ; (ii) due to the super-high capacity of Si (4200 mA h g −1 ), Li x Si alloy anode could also illustrate high specific capacity even at a small percentage of pre-stored lithium, that is, Li 4.4 Si shows a capacity of 2000 mA h g −1 . 44 Most of the Li-Si electrodes were obtained by electrochemical lithiation of Si-based electrodes, [45][46][47] which is very difficult to employ in practical applications. 46 But recently, two other methods to synthesize the Li-Si alloy have been reported: one is a pressing plus heat-treatment process as shown in Figure 1A, 46 another one is ball milling method ( Figure 1B) in argon atmosphere. [48][49][50] Apart from the Li x Si alloy, the lithium alloy anodes of other IVA group elements has also been widely reported, such as Li-Sn and Li-Ge alloy anodes. 47,[52][53][54][55][56][57]61 Similar to the Li x Si alloy, the Li-Sn and Li-Ge alloy anodes also exhibit relatively high specific capacity. 47,57 While the Li-Sn alloy anode also shows its unique merits, including the fast interdiffusion of Li in Sn and the <500 mV separation between Li-Sn alloy formation and Li plating. Archer's group reported Li-Sn hybrid battery anodes created by depositing an electrochemically active Sn on a reactive Li metal electrode by a facile ion-exchange chemistry as shown in Figure 1C, leading to very high exchange currents and stable long-term performance. 55   anodes were shown to be stable at 3 mA cm −2 and 3 mA h cm −2 . While in contrast to Si, Ge has the benefit of forming a minimal amount of native oxide in its outermost layer and the diffusivity of lithium in Ge is 400 times greater than that of lithium in Si at room temperature, but due to the high cost of Ge, Li-Ge alloy anodes have not gained much attention. 33 Li-Al alloy anode showed higher stability in the air, carbonate-based electrolyte and the electrolyte with LiNO 3 additive. 62,63 Also, Al alloying with Li exhibits much smaller volume change (≈96%) compared with other alloy anodes, such as Li-Si (320%) and Li-Sn (260%) alloy anodes. 64 In addition, Li diffusion coefficient in Li-Al alloy (6.0 × 10 −10 Ω −1 ⋅cm −1 ) exceeds that in bulk Li metal (5.69 × 10 −11 Ω −1 ⋅cm −1 ). 65 LiB alloy is widely used as an anode in thermal batteries, which can be regarded as free metal lithium metal filled in the fibrillar network framework of Li/B compound (Li 7 B 6 ), 59,60 such a porous structure can increase the specific surface area and ensure the even distribution of Li ions. The discharge potential of Li 7 B 6 is over 0.4 V (vs. Li/Li + ), thus when Li-B alloy is used in metal lithium batteries, its free metal lithium participates in electrochemical reactions preferentially. 59 Additionally, Li 7 B 6 has a good conductivity (1.43 × 10 3 Ω −1 cm −1 ) and a high Li ion diffusion rate compared to metallic lithium. 59 Thus, in 2013, Yang's group first investigated Li-B alloy as anodes for lithium/sulfur batteries. 59 It is because of the above advantages, Li-B alloy has better behavior in restraining the formation of dendritic lithium, reducing the interface impedance of electrode, and improving the cycle performance of the battery. In Li-In alloy electrodes, Archer's group found the interfacial resistance of the resultant Li-In alloy electrode was significant lower than that of the pristine Li metal, which allowed Li ions to diffuse along the surface to form uniform deposition on the hybrid electrode. 66 As a result of the enhanced interfacial ion transport mechanism, compact and uniform electrodeposition for the Li-In alloy anode at long time scales has been realized. The Li-In hybrid anodes in full cells employing high-loading commercial cathodes (lithium titanium oxide (LTO) and nickel manganese cobalt oxide) showed that the electrodes can be cycled stably for over 250 cycles with close to 90% capacity retention. Recently, Adelhelm investigated the different In/Li ratio on the performances of Li-In anode. 67 The right In/Li ratio, that is, 1.27:1, enabled stable lithium insertion/deinsertion in symmetrical cells for at least 100 cycles; while too much lithium in the electrode lead to a drop in redox potential combined with a rapid build-up of interface resistance. Li 3 Sb and Li 3 Bi alloy anodes have also been investigated in metal lithium batteries as early as the 1970s, 68,69 however, there are not too many group VA lithium alloys used directly as anodes in metal lithium batteries. 42,70,94,95 Mainly, that is because of the higher toxicity of some VA elements, such as Sb and As, and the smaller gravimetric capacity of Bi. 33 In addition, the synthesis conditions of VA lithium alloys are highly demanding. Take for example the Li-Sb alloy, apart from complex prelithiation with Sb, another method involves the electrolysis of a molten LiCl-KCl eutectic mixture with a liquid antimony cathode at high-temperature. 70 In contrast, the Bi/Sb-based nanocomposites and Bi/Sb-based intermetallics could be easily produced at a large-scale, while also demonstrating good electrochemical performances when used as anode materials for LIBs. 33 Therefore, during the past 40 years, there has been no significant development of the group VA lithium alloys as the anodes in metal lithium batteries.
Li-Na alloy can supply Li + on stripping and thus ensure the electrostatic shield effect of Li + . 73 Also, the Li-Na alloy would not sacrifice the specific capacity of the anode because Li and Na metals exhibit similar reaction activities as well as electrolyte compatibility of Li + and Na + . 73 However, developing a Li-Na alloy anode might be difficult because of volume expansion, 71-73 which causes SEI damage, large internal resistance and low Coulombic efficiency. Recently, Zhang's group reported a Li-Na alloy anode used in Li-O 2 batteries. 73 By optimizing the Na/Li value of the alloy, a dendrite-suppressed, oxidation-resistant, and crack-free Li-Na alloy anode could be obtained, 73 thus realizing an alloy anode with a long cycle life.
The Li-Mg alloy is advantageous because of the generally lower reactivity of Li (or relatively low Li activity), the large solid solution range, the mechanical integrity of Mg framework, and a relatively large diffusion coefficient of Li in Mg (∼10 −7 Ω −1 cm −1 for the Li-Mg alloy produced by vapor deposition). 74,[76][77][78]96 Mg alloying can increase lithium utilization when no external pressure is applied, while pure lithium metal is superior for setups that allow stack pressures in the MPa range. 74 The appropriate amount Mg, that is, 10 at%, introduced into Li metal anode can also effectively prevent contact loss. 74 Due to these various advantages, recently, Gao's group reported Li-Mg alloy as an anode for Li-S batteries. 75 Compared to the metallic Li anode, the Li-Mg alloy showed remarkably improved stability at the surface and in the bulk during cycling as shown in Figure 1D,E. They also found after Li stripping, a conducting Li-poor Li-Mg alloy matrix was formed, facilitating subsequent plating and diffusion of Li ions.
Even the theoretical capacity of Li-Zn alloy is not as high as Li-Si, Li-Sn, Li-Ge, Li-Sb alloys, and so on, but the volume expansion of Li-Zn alloys is not obvious when used in Li storage. 81 Chen et al reported a Li-Zn alloy synthesized by depositing Li on the Zn substrate precursor at a constant current density of 0.05 mA cm −2 until the potential reached 0 V (vs. Li/Li + ). 81 The efficiency of Li deposition/stripping on the Li-Zn alloy anode remained high at 96.7% after 400 cycles at a current density of 0.1 mA cm −2 and 250 cycles at the current density of 0.2 mA cm −2 .
Different from the Si and Sn that experienced reconstitution reaction with lithium to form alloys, Au and Ag, as two typical noble metals, involve solid-solution reactions with Li to form LiAu x and LiAg x alloys. 51,82 The solid-solution reaction involves much less structural changes than their counterparts (e.g. Si and Sn) in the lithiation-delithiation process, and can therefore take place with a low charge-discharge voltage hysteresis at a potential that is very close to that of the Li/Li + redox couple and eliminate the nucleation barriers. 82 For example, in 2016, Cui's group has found that Au, Ag, Zn, and Mg have good solubility in Li, and once fully lithiated, exhibited zero overpotential during deposition of Li as shown in Figure 2A. 51 The materials Al and Pt have relatively small solubility in Li metal and show small but observable overpotential for Li nucleation (5 mV for Al, 8 mV for Pt); Materials showing no solubility (Cu, Ni, C, Sn, Si) in lithium were also tested. As shown in Figure 2B, all five materials show a clear overpotential for Li metal nucleation. According to these vital findings, they designed Au nano particles distributed inside the hollow carbon spheres to selectively nucleate and grow Li metal inside carbon nanoshells during electrochemical deposition, as shown in Figure 2C.
Even though the binary Li-containing alloy systems, such as Li-Al, Li-Si, Li-Sn, Li-Ge, Li-Sb, and so on, possessed a larger capacity than commercialized graphite, these binary lithium alloy electrodes usually display rather poor cycle performance due to catastrophic structural changes with large volume expansion. 87 To overcome this poor reversibility in such intermetallic lithium alloy electrodes, ternary Li-containing alloy compounds have been investigated. the Li-Mg binary alloy with good solid solution over a wide range of composition without phase transformation, provides high specific capacity, and Li-B alloy with porous structure can increase the specific surface area and its very negative potential close to pure lithium (ca. 20 mV vs. Li/Li + ) provides a basic condition for the high energy density. Pan et al, reported Li 2 MgSi as a novel anode for Li-ion batteries. 89 Directly using Li 2 MgSi as an anode material can prevent the dissociation of metallic Mg and/or Li-Mg alloy from Mg 2 Si, and the pre-lithiated Li 2 MgSi is likely to reduce the stress/strain during delithiation/lithiation. In addition, by constructing a ternary alloy that contain inactive materials, such as LiCuSn and LiCuSb alloy, 87,90 the volume variation of active elements could be buffered by inactive medium when active elements are dispersed uniformly in situ or ex situ into the matrix of inactive components at nanoscale, as in the cases of active/inactive composites and intermetallic compounds. 97 Compared to plenty of binary Li-containing alloys anodes, there are few reports on multi-component lithium alloy anodes. 35,[85][86][87][88][89][90][91][92][93]98 One reason is the much lower specific capacity of the ternary lithium alloys, that is, Li 4.4 Ge 0.67 Si 0.33, 86 Li 0.25 CuP, 85 and so on, compared to the binary lithium alloys; however, the most likely reason is that it is rather difficult to prepare high-purity multi-component lithium alloys, 89 as a result, the theoretical capacities are hard to calculate while the electrochemical mechanisms are difficult to fully understand. 87 In order to better understand the advantages and disadvantages of the various Li-containing alloy anodes, the summaries have been made in Figure 3 and Table 1. As can be seen, there are no perfect lithium alloys used as the anodes. But from Figure 3, we can see that lower discharge plateaus and higher lithium ratios in the alloys could contribute to higher theoretical capacities, which give guide us toward the synthesis of optimized lithium-containing alloys as the anode. However, it is impossible to solve the problem simply by replacing lithium metal with lithium alloys. Thus, combining lithium alloys with other materials, such as graphene, polymers, and so on, to overcome the weak links of lithium alloys has been proposed, which we will discuss in the Part 3.  Figure 4A. 38 With the protection of graphene sheets, the large and freestanding Li x M/graphene foils are stable in different air conditions. The representative Li x Si/graphene foil maintained a stable structure and cyclability in half cells (400 cycles with 98% capacity retention), and when paired with high-capacity Li-free V 2 O 5 and sulfur cathodes, stable full-cell cycling could also be achieved. Additionally, the alloy electrodes have a high reduction potential, leading to low energy density. To overcome drastic volume variation during Li insertion/extraction cycles, apart from preparing superfine alloy particles that have small absolute volume variation or constructing ternary alloy that contains an inactive metal to inhibit the great volume expansion, the lithium alloys can also be encapsulated in a flexible and elastic polymer matrix. For example, Cui's group reported a polymer supported Li-Zn alloy structure as shown in Figure 4B. 99 They used the atomic layer deposition (ALD) method to deposit the ZnO on the polymide (PI) fiber. The core-shell PI-ZnO matrix was put into contact with molten Li, ZnO reacted with molten Li to form Li-Zn alloy, and simultaneously extra Li can be drawn into the polymer matrix, affording a Li-coated PI electrode. Thanks to the polymer shell, the lithium alloys crack and pulverization can be alleviated.

Lithium alloy matrices
By confining Li in a 3D structure, the infinite volume change during cycling could be eliminated. 100 Furthermore, the high surface area provided by 3D structures can further lower the localized current density and enable a more stable plating/stripping process. 25 Confining the lithium metal in 3D carbon materials is a common strategy, however there are also some pioneering works which use 3D lithium alloy matrices to host the lithium metal. 25,36,101 In 2014, Zhang's group first employed a 3D lithium alloy (Li 7 B 6 ) fibrous matrix for ultra-stable lithium-sulfur batteries. 36 The 3D nanostructured Li 7 B 6 framework with high surface area and enough volume space, could not only decrease the areal current density, but also adequately accommodate the electrolyte and re-deposited Li to stabilize the concentration of Li ions. By employing this Li@Li 7 B 6 anode, the Li-S batteries could stability cycle to 2000 cycles. Recently, Yan and his co-workers reported a 3D Mg doped LiB skeleton for hosting the metallic lithium and inhibiting the lithium dendrite growth as shown in Figure 5A,B. 101 The 3D LiB skeleton could significantly reduce volume variation during Li electrochemical dissolution/deposition process. Its superior lithiophilic and conductive characteristics could also contribute to the reduction of the local current density and homogenization of incoming Li + flux. More importantly, Yan et al, used the Density Functional Theory calculation to prove that the doping of Mg element to the 3D LiB skeleton could enhance the adsorption energy of Li. And the remaining Li-deficient Li-Mg alloy formed after Li stripping can help connect LiB fibers to stabilize the whole skeleton and lower interfacial resistance, which could effectively inhibit the lithium dendrite growth as shown in Figure 5C,D.
Additionally, constructing 3D lithium alloy-based composites to deposit lithium has been developed. For example, our group developed a 3D silver nanowire (AgNW) and graphene-based hierarchical host (3D-AGBN). 84 Fast and uniform electron transportation is guaranteed throughout the continuous nanonetwork of high-conductivity AgNWs since the Li ions will be favorably reduced on the surface of AgNWs to form Li-Ag alloys and induce the Li deposition contributing toward a low overpotential. Therefore, the Li deposition can be directed within the entire scaffold ( Figure 6A) to form a smooth Li layer covering the AgNW nanonetwork after 1 mA h cm −2 plating as shown in Figure 6B,C. Yang's group first synthesized 3D metal-organic frameworks derived carbon with abundant Zn clusters and, after infusion with lithium, the highly active nano Zn clusters could react with lithium for form the Li-Zn alloy as shown in Figure 6D. 83 Therefore, a 3D conductive carbon supported Li-Zn alloy structure was realized, which enabled the homogeneously distributed electric field and Li ion flux. The Li-Zn alloy layer rendered the matrix with good affinity toward lithium, and acted as a buffer layer for the following Li plating, effectively eliminating nucleation barriers. Zhang and his co-workers proposed another kind of 3D carbon supported Li-Zn alloy structure by infiltrating lithium into carbon cloth decorated with zinc oxide arrays as shown in Figure 6E. 102 The X-ray diffraction (XRD) spectrum as shown in Figure 6F confirmed the formation of LiZn alloy in Li-CC@ZnO. Such a LiZn alloy formation in the Li plating process further induced dendrite-free Li deposition. As a result, a low overpotential of ∼243 mV over 350 cycles at a high current density of 10 mA cm −2 was achieved (shown in Figure 6G), compared to the greatly fluctuating voltage and fast short circuit in the cell using bare Li metal. In addition to using the 3D alloy matrices to host metallic lithium, researchers have recently paid attention to various 3D Cu-alloy current collectors. After the 3D Cu current collector is modified by active metals, such as Al, 103 Sn, 100 Zn, 104 Au, 24 Ga-based liquid metal alloys, 105 and so on, it can form Li alloys after in-situ electrochemical deposition Li which then serve as seeds to induce the formation of non-dendritic Li via lowered Li nucleation overpotential and interfacial energy. For example, Guo's group reported a 3D Cu fibers grown on the Cu foil coated with a thin Al layer, 103 forming a 3D Cu@Al hybrid structure; after an initial discharge process above 0 V versus Li + /Li, the thin Al layer reacted with Li to generate a binary Li-Al alloy phase, which functioned as the lithiophilic sites. The Li nucleation and growth promoted by the Li-Al alloy layer realize a dendrite-free Li anode as shown in Figure 7A,B. In addition, by controlling the discharge process, the active Li stored in the form of Li-Al alloy could act as a Li resource to compensate for any irreversible Li loss during cycling. Abruña et al, reported a Zn coated Cu foil current collector. 104 During the Li plating process, the Li-Zn alloy buffer layer on the Cu foil surface would regulate the nucleation and growth of Li metal as shown in Figure 7C. As expected, the Coulombic efficiencies of plating/stripping were enhanced and nucleation overpotential was greatly reduced. They also investigated the electrochemical performances of high-energy-density Li-S full cells by using the Zn coated Cu foils as Li metal current collectors. As shown in Figure 7D, the cycling stability of these two kinds of batteries was significantly enhanced.
Trapping Li into a three-dimensional (3D) conductive host to construct 3D-Li is an effective strategy to suppress the growth of Li dendrites. However, the increased contact area between 3D-Li and electrolyte unavoidably induces more F I G U R E 8 SEM images for (A, C) Li electrode and (B, D) Li-Na alloy electrode with a Na/Li molar ratio of 6 after five stripping/plating cycles in 0.5 M NaCF 3 SO 3 /TEGDME or NaCF 3 SO 3 /DOL/TEGDME electrolyte. Source: Reproduced from Reference 73 by permission from Springer Nature side reactions to further deteriorate the electrochemical performance of lithium metal batteries. In order to keep the advantages of 3D lithium-alloy matrix as well as reduce the side effects, Qu's group recently constructed a paraffin wax (PW) coating Au-graphene/Cu foam current collector ( Figure 7E,F). 24 Such a unique structure was able to effectively avoid the growth of Li dendrites and formation of "dead Li" during the Li plating process, which could be attributed to these merits: (i) Au could react with Li to form Li x Au alloys that lowered Li nucleation overpotential and interfacial energy to effectively inhibit the formation of dendritic Li; (ii) 3D lithiophilic graphene decreased the local current density and enabled the homogeneous growth of 3D Li; (iii) Cu skeleton not only afforded interconnected pores to accommodate the volume variation of 3D-Li, but also served as a robust support to avoid the collapse of overall electrode especially during fast plating/stripping processes; (iv) the PW protection layer confined Li during the plating and stripping to mitigate the corrosion of electrolyte and depress the formation of Li dendrites and "dead Li."

Artificial protective layers for lithium alloy anodes
Unfortunately, even when a 3D matrix to host the Li-containing alloys is used, the materials still suffer from poor SEI stability, resulting in unsatisfactory electrochemical performances. 47 Therefore, constructing an artificial protective layer for lithium alloys anodes has been proposed. For example, Cui's groups reported two methods to construct an artificial-SEI layer from LiF to protect Li x Si alloy nanoparticles via reducing 1-fluorodecane and fluoropolymer CYTOP, respectively. 106,107 Ci et al, reported that a Li-O 2 coin cell with the LiAl x anode experienced a high-current pretreatment, 108 as a result, the SEI film (including Al 2 O 3 , LiF, ROCO 2 Li, LiOH, and Li 2 CO 3 ) formed after the pretreatment process facilitated the uniform Li + shuttling during the following Li plating/stripping process and stabilizes the LiAl x anode interface even after hundreds of cycles. The LiAl x anode in lithium oxygen batteries could increase cycling to 667 cycles under a fixed capacity of 1000 mA h g −1 compared to 17 cycles of LiAl x anode without pre-treatment. Recently, Zhang's group used a Li-Na alloy and 1,3-dioxolane (DOL) as anode and additive, respectively, to control dendrite growth and buffer the volume expansion of the alloy anode. 73 The DOL additive could react with Li-Na alloy in situ to form a robust and flexible passivation film that suppresses dendrite growth, buffers alloy anode volume expansion, and prevents cracking. As shown in Figure 8A-D, only the Li-Na alloy electrode with DOL additive can effectively suppress dendrite growth and will not crack after cycling. While a more common strategy to form an artificial SEI layer is using Li salts to react with lithium alloys in situ. It is easy to realize thin lithium alloy layers on the surface of Li metal by the reduction of metal salts in situ as shown in Equations (1) 109 : More importantly, the metal layer could immediately undergo a reaction with the underlying lithium to form a Li alloy as shown in Equation (2) 109 : For instance, in 2017, Nazar's group reported Li-rich composite alloy films synthesized in situ on lithium by the above two reactions. 109 Due to the lithium alloy and lithium salt composites together forming an artificial SEI layer to protect the lithium anode, it can sustain electrodeposition over 700 cycles (1400 h) of repeated plating/stripping at a practical current density of 2 mA cm −2 . Also, a 1500 cycle-life cell was realized when paired with a Li 4 Ti 5 O 12 positive electrode. Since then, various lithium alloy-lithium salt artificial SEI layers have been proposed, which are summarized in Table 2. It can be observed that metal halides are the most popular candidates to react with lithium and form inorganic composite artificial SEI layers.
However, these inorganic films are often thin, brittle, and prone to cracking due to the volume effect of the Li anode during extended cycling. 31 Therefore, researchers developed a series of flexible polymer SEI layers to prevent the formation of cracks or pinholes in the SEI layer. [118][119][120][121][122][123][124][125][126][127] With the polymer SEI layers, uniform Li deposition could be achieved at high current densities. The dendrite-free deposition of lithium metal leads to extended electrochemical cycling and stable coulombie efficiency. Recently, due to the unique advantages of the polymer SEI, polymer-lithium alloy hybrid artificial SEI layers 128,129 have also been developed. For example, Xie et al reported a poly(tetramethylene ether glycol) (PTMEG)-Li/Sn alloy hybrid layer on the Li metal surface. 128 The possible mechanism for forming the artificial SEI layer is shown in Figure 9A. Such a hybrid artificial layer not only provided a fast lithium-ion transport path on account of the Li/Sn alloy with ample Li vacancies, but also has a strong affinity for Li, since PTMEG is similar to poly(ethylene oxide) with a large number of C-O bonds. As a result, after only 20 cycles, many dendrites and dead Li were observed for the pristine Li, whereas the treated Li presents a smooth and compact deposition of Li + as shown in Figure 9B,C.
Moreover, researchers have also developed various kinds of lithium-containing alloy artificial protective layers. The protective lithium alloy layers are generally thin films formed on the Li metal surface via chemical/electrochemical pre-treatment, 73,108-117,128,129 mechanical pressing, 130,131 magnetron sputtering, 62 and so on, which are allow ionic

Electrolytes for lithium alloy anodes
Using liquid electrolytes for the lithium alloy anodes is mainstream in the past decades, because the liquid electrolytes have various advantages, that is, high ionic conductivity, and easily tunable electrolyte components, and so on. According to the massive literature reports, 45,73,[106][107][108][109][110][111][112][113][114][115][116][117][128][129][130][131][132] the most common liquid electrolyte systems for the lithium alloy anodes are restricted to two systems: ethers and carbonic esters. The most common representatives of these two systems are lithium bis(trifluoromethanesulfonyl)-imide dissolved in DOL/1,2-dimethoxyethane and LiPF 6 dissolved in ethylene carbonate (EC)/diethyl carbonate, respectively. On this basis researchers investigated other additives or changed the composition of the solvent to further enhance the lithium anode electrochemical performances. For example, as illustrated in Table 2, by adding the lithium halides salts into the basic electrolyte system to construct the artificial SEI film, the lithium alloy anodes were protected well and the lithium dendrites were effectively inhibited. However, the disadvantages of Li batteries based on the liquid electrolyte are also obvious, such as easily flammability and explosivity. Therefore, recently, all-solid-state Li batteries have received increased attention due to their incombustibility, dendrite blocking ability, and stability over large potential windows. 133 However, the way to realize high safety, high power density, and high energy density for all-solid-state Li batteries is still a challenge.
Recently, Hu's group creatively combined the lithium alloy anode with garnet electrolyte, in which the component has garnet crystal structure and high Li-ion conductivity, to solve the problems of all-solid-state Li batteries. [133][134][135][136] For example, in 2017, they first proposed a new methodology for reducing the garnet/Li-metal interfacial resistance by forming a Li-Ge alloy. First a thin Ge layer (20 nm) was evaporated onto the garnet pellet (Li 7 L a3 Zr 2 O 12 ) by an electron beam evaporation system, then a small piece of Li metal disc was put on the Ge-modified garnet. After heating, the Li-Ge alloy anode was be formed. As the effective contact area between Ge-modified-garnet and Li-metal anode increased more than eight times compared to bare garnet due to the alloying reaction between Li and Ge as shown in Figure 10A, the Li/Ge-modified garnet delivered a small interfacial resistance of 115 Ω cm 2 , far smaller than the Li/bare garnet (≈900 Ω cm 2 ). When it was assembled with LiFePO 4 to form an all-solid-state Li battery as shown in Figure 10B, the cycling performance of this all-solid-state Li battery has been improved significantly, even comparable with the full batteries using the liquid electrolyte as shown in Figure 10C. Even the Coulombic efficiency was better than the full cells using liquid electrolyte in Figure 10D.
Since then, the Li-Zn, Li-Sn and Li-Mg alloy modified garnet solid electrolytes have been subsequently developed. [133][134][135] Such a facile surface treatment on garnet electrolyte with forming lithium alloys method offered a simple strategy to solve the interface problem in solid-state lithium metal batteries.

Lithium alloy modified separators
A multifunctional separator achieved through coating a thin electronic conductive film on the Li facing side of the conventional polymer separator could contribute to Li dendrite suppression and an improved cycling stability. 137 Recently, Li and  138 The produced Li-Pb alloy armor between the separator and Li anode not only allowed a uniform electric field across the interface but also mitigated the Li metal nucleation, and therefore suppressed the dendrite growth during plating. As a result, the Li/Li symmetric cells and LiFePO 4 /Li cells with these PZT-pretreated PP separators exhibit significantly improved Coulombic efficiency and cycling life as shown in Figure 11A,B. Until now, directly employing lithium alloys to modify the separators to enhance the batteries performances is rarely reported, mainly because of the following reasons: (i) the modification process requires high precision; (ii) the complex modification process will easily cause the separators to fracture and make the batteries short circuit; (iii) the energy density of the batteries will be sacrificed by using the lithium alloy modifications.

PERSPECTIVE ON LI-CONTAINING ALLOYS FOR HIGH-SAFETY AND HIGH-ENERGY DENSITY BATTERIES
The introduction of Li-containing alloys seems to be a fundamental tool to resolve the safety hazards caused by lithium dendrites. But there are still many unanswered questions, for instance, the protective mechanisms for different alloys are not clearly understood 96 ; the Li-containing alloy materials can effectively improve the state of first-layer Li deposition, however, the substrate will revert to pure Li as deposition continues 54 ; the Li-containing alloy anodes still face the great volume change and serious side reactions during the striping/plating process, and so on. Therefore, these scientific problems are still urgently awaiting our solutions.
The challenges faced by lithium metal anodes cannot simply be solved by using a lithium metal alloy anode, lithium alloy artificial SEI film, or lithium alloy to modify the current collector, separator or electrolyte. Only by combining multiple strategies can the problems caused by lithium anode be completely solved. Here we propose several suggestions (as shown in Figure 12) for the future research of lithium-containing alloys employed in Li metal batteries: 1. As the binary Li-containing alloy either has high reaction activity (i.e., Li-Na), great volume change (i.e., Li-Si, Li-Sn), low energy density (i.e., Li-Bi, Li-Zn), or high cost (i.e., Li-Ag, Li-Cu) more consideration can be directed into ternary/multicomponent lithium alloys, which can complement each other through multiple components. However, the presence of additional metals that are not directly involved in electrochemical reaction results in additional weight and volume and thus cause the specific energy density to be reduced compared to using the pure lithium or binary lithium alloy anodes. 139 Another disadvantage is that a substantial change in specific volume upon charging and discharging the alloy electrode reactants still exists, and can lead to a loss of electrical contact, and thus capacity loss. 139 In addition, the complexity and cost of preparing ternary/multicomponent lithium alloys also need to be considered.

F I G U R E 12
The potential solving strategies toward the future high-safety and high-energy-density metal batteries based on the lithium alloys anodes 2. Considering the substantial volume change exists in lithium-containing alloy anodes, constructing nanostructures to host lithium alloys or preparing lithium alloy-based composites is also a good choice. 140,141 By encapsulating the lithium alloy into special nanostructures, that is, 3D graphene, 142 carbon nanotube (CNT), 141 and so on, or using polymer coatings, 24 the volume change of lithium alloy has been effectively eliminated. Due to these nanostructured hosts or extra components, Li-containing alloy anodes will be more stable in the organic electrolyte and lithium dendrites would also be effectively inhibited. 3. Constructing an artificial protection/SEI layer on the lithium alloy anode, that is, recently Won II Cho et al, reported a Li-Al anode protected by a Langmuir-Blodgett artificial SEI composed of MoS 2 . 143 Such a MoS 2 artificial SEI layer exhibited a combination of a high Li binding energy, molecular smoothness, and low barrier to Li adatom diffusion, which favors efficient binding of Li and transport away from the electrode/electrolyte interface as well as favors stable and reversible Li migration of the MoS 2 coated Li-Al anode. As a result, the MoS 2 coated Li-Al alloy anode exhibited highly reversible and stable Li migration during recharging of the cells compared to the Li-Al alloy anode without the MoS 2 coating. 4. Developing suitable electrolytes and separators for lithium-containing alloy anodes. Modification of electrolyte or separators with lithium alloys will limit their application in lithium metal batteries, as electrolyte components have significant influences on the electrochemical performances of electrodes, no matter anode or cathode. For example, the absence of volatile or flammable compounds is expected to make solid electrolytes safer than their liquid counterparts at elevated temperatures. 11 Additives can decompose, polymerize or adsorb on the Li surface, modifying the physico-chemical properties of the SEI and therefore regulate the current distribution during Li deposition. 15,22,144 Solvents, that is, ionic liquids, show an exciting role in improving the low temperature performances of batteries. 145 Lithium salts could not only stabilize the spontaneous SEI films, but also control the nucleation and growth of metallic lithium, thus enhancing the stabilities of lithium anodes during the stripping and plating processes. 20,146 While separators serve as a physical barrier between electrodes, traditional polyolefin-based separators easily suffer a "shut-down" problem when penetrated by lithium dendrites or exposed to overheating and/or overcharging. 11 A suitable separator for lithium-containing alloy anodes should have good thermal stability and function to suppress the lithium dendrites. Additionally, a suitable separator may also have the function to inhibit the polysulfide shuttling in Li-S batteries and Li-air batteries. Therefore, a better strategy is combining the lithium alloy anodes with various functional electrolytes and separators rather than use it to modify the electrolyte or separators.
Finally, we want to emphasize that more efforts should be devoted to developing the Li-containing alloys anodes suitable for Li-S batteries and Li-air batteries, as first, the energy density of LIBs is limited, which could not meet the future electric vehicles and grid energy storage markets; second, due to the higher reduction of polysulfide anions in Li-S batteries and higher oxidation of superoxide radical anions in Li-air batteries, some of Li-containing alloy anodes may be used stably in LIBs but are not stable in Li-S and Li-air batteries. Therefore, Li metal protection in Li-S and Li-air batteries is more complicated compared with lower energy density LIBs.
Even though there are still of challenges for the lithium metal batteries, they are regarded as the most promising solution for long-range and low-cost electrical vehicles. 147 We believe that through our unremitting efforts, the Li-containing alloy anodes will show great potential in the high-safety and high-energy-density lithium metal battery systems.