Concentrated electrolytes for rechargeable lithium metal batteries

Traditional lithium-ion batteries with graphite anodes have gradually been limited by the glass ceiling of energy density. As a result, lithium metal batteries (LMBs), regarded as the ideal alternative, have attracted considerable attention. However, lithium is highly reactive and susceptible to most electrolytes, resulting in poor cycle performance. In addition, lithium grows Li dendrites during charging, adversely affecting the safety of LMBs. Therefore, LMBs are more sensitive to the chemical composition of electrolytes and their relative ratios (concentrations). Recently, concentrated electrolytes have been widely demonstrated to be friendly to lithium metal anodes (LMAs). This review focuses on the progress of concentrated electrolytes in LMBs, including the solvation structure varying with concentration, unique functions in stabilizing the LMA, and their interfacial chemistry with LMA.


Introduction
As the third element of the periodic table, Li metal has an extremely low density (0.534 g cm −3 ) and a high specific capacity (3860 mAh g −1 ), about 10 times that of graphite anodes, combined with its low electrochemical potential (−3.040 V vs. SHE) [1][2][3][4]. Therefore, Li is called the 'holy grail' in the field of energy storage and has high application prospects [5]. However, at the same time, LMBs face several application challenges. Firstly, compared with traditional graphite anodes, * Author to whom any correspondence should be addressed.
Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. lithium has higher reactivity to electrolytes and more severe side reactions, resulting in lower Coulombic efficiency (CE). Secondly, during cycling, lithium has a much more considerable volume change than that of graphite, making a solid electrolyte interphase (SEI) form on the surface of the Li anode, repeatedly generated and broken during cycles. This can lead to continuous consumption or even depletion of electrolytes, deteriorating the lifecycle [6,7]. Thirdly, lithium metal anodes (LMAs) suffer from the growth of lithium dendrites during the cycling process, piercing the separator, which causes a short circuit and may lead to a fire [8][9][10]. In addition, making sure the high energy density, high-voltage cathodes, and high capacity cathodes match with Li metal anode brings more challenges for electrolytes, which have to meet a wide electrolyte window and robust chemistry interphases (CEI and SEI) [11].
In commercial lithium-ion batteries (LIBs), the salt concentration of electrolytes is usually fixed in the range of

Future perspectives
Lithium metal batteries (LMBs) are promising for high energy density batteries. The emerging concentrated electrolytes have many positive functions, including favorable for forming salt-derived inorganic rich interphases, inhibiting cathode dissolution, suppressing Al current collector corrosion, and low flammability and high thermal stability, beneficial for the electrochemical performance and safety of LMBs.
1.0 M-1.5 M, at which the electrolytes have the highest ionic conductivity, benefiting ionic transportation in the solution [12,13]. However, commercial carbonate-based electrolytes with such low concentrations are not stable with Li anodes. For example, 1 M LiPF 6 in EC/EMC (1:1 wt%) + 2% VC has high CE and good cycling stability in LIBs, but due to the high reducibility of lithium, it is prone to side reactions with Li metal anode, generating unstable SEI, which cannot form robust passivation film, resulting in the capacity decaying rapidly in a short period [14]. Therefore, electrolytes need to be further optimized for LMB applicationss.
In recent years, researchers have found that when the salt concentration in electrolytes rises to a certain threshold, the chemical structure of the solution changes drastically, the free solvent molecules and solvent-separated ion pair (SSIP) decrease, and the concentration of contact ion pair (CIP) and aggregate (AGG) increases; these are called high concentration electrolytes [15][16][17]. Therefore, compared with traditional electrolytes, concentrated electrolytes have excellent physical properties, such as high Li + transference number [18,19], high thermal stability, fast electrode reaction, and resistance to Al current collector corrosion. In addition, emerging concentrated electrolytes have been proven effective in improving the electrochemical performance of LMBs; the internal solvation structure of electrolytes varied with the increasing salt concentration, improving a series of properties, such as its reduction stability with LMA and oxidation stability with high-voltage cathodes. Therefore, compared with traditional dilute electrolytes, the concentrated electrolyte is more suitable for LMBs.

Solvation structure of concentrated electrolytes
In 2013, Suo et al first proposed the 'solvent-in-salt' concentrated electrolyte and applied it to LMBs (figures 1(a) and (b)) [20]. Compared with the traditional low-concentration 'saltin-solvent' electrolytes (also known as dilute electrolytes), when the weight ratio and volume ratio of salt/solvent are >1 (corresponding to the yellow region (A), blue region (B) and green region (D) in figure 1(b)), the physical and chemical properties of electrolytes change significantly. This phenomenon occurs because the solvation structures of concentrated electrolytes and dilute electrolytes are different [21]. This review starts from the micro perspective, using the salt/ solvent molar ratio to divide concentrated electrolytes and dilute electrolytes [22]. As shown in figure 1(c), the liquid electrolyte contains cations, anions, and solvent molecules that form three solvates, including SSIPs formed by the coordination of Li + and solvent molecules, CIP and AGG formed by Li + and anions. Among them, the anions in SSIP do not enter the Li + inner solvation sheath and have no direct interaction with Li + . CIP is an Li ion and anion tightly bound, and AGG is formed by the tight combination of one anion and two or more Li ions. The anions and solvent molecules are Lewis bases, and Li + is a Lewis acid. As a competitive reaction, anions and solvent molecules interact with Li ions to form a solvation sheath [21,23]. When the salt concentration is low, Li ion coordinates with solvent molecules in the electrolyte to form a double-layer solvation sheath structure due to the strong positive charge. The primary and secondary solvation sheath usually consists of quadruple coordination structures formed by Li + and solvent molecules, which are stable in organic solvents [24][25][26]. In such cases, anions are excluded from the solvation sheath, and the electrolyte system mainly consists of SSIP and free solvent molecules called the 'salt-in-solvent' structure, known as the dilute electrolyte [27]. When the concentration increases until the salt/solvent molar ratio reaches 1:4, the number of solvent molecules generates all primary solvation sheaths. With the increased salt concentration, the number of free solvent molecules decreases rapidly, and the secondary solvation sheath ceases. This electrolyte structure is the 'salt-solvated' structure. When the concentration rises to the salt/solvent molar ratio ≫ 1:4, its physical and chemical properties change considerably with the diluted electrolyte. At this time, the amount of solvent is not enough to fill the Li ion solvation sheath, and the anions are associated with Li ions under the Coulomb interaction and enter the solvation sheath. As a result, the number of SSIPs is significantly reduced, the solvation structure is mainly CIPs and AGGs, and the free solvent molecules almost disappear [28,29]. The electrolyte system takes the 'solvent-in-salt' structure, considered to be concentrated electrolytes.
In addition to using the salt/solvent molar ratio as the concentration unit, two other kinds of concentration are also used, including m and M, where m represents 1 mol salt per liter solvent and M corresponds to 1 mol salt per liter solution. Furthermore, based on the density (ρ solvent ) and molar mass (M solvent ) of the solvent, m and the salt/solvent molar ratio can be converted by equation (1).

Multifunction of concentrated electrolyte in LMBs
Generally, concentrated electrolytes have many positive features, including favorability to form a salt-derived inorganic rich interphase, inhibiting cathode dissolution, suppressing Al current collector corrosion, and low flammability and high thermal stability, beneficial for LMBs (figure 2). In the following section, we will discuss these individually.

Salt-derived inorganic rich interphasial chemistry
Compared to traditional dilute electrolytes, concentrated electrolytes can form a salt-derived inorganic-rich interphase on the anode surface (SEI, solid electrolyte interphase) and cathode (CEI, cathode electrolyte interphase), slow down the side reactions, and improve reduction stability and oxidation stability. The high reduction stability of concentrated electrolytes originates from its unique solution structure and properties. According to Coulomb's law, negatively charged anions cannot enter the inner Helmholtz layer on the anode surface due to the repulsive force, and the chance of a reduction reaction is minimal. The solvent molecules are electrically neutral, have a high concentration of dilute electrolytes, and are enriched in the inner Helmholtz layer. Therefore, the SEI formed by the dilute electrolytes is derived from solvolysis, and it contains many organic phases. This SEI is thick and bulky, which cannot well reduce the side reaction between the electrolyte and lithium metal, and is prone to significant expansion and contraction during the electrochemical cycle, leading to the rupture of the SEI, which in turn leads to dendrite formation and dead lithium growth, ultimately decreasing the CE of the battery [36,37]. With increasing concentration, the concentrated electrolyte no longer contains many free solvent molecules but is replaced by high concentrations of CIPs and AGGs. Anions weaken the Coulomb repulsion by binding to Li + . In addition, concentrated electrolytes can compress the Helmholtz layer on the anode surface. This series of combined actions can force the solvated anions to enter the inner Helmholtz layer and then generate salt-derived SEI on the surface of LMA [38,39]. The inorganic phase in these SEIs increases and often contains a higher content of LiF, which improves the SEI's mechanical strength, density, and flexibility and has an excellent protective effect on the LMA, thereby improving the lifecycle of LMBs [16,40]. For example, Suo et al used 7 m concentrated electrolyte prepared from fluorine-donating salt LiFSI and solvent FEC to form a dense and ductile SEI on the surface of the Li anode. As shown in figures 3(a) and (b), the brittle solvent-derived SEI generated in traditional dilute electrolytes cannot inhibit Li dendrite growth. However, the LiF-rich SEI generated from the high-fluoride-concentrated electrolyte cannot tunnel electrons well and has high surface energy, which is conducive to uniform Li deposition and prolonging the lifecycle [5,32,36]. The high reduction stability of concentrated electrolytes is explained from the energy perspective. In dilute electrolytes, the LUMO level of the solvent molecule is generally lower than that of the free anion, making it easier to obtain electrons on the anode surface to generate SEI through a reduction reaction [41]. However, in the concentrated electrolyte, many anions undergo association reactions to generate CIPs and AGGs whose LUMO level is lower than that of the solvent (figures 3(c) and (d)), resulting in the solvated anions prior to being reduced to generate salt-derived SEI in concentrated electrolyte [33,42,43]. Wood et al [44] compared the dQ dV −1 and XPS data of the low concentration (0.5 M, 1.2 M) and high concentration (3.6 M) LiFP 6 salt dissolved in a mixed solvent EC/EMC (3:7 w/w) +2 wt% VC cycling in a Cu||LiFePO 4 full-cell. It was found that SEI containing inorganic components such as LiF are generated in concentrated electrolytes, which was judged as the salt-derived SEI by analyzing the decomposition products of solvents and salts (figures 3(c) and (d)). Subsequent battery aging tests also proved that the inorganic-rich SEI could prolong battery life. In addition, concentrated electrolytes such as 3.6 M LiFP 6 in EC/EMC (3:7 w/w) + 2 wt%VC [44], 3 M LiFSI in G 4 /EC [45], 5.5 M LiFSI in DMC [46], 1 M LiTFSI + 2 M LiFSI + 3 wt% LiNO 3 in DME/DOL (1:1 v/v) [47] and 0.52 LiFSI-1AN-0.09VC (in mol) [48] prepared by other researchers were tested in lithium batteries, and they found that their CE is higher than the corresponding dilute electrolyte. SEM and XPS found the corresponding salt-derived inorganic-rich SEI on the cycled Li foil, which improved the cycling stability. Concentrated electrolytes also improve oxidative stability. Due to the increasing demand for battery energy density, high-voltage NCM ternary cathode materials are widely used in LMBs. However, traditional commercial electrolytes have poor tolerance to high voltage, and the CEI formed by free solvent molecules on the cathode surface has a high internal resistance, which is not conducive to long-term battery cycling [49][50][51]. At the same time, it cannot prevent the transition metal (TM) dissolution of the cathode at high voltage, which not only destroys the structure of the cathode but also causes the dissolved TM cations to be solvated in the electrolyte and migrate to the anode surface, destroying the SEI layer [40,52]. This will eventually lead to the irreversible decay of battery capacity. In concentrated electrolytes, many CIPs and AGGs make the inner Helmholtz layer on the positive side mainly composed of anions, and the cathode side is positively charged. The anions are more readily adsorbed on the cathode surface by Coulomb attraction, thus forming a dense CEI rich in Li 2 O, Li 3 N, and LiF [53,54]. Furthermore, from the energy level perspective, all solvent molecules are solvated in the concentrated electrolyte. It was found that compared with free solvent molecules, the HOMO level of solvated solvent molecules is reduced, which means that the salt-derived inorganic-rich CEI is easier to generate. The oxidation potential becomes positive, so it is more conducive to enhancing the stability of concentrated electrolytes at high voltages. Ren et al [34] prepared LiFSI-1.4DME (in mol) and tested it in Li||NCM333 half-cells at high voltage, reducing the HOMO level of DME by increasing the salt concentration, and FSIdecomposed LiF, Li x SO y , Li x NO y and other inorganic-rich CEI on the cathode side (figures 3(e)-(g)). Such a dense CEI can isolate the contact between the active cathode and DME, inhibit the dissolution of TM cations, and enable the LMB to maintain 92% capacity after 500 cycles at 4.3 V. In addition, if the concentration continues to increase, the oxidative stability of the electrolyte will continue to improve. Fan et al [35] performed electrochemical tests using 10 M LiFSI in EC/DMC (1:1, v/v) in a 4.6 V high Ni Li||NCM622 cell. The high-voltage stability of the battery was greatly improved by forming salt-derived fluorinated CEI on the cathode, with 86% capacity retention after 100 cycles (figure 3(h)).

Inhibiting cathode dissolution
The concentrated electrolyte can also inhibit the dissolution of cathodic TMs [31,52]. TM dissolution can be roughly divided into three steps. The first is the oxidation of metals to cations, a step that occurs more readily when the cell voltage is higher. The second step is the coordination of TM cations with free solvent molecules or anions to form a solvation structure in electrolytes. The third step is the diffusion of solvated cations in electrolytes and adsorption to the anode surface. Concentrated electrolytes inhibit TM dissolution mainly by inhibiting the second and third steps. A large amount of free solvent in the dilute electrolyte can promote the solvation of TM n+ . However, with increased salt concentration, there are few free solvents and free anions in concentrated electrolytes, and the solvation step (second step) of TM n+ is inhibited. At the same time, the high association of ions in concentrated electrolytes makes the 3D network structure formin the solution, and the viscosity of the electrolyte increases rapidly, which is not conducive to the shuttle of solvated TM n+ in the solution phase. In addition, concentrated electrolytes can also isolate contact between the electrolyte and cathode by forming a dense saltderived CEI on the cathode side, thereby inhibiting TM dissolution. Liu et al [55] increased the concentration of conventional carbonate electrolytes to enable stable cycling in LMB and prepared 3 M LiPF 6 in EC/EMC/DMC (1:1:1, v/v/v) for cycling in Li||Li 1.2 Ni 0.15 Fe 0.1 Mn 0.55 O 2 cell; the capacity retention was close to 100% at a high rate (5 C) and high voltage (4.8 V). As shown in figures 4(a) and (b), LiPF 6 tends to decompose to form a dense and uniform CEI in concentrated electrolytes. Compared with organic-rich CEI formed by carbonate decomposition, it can inhibit TM dissolution and stabilize the cathode structure. Figure 4(c) shows that the dissolution of TMs on the cathode side is significantly reduced in concentrated electrolytes.
In addition, concentrated electrolytes also show excellent performance suppressing the polysulfide shuttle phenomenon of the polysulfide Li 2 S x (x > 2) generated by side reactions in Li||S batteries. The polysulfide dissolved in the sulfur cathode undergoes reduction reactions at the LMA to form lowvalent compounds, which can be oxidized back to the cathode again. Due to these processes, the CE is reduced, and the battery is significantly self-discharged during the standing process, hindering the application of Li||S batteries. In concentrated electrolytes, the concentration of Li + is greatly increased, which inhibits the dissociation and dissolution of polysulfides. At the same time, the increase of electrolyte viscosity also reduces the diffusion coefficient. This hinders the diffusion of polysulfides in the electrolyte, thereby inhibiting the polysulfide shuttle phenomenon and greatly improving the performance of Li||S batteries [20,56,57]. Suo first proposed a high-concentration LiTFSI-based ether electrolyte to inhibit polysulfide dissolution and anionic convection near the electrodes in Li||S batteries, thereby suppressing the shuttle effect and lithium dendrite formation, finally prolonging the  [58] also prepared 5 M LiFSI in 1,3-dioxolane (DIOX)/DME (1:1, v/v) in Li||S batteries. By reducing the solubility of the polysulfides at high concentrations and FSIdecomposition at the cathode to form a smooth protective film that can hinder the passage of polysulfides, the shuttling effect is suppressed, thereby retaining 77% capacity after 1000 cycles.

Suppressing Al current collector corrosion
Concentrated electrolytes have been demonstrated to suppress the corrosion of the Al current collector attacked by organicbased lithium salts (LiTFSI, LIOTf and so on). Traditional commercial electrolytes use LiPF 6 as the salt reacts with a small amount of H 2 O to generate HF, forming a dense AlF 3 passivation film on the Al surface to inhibit Al corrosion in LIBs ( figure 5(a)) [33,59]. However, organic-based lithium salt, taking the LiFSI for example, fails to form AlF 3 passivated Al current collectors ( figure 5(b)), but the corrosion of Al current collectors is suppressed by increasing the concentration. The inhibition of Al corrosion by concentrated electrolytes is consistent with the inhibition mechanism of TM dissolution, which inhibits Al 3+ solvation by reducing free solvent molecules and anion concentrations in the electrolyte and inhibits Al 3+ diffusion through high viscosity and a 3D electrolyte network (figure 5(c)) [30,60,61]. Furthermore, the dense passive film formed by concentrated electrolytes on the electrode surface is also conducive to inhibiting Al corrosion [49,62]. Suo et al configured 7 m LiFSI in FEC fullfluorine concentrated electrolytes to change the Li + solvation structure and reduce the contact probability between solvent molecules and Al, thereby inhibiting Al collector corrosion ( figure 5(d)). The high-fluoride-concentrated electrolytes form a large number of wide band gap LiF on the cathode side, which hinders electron tunneling and which is also beneficial to Al anticorrosion [32]. Yamada et al also successfully overcame Al corrosion (figure 5(e)) and TM dissolution at high voltage by using a high concentration ester electrolyte LiFSI-1.1 DMC (in mol) [24,31,63]. In addition, Luo et al [64] observed the Al corrosion of 1.0 M and 5.0 M LiFSI in EC/DMC (3:7, v/v) at high voltage and found that anions and solvent molecules in solution were mainly involved in the formation of CIP and AGG at high concentrations, reducing the concentration of free FSIand suppressing Al corrosion. The capacity decay rate in a 4.45 V Li||LiCoO 2 battery is only 0.53%.

Low flammability and high thermal stability
Concentrated electrolytes have high thermal stability. Commercial dilute electrolytes mainly use carbonate solvents, which are prone to side reactions with lithium, generating loose SEI on the surface, and then growing dendrites that pierce the separator and cause battery short circuits. In addition, carbonates, especially linear carbonates such as DMC and EMC, are highly volatile and flammable [67][68][69]. However, due to the change of solution structure in concentrated electrolytes, the number of free solvent molecules decreases, and the contact with lithium decreases [15,22,70]. Moreover, concentrated electrolytes can form an anion-derived SEI passivation film on the surface of LMA, which is denser and more stable than the SEI formed by traditional electrolytes. As shown in figures 6(a)-(c), Lee et al formulated 4 M LiFSI in PC/FEC (93:7 v/v). They applied it to a Li||NCM811 full-cell by using a favorable coordination structure of Li + -FSIsolvent clusters at high concentration and FEC-optimized SEI; a flame-retardant high-voltage LMB with excellent electrochemical performance was achieved (energy density up to 679 Wh kg −1 and capacity remains stable at high temperature) [65,71]. In addition, concentrated electrolytes can further reduce flammability by using a series of nonflammable solvents, such as phosphates or sulfones, which are traditionally difficult to apply to LMBs due to their high reactivity with LMA [61,[71][72][73]. Zhang et al [66] used highconcentration lithium salt LiFSI and nonflammable solvent trimethyl phosphate (TMP) to prepare different concentrations of electrolytes and measured TGA. They found that the thermal stability of electrolytes increased significantly with concentration ( figure 6(d)). In addition, 4 m LiFSI in TMP was used as the nonflammable electrolyte in a Li||LiFePO 4 cell, and the capacity did not decline significantly over 90 cycles. Shiga et al [74] also used a self-extinguishing concentrated fluorinated phosphate electrolyte LiFSI-TEEP (1:2 in mol) in a high-voltage Li||LiNi 0.8 Co 0.15 Al 0.05 O 2 half-cell. The thermal stability is improved by the multiple effects of high concentration, high thermal stability phosphate, and solvent fluorination. The battery retained about 80% capacity after 100 cycles at 80 • C, and the electrolyte could not be ignited at 700 • C, which improved the safety performance of LMBs.

The reversibility of LMA in concentrated electrolytes
CE is an important parameter in measuring the performance of electrolytes because it can significantly affect the lifecycle of the battery, and the CE expression is where Q discharge and Q charge represent the amount of charge passed in the external circuit during the discharge step and charge step, respectively. For commercial LIBs, since the anode is graphite, all active lithium species are provided by the cathode, so the CE accurately reflects the active lithium loss of the battery in each cycle. Therefore, capacity retention can be expressed as follows: where i is the electrochemical cycle numbers and ACE is the average Coulombic efficiency. However, for LMBs such as NCM||Li half-cells, there is usually an excess of active lithium species on the anode side, so the lithium loss on the anode during cycling is invisible. In this case, CE cannot continue to reflect the net Li loss in the cycle, so capacity retention cannot be accurately calculated using equation (3). However, CE can still reflect whether the cathode is prone to side reactions with the electrolyte, such as the oxidative decomposition of the electrolyte and the cathode degradation, in order to evaluate the stability of the electrolyte to the cathode [75][76][77]. The CE of the Li||Cu half-cell is also an important indicator used to reflect electrolyte quality. The CE reflects the reversible active Li on Cu, but it indirectly reveals the Li loss caused by the formation of dead Li and SEI and the compatibility of the electrolyte with lithium. In addition, the anode-free LMB (AFLMB), such as the NCM||Cu cell, is similar to LIB; that is, the active Li is completely provided by the cathode. Therefore, capacity retention can be calculated with CE by using equation (3), and then the electrolyte performance can be evaluated [78]. In this paper, the cycling performance of different concentrated electrolytes in LMB in recent years is listed and summarized in table 1. This table also summarizes the CEs of different electrolytes in Li||Cu cells and LMBs and annotates different test conditions as much as possible.
Due to the different cathodes of full cells, the voltage and specific capacity that can be provided are also inconsistent, and the CEs and cycle performance of full cells are challenging to normalize for comparison. Therefore, this paper summarizes the CEs of the different concentrated electrolytes in Li||Cu half-cells of table 1 in figure 7. Where the same abscissa represents the same kind of solvent, and the same color point represents the use of the same salt.

Solvents used for concentrated electrolytes
In LMBs, ether solvents are widely used due to their stability in lithium anodes. However, the oxygen atom in the ether group contains two lone pairs of electrons, which is alkaline and nucleophilic, and is easily oxidized when there is hydrogen on the α carbon of the ether group, so ether solvents are unstable at high voltage. In a dilute ether electrolyte, many free solvent molecules in solution are prone to side reactions with high-voltage cathodes and oxidative electrolyte decomposition, which lead to a decrease in lifecycle. The concentrated electrolytes improve the stability of ether electrolytes at high potential through the change of solvent structure. Wan et al [99] prepared electrolytes of 1 M LiFSI in DME and 4 M LiFSI in DME, respectively, and measured their CEs in Li||Cu cells, finding that the CE corresponding to 1 M LiFSI in DME was only 88.9%, which is far lower than 99.3% of 4 M LiFSI in DME. Compared with the traditional diluted electrolytes, the concentrated electrolyte has higher CE of lithium anode because the high concentration not only changes the solvation structure and reduces the number of free solvent molecules, but also manipulates a dense, high mechanical strength, and superior Li ion conducting SEI on the surface of LMA, making the deposited Li particle growing up with the small surface, thereby reducing the occurrence of side reactions. Furthermore, during lithium deposition, continuously consuming Li ions on the surface of LMA forms a Li ion concentration gradient. The concentrated electrolytes have a high Li ion concentration guaranteeing mass transport and a high Li ion transference number reducing the space aniondepletion layer, thus effectively inhibiting the growth of lithium dendrites and improving the CE of LMA [20,23,33,88]. Suo et al [20] prepared a high-concentration bis-ether electrolyte 7 m LiTFSI in DME/DOL (1:1 v/v) and obtained a high Li + transference number of 0.73 and a high conductivity of 0.814 mS cm −1 . A high CE of ∼100% is obtained in Li||S cells while maintaining a high capacity of 74% after 100 cycles. In addition, Lin, Suo et al [11] used 6 m LiFSI in DME in an AFLMB, the anode was Cu, and the cathode was Li 1.37 Ni 0.8 Co 0.·1 Mn 0.1 O 2 (Li 1.37 NCM811) after prelithiation. By embedding more Li into the cathode, the specific capacity and energy density are effectively improved. An effective specific energy of 447 Wh kg −1 is obtained, and a high capacity retention rate of 84% is achieved after 100 cycles. Furthermore, this electrolyte achieves a high initial CE of 93.2% in Li||Cu half-cells, which exceeds 99% CE after 10 cycles. This demonstrates ultra-high Li cycling stability and opens up new ideas for improving the energy density of LMBs. It can be found from figure 7 that the concentration of carbonate electrolytes is relatively higher than that of highconcentration ether electrolytes. The C=O group contained in carbonate is more accessible to obtain electrons than ether, so the carbonate solvent is easier to react with Li, and a higher concentration is required to stabilize the electrolyte. The figure shows that the high concentration significantly improves the CE of LMBs, comparable to or even exceeding that of ether electrolytes, indicating the excellent electrochemical performance of high-concentration carbonate electrolytes. Suo et al [32] used 7 m LiFSI in FEC for 5 V Li||LiNi 0.5 Mn 1.5 O 4 full cells (N/P ratio = 1.4) by using the fluorinated carbonate solvent FEC to form LiF-rich SEI, stabilize the LMA while enhancing the electrolyte oxidative stability, and overcome Al corrosion at high voltages. The high capacity retention of 78% and an energy density of 583 Wh kg −1 were achieved in 130 cycles, and the average CE was as high as 99%. Yamada et al also prepared a high-concentration carbonate electrolyte LiFSI-1.1 DMC (in mol). It was found capable of operating stably in graphite||LiNi 0.5 Mn 1.5 O 4 cells at a 4.6 V high voltage (>90% capacity after 100 cycles), and in Li|| LiNi 0.5 Mn 1.5 O 4 half cells (5.2 V), it can still maintain 95% of initial capacity after 100 cycles, and CE is close to 100% [31].
At the same time, the selection of components in concentrated electrolytes is more diversified due to the change in its solution structure, and other types of solvents besides ethers and carbonates are also selected. In 2018, Zhang, Xu et al [98] used LiFSI and sulfone solvent TMS to mix a concentrated electrolyte at a molar ratio of 1:3. Because of the high oxidation resistance of the S=O bond, sulfone compounds can remain stable at high voltage. Applying this electrolyte to a Li||NCM333 full cell with limited lithium (50 µm), it has an extremely high CE of over 99.8% in 300 cycles while possessing about three-quarters capacity retention. In addition, the safety performance of LMB can be improved by using a nonflammable phosphate solvent. Zeng et al [22] prepared high-concentration phosphate electrolyte LiFSI-2TEP (in mol) + 0.05 M LiBOB + 5% FEC (by volume) for application in Li||LiCoO 2 cells with a reversible specific capacity of 135 mAh g −1 , while it has a high ACE of 99.7% and cycle stability (88% capacity retention after 350 cycles). This is because phosphate solvent molecules are mainly coordinated with Li + at high concentrations and has extremely low reactivity to LMA. ACE exceeds 99% in Li||Cu cells, equivalent to high-concentration ether electrolytes.

Salts used for concentrated electrolytes
The effects of different salts on battery cycle performance were analyzed. The salt used in LIBs is mainly LiPF 6 , which is friendly to graphite anodes, has high ionic conductivity, and can passivate Al current collectors. However, LiPF 6 easily reacts with a small amount of H 2 O to form HF, which can corrode the SEI layer on the surface of LMA, making the LMB unstable, and having poor thermal stability [100]. Therefore, LiPF 6 is not suitable for application in lithium battery electrolytes. Salts commonly used in LMBs, such as LiFSI, LiTFSI, LiBOB, and LiDFOB, contain elements such as F, N, and B, which can form inorganic phases such as LiF, Li 3 N, Li 2 O, etc [47] on LMA surface, thereby effectively protecting lithium and improving the cycle stability of the battery. As shown in figure 7, the CE of electrolytes using LiFSI is generally at a high level because LiFSI is relatively stable and has excellent electrical conductivity and thermal stability while generating more LiF [85]. LiF has the largest bandgap (13.6 eV) and the widest electrochemical window, and it also has high surface energy and a low surface diffusion barrier of Li adatoms because the Li atom and F atom have small lattice constants and considerable electronegativity differences, which is beneficial to the protection of LMA [32,83].

Additives
Putting additives in concentrated electrolytes can improve electrochemical performance. For example, it has been found that the performance of electrolytes added with LiNO 3 as an additive is also excellent. As a film-forming additive, LiNO 3 can form Li 2 O, Li 3 N, and LiN x O y , etc [101,102]. The combination with LiF formed by salt decomposition is conducive to stable and dense SEI formation and can change the deposited lithium morphology from dendritic to spherical, thereby stabilizing LMA. At the same time, due to the charge effect, NO 3 accumulates in a large amount on the cathode side, thereby removing solvent molecules from the electric double layer and improving the high-voltage resistance [103]. In addition, the higher the NO 3 concentration, the more favorable it is to broaden the oxidation window of the electrolyte. Kim et al also used LiFSI to inhibit the agglomeration of LiNO 3 in DME and then prepared a double-salt concentrated electrolyte 2 M LiFSI + 2 M LiNO 3 in DME, in which LiFSI can form LiFrich SEI, and LiNO 3 can decompose on the Li anode surface to form Li 2 O and Li 3 N and are able to change the deposited Li morphology from dendritic to spherical, thereby stabilizing the LMA, cycling 425 times with an ACE of 98.5% in Li||Cu cells. Meanwhile, 47.3% capacity was maintained in Cu||NCM622 AFLMB in 100 cycles [84,86]. In addition, additives such as LiDFOB, FEC, VC, etc, also help to form a good layer on cathode and anode surfaces. The B-O bond of LiDFOB is weak and easy to break, so a flexible passivation film containing a B component is preferentially formed, inhibiting Al corrosion and improving the high-voltage resistance [89]. FEC has a low HOMO level, which is beneficial for improving oxidative stability. It also has a low LUMO level and can preferentially decompose to generate LiF, which can synergize with LiDFOB to make a more uniform CEI [97,104]. VC is an excellent film-forming additive that can decompose to form polycarbonate and Li 2 CO 3 to protect the LMA, and is widely used in electrolytes [48,105,106].

Summary
A comparison of concentrated and conventional dilute electrolytes is shown in figure 8, including some properties of concentrated electrolytes beneficial for LMBs that have been discussed in the above section. Herein, we want to note some adverse impacts of increasing concentration on the batteries that should not be ignored. First, the viscosity of concentrated electrolytes is higher, resulting in lowering the ionic mobility that would reduce the ion diffusion rate and decrease the ionic conductivity of electrolytes, adversely affecting the kinetic of batteries [107]. Second, the wettability of concentrated electrolytes with electrodes and separators becomes worse and weakens the transfer speed of Li + between phases, which is not conducive to uniform lithium deposition during cycling and is prone to lithium dendrites. Third, concentrated electrolytes bring about cost increases that would be a challenge for their technical and economic performance. For traditional dilute electrolytes (1 M), 70%-90% of electrolyte costs comes from lithium salts. In contrast, the amount of salts in concentrated electrolytes is several times that of traditional electrolytes (for example, salt content in 5 M LiFSI in DMC is almost 6 times that of 1 M LiPF 6 in EC/DMC) [31]. In summary, concentrated electrolytes have many merits regarding electrochemical performance and safety of LMBs, including being favorable for formation of a salt-derived inorganic rich interphase, inhibiting cathode dissolution, suppressing Al current collector corrosion, and low flammability and high thermal stability. However, at the same time, we should also realize their shortcomings as far as the cost and kinetic energy, and make more efforts to address them before their application.