Ion transport and structural design of lithium-ion conductive solid polymer electrolytes: a perspective

Solid polymer electrolytes (SPEs) possess several merits including no leakage, ease in process, and suppressing lithium dendrites growth. These features are beneficial for improving the cycle life and safety performance of rechargeable lithium metal batteries (LMBs), as compared to conventional non-aqueous liquid electrolytes. Particularly, the superior elasticity of polymeric material enables the employment of SPEs in building ultra-thin and flexible batteries, which could further expand the application scenarios of high-energy rechargeable LMBs. In this perspective, recent progresses on ion transport mechanism of SPEs and structural designs of electrolyte components (e.g. conductive lithium salts, polymer matrices) are scrutinized. In addition, key achievements in the field of single lithium-ion conductive SPEs are also outlined, aiming to provide the status quo in those SPEs with high selectivity in cationic transport. Finally, possible strategies for improving the performance of SPEs and their rechargeable LMBs are also discussed.


Introduction
Lithium-ion batteries (LIBs) are able to carry electric energy in chemical forms by virtue of the reversible (de)intercalation of Li + -ions (Li + ) into hosting materials, which are known as 'rocking-chair batteries' [1]. Amongst existing rechargeable battery technologies, LIBs have achieved widespread applications spanning from portable electronic devices (i.e. computer, communication, and consumer electronics) to electric vehicles, due to their high operating voltage (ca. 4.0 V vs. Li/Li + ) and energy density (ca. 300 Wh kg −1 ), relatively long cycle life (>300 cycles), and negligible memory effect [2].
The prevailing LIBs are built with nonaqueous liquid electrolytes comprising lithium hexafluorophosphate (LiPF 6 ) as conductive salt and organic carbonates (e.g. ethylene carbonate (EC), and methyl ethyl carbonate, etc) as solvents, with the addition of small amounts of functional electrolyte additives (i.e. vinylene carbonate, and fluoroethylene carbonate, etc) [3,4]. Generally, the as-formulated non-aqueous liquid electrolytes present sufficient ionic conductivities (ca. 10 −2 S cm −1 ) at room temperature and excellent chemical and electrochemical stabilities (>4.2 V vs. Li/Li + ) [3], enabling the operation of LIBs with decent power capability. However, the LiPF 6 -based nonaqueous liquid electrolytes show poor chemical stability, exacerbating the cycling performance of LIBs during long-term operation, particularly at elevated temperatures (>40 • C) [5]. Effectively, at the industrial level, LiPF 6 is produced via the reaction of lithium fluoride (LiF) and phosphorus pentafluoride (PF 5 ) in anhydrous hydrogen fluoride solvent [6], in which trace amounts (ppm level) of protic impurities (e.g. hydrogen fluoride) are inevitably inherited from the preparative procedures. It is reported that these protic impurities are responsible for triggering the chemical decompositions of LiPF 6 -based electrolyte solutions, generating highly toxic substances, such as phosphorus trifluoride (O=PF 3 ), and organophosphorus compounds (O=PF 2 OR, R=alkyl) (figure 1(a)) [5].
In addition to the chemical instability of LiPF 6 , low flash points (e.g. T flash = 18 • C for DMC [7]) and potential leakage of organic carbonates also impose critical concerns on the inherent safety of electrolyte materials and their rechargeable batteries [7], particularly with the implementation of highenergy electrode materials such as lithium metal and its alloy [8]. Therefore, traditional LIB technologies are unlikely to meet the requirements of the emerging fields with increasing energy density (>400 Wh kg −1 ) and safety demands, such as power and energy storage and smart grids [9]. Thus, there is an urgent need to develop next-generation battery technologies with low-cost, high-energy density, long cycle lifespan, and high intrinsic safety [10].
With the replacement of organic liquid electrolytes with solid-state electrolytes, solid-state lithium metal batteries (SSLMBs) have been deemed as a feasible solution to enhance the inherent safety and energy density of the contemporary LIBs, due to the elimination of volatile components and the possible utilization of high-capacity electrode materials (e.g. lithium metal electrode: 3860 mAh g −1 ; figure 1(b)) [11]. Presently, research activities related to SSLMBs have become increasingly important [12].
In 1973, Wright and co-workers [13] reported that the mixtures of poly(ethylene oxide) (PEO) and several alkali metal salts (such as potassium thiocyanate (KSCN) and sodium thiocyanate (NaSCN)) afford ionically conductive plastic materials at higher temperatures (>60 • C) [13]. Later, Armand et al [14] realized the profound significance and potential application of these ionically conductive polymers for developing SSLMB technologies, thus proposing their utilization as solid electrolytes for SSLMBs. Briefly, lithium-ion conductive SPEs possess several intriguing merits, including greatly enhanced safety, ease of processing, simplicity of cell construction and assembly, and so on [15]. Therefore, the research on SPE-based SSLMBs has attracted extensive attention from the academic and industrial sectors (figure 1(c)).
Polymer electrolytes (PEs) are a kind of ionic conductive material utilizing high molecular weight polymer as a matrix. Generally, according to the chemical compositions, PEs can be sorted into four different families, including solid polymer electrolytes (SPEs), composite solid polymer electrolytes (CSPEs), plasticized polymer electrolytes (PPEs, liquid plasticizer content <50 wt%), and gel polymer electrolytes (GPEs, liquid plasticizer content >50 wt%) (figure 1(d)) [16]. Generally, the chemical compositions of SPEs are relatively simple, containing only lithium salts and polymer matrices. For CSPEs, a certain amount of inorganic solid fillers is introduced to improve the transport properties of lithium ions and the mechanical strength of electrolyte membranes [17]. The latest research progresses in this domain have been scrutinized in relevant literature [18,19]. For PPEs and GPEs, liquid plasticizers are added to promote the segmental motions of polymers and better carry ionic species. Effectively, these two kinds of PEs behave mostly like conventional liquid electrolytes depending on the contents of liquid components [16]. Note that the introduction of small molecule compounds accelerates chemical and electrochemical parasitic reactions at the interphases/interfaces between electrolyte and electrode materials, decreasing the cycle life of corresponding SSLMBs, despite their capability in improving ionic conductivities [20].
In this perspective, recent advances in lithium-ion conductive SPEs are briefly discussed, with main attention paid to the ion transport in SPEs, emerging conductive lithium salts, and polymer matrices utilized. In addition, the status quo of a special type of SPE with lithium-ion transference number (T Li + ) close to unity, single lithium-ion conductive SPEs (SLIC-SPEs), is also reviewed. Design strategies and future directions for developing robust SPEs and their SSLMBs are provided.

Ion transport in SPEs
A comprehensive understanding of the transport phenomena of ionic species is of supreme importance for designing highperformant SPEs. Figure 2(a) outlines several key findings in elucidating the ion transport in SPEs at the microscopic level over the past 40 years. As early as the 1970s, Armand et al [14] anticipated that the segmental motion of polymeric backbones is related to the transport of ionic species, particularly for Li + cations, as shown in figure 2(b). In the 1980s, with solid-state nuclear magnetic resonance techniques, Berthier et al [21] demonstrated that ion transport of a PEO-based SPE system (typically 'salt in polymer') occurs primarily in the amorphous region therein. These early studies provide implicit microscopic images of the ion conduction processes in SPEs [22,23]. From another perspective, Stoeva et al [24] proposed that the crystalline phases of PEO-based SPEs are ionically conductive with well-defined microstructures ( figure 2(b)). For the crystalline LiAsF 6 /PEO electrolytes, relatively rapid diffusion of Li + ions is realized via the hopping of ionic species without involving the segmental motion of PEO chains (diffusion paths indicated by a pink circle in figure 2(b)), the PEO matrix remains 'immobilized' during the ionic conduction processes. In sharp contrast, the ionic conductivities of crystalline LiAsF 6 /PEO electrolytes are nearly one order of magnitude higher than those of amorphous LiAsF 6 /PEO electrolytes [24]. Yet, these crystalline SPEs are likely to achieve high ionic conductivities with low molecular weight PEO (< 5000 g mol −1 ), which could hardly afford self-standing films and thereby hinder their practical applications in lithium batteries [24][25][26].
Since the 1990s, Angell et al [27] systematically studied the ion transport behaviors of the PEO-based electrolyte systems and revealed that the Li + transport is highly coupled with the movement of polymer chain segments. To quantify the relational degree between the ion transport behavior and the movement of polymer chain segments, the concept of decoupling indices (R τ ) was proposed, as mathematically described by equation (1): where R τ is the decouple indices, τ s is the macroscopic structural relaxation time of the matrix glass (in seconds), and τ σ is the conductivity relaxation time (in seconds) [27,28]. Traditional SPEs comprising polyether-type matrices (e.g. PEO) and common lithium salts (e.g. lithium perchlorate (LiClO 4 )) are classical 'coupling' systems (generally, R τ < 1) [28,29]. In this scenario, ionic transport is highly correlated with the segmental movement of the polymer chains within the amorphous region (figure 3(c)) [30], and ion mobility is closely related to temperature change, e.g. the ion conductivity decreases rapidly to about 10 −14 S cm −1 when lowering the temperature close to glass transition temperature (T g ) [28]. In 1993, Angell et al [31] proposed the concept of 'polymer in salt' electrolytes, in which a large amount (>50 wt%) of lithium salt with low melting point and high dissociation characteristics is utilized to form 'decoupled' SPEs systems. In this scenario, the Li + transportation does not depend on the segmental motions of polymer chains [31].
According to the different transport behaviors of ionic species, SPEs could be generally categorized into 'coupled' and 'decoupled' systems [32]. For the 'coupled' systems, Li + transport is highly correlated with the motion of the polymer chain segments (figure 3(c)). For the 'decoupled' systems, the numbers of cation/anion clusters increase with increasing salt concentration, and the aggregated cation/anion clusters are interconnected with each other, favoring the formation of an ionic conductive network which could provide a fast conduction channel for Li + transport [28,33] (figure 3(c)). In general, the decoupled indices of the 'decoupled' system could be as high as 10 13 [27]. Unfortunately, most available lithium salts are unlikely to meet the stringent requirement imposed by 'decoupled' SPEs including low melting point and extremely high dissociation. Besides, increasing salt concentration also sacrifices the mechanical properties of the as-formed SPEs [34][35][36]. Therefore, the development of the 'coupled' SPEs systems tend to be more rapid than that of the 'decoupled' SPEs ones.
To date, the 'coupled' SPEs are the most widely studied attributed to their easy processing and good compatibility with high-energy electrode materials. Among them, PEO and its derivatives are the most thoroughly studied matrices [37][38][39][40][41][42][43], which is ascribed to the features below: (a) the oxygen atom on the repeat unit of ethylene oxide (-CH 2 CH 2 O-, EO) owns strong donicity, which can form complexes with metal ions, thus promoting the dissolution of alkali metal salts, and realizing Li + transport via coupling and decoupling of polymer chain segments [30]; (b) traditional PEO-based SPEs own several advantages, including low density (ca. 1.2 g cm −3 ), good chemical stability, and low cost, and can better inhibit lithium dendrites growth in SSLMBs [44,45].
The neat PEO is a semi-crystalline helical polymer, possessing a certain degree of crystallinity (> 60%), due to its regular and highly ordered structure [46,47]. As mentioned above, the ion transport of SPEs relies heavily on the segmental motion and local relaxation of polymeric chains, and Li + transport mainly occurs in the amorphous region of SPEs [48][49][50][51]. Consequently, the ionic conductivity of PEObased SPEs is generally lower than 10 −5 S cm −1 at room temperature [52], which hinders its large-scale application in SSLMBs. Reducing the crystallinity of traditional PEO-based SPEs has become a hot research topic in the field of SPEbased SSLMBs, and various approaches have been assessed, including structural modifications of PEO, and doping with inorganic materials [44,53], and so on (see section 4.1 for detailed discussion).

Developing robust conductive lithium salts
Generally, the conductive lithium salt not only acts as the source of charge carriers for SPEs, but also participates in the construction of electrode-electrolyte interphases/interfaces via chemical and/or electrochemical reactions [54]. Therefore, the composition and chemical structure of conductive lithium salts have a critical impact on the fundamental properties of SPEs. Generally, ideal conductive lithium salt should contain several traits (figure 3(a)), including solubility, interfacial compatibility, chemical stability, aluminum corrosion, etc.
To form ionic conductors, the breakdown of ionic bonds between Li + cation and anions in the presence of electrondonating polymer matrices tend to be of higher priority. Indeed, the dissociation process is determined by the lattice energy of salt, the cohesive energy of polymer, and the solvation energy thereof ( figure 3(b)). For polymers, the cohesive energy density (CED) is mathematically expressed as [55]: where H vap stands for the heat of vaporization, R and T represent the respective ideal gas constant and the absolute temperature, V is the molar volume, and δ stands for the solubility parameter, a semi-quantitative measure of the polarity of the repeat units.
In addition, the microscopic viscosity of the SPEs system plays an important role in dictating the transport properties of ionic species therein [56]. And the viscosity of SPEs is highly correlated with the free volume provided by anions. Therefore, anions with high structural flexibility are necessary for building high-performance SPEs [52]. Presently, various kinds of anions have been introduced into Li-ion conductive SPEs, including halide, carboxylate, sulfonate, and imide anions (figure 3(c) and table 1) [57].

Sulfonimide-based lithium salts
Sulfonimides, particularly perfluorinated sulfonimide anions, [(R F 1 SO 2 )(R F 2 SO 2 )N] − , are one of the most noticeable anions for Li-ion conductive SPEs, attributed to their low affinity toward Li + and high structural flexibility [75,76]. Among which, bis(trifluoromethanesulfonyl)imide anion ([(CF 3 SO 2 ) 2 N] − , TFSI − , figure 4(a)) firstly prepared by Meussdorffer and co-workers [77] in the acid form in 1972, appears to be the most investigated candidate representative anions. In the case of TFSI − anion, the sulfonimide anion center owns several resonance structures, allowing the delocalization of negative charges on the nitrogen atom to the four oxygen atoms. Additionally, the strong electron-withdrawing ability of CF 3 groups could further lower the Lewis basicity of the anion (Gutmann donor number = 5.4 [75]), affording low dissociation energy of the Li + cations. The interconversion between different conformations of TFSI − anion occurs with extremely low energy barriers (< 5 kJ mol −1 [78]), which endows its large free volume. These key properties facilitate its dissolution and dissociation in polymer matrices, promoting the rapid transport of ionic species ionic conductivities (e.g. 1 × 10 −4 S cm −1 , 80 • C). Presently, LiTFSI has been deemed as a benchmark salt for screening the new anions for Li-ion conductive SPEs [45].
In recent years, other kinds of lithium sulfonimide salts have been employed as conducting salts for SPEs. Lithium fluorosulfonimide salts containing fluorosulfonyl (FSO 2 -) group have become an interesting family, owing to their unique capability in building stable solid electrolyte interphase/interface (SEI) layer on various kinds of   figure 4(a)), one representative example of lithium fluorosulfonimide salts, was synthesized in acid form by Appel in the early 1960s [79]. The anion was proposed as a candidate for battery application by Armand in the 1990s [80]. In the past years, our groups have systematically investigated the properties of various kinds of LiFSI-based SPEs, including a wide array of polymeric matrices (e.g. PEO [52] and poly(ionic liquids) (PILs) [81]), which show much higher ionic conductivities and better chemical and electrochemical stabilities on electrode materials than those of the LiTFSI-based ones [69]. It has been demonstrated that the rotation barriers of fluorosulfonimide anions are lower than the symmetric perfluorinated sulfonimide anions (e.g. 0.9 kJ mol −1 for (fluorosulfonyl)(pentafluoroethanesulfonyl)imide ([(FSO 2 ) (n-C 4 F 9 SO 2 )N] − , FNFSI − ) vs. 6.3 kJ mol −1 for bis (perfluoroethanesulfonyl)imide ([(C 2 F 5 SO 2 ) 2 N] − , BETI − ) [82]), endowing the formers with better structural flexibility and stronger plasticizing ability [82]. Besides, the S-F bond in the fluorosulfonimide anions tends to be more electrochemically active compared to the C-F bonds in the symmetric perfluorinated sulfonimide anions, which may undergo electrochemical decompositions prior to the reductions of polymer matrices (ca. 1.0 V vs. Li/Li + , figure 4(b)) [54]. The as-formed decomposition products (especially LiF) of the fluorosulfonimide anions favor the formation of dense and electronic insulating SEI layers, thus preventing continuous decompositions of electrolyte components [54,83].
With the extension of the perfluorocarbon side chain, lithium fluorosulfonimide salts show enhanced compatibility with Li • anode significantly improved stability toward (electro-)chemical oxidation, which is related to the decomposed products of the longer perfluoroalkyl chains which may improve the stability of electrode-electrolyte interphases [57]. As a result, a prototype Li • ||lithium iron phosphate (LiFePO 4 , LFP) cell with LiFNFSI/PEO at a molar ratio of EO unit to Li + (hereafter abbreviated as [EO]/[Li + ], by mole) of 20 showed excellent cyclability (capacity retention: > 80% at cycle 500) [57]. However, the SPEs based on lithium perfluorinated sulfonimide salts suffer from rapid anionic migrations due to the negligible interactions between anion and polymer matrices (compared with the Li + dipole interactions). Typical T Li + values for these electrolytes are close to 0.2. The rapid transport of anionic species induces concentration gradients, which causes undesired concentration polarizations and leads to inferior utilization of active materials of composite electrodes [84,85].
Through the attachment of anionic moieties onto the polymer matrices (or inorganic macromolecules), the transport of anionic species could be nearly eliminated, as extensively discussed in section 4. Alternatively, enhancing the interactions between anions and polymer matrices via non-covalent bonds could also slow down the migration of anions. Researchers also focus on other ingenious and effective approaches to capture anion by introducing hydrogen-bond in conductive lithium salt structure. Oteo et al [86] reported the utilization of a non-perfluorinated sulfonimide anion, lithium  [65]. Besides, one could also count on other interactions between anion and polymer matrices, e.g. dipoleion interaction, dipole-dipole interaction [68], π-π stacking [67], etc (see the schematic diagram in figure 4(c)) By replacing an oxygen atom in TFSI − anion with strong electron-withdrawing group CF 3 SO 2 N=, the interactions between Li + cations and sulfonimide anions could be further reduced. The as-obtained anion, (trifluoromethane(S-trifluoromethanesulfonylimino)sulfonyl) figure 4(a)), shows extremely low affinity toward Li + cations, and high ionic conductivity for its PEO-based SPEs compared to the LiBETI-based ones [69]. This result clearly suggests the important role of negative charge delocalization in achieving highly Li + -ion conductive SPEs [69].

Other emerging conductive lithium salts
In addition to the sulfonimide-based conductive lithium salts, several other kinds of conductive lithium salts have also received extensive attention in the domain of SPE-based SSLMBs. For example, lithium bis(oxalate)borate (LiBOB) and lithium difluoro(oxalate)borate (LiDFOB) have been investigated by several research groups ( figure 5(a)), owing to their ability in forming robust SEI layers on Li • anode to suppress the continuous reductive decompositions of polymer matrices [87].
In another example, Zhang et al systematically characterized the effect of lithium tricyanomethanide (LiC(CN 3 ), LiTCM, figure 5(b)), as a fluorine-free conductive salt, on the properties of PEO-based electrolytes [72]. LiTCM/PEO provides lower polarizations for Li • ||Li • symmetrical cell as compared to the LiTFSI/PEO reference system, despite its slightly lower ionic conductivities resulting from higher glass transitions (T g : −27 • C (LiTCM/PEO) vs. −30 • C (LiTF-SI/PEO)). The improved interfacial compatibility would be attributed to the reductive polymerizations of the TCM − anion, generating a highly Li + conductive graphene-contained SEI layer therein [72].
Overall, since the initial utilization of SPEs for SSLMBs, the sulfonimide-based lithium salts have become a research hotspot due to their unique properties such as highly flexible structure, low charge density, ease in structural modifications, etc. Remarkable achievements have been made in terms of the design of innovative lithium salts for SPEs; yet, mechanistic understandings of the role of certain functional groups in SSLMBs, especially the interphases formed between electrode and SPEs with different kinds of anions, are still needed.

Developing advanced polymer matrices
The characteristics of polymer matrices have a profound influence on the performances of SPEs [88][89][90], and the fundamental physical properties of some popular polymer matrices are summarized in table 2. As a reference polymer, the T g of PEO is relatively low (T g = −64 • C), a sign of rapid segmental motion of EO units above room temperature [89]. By replacing the C-O linkage with Si-O bond, one may further decrease the T g value. For example, poly(dimethylsiloxane) (-(Si(CH 3 ) 2 O) n -, PDMS) shows an extremely low glass transition behavior at −127 • C; yet, the practical application of PDMS-based SPEs is hindered by its high cost, difficulty in process, and adverse side effects (i.e. the hydrolysis of Si-O-Si to silanol, and the spontaneous condensation at room temperature) [90,91]. Through the incorporation of polar groups (e.g. nitrile, carbonyl), one may promote the dissolution of metal salts; however, the strong van der Waals interactions between these groups drastically increase the CED of the neat polymer and the T g values. In a typical example, high-molecular-weight poly(acrylonitrile) (PAN) shows a high glass transition at 125 • C (table 2) and CED (620−900 J cm −3 mol −1 ), which could barely solvate common lithium salts in the absence of small molecular solvents [92,93]. Therefore, with commercially available polymers, it is rather difficult to obtain highly conductive and self-standing SPEs membranes.
The design of the molecular structures of polymer matrices is one of the most effective methods to boost the ionic conductivity at room temperature and anti-oxidation properties of PEO-based SPEs [53,87,[94][95][96][97]. Here, the research progress related to some emerging polymer matrices is discussed in the following section, including (a) Jeffamine-based amorphous polymers, and (b) polycarbonate and its derivatives. Note that, except for neutral polymers without any ionic groups, there has been a growing interest in utilizing polymerized ionic liquids as matrices for SPEs [81,[98][99][100][101]. The progress in this domain has been scrutinized in recent review articles [101] and will not be discussed in the present work.

Amorphous polyethers
Jeffamine ® is a kind of commercial polyether amines, terminated with primary amino groups and containing ethylene oxide (EO), propylene oxide (-CH(CH 3 )CH 2 O-, PO), or a mixture of EO/PO [102]. Utilizing Jeffamine moiety in building SPEs delivers several advantages, including (a) the repeat unit of EO/PO could effectively dissolve and dissociate lithium salt due to the strong donicity of EO/PO units, (b) the structural disorder of EO/PO could efficiently inhibit the crystallization processes; and (c) the primary amino group at the end could undergo condensation polymerizations with anhydride, thus allowing facile regulation on the topological structure of the polymer matrices. Consequently, Jeffamine ® compounds provide a simple and effective approach for the efficient preparation of novel polymer matrices (figure 6(a)) [102]. The basic physical properties of some Jeffamine-based SPEs are collected in table 3.    In 1992, Benrabah et al [103] synthesized a series of polyamide compounds through the polycondensation reaction between Jeffamine (e.g. Jeffamine ED-600, ED-900, terephthaloyl chloride (TAT), etc) and acyl chloride, aiming to lower the crystallinity of polyether-based SPEs (table 3, entry 1). It is reported that the PO units in Jeffamine-type polymers could effectively restrain the crystallization of SPEs, allowing the LiTFSI-based SPEs to achieve improved ionic conductivities (ca. 10 −6 S cm −1 at 30 • C) at temperatures below the melting points of PEO [103].
Subsequently, Aldalur et al [104] reported a novel class of comb-like polymer matrices comprising polyether amine oligomer side chains (i.e. Jeffamine compounds) and poly(ethylene propylene maleic anhydride) backbone. It is noteworthy that the synthesis route of the matrices is simple and scalable. Combining the high degree of configurational freedom and flexibility of the PO/EO units in Jeffamine compounds (i.e. Jeffamine M-600, Jeffamine M-1000, and Jeffamine M-2070, table 3), a series of new polymer matrices with good elastic and amorphous properties have been obtained [104]. Differing from linear PEO-based ones, the comb-like SPEs containing Jeffamine show high ionic conductivity (LiTFSI/Jeffamine, 4.5 × 10 −5 S cm −1 at room temperature) and excellent electrochemical stability ( figure 6(b)). Furthermore, the cycling stabilities of Li • ||LFP and Li • ||S cells are remarkably enhanced, proving the feasibility of utilizing Jeffamine compounds as a building block for high-performance SPEs [104].
Based on the previous synthesis and screening of Jeffamine-type polymers, Aldalur et al [105] combined LiFSI with the amorphous PMA-Jeffamine polymer matrices, attempting to improve ionic conductivities under room temperature and the interfacial stability between electrolyte and lithium metal anode. The electrolyte comprising of LiFSI/Jeffamine-poly(ethylene-alt-maleic anhydride), (PEaMA) shows high ionic conductivities at ambient temperature (e.g. 1.8 × 10 −4 S cm −1 at 30 • C, table 3, entry 7), and enabled long-term cycling of Li symmetric cells (exceeding In addition, the synthesis procedures have a remarkable effect on the properties of Jeffamine-type polymers [106]. Usually, for the solvents with low dielectric constants (e.g. trifluorotoluene), the as-obtained Jeffamine-type polymers are highly entangled, behaving like rubber. However, for the solvents with a high dielectric constant (e.g. N, N-dimethylformamide), the starting material PEaMA could be well dissolved and the as-obtained polymers are somehow flowable, as shown in figure 6. The flowable polymer electrolytes (FPEs) own several advantages: (a) high ionic conductivities at room temperatures (e.g. 1.4 × 10 −4 S cm −1 under 30 • C) owing to the low T g and highly amorphous characteristics, and (b) improved chemical and electrochemical compatibility towards lithium anode due to the better adhesion properties ( figure 6(b)). It has been demonstrated that, with Jeffamine FPEs as an artificial layer, the cycle life of the Li • ||LiFePO 4 cell is extended in comparison with its counterparts containing PEO (figure 6(e)) [106].
Besides high ionic conductivity, the high-strength property is also essential for the processing of SPEs in SSLMBs. Grafting PS onto a Jeffamine backbone to prepare copolymers (Jeffamine-PS) [107], or blending the PEMa-Jeffamine matrices with PVDF nanofibers [110], can effectively improve film-forming ability with little expense at ionic conductivities, e.g. 7.9 × 10 −5 S cm −1 for Jeffamine-PS copolymers under 40 • C, and ca. 10 −4 S cm −1 for the PEMAa-Jeffamine/PVDF blended electrolyte under 30 • C [110]. Compared with semi-crystalline PEO-based ones, Jeffamine-based self-standing SPEs display relatively high ionic conductivities even under room temperature and enhance chemical and electrochemical properties, which are promising alternatives to PEO-based SPEs for SSLMBs.
Additionally, the Jeffamine-based compounds can be used as other components of batteries, such as polymer adhesives, oxidized active materials, and interface coatings, thus improving the mobility of Li + ions, electrochemical performance, and chemical and electrochemical compatibility toward lithium electrodes [102].

Polycarbonate and its derivatives
Compared to polyether matrices, polycarbonates have attracted extensive attention from the battery community, due to their stronger oxidation resistances than polyether-type polymer matrices [111]. Commonly, polycarbonate-type polymers utilized for SPEs include poly(ethylene carbonate) (PEC), poly(propylene carbonate) (PPC), and poly(trimethylene carbonate) (PTMC). Interestingly to note that the values of T g for these three kinds of polycarbonate matrices are much higher than that of PEO, i.e. T g = 18 • C (PEC) vs.
, indicating that the segmental motions in these polymers are rather difficult [111][112][113]. Meabe et al [114] comparatively investigated the ion transport mechanism of SPEs based on polyether and polyesters comprising LiTFSI as the lithium salt and the blend of poly(ε-caprolactone) (PCL) and PEO as the polymer matrices. It is reported that LiTFSI is inclined to coordinate with the carbonyl group (-OC=O-) in PCL backbone in the case of PCL content >50 mol%, and there are plenty of compact ion pairs. Lithium salt can be effectively dissolved, and Li + cation is preferentially coordinated with oxyethylene units in the PEO structure, in case of PCL content <50 mol%. Based on these understandings of coordination and phase separation, the ionic conduction mechanism of polyether/polyester hybrid systems is elucidated (figure 7) [114].
In addition, the polycarbonate-based SPEs show high ionic conductivity compared with polyether-based ones even at room temperature [111,115,116], for example, 3.0 × 10 −4 S cm −1 under 20 • C for cellulose nonwoven/PPC reported by Cui and co-workers [115], and 10 −4 S cm −1 for LiFSI/PEC under 30 • C reported by Tominaga et al [117]. Further investigations suggest that low molecular weight components (e.g. PC and EC), originating from the chemical decompositions of PPC and PEC, are responsible for the unexpectedly high ionic conductivities observed for the polycarbonate-based SPEs (figure 8) [118,119].
Note that the high interfacial reactivity occurred between polycarbonate and lithium electrode, which affects their chemical and electrochemical stability to some extent. Wang et al [120] revealed that significant side reactions between PPC electrolyte and lithium electrode at the elevated temperature (80 • C) generate liquid components such as PC (as shown in figure 8), which brings great safety risks. Therefore, to date, polycarbonate-based SPEs are still under basic research in the laboratory, which cannot meet the application requirements of SSLMBs at current stage.

Single lithium-ion conductive SPEs
Typically, the classic SPEs obtained with discrete anions are typical dual-ion conductors (T Li + < 0.4), in which both negative and positive charges could migrate under the electrics field [84,121]. During charge/discharge cycles, the migration of anionic species (in opposite directions vs. cationic species) gradually causes concentration gradient and internal polarizations of redox reactions, which finally accelerates dendrite growth and parasitic reactions at electrode-electrolyte interphases/interfaces [122,123]. Therefore, the selectivity of cation transport is of vital importance for the stable operation of SSLMBs [124].
To suppress or even eliminate the migration of the negative charges in SPEs, a new type of SPEs has been suggested, which is known as single-ion conductive SPEs (SLIC-SPEs). Generally, SLIC-SPEs with the values of T Li + close to unity are majorly obtained by three approaches (figure 9) [17,122,124,125]: (a) chemically grafting the anions on polymeric backbones (figure 9(a)); (b) covalently bonding the anions of lithium salts on the inorganic backbone ( figure 9(b)), and (c) incorporating anion acceptors to cage the anions in dual-ion conductive SPEs (figure 9(c)). Currently, several excellent reviews have systemically discussed the SLIC-SPEs built from the latter two approaches (utilizing inorganic backbones and anion acceptors [126,127]), and we will mainly focus on the SLIC-SPEs made from the first method, i.e. attaching anions to polymeric backbones in chemical means ( figure 9(a)).
For typical dual-ion SPEs, lithium salt and polymer matrices are indispensable. Inheriting the same concept, the research activities in the polymer backbone-based SLIC-SPEs could be briefly presented in two aspects: (a) rational design of the anionic center, aiming to improve the dissociation of lithium ions and thereby provide higher concentrations of active ions; and (b) regulating the topological structures of polymer backbones, with the objective of facilitating rapid ion transport through ionic sites [124]. Some representative SLIC-SPEs and their basic properties are also summarized in table 4.
The key achievements in both aspects are presented in the following sections.

Rational design of anionic center
Since the 1980s, the effect of anionic structures on Li + conductivities of SLIC-SPEs has been continuously investigated by various research groups [124]. Early attempts carried out by Tsuchida et al [128] focused on a polymeric lithium salt based on carboxylate anions (−CO 2 − ). These carboxylatebased SLIC-SPEs presented extremely low ionic conductivities even at high temperatures (ca. 10 −8 S cm −1 under 60 • C), owing to the strong affinity of carboxylate anions towards Li + cations [124]. Afterward, Bannister et al [129] suggested the incorporation of perfluoroalkyl chains, and the as-obtained SLIC-SPEs showed nearly two orders of magnitudes improvement in ionic conductivities (10 −6 S cm −1 under 60 • C), as compared to those based simple alkyl carboxylate groups. This suggests that replacing hydrocarbon alkyl groups with perfluoroalkyl groups can facilitate lithium-ion dissociation of carboxylate anions, thus improving the ionic conductivities of SLIC-SPEs.

Regulating the topological structure of polymer backbones
For polymer backbone-based SLIC-SPEs, PEO is utilized as a polymer matrix to facilitate Li + transportation for those polysalts without solvating units for lithium ions; yet, the high degree of crystallinity of PEO greatly slows down the migrations of ionic species under ambient temperatures (<60 • C) [123]. Therefore, the topological structures of polymer backbones are also key factors determining the ionic conductivities of SLIC-SPEs [122,123]. Polymerizing salt monomers with soft monomers is an effective pathway to decrease the degree of crystallinity and T g , and accelerate Li + transport [122,135,136]. Besides, copolymerization could also circumvent the possible phase separation of the blended SLIC-SPEs, and improve not only the long-term durability of electrolyte membranes but also interphase contact between electrodes and electrolytes [122,137].
The common copolymerization methods are random copolymerization, block copolymerization, and homopolymerization and so on (figure 11) [124]. Random copolymerization can decrease the degree of crystallinity of SLIC-SPEs and T g , thus, promoting Li + transport among chain segments. In addition, the performance of polyanion type SPEs can be further improved with substituting high crystallinity PEO with a flexible oligomeric EO.
Tsuchida et al [140] reported copolymerized SLIC-SPEs via copolymerization of lithium methacrylate monomers with oligo(oxyethylene) methacrylate monomers (figure 11(c)), which displayed ionic conductivities of 1.6 × 10 −7 S cm −1 even under room temperature. Park and co-workers [133] reported a similar copolymerized SLIC-SPEs by copolymerization lithium acrylamide hexanoate monomer and oligomeric (oxyethylene) methacrylate monomer, of which the conductivity was 1.5 × 10 −7 S cm −1 under room temperature. For the obtained two irregular copolymerized SLIC-SPEs, the conductivity reaches up to ca. 10 −7 S cm −1 . That is, for the sulfonate anion based SLIC-SPEs, the ionic conductivity could also be elevated by copolymerizing with the oligomeric EO segments. Actually, two orders of magnitude higher conductivities can be obtained in case fluorinated polymer monomers are applied [135,136].
Our group [120] reported several kinds of amorphous SLIC-SPEs (Li[PSTFSI-co-MPEGA]) via copolymerizing with different LisTFSI with methoxy polyethylene glycol acrylate (MPEGA) ratios. The ionic conductivity of the Li[PSTFSI-co-MPEGA] copolymer electrolytes are higher by 1-3 orders of magnitude than these of LiPST-FSI/PEO blended electrolytes (7.6 × 10 −6 S cm In short, compared to the traditional double-ion conductors, the SLIC-SPEs are the typical single-ion conductors, and show several traits: (a) relatively high T Li + (close to 1); (b) impressing the growth rate of lithium dendrite attributed to effectively avoiding concentration polarization caused by anion migration; (c) impressing parasitic reactions b/w electrolytes and electrodes, particularly with lithium electrode, thereof, reducing the accumulation of SEI film products (i.e. produces a thin, dense and stable SEI interphase) and (d) robust chemical and electrochemical stability under high voltage region as well as the improved capacity of SSLMBs.

Conclusion
Compared to the commercialized liquid electrolyte, SPEs have several advantages, including ease of process, and intrinsic safety. Although the SSLMBs have been exemplarily applied, continuous efforts are still needed to further improve intrinsic safety and compatibility with electrode materials, as detailed below: (a) For current LiTFSI/PEO systems, conductive lithium salt structure, R F 1 −SO 2 −N (−) −SO 2 −R F 2 (R F 1 , R F 2 ; R F 1 = n-C x F 2x+1 , R F 2 = n-C x F 2x+1 ), could be further modified to extend the degree of negative charge delocalization, in the hope of lowering the crystallinity degree of PEO and promoting the transport of Li + ions [141]. (b) Compared with the linear PEO matrix, the comb polymers with multi-branch chain structure show lower T g and degree of crystallinity, which is contributed to amorphous SPEs at ambient temperature. Note that overcoming polymer crystallization can not only improve the conductivity of SPEs under ambient temperature but also enhance the stability of the electrode and electrolyte interphase. (c) For SLIC-SPEs systems, improving the dielectric constant of the polymer matrices seems to be an effective strategy to promote the dissociation of Li + , and thereby improve the Li + transport capability therein. In addition, one may also attain high T Li + values by grafting anions to polymer nanoparticles via semi-batch emulsion polymerization [142].
In short, by the structure design of conductive lithium salt and polymer matrices, the ion transport characteristics and physical and electrochemical properties of SPEs could be significantly enhanced. Bridging the research activities between academia and industrial sectors could certainly promote the pragmatic development of SPEs.