1 Introduction
Electric vehicles and smart wearable devices are becoming much popular in daily life, pushing urgent demand of advanced energy storage devices with higher energy density and safety performance [1]. Unfortunately, limited by the intrinsic low theoretical specific capacity of graphite anode (372 mAh/g), commercial lithium-ion batteries struggle to meet the growing requirement of these devices. Consequently, as the “holy grail” electrode, lithium metal anode is gathering ever-increasing attentions due to its ultrahigh theoretical specific capacity (3860 mAh/g) and low redox potential (–3.04 V versus standard hydrogen electrode), which could possibly be used in the next generation of lithium-based batteries [2–4]. However, the application of LMBs is still restrained by lacking mature electrolyte systems [5]. For instance, the commercial organic electrolytes usually suffer from severe safety risks like leakage and flammability [6,7], which have been reported in newspaper for many times. Simultaneously, many commercial organic electrolytes possess limited electrochemical stability to lithium metal anode [8], leading to severe side reactions and mechanically unstable solid electrolyte interphase (SEI) layers which will continuously consume the active lithium and liquid electrolytes. What is worse, the resulting Li dendrite growth and uncontrollable volume expansion of Li metal anode can further accelerate side reactions in turn and finally give rise to destructive internal short-circuit, thermal runaway, or even spontaneous combustion of the whole batteries [9,10]. Due to the fact that all these disadvantages hinder the practical application of liquid organic electrolyte, LMBs with SSEs have gradually become research hotpots because of their superior safety merits such as enhanced mechanical rigidity and nonflammability [11].
SSLMBs are composed of lithium metal anode, SSEs, and cathodes, in which liquid organic solvents are rarely used. To guarantee their energy density and safety performance, the preparation of high-performance SSEs is obviously becoming the top priority. In this case, efforts have been made to design new materials and study the internal mechanism of SSEs. According to their chemical composition, the SSEs used in lithium metal batteries mainly include inorganic solid electrolytes (ISEs) and solid polymer electrolytes (SPEs). The inorganic ceramic electrolytes, for example, garnet [12], argyrodite [13], NASICON [14], and perovskite [15] have all been studied for their favorable ionic conductivity, but they are still confronted with unsatisfactory mechanical performance with fragile property, not conducive to large-scale processing and favorable long-term cycling. The SPEs like PEO, PAN, and PMMA, on the contrary, exhibit enhanced flexibility and interphase wettability with the electrodes [16], while they are usually faced with poor room-temperature ionic conductivity [17,18], lower lithium-ion transference number [11], and limited electrochemical window [19]. Therefore, combining SPEs and inorganic fillers to synthesize CSEs has caught researchers’ eyes to address these shortcomings, where ceramic fillers can be dispersed into polymer matrix as they can reduce the crystallinity of polymers and enhance the ionic conductivity, mechanical performance, as well as the electrochemical stability. Moreover, based on the ability to conduct lithium ions, inorganic fillers mainly contain active fillers and passive fillers, among which active fillers are believed to be more effective in enhancing the ionic conductivity of CSEs by providing additional ion-conducting pathways. Therefore, the CSEs with active fillers are expected to take full advantages of both SPEs and ISEs, instead of their counterpart with passive fillers [20].
According to the ceramic mass fraction, CSEs can be divided into two categories. If the content of inorganic fillers is less than 50 wt.%, they are usually described as PR electrolytes. Otherwise, they are classified in CR electrolytes [21]. Although CSEs combining the merits of both ISEs and SPEs are considered as one of the most promising candidate electrolytes for secondary SSLMBs, there are still several pending problems for CSEs design which should be fully understood.
First, the lithium-ion conduction pathways may be distinct in various CSEs because of their diverse ceramic categories and internal structure [22], which deserves comprehensive consideration. For CSEs with inorganic active materials, the lithium-ion transport mainly occurs in the bulk of ceramics and polymer phase, as well as the interphase between them. Significantly, the smooth Li+ pathways always exhibit dynamic change among these three regions whose interactions directly determine the overall ionic conductivity of CSEs. Therefore, in electrolyte design, the contribution of different pathways for ion transport is expected to be systematically studied to achieve efficient ion transport channels by regulating the internal structure or interaction intensity.
In addition to the kinetic properties of CSEs, interface compatibility is also an important factor which can dramatically affect the electrochemical properties and should not be overlooked [23]. For the complete solid-state battery system, there usually exists the issue of interface compatibility around the interphase region between polymer and the ceramic phase and the contact area of the electrolyte and both electrodes. Notably, the former usually causes sluggish ion transport through the interphase and severe agglomeration of ceramic components especially with granular structure, while the latter is closely related to the redox decomposition of electrolyte, the consumption of active materials, and the destruction of electrode structure [24]. In short, the introduction of coupling agents, additives, or specific buffer layer may be conducive to tackling these tough issues.
The mechanical properties of CSEs should also be focused on, as they have significant impacts on the service ability and safety behavior of the whole SSLMBs. At present, the addition of ceramics has been proposed as an important and effective strategy to enhance the mechanical strength of CSEs [21]. However, according to Monroe and Newman’s model, only electrolytes possessing sufficient ceramic content and at least twice shear modulus over lithium metal can effectively suppress the Li dendrites [25]. Therefore, the strength, size, and addition amount of ceramic components need to be carefully considered. It must be pointed out that excessive ceramics content can also deteriorate the self-standing and film-forming abilities of CSEs, leading to the destruction of the electrolyte structure and the reduction of processability, such as chap or brittle fracture.
Herein, recent advances and limitations of both PR and CR systems with typical inorganic components are summarized and compared. In particular, the factors affecting the above-mentioned issues of CSEs, such as the categories and structure of ceramics, are systematically discussed in order to promote more rational design of the effective preparation process (Fig.1). The possible development directions in the future for PR and CR systems are respectively predicted, and the remaining challenges, promising strategies and design principles for subsequent academic research and industrialization are presented. This review is expected to render a deeper insight into the favorable CSEs and further promote the commercialization of SSLMBs.
2 PR electrolytes
As for PR electrolytes with low ceramic content, ceramics is usually dispersed in the polymer matrix as inorganic fillers. According to the crystal structure and chemical components, active fillers can be mainly divided into garnet-type, perovskite-type, NASICON-type, sulfide, and hydride. Based on their physical structure, they can be divided into particles, nanofibers, nanosheets, and frameworks. In this part, advanced PR electrolytes with various categories and structural designs of fillers are systematically discussed.
2.1 Categories of fillers
2.1.1 Garnet-type fillers
Garnet-type inorganic fillers have been widely used in CSEs because of their high ionic conductivity, excellent chemical stability with Li metal, reliable security in the air, and wide electrochemical stability window (0–6 V versus Li+/Li). The typical chemical formula of garnet is Li7La3Zr2O12 (LLZO) with both tetragonal and cubic phase. It is worth mentioning that the cubic phase LLZO exhibit a higher ionic conductivity but is thermodynamically unstable [26]. To improve its thermodynamic stability, metal cations such as Al3+, Ga3+, Ti4+, Nb5+ and Ta5+ have been doped into the lattice [27,28], which can increase the lithium-ion vacancy and reduce the free energy of the entire system.
For this type of PR electrolytes, the internal interactions with lithium salts can be significantly improved by adding garnet-type fillers, which can further give rise to optimized electrochemical performance. In the electrolyte with mobile anions, free anions tend to move in the opposite direction to cations, becoming a barrier for cation transportation in an applied electric field. Therefore, a large concentration gradient of lithium ions is formed from the bulk electrolyte to the anode surface. More seriously, such a gradient becomes more predominant with the time of Li deposition extending [29]. However, the addition of Li6.75La3Zr1.75Ta0.25O12 (LLZTO) nanoparticles could bring an anion-immobilization effect in the flexible PEO/LLZTO (PLL) composite electrolyte (Fig.2). Surprisingly, anions in this PLL electrolyte are tethered by the ceramic fillers, inducing a uniform distribution of space charges and lithium ions which contribute to a dendrite-free lithium deposition. The dissociation of anions and lithium ions also help to reduce the polymer crystallinity, rendering stable and fast transportation of lithium ions.
Moreover, garnet-type fillers can also strengthen the interaction between polymer and ceramic by the Lewis acid-base effect, which has been found in PVDF matrix with LLZTO [30]. Systematic characterizations and calculations reveal that La atoms in LLZTO could couple with the N atoms and C=O groups of typical solvent molecules like DMF along with the N atoms in high-electron-density state, which behave like a Lewis base and led to partial dehydrofluorination of PVDF skeleton (Fig.2). Partially modified PVDF chains activate the interactions between the PVDF matrix, lithium salt, and LLZTO fillers (Fig.2, Fig.2), leading to a significantly improved performance of the electrolyte. Similar dehydrofluorination reaction resulting from the Lewis acid-base effect is also observed between Li6.75La3Zr1.75Nb0.25O12 (LLZNO) and the PVDF polymer (Fig.2–Fig.2) [31]. Additionally, benefitting from the strongly coupled effects via interfacial chemical reactions and the synergistic effects between LLZNO and PVDF, the LLZNO-based CSE membrane present an excellent thermal stability and electrochemical performance at both ambient and high temperatures.
2.1.2 NASICON-type fillers
NASICON-type inorganic fillers have attracted wide attention due to their chemical stability in ambient environment, high ionic conductivity at room temperature, and wide electrochemical stability window. Notably, when the element Na is replaced by Li, NASICON-type inorganic fillers can still retain their original crystal structures and turn into lithium-ion-conduct inorganic fillers. Similar to garnet-type electrolytes, substitution of metal cations with various valences in the framework for NASICON-type host materials (such as LiTi2(PO4)3 and LiGe2(PO4)3) is a valid approach to optimize their electrochemical performance [32]. The typical chemical formulas of NASICON-type inorganic fillers are Li1+xAlxTi2–x(PO4)3 (LATP) and Li1+xAlxGe2–x(PO4)3 (LAGP).
Benefitting from the high ionic conductivity of NASICON-type ceramics, the combination of inorganic fillers and polymer matrix can effectively improve the overall ionic conductivity of CSEs. In Fig.3, CSE membranes with 10 wt.% LATP in PVDF-HFP matrix show the highest ionic conductivity of 2.3 × 10−4 S/cm at room temperature, which is three times higher than that of bare polymer electrolyte (7.1 × 10−5 S/cm) [33]. The enhancement of ionic conductivity is mainly attributed to the modification of polymer around the LATP particles, which could create an interfacial region that is the key to reducing the activation energy for cation transport and enhancing the ionic conductivity (Fig.3) [34]. However, a higher ceramic content is not always positive correlated with the ionic conductivity, as the agglomeration problem can also be noticed with the increase of ceramic content (Fig.3(c)), which significantly reduces the ability of the interphase to expand into adjacent bulk polymer, and consequently the interphase volume fraction and conductivity.
In addition to normal polymer matrix like PEO or PVDF, NASICON-type also exhibit a satisfying compatibility with other polymers, such as polymerized ionic liquids (PILs, Fig.3) [35]. These researches prove that the ionic conductivity could also be influenced by polymer matrix, which inspires researchers to focus more on the selection of rational polymer materials.
2.1.3 Perovskite-type fillers
Perovskite-type inorganic fillers usually exhibit a desirable chemical stability to moisture or air, and an excellent thermal stability, whose typical chemical formula is Li3xLa2/3–xTiO3 (LLTO) with a cubic phase crystal structure [15]. However, the Ti4+ in the LLTO can be reduced to Ti3+ by lithium metal in a low-voltage condition, resulting in the consume of active lithium metal, increased electronic conductivity of LLTO, and destruction of the crystal structure, leading to a restrained large-scale application to some extent.
To resolve such a catastrophic problem, various strategies like surface coating and element substitution have been proposed to strengthen their stability. For instance, biodegradable polydopamine (PDA) layers can be used as the interface coating on LLTO particles by a facile wet chemical method [36]. For the composite electrolytes composed of PDA-coated LLTO particles and PVDF matrix (Fig.4), the modified inorganic phase delivers a superior stability against lithium metal anode. Besides, owing to the abundant catechol and imino structures in PDA which strengthen the interfacial bonding between LLTO filler and the PVDF matrix, the as-prepared PR electrolyte membrane performs an outstanding flexibility, superior interfacial stability, high ionic conductivity, and wide electrochemical window.
Except for surface coating, substituting Ti4+ with other high valence metal cations is another effective way to enhance the stability of perovskite-type fillers. The substitution of metal cations like Ta or Hf could increase the lithium-ion vacancies and promote the ion transmission. Moreover, some elements are also proved to bond with anions. For instance, Ta5+ from perovskite-type Li3/8Sr7/16Ta3/4Zr1/4O3 (LSTZ) filler is able to bond strongly with F– in the TFSI– anion of LiTFSI (Fig.4–Fig.4) [37], thus facilitating the release of free lithium ions from LiTFSI, which significantly increases both the ionic conductivity and the lithium transference number. Similar conclusion is proved in a flexible CSE membrane combining highly conductive Li0.38Sr0.44Ta0.7Hf0.3O2.95F0.05 (LSTHF), PVDF and LiClO4 [38]. Notably, as shown in Fig.4(e) and Fig.4(f), the substitution of Hf could decrease the hopping barrier of lithium ions (0.35 eV), making a great contribution to the remarkable ionic conductivity of CSEs.
2.1.4 Sulfide fillers
Compared with the oxide electrolyte, the binding of lithium ions to the sulfide-type host framework is weakened with the reduced ion transport energy barrier, which is mainly due to the larger ionic radius and weaker electronegativity of S2–. Thus, sulfide-type ISEs possess supreme ionic conductivity among all kinds of ISEs. In addition, their lower grain boundary resistance and remarkable machinability can simplify the fabrication process. According to their different crystal structure, sulfide-type ISEs can be divided into the thio-LISICON-type [39], argyrodite-type [40], and LGPS-type [41], of which the LGPS-type ISEs have the highest ionic conductivity (25 mS/cm at room temperature) [42]. Unfortunately, sulfide-type electrolytes are extremely sensitive to the moisture in the air, tending to produce toxic hydrogen sulfide and decrease the ionic conductivity [43]. Moreover, their poor chemical stability to lithium metal anode and narrow electrochemical windows will further exacerbate side reactions and accelerate dendrite growth in SSLMBs [13].
Luckily, polymer/sulfide CSEs have indeed offered an opportunity to tackle these tough issues, as polymer matrix can protect the sulfide-type fillers from reacting with moisture, lithium metal or high-voltage cathodes, and enhance their flexibility at the same time. By simply embedding Li7PS6 (LPS) into PVDF-HFP matrix, the hydrolysis reaction between LPS with moisture in ambient environment and the side reactions with electrode materials in solid-state batteries is prevented (Fig.5), resulting in ameliorative chemical and electrochemical stabilities [44]. In turn, the presence of LPS is able to facilitate the ion conduction by creating more amorphous regions, thus enhancing the electrochemical behavior of the CSE (Fig.5–Fig.5).
In addition to the tough issue of undesirable thermodynamic stability, the limited interphase compatibility in the sulfide-based PR electrolytes, which can lead to severe agglomeration of particles and block the lithium-ion migration, should also be noticed. In this case, it is noted that introducing surface modification additives is an effective way to ameliorate the interaction between polymer matrix and inorganic fillers. As for the typical PR electrolyte based on PVDF-HFP and Li10GeP2S12 (LGPS), perfluoropolyethers (PFPEs) can work as a bifunctional adjuvant (Fig.5) [45]. Particularly, PFPEs with a low molecular weight facilitate stable dispersion of LGPS in casting solution, which is ascribed to their strong electronegativity of C-F bonds. Moreover, PFPEs with a high molecular weight function as an interfacial stabilizer, leading to a dramatically improved interfacial compatibility with Li anode by in situ forming a LiF-rich SEI layer. Similar interaction enhancing effect is achieved by a commercialized silane coupling agent, (3-chloropropyl) trimethoxysilane (CTMS), which is used as a bridge builder to realize the chemical bonding interaction of LGPS with polymer phase (Fig.5) [46]. With the help of CTMS, an in situ coupling reaction is held between LGPS and polymer segments. Notably, LGPS are integrated tightly with polymer matrix, which cannot only eliminate the phase boundary between EO chains and LGPS, but also reduce the grain boundary resistance among LGPS particles. According to the researches above, introducing effective additives with specific functional groups or structures would be a universal strategy to enhance the interphase compatibility, which may also be applied to other CSEs.
2.1.5 Hydride fillers
Since LiBH4 is proved to achieve a high ionic conductivity of 10−3 S/cm at around 120°C [47], hydride ISEs have caught researchers’ eyes, which are composed of metal cations like Li+, Na+, and complex anions like BH4−, NH2–, B12H122–, etc. Benefitting from their unique chemical components, hydrides possess a better stability with lithium metal anode, lower density, and favorable processing performance. However, their practical application is still hindered by the limited room-temperature ionic conductivity. Meanwhile, due to the strong reducibility of H, hydrides exhibit a poor compatibility with high-voltage cathodes and a high sensitivity with moisture [48]. To overcome these obstacles, strategies like ion substitution, second-phase incorporation, and interface engineering have been applied to enhance their performance [49].
Besides, combining hydrides and polymers has also been proved to be an effective way to enhance the ionic conductivity and chemical stability of electrolytes. Prepared by a simple solution-casting method, the PEO10-Li4(BH4)3I electrolyte exhibits a relatively high lithium transference number of 0.45, which is probably attributed to the interaction between BH4− and terminal hydroxyl groups of PEO (Fig.6) [50]. Simultaneously, its electrochemical window can reach about 3.6 V, able to serve as the electrolyte candidate for Li-S batteries (Fig.6). Likewise, by simply adding Li2B12H12 into PVDF-LiTFSI matrix, a higher electrochemical window of 4.0 V can be obtained and match with LiFePO4 cathode [51]. Furthermore, the strong adsorption force between TFSI− and Li2B12H12 can further enhance the immobilization of TFSI−, resulting in the remarkable Li+ conduction properties inside the electrolyte (Fig.6).
It can be concluded that hydrides possess a great potential for SSLMBs on account of their superior compatibility to lithium metal, but efforts still have to be made to further enhance their high-voltage stability. Therefore, the combination of hydrides and polymers with high-voltage tolerance like PAN and PVDF would be a promising tactic, and the application of interface stabilizing additives also needs more attention.
The lithium-ion migration mechanism has a guiding significance for the structural design of CSEs. Based on the above-mentioned researches, lithium ions can migrate via three types of pathways in CSEs: the polymer phase, the ceramic phase, and the interphase region [21,45]. For the PR electrolytes with a lower ceramic content, the isolated particles in the polymer matrix can provide continuous lithium-ion channels neither inside the particles nor at the polymer/ceramic interphase. As a result, the ion conduction in the CSEs mainly rely on the pathways through polymer phase Therefore, both passive and active fillers can promote the ion conduction mainly by decreasing the crystallinity of polymer (Fig.7, Fig.7). The contribution of intrinsic conductivity from bulky ceramics to the overall ionic conduction of CSEs is severely limited.
Following the increase of ceramic content, bulk interphase conduction can be achieved in the continuous interphase region formed by connected ceramic particles. In this case, CSEs possessing active or passive fillers with continuous ion conduction through abundant interphase can afford much a higher ionic conductivity [52]. The diversity of their ionic conductivity could be attributed to the additional lithium-ion pathways brought by active fillers (Fig.7, Fig.7).
It is concluded that the ionic conductivity of PR electrolytes with active fillers is influenced simultaneously by the ion transport ability of ceramic fillers, the interphase region, and the polymer phase. Consequently, attentions must be paid to both bulky phase and interphase conduction during the design of structural inorganic fillers. Designing inorganic fillers with special structures would become an effective way to build continuous lithium-ion migration pathways in both the interphase region and the ceramic phase.
2.2 Structural design of fillers
Based on the above discussion, there are still many deficiencies for granular ceramic fillers. First, the notorious agglomeration of nanoparticles can easily be observed for CSEs with a high ceramic content. In addition, in the PR electrolytes, the isolation of ceramic particles will interrupt the continuity of the interphase and limit the conductivity. The above-mentioned problems will result in a catastrophic phase separation and blocked lithium-ion migration through the bulky electrolyte. Hence, exploiting versatile inorganic active fillers with specific structures to tackle these tough issues of CSEs has received widespread attention.
2.2.1 Nanofibers
Inorganic fillers with 1D structure like nanofibers (NFs) or nanowires (NWs) usually possess a high aspect ratio, which can provide better mechanical strength characteristics than the corresponding nanoparticles. Besides, they also possess an excellent flexibility and prolonged lithium-ion pathways compared with nanoparticles. When fabricated into an interlacing network, the excellent flexibility and ionic conductivity of 1D fillers can significantly improve the electrochemical performance of CSEs.
Researchers have found that the NFs in CSEs can provide continuous lithium-ion pathways at the interphase and interactions with polymer, which is influenced by their dispersion state in the polymer matrix (Fig.8). The LLZTO NFs with an appropriate concentration could connect with each other, providing a large interphase with continuous and rapid pathways for ion transport [53]. Meanwhile, the LLZTO NFs also assist in preventing the recrystallization of polymer matrix such as PEO and enhancing the ion-conducting ability in polymer phase. Moreover, the polymer-in-salt based PEO electrolyte is able to facilitate the lithium-ion conduction by immobilizing the TFSI− anions in the form of anion clusters and decreasing the crystallinity of PEO. Except for PEO, active fillers like LLZO NFs can also have intense interactions with PAN. Yang et al. elucidated that LLZO NFs partially modified the PAN matrix and created preferential pathways for Li+ conduction through the modified polymer regions, thus significantly enhancing the ionic conductivity (Fig.8–Fig.8) [54]. However, the interaction mechanism between them still needs deeper investigation.
Similar to the research of LLZTO particles [30], there also exists a Lewis acid-base effect in the PVDF/LLTO NFs composite electrolyte [55]. As shown in Fig.8(e)–Fig.8(h), the LLTO NFs mixed with NMP produce a Lewis-like alkaline and further induce the partial dehydrofluorination of PVDF, which enhances the interactions between PVDF, LiTFSI, and LLTO NFs and consequently delivers an optimum behavior. These researchers also elucidate that the Lewis acid-base effect is probably ubiquitous in the PVDF/LLZTO based systems, regardless of the structure of LLZTO.
However, the anisotropic characteristic of NFs is not noticed by the above-mentioned researchers. To further enhance the ionic conduction, the orientation of NFs would be a nonnegligible decision factor. Compared with randomly dispersed LLTO NFs, well-aligned LLTO NFs is proved to possess a higher ionic conductivity (Fig.8–Fig.8) [56]. The well-aligned NFs dispersed in polymer matrix with minimal overlap could increase the volume fraction of polymer/ceramic interphase to the most extent and optimize the lithium-ion pathways. It is proved that well-aligned LLTO NWs exhibit an enhancement of ten times in ionic conductivity compared with randomly dispersed NWs. Such appreciable conductivity improvement reveals that the orientation optimization of NFs can be regarded as an effective and universal tactic to promote the ion migration, which may be applied to other orientated materials.
2.2.2 Nanosheets
Compared with nanoparticles and nanowires, nanosheets (NSs) inorganic fillers in the PR system have been rarely reported in the past few years, probably because of the lacking of mature preparation methods. Actually, 2D structured NSs possess the characteristics of an ultrathin layered structure and a large aspect ratio, which can lead to larger active interphase regions with polymer matrix as well as the ceramic phase for ion conduction, and further improve the ionic conductivity and mechanical strength of CSEs.
In general, 2D fillers can afford to optimize the ion transfer path in the bulky CSEs. For example, garnet-type LLZNO NSs obtained through a coprecipitation method with graphene oxide template could be used as the fillers in the PEO matrix [57]. Compared with nanoparticles, LLZNO NSs construct an interconnected structure that offers a continuous lithium-ion transport pathway in the CSE, providing a higher ionic conductivity and structural robustness (Fig.9). Similarly, due to the abundant and fast interphase lithium-ion transfer channels and sheet-like obstacles shield established on LLZAO NSs, the corresponding CSE exhibits an enhanced ionic conductivity and a promoted capability to suppress the growth of lithium dendrites (Fig.9) [58]. In comparison with branch-like fillers, LLZAO NSs exhibits an excellent continuity on both the interphase and ceramic phase. Briefly, although NSs have significant advantages, such as stronger mechanical strength or more continuous lithium-ion pathways compared with NFs, subsequent efforts are still essential to develop more effective fabrication methods and reveal the interaction mechanism of NSs in CSEs. In addition, it is worth noting that researches on the orientation of NSs may become an important hotspot to enhance the ionic conductivity.
2.2.3 Frameworks
Compared to the NFs and NSs, self-standing 3D structured frameworks possess obvious advantages. The pre-fabricated frameworks via the sol-gel method [59], have an excellent mechanical performance, which can play the role of mechanical support and restrain the dendrite growth. The interconnected porous frameworks can simultaneously avoid particle agglomeration and provide continuous lithium-ion pathways in the ceramic phase and the interphase region (Fig.10–Fig.10).
Except for the physical inhibition, chemical inhibition of lithium dendrite could also be achieved by chemical modification of frameworks. For example, the 3D fluorinated LLTO (F-LLTO) NFs network, which is hybridized with PEO (Fig.10(c)) could not only provide continuous lithium-ion pathways, but also react with active lithium metal and form a LiF-rich SEI [60]. Surprisingly, the physical protection of the framework and the chemical protection of the LiF-rich SEI layer synergistically suppress the dendrite growth on the lithium metal anode.
In addition to the chemical modification, the space orientation of ceramic frameworks also has a great influence on the electrochemical performance of CSEs. Similar to the above-mentioned aligned NFs, vertically aligned frameworks can optimize the lithium-ion transmission path, homogenize the Li+ flux in the bulky electrolyte, and strengthen the mechanical performance (Fig.11(a)) [61]. The ice-templating method, for instance, has been proved to be a facile way for fabricating vertically aligned LAGP with a porous structure. The aligned ceramic phase in the CSE allows fast conduction of lithium ions, which is expected to be applied to other nanoparticles in the future. Based on the above-mentioned researches, upgrading traditional fillers with a 3D nanostructured framework may be a feasible way to improve the ionic conductivity and regulate the deposition morphology, as well as the safety behavior.
2.3 Brief summary
In brief, as an early developed mature electrolyte, PR systems have been studied for decades. Advanced technologies like machine learning and 3D printing have also been used in choosing and preparing materials [63,64]. As a result, varieties of active fillers are applied to PR electrolytes (as shown in Tab.1), which could have the inhibiting effect of crystallization and partial modification effect in the local regions around the nanoparticles. Benefitting from their high flexibility and safety, the PR system electrolyte have been widely used in LMBs with different materials or even in open systems like Li-air batteries (Fig.11) [62].
Tab.1 Summary of recently reported PR electrolytes |
| Structure | CPEs | Ionic conductivity | Electrochemical window (vs Li/Li+) | Li+ transference number | Activation energy | Ref. |
|---|---|---|---|---|---|---|
| Particle | 40 wt.% LLZTO-PEO | 1.12 × 10−5 S/cm at 25°C | 5.5 V | 0.58 | 0.83 eV | [29] |
| 10 wt.% LLZTO-PVDF | 5 × 10−4 S/cm at 25°C | – | – | 0.2 eV | [30] | |
| 7.5 wt.% LLZNO-PVDF | 9.2 × 10−5 S/cm at 25°C | 4.6 V | – | 0.48 eV | [31] | |
| 10 wt.% LATP-PVDF-HFP | 2.3 × 10−4 S/cm at 25°C | – | – | – | [33] | |
| 10 wt.% LATP-PEO | 1.7 × 10−4 S/cm at 20°C | – | – | – | [34] | |
| 10 wt.% LATP-PIL | 7.78 × 10−5 S/cm at 30°C | 4.55 V | 0.21 | 0.5 eV | [35] | |
| 15 wt.% LLTO@ PDA-PVDF | 1.1 × 10−4 S/cm at 30°C | – | – | 0.29 eV | [36] | |
| 20 wt.% LSTZ-PEO | 5.4 × 10−4 S/cm at 25°C | 5.25 V | 0.43 | – | [37] | |
| 10 wt.% LSTHF-PVDF | 5.3 × 10−4 S/cm at 25°C | 4.8 V | 0.5 | 0.126 eV | [38] | |
| 10 wt.% LPS-PVDF-HFP | 1.1 × 10−4 S/cm at 25°C | – | – | 0.41 eV | [44] | |
| 20 wt.% LGPS-PVDF-HFP | 1.8 × 10−4 S/cm at 25°C | 4.83 V | 0.68 | – | [45] | |
| 3 wt.% LGPS-PEO-PEG | 9.83 × 10−4 S/cm at 25°C | 5.1 V | 0.68 | 0.26 eV | [46] | |
| PEO10-Li4(BH4)3I | 4.09 × 10−4 S/cm at 70°C | 3.6 V | 0.45 | – | [50] | |
| 1 wt.% LBH-PVDF | 1.43 × 10−4 S/cm at 25°C | 4.0 V | 0.34 | – | [51] | |
| Nanofiber | 10 wt.% LLZTO-PEO | 2.13 × 10−4 S/cm at 25°C | 4.9 V | 0.57 | 0.45 eV | [53] |
| 15 wt.% LLTO-PVDF | 5.3 × 10−4 S/cm at 25°C | 5.1 V | – | – | [55] | |
| 30 wt.% N-LLTO-PVDF-HFP | 3.8 × 10−4 S/cm at 25°C | 4.9 V | 0.42 | 0.29 eV | [65] | |
| 3 wt.% aligned LLTO-PAN | 6.05 × 10−5 S/cm at 30°C | – | 0.42 | 0.77 eV | [56] | |
| 15 wt.% LLTO-PEO | 2.4 × 10−4 S/cm at 25°C | 5.0 V | – | 0.4 eV | [66] | |
| Nanosheet | 15 wt.% LLZNO-PEO | 3.6 × 10−4 S/cm at 25°C | – | – | 0.33 eV | [57] |
| 20 wt.% LLZAO-PEO | 4.3 × 10−5 S/cm at 25°C | 5.3 V | – | – | [58] | |
| Framework | 44 wt.% LLTO-PEO | 8.8 × 10−5 S/cm at 25°C | 4.5 V | – | 0.64 eV | [59] |
| Vertical aligned LAGP-PEO | 1.67 × 10−4 S/cm at 25°C | – | 0.56 | 0.45 eV | [61] | |
| Vertical aligned LATP-PEO | 5.2 × 10−5 S/cm at 25°C | – | – | – | [67] | |
| F-LLTO-PEO | 5 × 10−4 S/cm at 25°C | 6 V | 0.5 | – | [60] | |
| LLZO-PEO | 9.2 × 10−5 S/cm at 25°C | 5.1 V | – | – | [62] |
However, the PR system also has obvious shortcomings and limitations. As the content of inorganic fillers is relatively low in PR electrolytes, the isolation of ceramic particles will interrupt the continuity of the interface, which is detrimental to further improve the ionic conductivity. Besides, the high polymer content usually leads to the lower modulus of PR electrolytes, which cannot effectively inhibit the dendrite growth. Moreover, the intrinsic disadvantages of polymers may limit the thermal and electrochemical stability of PR electrolytes. To overcome such an intermittent transmission path, structural design has been applied to prepare fillers with different structures like NFs, NSs, and frameworks. Although structure design is proved to be an effective way to provide continuous lithium-ion pathways, granular fillers are still irreplaceable for now mainly due to their easier preparation process and balanced performance. Consequently, CR electrolytes with an ultrahigh ceramic content are getting more attention due to their enhanced mechanical strength, thermal stability, and fully percolated framework.
3 CR electrolytes
Different from the PR systems mentioned above, CR electrolytes with less polymer content allow lithium ions to transmit mainly across the ceramic phase. When the concentration of the active ceramic phase exceeds the threshold, a fully percolated framework that consists of connected ceramics can offer abundant lithium-ion expressways and promote the swift ion migration [11]. The conversion of ion transport pathways following the increasing ceramic content has been proved in typical PEO/LLZO electrolyte (Fig.12) [22].
Except for the discussion of ion transport pathways, researchers also emphasize that the PR electrolyte is more suitable for small-scale flexible energy storage devices owing to its higher flexibility and lower cost, while the CR electrolyte with a high mechanical strength and safety is more applicable to large batteries used in electric vehicles (Fig.12, Fig.12) [21]. Consequently, the intrinsic ionic conductivity of ceramic frameworks, as well as their structures and space orientation have a great influence on the electrochemical performance of CR electrolytes. Besides, the internal interaction mechanisms and different preparation methods are also the critical research objects. In this section, the categories and fabrication methods of inorganic frameworks in the CR electrolytes are systematically discussed.
3.1 Categories of frameworks
3.1.1 Garnet-type
Due to the ultrahigh ceramic content in CR electrolytes, the ion transmission mainly occurs in the ceramic phase. As there still exist polymers in the gap of ceramic, the ion migration at the polymer/ceramic interphase still matters to the bulk ion conduction. To strengthen the interaction between the ceramic and the polymer, and to optimize the interphase conduction, specific polymers are applied to the CR electrolytes. As shown in Fig.13–Fig.13, a uniform PAN coating on the garnet-type LLZTO surface could be obtained via a ball milling method, which is accompanied by the dehydrocyanation of polymer backbone and the generation of local conjugated structures [68]. Such unique internal structure and ultrahigh ceramic content facilitate the interparticle Li+ transport in the ultrathin composite electrolyte (< 10 μm).
Except for the lithium-ion transport kinetics, the agglomeration of ceramic can easily be observed in CR systems due to the poor interphase compatibility between concentrated particles and polymer matrix. To further strengthen the organic/inorganic interactions for CR electrolytes, the LLZTO phase can be modified by the PDA coating layer with a thickness of 4 to 5 nm in consideration of the superior wetting capability of dopamine (Fig.13) [69], leading to a uniform dispersion in the polymer matrix. The dual wetting capability of PDA on both the organic and the inorganic phase provides a strong interphase bonding, which effectively facilitates lithium-ion transfer through the electrolyte and enhances its thermal stability. In fact, similar to the above-mentioned silane coupling agent CTMS [46], dopamine could also play the role of surface modification additive, which proposes a universal and reliable method to optimize the interphase compatibility of CR systems.
To promote the practical application of garnet-based CR electrolytes, innovative large-scale fabrication technologies should be noticed as well. In consideration of their high ceramic content, mechanical mixing would be a proper strategy for facile preparation with a high homogeneity. Wang et al. [70] developed a solvent-free shear mixing method to fabricate self-standing CR electrolyte membranes (Fig.13). With only 0.5 wt.% PTFE as the binder, the mixture can be easily interwoven into a malleable and elastic flake, and further pressed into a freestanding membrane with a desired thickness via direct cold pressing. It is worth noting that this solvent-free shear mixing method can be applied to the preparation of other ceramics, and may even be extended to other metal batteries.
3.1.2 NASICON-type
The NASICON-type ceramic such as LATP and LAGP are widely used in CSEs, which demonstrates a good adaptability with different polymers. However, there is still a lack of researches on different ceramic content or distributions for NASICON-based CR electrolytes, which can greatly influence the electrochemical performance. For example, by comparing CR electrolyte with different LiSn2(PO4)3 (LSP) contents, Ahmed et al. have found that the lowest bulk resistance could be obtained with a LSP content of 70 wt.% (CCE-70, Fig.14–Fig.14), whose ionic conductivity is even higher than pure LSP pellet. Such a high ionic conductivity may be assigned to the dense internal structure of CCE-70, while pure LSP exhibits an anti-sintering behavior like cracking and poor inter-grain contacts [71]. Notably, the composite pellet with an ultrahigh LSP content (90 wt.%) is also fabricated by means of hot-pressing [72], yet the enhancement on the ionic conductivity and activation energy are not obviously observed. These researches indicate that the ceramic content is not always positively correlated with the ionic conductivity, and an optimal mass fraction of ceramics should be found during the process of design and preparation.
Except for the mass ratio, the electrochemical performance is also limited by the space dispersion of ceramics, which is always ignored. For example, unlike traditional CSEs with a smooth surface morphology, a free-standing PVDF-HFP/LATP composite membrane with two structurally different surfaces is applied to Li-Oxygen/Air batteries [73]. It possesses a LATP-enriched coarse surface on the upper side and a polymer-enriched smooth surface on the bottom (Fig.14–Fig.14). Such a Janus hierarchical structure is particularly suitable for Li-Oxygen/Air batteries: the smoother polymer-enriched surface can effectively protect LATP from reacting with lithium metal anode, while the coarse porous LATP-enriched surface offers numerous micro-channels not only for the oxygen/air diffusion pathway but also for the deposition of discharge product such as Li2O2. This hierarchical structure design is enlightening for optimizing the electrochemical performance of CSEs, which may be much proper for practical application. Despite the fact that many researches about NASICON-based CR electrolytes have been reported, the interaction mechanisms and design principles between ceramic and polymer phase still need further investigation.
3.1.3 Perovskite-type
Compared with garnet-type and NASICON-type ceramics, perovskite-based CR electrolytes have seldom been studied, which are mainly limited by its instability to lithium metal. Furthermore, the dependence on the supporting layer in the preparation process will increase its thickness, further inevitably reducing the overall energy density. To achieve a large-scale and optimized fabrication of perovskite-based CR electrolyte, the relatively mature method is the combination of tape-casting and sintering methods (Fig.15) [74]. During the preparation process of CR electrolytes, the tape-casting procedure is able to control the thickness by a height-adjustable scraper cooperating with PET release film, while sintering can reduce the grain boundary and porosity. With the aid of freestanding and dense electrolyte film, Li+ ions only need to transfer through several grains from side to side, significantly enhancing the ionic conductivity. This tape-casting and sintering method is probably widely available for the preparation of other composite electrolyte films.
3.1.4 Sulfide-type
Due to the extreme instability of sulfide-type ceramics to moistures, lithium metal anode, and high-voltage cathodes, researches on sulfide-based CR electrolytes are still in an insufficient state. What is worse, the high sulfide content will weaken the protection from polymer and exacerbate their instability to electrodes. In consideration of these problems, introducing additives, protect coatings, or interlayers would be feasible approaches to enhance the thermodynamic stability. For example, for the thin CR film composed of LGPS, PTFE and nylon mesh [24], to realize physical isolation between LGPS and the electrodes, an interlayer composed of TEGDA, BA copolymer is introduced to the anode side, and a LiNbO3 layer is coated on NCM532 to ensure the stability at the cathode side (Fig.15(b)). It turns out that the as-prepared battery system manifests a desirable operability, chemical stability, and a high energy density. In the future, specific materials possessing a high reduction stability with lithium metal such as polyethers could be used at the anode side, while at the cathode side, high-voltage durable polymers or plastic crystals may be further applied. Moreover, compounding with other electrolyte systems which exhibit a desirable redox stability is also a feasible strategy for electrolyte design.
3.2 Methods for producing ceramic framework
3.2.1 Electrospinning
For the above-mentioned CR electrolytes using ceramic particles, the preparation methods such as ball milling or tape-casting are relatively simple, leading to a non-oriented internal structure. However, owing to the anisotropic characteristics of ceramics with structural design, it would be more complex to fabricate CR electrolytes with ceramic components of NFs or NSs.
Electrospinning is widely used to fabricate fiber frameworks with an intertwined structure due to its convenient, controllable, and low-cost advantages. By forming an electrostatic field between the capillary tube and the collector, precursor solvent can be accelerated by the electrostatic force and form continuous nanofibers on the collector. For example, a LATP NFs framework is prepared via electrospinning and compounded with a single-ion conductor MEEP/LiSTFSI (Fig.16) [75]. The interconnected LATP framework provides continuous expressways for Li+ transmission, as well as a mighty mechanical supporting in electrolyte. Besides, the single-ion-conducting MEEP could immobilize anions by integrating a cross-linkable alkene group within the anion [76].
It is worth noting that electrospinning is always followed by high-temperature calcination to remove extra organic components in the precursor solvent. Furthermore, to prepare fiber frameworks with a better performance, specific factors should be taken into account: First, the viscosity of precursor solvent must be chosen carefully, as it is always positively correlated with the diameter of fibers. Next, although a larger electric field intensity can accelerate the precursor solvent with a higher speed, an exorbitant speed will destroy the continuity of fibers, which will further interrupt the continuity of ion transmission pathways. Finally, by changing the collector from static to rotating, randomly dispersed fibers will become vertically aligned, which can effectively enhance the electrochemical performance of electrolytes [56].
3.2.2 Solution dispersion
Compared with the electrospinning method, solution dispersion possesses a loose requirement for equipment with a more simplified preparation process. Unfortunately, the agglomeration problem will be much severe for the solution with a high ceramic content, eventually leading to a blocked Li+ migration in the CR electrolytes. To inhibit the agglomeration, surface modification additives like PDA and CTMS have been extensively reported, which can strengthen the interaction between the ceramic and the polymer phase, and thus enhance the interphase compatibility. Similarly, a silane coupling agent 3-(trimethoxysilyl)propyl methacrylate (KH570) could also be used to modify the LLZAO component (Fig.16) [77]. In the resultant CSE, silane-decorated LLZAO NFs are cross-linked along with the polymerization of monomers. This controlled fabrication of composite structures leads to a well-percolated network, forming continuous three-dimensional Li+ conductive pathways inside the CSE with a reduced activation energy. To improve the dispersion state of ceramics, the organic/inorganic compatibility between ceramics and dispersion mediums should be noticed in the future.
3.2.3 Vacuum filtration
Compared with ceramic particles or NFs, multilaminar CR electrolytes based on NSs usually possess an enhanced mechanical strength and a larger ion-conducting area. In particular, the unique multilaminar structure can be obtained via the vacuum filtration method, which can afford a durative force vertical to the filter membrane. With the help of the force, NSs can spontaneously stack with each other layer by layer, forming an overlapped multilaminar structure which can provide continuous ion-conducting expressways and resist the penetration of lithium dendrites. For example, the as-prepared PEO/LLZO NSs composite film (LLISE) through the vacuum filtration process possess a dense internal structure and a shortened lithium-ion diffusion distance (Fig.17) [78]. Meanwhile, multilaminar LLZO NSs in the LLISE optimize their compressive strength (3.2 GPa) with a thickness of only 20 μm, which effectively inhibits the growth of lithium dendrite. To further optimize the internal structure, hot pressing or cold pressing could be applied to obtain a denser electrolyte membrane.
3.2.4 Sol-gel synthesis
Just as the PR electrolytes, the conventional sol-gel synthesis is also widely used to prepare 3D frameworks for CR electrolytes. As a mature preparation method, ceramic frameworks synthesized by sol-gel process always possess a high homogeneity and an interconnected structure. For instance, the 3D garnet framework fabricated via nanostructured hydrogel can provide long-range and continuous lithium-ion pathways owing to its percolated network (Fig.17) [79]. Moreover, the high content of 3D garnet framework improved the thermal and electrochemical stability as well as the interfacial compatibility against lithium metal, resulting in a smaller resistance than the conventional CSEs with nanoparticles.
However, it is still a challenging task for traditional sol-gel synthesis to optimize the structure of frameworks with a higher specific area. Innovated by the special structure and strong ability to absorb the liquid of sponge, a template-induced sol-gel method has been successfully developed [80]. As shown in Fig.17, using sponge as the template, a 3D LLZAO framework is fabricated with a porous structure. Benefitting from the high proportion of LLZAO framework and its sponge-like structure, the composite electrolyte exhibits a good flexibility, improves thermal stability, and broadens electrochemical window. This template-induced strategy is illuminating for designing frameworks with a unique structure.
3.3 Brief summary
Among lately developed composite systems, CR electrolytes are gathering ever-increased attention. Due to the high mass fraction of ceramics, CR electrolytes possess distinctive internal structures and percolated networks, as well as an enhanced electrochemical and mechanical performance. In this case, the categories of ceramics and different fabrication methods of frameworks have been systematically discussed above (Tab.2). Meanwhile, it is worth noting that polymers are also playing irreplaceable roles in the CR systems: (1) Polymers like PEO and PTFE can work as the binder to stick ceramics together, endow electrolyte with self-standing characteristic and ideal mechanical properties. Simultaneously, unlike rigid ISEs, CR electrolytes can deliver a satisfying flexibility benefitting from shapeable polymers, which promotes the large-scale fabrication of electrolytes and accelerates their practical application. (2) As for ceramics with a poor thermodynamic stability like the perovskite-type and the sulfide-type, polymers can protect them from side reactions with lithium metal or moisture by physical isolation inside the CE electrolyte. Moreover, specific polymers can also react with both electrodes, forming stable SEI/CEI layers and stabilizing battery operation. (3) Polymers can interact with ceramics and then the as-obtained interphase region can act as the important pathway for lithium-ion conduction. The interactions between polymers and ceramics, such as the Lewis acid-base effect can further accelerate the ion conduction around the interphase region. Furthermore, specific polymers can also provide an anion-immobilizing effect and promote the dissociation of lithium salts, thus improving the lithium transference number (Fig.18).
Tab.2 Summary of recently reported CR electrolytes |
| CPEs | Ionic conductivity | Electrochemical window (vs Li/Li+) | Li+ transference number | Activation energy | Ref. |
|---|---|---|---|---|---|
| 85 wt.% LLZTO-PEO | 6.24 × 10−5 S/cm at 25°C | 5 V | – | – | [21] |
| 94.3 wt.% LLZTO@PAN | 1.1 × 10−4 S/cm at 60°C | 4.35 V at 60°C | 0.66 at 60°C | 0.36 eV | [68] |
| 99.5 wt.% LLZTO-PTFE | 5.2 × 10−4 S/cm at 25°C | – | – | – | [70] |
| 80 wt.% PDA@LLZTO-PEO | 1.1 × 10−4 S/cm at 30°C | 4.8 V | – | – | [69] |
| 70 wt.% LLZTO-PVDF | 2 × 10−4 S/cm at 25°C | 5 V | 0.66 | 13.45 kJ mol−1 | [81] |
| 80 wt.% LLZTO-PVDF | 1.08 × 10−4 S/cm at 60°C | 4.7 V | – | 0.39 eV | [82] |
| 80.4 wt.% LLZTO-PTFE-SN | 1.2 × 10−4 S/cm at 25°C | 4.8 V | 0.53 | – | [83] |
| 80 wt.% LLZO-PCL-PTMC | 1.31 × 10−4 S/cm at 60°C | 5.4 V | 0.84 | 0.22 eV | [84] |
| 52.5 wt.% LLZTO-ACN-DMC | 3.1 × 10−3 S/cm at 25°C | 4.7 V | 0.67 | – | [85] |
| 90 wt.% LLZTO-PVDF | 3.7 × 10−4 S/cm at 25°C | 5 V | 0.39 | – | [86] |
| 75 wt.% LAGP-PCL | 1.7 × 10−4 S/cm at 30°C | 5 V | 0.3 | 21 kJ mol−1 | [87] |
| 50 wt.% LATP-PVDF-HFP | 1.02 × 10−4 S/cm at 25°C | 4.5 V | – | – | [73] |
| 90 wt.% LiSn2(PO4)3-PEO | 3.73 × 10−5 S/cm at 25°C | – | – | 0.378 eV | [71] |
| 70 wt.% LiSn2(PO4)3-PEO | 3.48 × 10−5 S/cm at 27°C | – | ~0.39 | ~0.34 eV | [72] |
| 70 wt.% LATP-PEO | 4 × 10−5 S/cm at 25°C | – | – | 0.77 eV | [88] |
| 73.9 wt.% LLTO-PVP-BBP | 2.0 × 10−5 S/cm at 25°C | – | – | – | [74] |
| 99.5 wt.% LPSCl-PTFE | 1.7 × 10–-3 S/cm at 25°C | – | – | – | [70] |
| 99 wt.% LGPS-PTFE | 3.6 × 10−4 S/cm at 25°C | – | – | – | [24] |
| 80 wt.% Li10SnP2S12-PEO | 4 × 10−5 S/cm at 25°C | – | – | 0.9 eV | [89] |
| LATP-MEEP | 1.9 × 10−3 S/cm at 60°C | 5.4 V | 0.94 | – | [75] |
| 70 wt.% s@LLAZO-PEGDA | 4.9 × 10−4 S/cm at 25°C | – | – | 0.29 eV | [77] |
| 91.3 wt.% LLZO-PEO | 1.04 × 10−4 S/cm at 25°C | – | – | 0.36 eV | [78] |
| 62 wt.% LLZO-PEO | 8.5 × 10−5 S/cm at 25°C | 5 V | – | – | [79] |
| 56 wt.% LLZAO-PEO | 2.51 × 10−4 S/cm at 25°C | 5.58 V | 0.53 | 0.36 eV | [80] |
However, researches about CR electrolytes are still at a shallow and incomplete level. Specifically, compared with garnet-type ceramics, CR electrolytes with perovskite-type and sulfide-type ceramic frameworks still need exploration, which may probably be restrained by their instability to lithium metal or moisture. Besides, further investigations into mature fabrication technologies of frameworks and the corresponding interaction mechanisms are required as well. The development of CR electrolytes will be inspiring for optimizing the performance of CSEs, even the overall SSLMB systems.
4 Layered designs of CSEs for practical application
Of all kinds of CSEs, electrolytes with a monolayer structure are the most pervasive, but the performance of single structured CSEs is often limited due to their unbalanced characteristics. PR electrolytes with a low ceramic content usually possess a better flexibility, homogeneous distribution of fillers, and decreased crystallinity of polymer matrix. CR electrolytes containing a higher ceramic content always exhibit a wider electrochemical window, a higher mechanical strength, and interconnected lithium-ion pathways [21]. Consequently, preparing bilayer or multilayer electrolyte membranes is an effective strategy to combine the excellent electrochemical properties of both PR and CR electrolytes and avoid their shortcomings. According to the number of layers, they are divided into bilayer structured and multilayer structured CSEs.
4.1 Bilayer structured CSEs
Bilayer structured CSEs usually consist of a PR electrolyte layer and a CR electrolyte layer, which is mainly prepared by using the effect of gravity field. For the conventional CSE system based on LLZO and PVDF-HFP, with the help of gravity during the one-step blade-casting process (Fig.19), LLZO particles spontaneously precipitate in the bottom layer of the membrane, which later become the CR layer of the CSE with a broadened electrochemical window and is designed to contact with cathodes. Naturally, the top layer of the membrane becomes the PR layer which is placed on the side of the lithium metal anode [90], mainly due to the flexibility and tight interphase contact. Unlike former research, another asymmetrical structure form of CSE has also been proposed [91], which integrates a CR layer on the anode side and a PR layer on the cathode side by utilizing the natural settlement of LLZAO nanoparticles during the battery assembly process (Fig.19). In such unique architecture, the rigid ceramic-rich layer is toward the Li metal and can effectively suppress the growth of Li dendrites, and the soft polymer-rich layer could wet the cathode effectively to endow a connected interface.
In comparison of these two researches, both the PR and CR layers in the bilayer structured CSEs are designed to contact with cathodes and anodes respectively, which show that the orientations of PR and CR layers in the solid-state LMBs should depend on the specific requirements. It is worth noting that the location of different electrolytes may have significant impacts on the interfacial compatibility of both the anode and the cathode. Therefore, actual demands for battery design can hardly be satisfied merely relying on bilayer structured CSEs.
4.2 Multilayer structured CSEs
As discussed above, the as-designed bilayer structured CSEs may suffer from the risk of balancing the behaviors of flexibility, mechanical strength, and the thermodynamic stability. Designing multilayer structured CSEs according to the properties of both electrolyte and electrode, as well as the actual service environment is regarded as the promising method to tackle these problems. Up to the present, the sandwich structure of CSEs with two PR electrolytes as the outer layers and a CR middle layer has become the mainstream research model. For instance, Huo et al. have revealed that the CR layer with larger LLZTO particles exhibits a higher mechanical strength while the PR layer with smaller LLZTO particles shows a better ionic conductivity, which inspires them to rationally design a sandwich-type CSE membrane with a mechanically strong CR interlayer and two flexible PR outer layers (Fig.20) [92]. Notably, the size of inorganic particles in the sandwich structure is closely related to the electrochemical property. For giant granular components, they can significantly inhibit the dendrite growth due to the favorable modulus and large coverage area while there may exist abundant gaps around the surface of cathodes, leading to a blocked ion migration. On the contrary, tight adhesion usually appears for the interphase region of fine particles and the cathode. Therefore, layered CSEs with hierarchical inorganic particles can successfully achieve both a dendrite suppression and an excellent interfacial contact with the electrodes (Fig.20–Fig.20).
In addition to the size of inorganic particles, for the multilayer structured CSE designed based on the above concept, the PR layers on both sides always possess the same polymer matrix and are difficult to deliver the optimal performance. Therefore, it is indispensable for designing PR outer layers with different polymer matrixes according to the actual scene. In consideration of this, a layer-by-layer coating technique is developed [93] to fabricate the sandwich structure CSE membrane with three different polymers (Fig.20). Specifically, oxidation-resistant PAN-based and reduction-resistant PEO-based PR outer layers are used to contact with the cathode and the anode respectively, while the intermediate layer is a PVDF-based CR layer filled with LLTO nanofibers (80 wt.%, Fig.20). Due to this unique three-layer design, a perfect combination of different polymer matrix is achieved, which simultaneously improves the compatibility between the CSE and the electrodes.
4.3 Brief summary
Although both PR and CR electrolytes have their intrinsic deficiencies for practical application, layered designs of CSEs combining PR and CR electrolytes can take advantage of each component layer and minimize their shortcomings. The performance of different CSEs with a layered structure are compared in Tab.3. In general, bilayer structured CSEs containing a PR layer and a CR layer can be simply prepared with the help of the gravity field. Depending on the specific characterization and requirements of the electrolytes, the orientations of PR and CR layers can be adjusted optionally. However, it is still a difficult task to meet the practical needs only relying on bilayer structured CSEs. Consequently, designing multilayer structured CSEs is more likely to satisfy the contradictory requirements in practice, such as suppressing Li dendrite by high modulus and ensuring a good interfacial contact by flexibility, as well as resisting oxidation and reduction simultaneously.
Tab.3 Summary of recently reported CSEs with layered structure |
| Layer number | CPEs | Ionic conductivity | Electrochemical window (vs Li/Li+) | Li+ transference number | Activation energy | Ref. |
|---|---|---|---|---|---|---|
| Bilayer | CR-PR bilayer LLZO-PVDF-HFP | 1.1 × 10−4 S/cm at 25°C | 4.5 V | 0.55 | 0.146 eV | [90] |
| Soft-tough asymmetric LLZAO-in situ polymerized TPGDA | 8.43 × 10−4 S/cm at 25°C | 5 V | 0.42 | – | [91] | |
| 80 vol.% LATP-PEO|PEO@ PVDF-HFP | 5.03 × 10−4 S/cm at 45°C | 4.9 V | – | – | [94] | |
| Multilayer | 20 vol.% 200 nm LLZTO-PEO| 80 vol.% 5 μm LLZTO-PEO|20 vol.% 200 nm LLZTO-PEO | 2.3 × 10−5 S/cm at 30°C | 5.03 V | – | – | [92] |
| 15 vol.% LLTO-PEO|80 vol.% LLTO-PVDF|15 vol.% LLTO-PAN | 2.82 × 10−4 S/cm at 25°C | 4.9 V | – | 0.271 eV | [93] | |
| PEO|LLTO-PEO|PEO | 1.6 × 10−4 S/cm at 24°C | 4.7 V | 0.48 | – | [95] | |
| 15 vol.% LLTO-PVDF|75 vol.% LLTO-PVDF|15 vol.% LLTO-PVDF | 4.7 × 10−4 S/cm at 25°C | 5 V | – | 0.301 eV | [96] |
In the future research, both the categories of ceramics and polymers should be in-depth analyzed during the design and preparation of multilayer structured CSEs. Furthermore, developing other sandwich structures with various locations of PR and CR layers is also a possible research direction. More flexible and scalable preparation processes of multilayer structured CSEs also need to be further explored.
5 Conclusions and perspectives
Compared with conventional ISEs and SPEs, CSEs are more competitive candidate electrolytes for the next generation of SSLMBs due to the combination of the advantages of both ISEs and SPEs. As an early-developed CSEs system, PR electrolytes with a low ceramic content usually possess a better flexibility. The addition of inorganic active fillers can decrease the crystallinity of polymer matrix, providing the enhancement of polymer chain segment motion with a better flexibility, an improved ionic conductivity, and a homogeneous distribution of fillers. More importantly, ceramic fillers are able to modify the local region around them through special interactions like the anion-immobilizing effect or the Lewis-acid-base effect, which can promote the dissociation of lithium salt and accelerate the transmission of lithium ions. However, the ceramic particles surrounded by the polymer matrix are isolated with each other due to the relatively low content, causing discontinuous lithium-ion pathways. On the contrary, CR electrolytes are able to provide smooth ion transport pathways with concentrated active particles contacting with each other, usually accompanied by the improvement of mechanical stability and safety behavior. However, the internal interface compatibility and external thermodynamic stability with cathode/anode may be unsatisfactory.
In this review, the recent progress of PR and CR electrolytes was discussed, mainly focusing on the ion migration in the bulk and interphase region of inorganic active materials with various chemical components and structural designs. Besides, the interphase compatibility and thermodynamic stability for different CSEs during the process of electrolyte design were emphasized. In addition, rational design concept and preparation method for the frameworks of CR electrolytes, and the multilayer structure of CSEs under practical application conditions have also been elucidated.
Although both PR and CR electrolytes exhibit a great application potential, several challenges still need to be addressed in future research. Certain noteworthy points are discussed as follows:
Multiple factors should be considered in the process of electrolyte design. First, the intrinsic structure and chemical properties of inorganic ceramic components have significant effects on their own conductivity and can further affect the overall conductivity of the CSE system. Moreover, the ion transport ability is closely related to the organic/inorganic interphase region. Therefore, the volume content of the interphase region and the organic/inorganic interphase compatibility cannot be ignored. Notably, the volume fraction mainly depends on the surface area of the inorganic ceramics while the interphase compatibility is usually attributed to the physical and chemical interactions, such as the electrostatic force, the percolation effect, the space charge layer effect, and the Lewis acid-base effect. In addition, as for the design of polymer phase, the species and concentrations of both polymer and lithium salts, as well as the space structure and functional groups need to get focused attention.
Functional design helps to maximize the electrochemical performance of CSE systems. Except for conventional nanoparticles, inorganic ceramic materials have various spatial morphologies, including nanowires, nanosheets, and three-dimensional frameworks. Besides, they may deliver different orientations, such as parallel or perpendicular to the electrolyte membrane, and even random spatial orientation. Both the morphologies and orientations can significantly affect ion transport efficiency, which should be paid more attention to. As for the polymer phase, they usually undertake the task of electrolyte binder, protection coating and constructing interphase region. Therefore, the selection of polymer phase should take into accounts their chemical compositions, electrochemical stability, and the interactions with ceramics.
For multilayer structured CSEs, researchers usually attach great importance to the ion transport inside each electrolyte layer but the ion transport between the two layers in contact with each other has always been grossly neglected. Moreover, the bonding strength of contiguous layers is also an important consideration in electrolyte design, which is intimately connected to the interfacial impedance and ion transport. To avoid the above issues, integrated design for multilayer structure, such as the layer-by-layer tape casting method may be an ideal choice, which eliminates the obvious interlayer porosity and poor contact, resulting in an enhanced compatibility. Other effective methods for the preparation of multilayer structured CSEs are still urgently needed. Furthermore, multilayered CSEs with different functional layers containing various ceramics or polymers could also be noticed as an optimizing strategy.
To meet the demands of industrial production for SSLMBs, large-scale preparation of various CSEs which can cover the characteristics of simple process and low cost are also necessary. In particular, nanofibers, nanosheets or frameworks have been proved to be satisfying inorganic components in PR and CR systems, which can provide continuous lithium-ion pathways and an enhanced mechanical support. However, further optimization of the existing technologies including electrospinning, solution dispersion, vacuum filtration, and sol-gel process is requisite. Other innovative manufacturing methods for fabricating these inorganic components with specific structures, for example 3D printing and bionic technology, are also urgently needed to achieve a better structural design and facile fabrication.