A Selective‐Transport Elastomeric Coating Regulating Hierarchical Solid Electrolyte Interphase for Low‐Temperature Lithium‐Metal Batteries

ABSTRACT

Enhancing the low‐temperature cycling performance of lithium metal batteries (LMBs) relies on the rational design of solid electrolyte interphases (SEIs). Conventional approaches typically involve tuning electrolyte compositions to indirectly generate SEIs dominated by organic or inorganic components. However, organic‐rich SEI fails to inhibit the growth of Li dendrites, compromising sluggish Li⁺ kinetics, and inorganic‐rich SEI suffers from mechanical brittleness at low temperatures, resulting in inadequate interfacial mechanical stability. Herein, we introduce a siloxane‐based elastomeric coating on the Li anode surface by leveraging its intrinsic solvent phobicity to achieve selective ion conduction, facilitating the formation of a LiF‐rich inner SEI, which synergizes with the elastomer to construct a double‐layer organic‐inorganic SEI. Theoretical calculations and experimental results demonstrate that such a double‐layer SEI combines mechanical flexibility enabled by organic components with promoted Li⁺ transport imparted by inorganic components, synergistically improving the cycling stability of LMBs under low‐temperature conditions. The target LMBs paired with industrial‐standard NCM811 cathodes deliver 99% capacity retention over 300 cycles at –25°C. Unlike indirect electrolyte modification approaches, our method enables direct manipulation of SEI structures and is compatible with various electrolyte systems.

A solvent‐phobic elastomeric coating is developed for selective ion conduction on the anode interface. By directing the formation of a double‐layered SEI featuring an organic outer layer and an inorganic inner layer, the coating imparts high chemo‐mechanical stability to the interface of low‐temperature Li‐metal batteries. Consequently, Li||NCM811 cells retain 99% of capacity over 300 cycles at –25 °C and 0.5 C.

1. Introduction

The deteriorated capacity of lithium‐ion batteries (LIBs) at low temperatures limits their practical applications in electric vehicles, aerospace equipment, and polar exploration [1, 2]. With the highest theoretical specific capacity (3860 mAh g⁻¹) and lowest redox potential (−3.04 V vs. the standard hydrogen electrode) [3], lithium (Li) metal has emerged as a promising anodic material by providing a potential pathway in elevating the energy density [4, 5]. Furthermore, the incorporation of high‐voltage and nickel‐rich cathodes (like NCM811 and NCM955) could achieve much higher energy density [6, 7, 8]. However, the reduced Li⁺ deposition kinetics and uncontrolled electrolyte decomposition thermodynamics at low temperature (<−20°C) render unstable formation of solid electrolyte interphases (SEIs) [9, 10, 11, 12], which results in serious growth of Li dendrites and rapid failure of coulombic efficiency [13, 14]. Up to now, developing a stable SEI still remains a fundamental challenge for low‐temperature operable LMBs.

Generally, SEI can be classified into organic‐rich and inorganic‐rich types according to its main constituents [15, 16, 17]. The organic‐rich SEI derived from the decomposition of solvents has been demonstrated to have mechanical flexibility and thereby contribute to the volume adaptability against Li anodes [18]. However, low ionic conductivity and mechanical strength of organic components have been proven to facilitate the formation and growth of Li dendrites, especially at low temperatures [19]. In previous investigations, researchers focused on constructing salt‐derived SEI that is rich in inorganic components, such as LiF, Li₃N, and Li₃B, via modifying organic solvents at the molecular level [20, 21, 22], designing locally high‐salt‐concentration electrolyte [23, 24], and incorporating functional additives [25, 26, 27]. Although the inorganic‐rich SEI exhibits higher mechanical stability and effective ion transport compared with the organic‐rich SEI, it still suffers from the risk of mechanical rupture at low temperatures due to its intrinsic brittleness (Scheme 1a) [28, 29]. It seems that constructing an SEI with integrated advantages of both organic and inorganic components is a reasonable approach for realizing low‐temperature LMBs with stable cycling performance. Furthermore, the present research tried to modulate the SEI components through indirect methods from the electrolyte aspect, while the organic solvents cannot avoid contact with Li metal anodes, and parasitic reactions still occur at low temperatures [30]. Nowadays, designing functional coating has been considered as a direct strategy to generate a stable interface between electrolytes and anodes in LMBs [31, 32]; however, the feasibility of protective layers in low‐temperature conditions remains underexplored [33].

SCHEME 1.

Schematic diagram of structural evolution of (a) conventional single‐layer (LiF‐rich organic‐inorganic hybrid layer) and (b) current double‐layer (organic polymer outer layer and inorganic inner layer) SEIs.

Herein, we report a novel engineered elastomeric interlayer to construct a distinct organic‐inorganic double‐layer SEI. This coating is designed with a solvent‐phobic polysiloxane backbone, which effectively suppresses parasitic reactions between organic solvents and Li metal, while maintaining mechanical extensibility to accommodate significant volume changes at low temperatures. Furthermore, the proportion of functional side chains was tuned for optimized surface energy, mechanical properties, and dielectric characteristics of the elastomeric layer. Notably, the solvent‐phobic property of the elastomeric coating facilitates the formation of an anion‐derived inorganic inner layer beneath the organic elastomeric layer, resulting in a hierarchical structure with a well‐controlled distribution of organic and inorganic phases. Such ordered SEI maximizes the synergistic effects of both components: the flexible outer layer effectively dissipates stress concentrations caused by Li anode volume changes, meanwhile the anion‐derived inorganic layer promotes uniform Li⁺ deposition kinetics (Scheme 1b). By leveraging this solvent‐phobic elastomeric coating, cPSiA₃F₇‐Li||NCM811 cells demonstrate excellent capacity retention of 99.3% after 300 cycles, even with a thin Li anode of 40 µm and an industrial‐standard NCM811 cathode (97.5% active‐material content) of 3 mAh cm⁻², at an operating temperature from −25°C to 25°C. Moreover, the high compatibility of the designed solvent‐phobic elastomeric layers with various electrolytes further reveals their potential as an innovative approach for developing practical low‐temperature LMBs.

2. Results and Discussion

2.1. Design of Elastomeric Layer

The elastomeric layer employing polysiloxane (PSiO) with low glass transition temperature (T_g) as backbone retained its intrinsic flexibility at low temperatures [34]. In addition, the insolubility of PSiO in most ester and certain ether solvents (Figure S1) indicated its capacity to prevent the permeation of organic solvents, thereby suppressing the decomposition of organic solvents on the surface of the Li anode. The PSiO grafted with fluorinated‐ester and carboxyl was fabricated by reacting with the 2,2,2‐trifluoroethyl acrylate (TFA) and acid acrylate (AA) to address the weak coupling problem between PSiO and lithium salts (Figure 1a). The high dielectric constant of the fluorinated ester group improved the ionic conductivity of grafted PSiO [2]. The carboxyl group not only spontaneously reacted with lithium metal to form a lithophilic interface of ‐COO^–Li⁺ [35], but also modulated the mechanical robustness of grafted PSiO through hydrogen bonding. A series of PSiOs grafted with carboxyl and fluorinated‐ester groups was synthesized, abbreviated as PSiO‐AA_x‐TFA_y, where “x” and “y” denote the molar ratios of carboxyl and fluorinated‐ester units in a single molecular chain, respectively. ¹H nuclear magnetic resonance (¹H NMR) spectra indicated the successful grafting of carboxyl and fluorinated ester groups, where the total grafting ratio of both units was obtained as 90% by integrative calculations (Figure 1b and S2; Table S1). The 10% leftover sulfhydryl group can be further cross‐linked to improve mechanical robustness for drastic volume changes of the Li anode at low temperatures. The cross‐linked PSiO‐AA_x‐TFA_y samples are abbreviated as cPSiA_xF_y.

FIGURE 1.

Synthesis and property analysis of elastomeric polymers. (a) Synthesis route and (b) ¹H NMR spectra of PSiO‐AA_x‐TFA_y. (c) DSC curves and (d) rheological frequency sweep of the cPSiA_xF_y samples. (e) The cartoon (top) and digital image (bottom) of the H‐cell experiment setup. The dot‐capped lines in the digital image represent the liquid level.

Temperature‐dependent Fourier transform infrared (FTIR) spectroscopy showed that the intensities of the two absorption peaks assigned to the free and hydrogen‐bonded C═O groups in cPSiA_xF_y molecules changed in opposite directions (Figure S3), consistent with the dissociation of hydrogen bonds upon heating. Further inspection of the characteristic ¹H NMR signal of the carboxylic acid protons revealed a gradual downfield shift as the carboxylic acid content increased (Figure S4). This shift can be attributed to the enhanced intermolecular hydrogen‐bonding interactions between carboxyl groups, which decrease the electron cloud density surrounding the protons. The T_gs of cPSiA_xF_y, which directly correlated with chain flexibility, were determined by differential scanning calorimetry (DSC). It showed that the hydrogen bonding originated from the carboxyl group in cPSiA_xF_y had a significant effect on the chain flexibility with the T_gs of cPSiA₁₀F₀ and PSiA₀F₁₀ being −32°C and −63°C, respectively (Figure 1c) [6, 36]. Although the T_gs of cPSiA_xF_y increased with the carboxyl contents, they were still far below the ambient temperature. As illustrated by the rheology tests of cPSiA_xF_y at 25°C (Figure S5), the cPSiA₀F₁₀ exhibited a liquid‐like behavior (storage modulus G′ < loss modulus G″) due to the absence of intermolecular hydrogen bonding, and other samples demonstrated typical elastic characteristics (G′ > G″). At −25°C, modulus crossover in the low‐frequency range was only observed for the cPSiA₃F₇ (Figure 1d), indicating its higher mechanical elasticity than other samples. This can be explained by the fact that the hydrogen bonding can significantly influence the mechanical properties of elastomeric polymers, while excessive hydrogen bonding limits the chain mobility at low temperatures.

The solvent‐phobicity of the cPSiA_xF_y coating was evaluated by contact angle measurements using electrolyte solvent (Figure S6). The solvent droplets rapidly spread on bare copper foil, resulting in an unmeasurable contact angle. In contrast, after coating with cPSiA_xF_y, the surface exhibited significantly reduced affinity toward solvents, with the contact angle increasing to 30.4°. The Li⁺ conduction capability of cPSiA_xF_y was further investigated by molecular dynamics simulations, which afford a Li⁺ diffusion coefficient of 7.14 × 10^{− 5} Ų ps^{− 1} in the elastomeric coating (Figure S7). The selective ion transport of cPSiA_xF_y was verified by an H‐cell device (Figures 1e and S8), of which the left side was filled with 4 mL of 2 M LiPF₆ electrolyte while the right side was filled with 7 mL solvent. The two sides were isolated with a commercial Celgard2325 separator or cPSiA₃F₇@Celgard (a Celgard2325 membrane coated with 100 mg of cPSiA₃F₇ on both sides), respectively. H‐cells were subjected to a refrigerator and kept at −25°C for 12 h. As illustrated by Figure 1e, for the H‐cell separated by Celgard2325, the liquid level on the right side exhibited a significant decrease after the above‐mentioned process, indicating the permeation of organic solvents through Celgard2325. For the H‐cell separated by cPSiA₃F₇@Celgard, the liquid level showed negligible change, corresponding to the repellency of cPSiA₃F₇ coating toward organic solvents. Simultaneously, the presence of LiPF₆ on the right side, as confirmed by ⁷Li NMR and ³¹P NMR (Figures S9 and S10), also indicated the permeability of ions (Li⁺ and PF₆ ⁻). To further clarify the effect of functional groups on the ion‐transport process, experiments were also performed at room temperature using cPSiA₁₀F₀@Celgard, cPSiA₃F₇@Celgard, and cPSiA₀F₁₀@Celgard, respectively. The ¹⁹F NMR spectroscopy (Figure S11) analysis showed that the salt concentrations in right‐side solutions were 0.020 M, 0.064 M, and 0.074 M for cPSiA₁₀F₀@Celgard, cPSiA₃F₇@Celgard, and cPSiA₀F₁₀@Celgard, respectively. These results reveal that, compared to carboxyl groups, the fluorinated ester groups promote kinetics for ion migration. The sieving effect of liquid electrolytes enabled by cPSiA_xF_y is essential for the formation of double‐layer SEI.

2.2. Investigation of Lithium Deposition Behavior

The cPSiA_xF_y‐Li formation involved drop‐coating the tetrahydrofuran solution onto the lithium‐metal foils, followed by UV‐triggered radical crosslinking (Figure S12). The depth and morphology of the elastomeric coating on the Li anode were observed using nano‐indentation and scanning electron microscopy (SEM), respectively. Taking cPSiA₃F₇‐Li as an example, the elastomeric polymer was uniformly coated on the Li anode with a thickness of around 0.79 µm (Figure S13). The elemental distribution characterized with energy dispersive X‐ray spectroscopy (EDS) (Figure S14) also exhibited that the F and O elements belonging to cPSiA₃F₇ were uniformly distributed on the elastomeric‐coated Li anode. The atmospheric erosion experiment provided direct evidence for the protective role of thin elastomeric layers on the Li anode (Figure S15). As revealed by attenuated total reflection FTIR spectroscopy (Figure S16a), a new peak at 1580.1 cm^{− 1} was assigned to the C═O vibration of the newly formed ‐COO^–Li⁺ species in the elastomer‐coated Li anode [37]. The formation of single‐ion conductive species (‐COO^–Li⁺) was expected to increase the cation transport number (t_Li+, Figure S16b) [38].

To compare the surface morphology after Li deposition, we assembled Li||Cu cells with cPSiA_xF_y‐coated Cu collectors and conducted Li deposition measurement with a capacity of 0.1 mAh cm⁻². As observed in Figures 2a and S17, the coverage of Li deposition significantly reduced after capping with cPSiA₁₀F₀ at both 25°C and −25°C. The coupling mechanism between the intrinsic properties (e.g., surface energy and dielectric constant) of polymer coating and Li‐deposition behavior was further investigated. The high carboxyl content in cPSiA₁₀F₀ led to the formation of intensive intermolecular hydrogen‐bonding interactions, which suppressed segmental mobility and polarizability, as reflected by the higher glass transition temperatures and lower dielectric constants (Figures 1c, 2b, and S18). As a result, Li⁺ transport was hindered. In contrast, increasing the content of fluorinated ester units enhanced the segmental mobility and polarizability of cPSiA_xF_y. The resulting higher dielectric constant improved its affinity toward LiPF₆ [39], thereby reducing the Li nucleation overpotential (Figure S19) and promoting denser Li deposition. Solubility tests further supported the enhanced LiPF₆‐dissociation capability of fluorinated ester units, indicating that TFA with fluorinated ester groups exhibited better compatibility with LiPF₆ than AA with carboxyl groups (Figure S20). Of note, due to the high fluidity of PSiA₀F₁₀ at 25°C, the dielectric constant cannot be measured, while its trend can be reasonably predicted based on the molecular polarization theories [40, 41]. Moreover, Li agglomeration around cPSiA₀F₁₀‐Cu collector might be attributed to the absence of carboxyl groups contributing to the mechanical robustness and formation of anchor sites [42]. Meanwhile, higher surface energy of elastomeric layers tended to form the rod‐like Li sediments (cPSiA₇F₃ and cPSiA₅F₅), while the low surface energy rendered spherical Li morphology (cPSiA₃F₇ and cPSiA₀F₁₀) (Figures 2b and S21; Table S2). Such a phenomenon was more obvious at −25°C due to the sluggish transport and reaction kinetics of Li⁺ [43], resulting in a decreased diameter of Li depositions and accelerated vertical growth (cPSiA₇F₃ and cPSiA₅F₅). Significantly, only cPSiA₃F₇ coating enabled the dense and homogeneous Li deposition at both 25°C and −25°C.

FIGURE 2.

The analysis of Li deposition and solvent‐phobic mechanisms for cPSiA_xF_y. (a) SEM images of 0.1 mAh cm⁻² of lithium electrodeposited on cPSiA_xF_y‐coated copper, scale bar: 1 µm. (b) Surface energies and dielectric constants of cPSiA_xF_y. The distribution of (c) potential and (d) Li⁺ concentration near the anode surface is influenced by the dielectric constant of cPSiA_xF_y. (e) Simulation models of the lithium–electrolyte interfaces on (left) Bare‐Li and (right) cPSiA₃F₇‐Li (with elastomeric solvent‐phobic coating). Li atom: green, C atom: brown, F atom: light blue, H atom: light pink, O atom: red, P atom: light fuchsia, S atom: yellow, Si atom: blue. Number of Li atoms migrating from electrode into electrolyte over time at −25°C. Energy versus time curves of Bare‐Li and cPSiA₃F₇‐Li based systems to reach equilibrium points at −25°C.

To investigate the dielectric‐constant effect on Li deposition behavior, a numerical model was developed using finite element simulation. As illustrated in Figure 2c, a higher dielectric constant led to a more uniform electric field distribution at the Li anode interface, implying homogeneous electrochemical deposition of Li⁺ on the anode surface. Owing to the strong polarity of ‐CF₃ groups, the normalized concentration difference of cPSiA₃F₇ was significantly lower than that of cPSiA₁₀F₀, as revealed in Figure 2d. Furthermore, Figure 2d revealed that the Li⁺ concentration gradient from bulk electrolyte to anode surface gradually diminished with increasing dielectric constant. Impedance analysis was also performed on Li symmetric cells using cPSiA_xF_y‐Li anodes (Figure S22), which revealed that the interfacial impedance of symmetric cells was also negatively correlated with the dielectric constants of cPSiA_xF_y. It indicated that high dielectric‐constant coatings can effectively mitigate the interfacial polarization at the Li anode, thereby contributing to the uniform Li deposition. Moreover, the cPSiA₃F₇‐Li enabled a Li symmetric cell with the high cycling stability over 2000 h at 25°C, representing the best electrochemical performance among all tested samples (Figure S23). Therefore, cPSiA₃F₇ was selected for further study.

To further elucidate the protective mechanism of the cPSiA₃F₇ layer on Li electrodes, we also conducted ab initio molecular dynamics (AIMD) simulations at −25°C and 25°C. Bare‐Li showed direct contact between Li atoms and liquid electrolyte (Figures 2e and S24), while the cPSiA₃F₇‐Li, covered with an elastomeric layer, separated the Li atoms and electrolyte. Statistical analysis on the number of migrated Li atoms as a function of time showed that, compared with the Bare‐Li system, cPSiA₃F₇ effectively blocked the direct contact between the electrolyte solvents and Li anode, thereby suppressing interfacial side reactions. As a result, only a small fraction of Li atoms in the cPSiA₃F₇‐Li system formed chemical bonds with electrolyte solvents and diffused into the bulk electrolyte, which mitigated the formation of dead Li (Figures 2e, S25, and S26). The time‐dependent energy evolution of the two systems further reflected the extent of interfacial reactions and the decomposition rate of the liquid electrolyte on the Li anode (Figure 2e). In comparison with the Bare‐Li system, the cPSiA₃F₇‐Li system exhibited a smaller energy variation and reached equilibrium more rapidly, suggesting that cPSiA₃F₇ suppressed electrolyte decomposition and stabilized the electrode/electrolyte interface at both −25°C and 25°C (Figure S27). The result also confirmed the high electrochemical stability of the cPSiA₃F₇‐Li system at both temperatures owing to its elastomeric solvent‐phobic nature.

2.3. Basic Electrochemical Test

The contribution of cPSiA₃F₇ coating to electrochemical performances was characterized by assembling lithium symmetric cells paired with modified and unmodified anodes, respectively. Combining the chronocurrent results and electrochemical impedance spectroscopy (EIS), the t_Li+ of cPSiA₃F₇‐Li‐based symmetrical cell was calculated as 0.71, much higher than that of Bare‐Li (0.48) (Figure 3a,b). Such enhancement originated from the in situ formation of lithophilic ‐COO^–Li⁺ interphase. In addition, density functional theory calculations demonstrated that the ‐CF₃ group reduced the electron density of ester groups through its inductive effect, thereby lowering the binding energy between the ester group and Li⁺ (Figure S28). This weaker Li⁺‐binding strength reduced excessive ionic trapping at coordination sites and thereby facilitated Li⁺ exchange/migration between neighboring sites [44]. To illustrate the stabilizing effect of the cPSiA₃F₇ layer on the Li anode, the impedance evolution of the cells was monitored at room temperature over 0–100 h (Figure S29), and the relative change was quantified via the formula (I₁₀₀ – I₀ )/I₀ (%). After 100 h, cPSiA₃F₇ coating reduced the impedance increase of Li||Li cells in liquid electrolyte from 169% to 24%, which could be attributed to the protection of solvent‐phobic coating (Figure S30).

FIGURE 3.

Electrochemical characterizations of the lithium symmetric cells with different electrodes. (a,b) cation transport numbers, (c) Tafel plots of Bare‐Li and cPSiA₃F₇‐Li, and (d) activation energies with inset Arrhenius behavior of the resistant. Voltage‐time profiles of Bare‐Li and cPSiA₃F₇‐Li (e) at different current densities of 0.1, 0.2, 0.3, 0.4, 0.5, and 1.0 mA cm⁻² (25°C) and (f) at constant current density 0.2 mA cm⁻² (−25°C). EIS plots after different cycles with inset images of voltage‐time curves of (g) Bare‐Li and (h) cPSiA₃F₇‐Li. (i) SEM images and digital photos after 50th cycle of Bare‐Li and cPSiA₃F₇‐Li, scale bar:20 µm.

Generally, the Li⁺ desolvation process has been widely recognized as the rate‐determining step governing the overall charge transfer kinetics, especially under low temperatures [45]. The effect of cPSiA₃F₇ coating on the charge transfer and activation energy (E_a ) of Li deposition was investigated by Tafel curves (Figure 3c) and temperature‐dependent EIS (Figures 3d and S31), respectively. The cPSiA₃F₇‐Li‐based cell exhibited lower charge transfer resistance (R_ct ) than that of Bare‐Li, which was extracted from the Nyquist plots at various temperatures (Figure S32). Meanwhile, cPSiA₃F₇‐Li exhibited a comparable exchange‐current density (i₀ ) to Bare‐Li (Figure 3c), implying that cPSiA₃F₇ contributed to the Li⁺ desolvation process without compromising the electrochemical reaction kinetics. According to the Arrhenius equation, the E_a of the cPSiA₃F₇‐Li was calculated as 53.72 kJ mol⁻¹, lower than that of Bare‐Li (56.73 kJ mol⁻¹) (Figure 3d). The result reflected the lower energy threshold for Li deposition in the cPSiA₃F₇‐Li system, implying faster ion transfer of the cPSiA₃F₇‐Li during the plating/stripping process [46].

Rate performance of symmetric lithium cells at current densities ranging from 0.1 to 1 mA cm⁻² further demonstrated the stabilizing effect of cPSiA₃F₇ for the Li anode. In Figure 3e, the polarization voltage of Bare‐Li fluctuated significantly with increasing current densities, eventually leading to a soft short circuit at both 0.5 and 1 mA cm⁻². On the contrary, the cPSiA₃F₇‐Li enabled superior cycling performance with stable polarization voltage at different current densities. In addition, cPSiA₃F₇‐Li stably cycled over 500 h at 0.2 mA cm⁻² under −25°C, whereas Bare‐Li exhibited a rapidly increased polarization voltage after 200 h (Figure 3f). As illustrated by the impedance evolution in Figure 3h, the symmetric cell with cPSiA₃F₇‐Li showed a slight increase in impedance after 50 cycles, while Bare‐Li exhibited an impedance reduction due to the micro short‐circuit caused by irregular Li deposits (Figure 3g) as evidenced by SEM (Figure 3i). The improved stripping/plating reversibility of cPSiA₃F₇‐Li was attributed to the selective‐transport coating, which enhanced ion transport kinetics (higher cation transport number and lower desolvation energy barrier) and promoted the formation of a chemo‐mechanically stable double‐layer SEI, which will be discussed later.

2.4. Insight Into SEI Structure

To understand the formation and evolution of SEI at low temperatures, Li anodes disassembled from Li||Li cells after 50 cycles at −25°C and 0.2 mA cm⁻², were collected for interphase analysis. X‐ray photoelectron spectroscopy (XPS) provided chemical components of the SEI on Bare‐Li and cPSiA₃F₇‐Li, respectively (Figure 4a,b). The F 1s spectra of both Bare‐Li and cPSiA₃F₇‐Li showed F‐C (687.8 eV) and F‐Li (685.3 eV) signals corresponding to organic and inorganic components [47]. For intrinsic SEI formed on Bare‐Li, the peaks in the F 1s and Li 1s spectra showed slight changes as sputtering time increased (Figure 4a). As for the double‐layer SEI formed on cPSiA₃F₇‐Li, the peak of F‐C species exhibited a significant decrease as sputtering time increased, indicating a gradient structure of SEI composition along with depth, which was also confirmed by the Li 1s spectra analysis (Figure 4b). The small amount of Li‐F (56.9 eV) species was observed on the surface of double‐layer SEI [48], and as the sputtering duration increased, the Li‐F content increased rapidly and became the main component. The Li_x‐OP_yR signal appeared in a deeper SEI region (Figure 4b), indicating that the bottom layer of the SEI within cPSiA₃F₇‐Li was predominantly composed of salt‐derived inorganic species. The quantitative compositions of organic (Li‐O₂CR, Li‐OR, Li_x‐OP_yR, and C‐F bonds) and inorganic species (Li‐F bonds) in both SEIs were summarized in Figure S33. The above results indicated that the intrinsic SEI on Bare‐Li was an organic‐inorganic hybridized layer, while the SEI formed on cPSiA₃F₇‐Li was a hierarchical structure with the organic layer on top and the inorganic layer at the bottom.

FIGURE 4.

Nano‐structural resolution of SEI after 50 cycles in different electrodes. (a,b) F 1s and Li 1s XPS spectra with etching depth profiles. (c,d) 3D views of Li⁻, LiF₂ ⁻, LiCO₃ ⁻ and CH₂O⁻ in the ToF‐SIMS sputtered volumes and (e, f) DMT modulus distribution with force mapping of SEIs from Bare‐Li and cPSiA₃F₇‐Li. Modeling analysis on the role of double‐layer SEI for the mechanical stability of the Li anode interface. (g) The stress distribution of Li dendrites modulated by the organic‐rich SEI (left), inorganic‐rich SEI (middle), and double‐layer SEI (right), scale bar: 1 µm. (h) Quantitative stress distribution of lithium dendrites in three SEIs.

Time of flight secondary ion mass spectrometry (ToF‐SIMS) also provided 3D profiles to understand the structural distribution of the SEI layer. Representative secondary ion fragments (Li⁻, LiF₂ ⁻, CHO₂ ⁻, LiCO₃ ⁻) were selected to analyze the chemical composition of SEI (Figure 4c,d). Among them, Li⁻ fragment represented the organic lithium‐containing component that was negatively ionized [49]. The LiF₂ ⁻ fragment corresponded to the inorganic LiF, which originated either from the reaction of the lithium‐containing product with fluoride or from the decomposition of LiPF₆. Additionally, the CHO₂ ⁻ and LiCO₃ ⁻ fragments were used to represent the decomposition products of organic solvents. Obtained from the 3D profiles and sputtering depth distribution profiles, the Bare‐Li had a large amount of Li^– within the SEI layer (Figures 4c and S34), while Li⁻ fragments from cPSiA₃F₇‐Li predominantly localized in the top layer (Figures 4d and S34). Notably, the distributions of CHO₂ ⁻ and LiCO₃ ⁻ in Bare‐Li were significantly higher than that of cPSiA₃F₇‐Li, indicating the extensive decomposition of organic solvents on Bare‐Li surface (Figure S34). The unstable organic components and harmful inorganic components were simultaneously incorporated into the SEI layer, leading to the poor low‐temperature cycling stability of the Bare‐Li‐based cells. By combining the analysis of XPS and ToF‐SIMS, we proposed that the SEI of cPSiA₃F₇‐Li maintained an organized structure with a hierarchical distribution of organic and inorganic components. The synergistic effect of organic and inorganic layers can significantly enhance the chemo‐mechanical stability of the electrolyte‐anode interface, particularly under low‐temperature conditions.

The modulus distribution and roughness of the SEIs were characterized by atomic force microscopy (AFM). As shown in Figure 4e,f, the modulus distribution of the SEIs formed on cPSiA₃F₇‐Li was narrower than that of Bare‐Li (the 3D map represented the color mapping of the modulus), indicating more homogeneous Li deposition behavior on cPSiA₃F₇‐Li. Meanwhile, compared to Bare‐Li (13.16 GPa), the relatively lower average modulus of cPSiA₃F₇‐Li (11.99 GPa) might be attributed to the top elastomeric layer. Especially, after the Li plating process, in contrast to the significant height gap of Bare‐Li, the cPSiA₃F₇‐Li showed relatively homogeneous morphology and low height fluctuation (Figure S35). It suggested that a double‐layer SEI with excellent rupture durability enabled uniform Li deposition under extreme conditions.

To further demonstrate the mechanical stability of double‐layer SEI during the cycling process, three Li anode interface models were established (Figure 4g): an organic‐rich SEI (low modulus in left), an inorganic‐rich SEI (high modulus in middle), and an organic‐inorganic double‐layer SEI (high‐modulus inner layer with low‐modulus outer layer in right). It revealed that during Li⁺ deposition, stress within the organic SEI was predominantly distributed within Li dendrites, suggesting good volume adaptability of the organic‐rich SEI (Figure 4g,h). However, due to the low modulus, the organic‐rich SEI was unable to suppress the Li‐dendrite growth (Figure S36). In sharp contrast, the inorganic SEI exhibited enhanced capability in suppressing Li dendrites (Figure S37), while stress accumulated at the dendrite tips as Li deposits (Figure 4g,h), possibly leading to the fracture of SEI. As expected, the double‐layer SEI can simultaneously suppress the dendrite growth and dissipate the stress build‐up via its gradient structures, ensuring the interfacial stability at the mechanical level (Figures 4g,h, and S38).

2.5. Evaluation of Full‐Cell Performance

The double‐layer SEI of cPSiA₃F₇‐Li with high chemo‐mechanical stability enabled satisfactory cycling performance of the LMBs under both low and ambient temperatures. Herein, the industrial‐standard NCM811 cathode (97.5% active material content) was paired with cPSiA₃F₇‐Li anode for the electrochemical‐performance test of cPSiA₃F₇‐Li‐based LMBs. As shown in Figure S39, the coin cell with cPSiA₃F₇‐Li and NCM811 exhibited good rate performance with stable coulombic efficiency (CE) at a high current density of 4 C (1 C = 200 mA g⁻¹). A stable cycling performance with a capacity retention of 81.2% over 300 cycles was obtained from cPSiA₃F₇‐Li||NCM811 cell at 0.5 C and 25°C (Figure S40). On the contrary, the Bare‐Li||NCM811 cell exhibited faster capacity decay with capacity retention of 80% after 166 cycles, due to the structural instability (Figure S41) of SEI (even if a LiF‐rich SEI layer was formed) during long‐term cycling (Figure S42). Significantly, the side reactions between the Li anode and organic solvents were effectively suppressed by the elastomeric solvent‐phobic cPSiA₃F₇ layer. The influence of side reactions was more significant in a cell assembled with a thin Li anode (40 µm) and a high mass loading cathode (3 mAh cm⁻²). The capacity of Bare‐Li‐based coin cell was rapidly decaying after only a few cycles, while the coin cells with cPSiA₃F₇‐Li exhibited no significant capacity decay with capacity retention over 84% after 100 cycles (retention areal capability over 2.5 mAh cm⁻², Figure 5a). To the best of our knowledge, such superior electrochemical stability was rarely achieved for the cells with such high mass loading/high active material content and thin Li anode (Table S3). More interestingly, the cPSiA₃F₇‐Li‐based coin cells using two other commercial liquid electrolytes also manifested good cycling performance with capacity retention over 82% and 90% after 100 cycles, respectively (Figure S43), illustrating the universality of cPSiA₃F₇ as a protective layer for the Li anode.

FIGURE 5.

The effects of the cPSiA₃F₇ layer on cycling stability of Li||NCM811 batteries. (a) Cycling performance of Li (40 µm)||NCM811 (3 mAh cm⁻²) cells with different anodes at 0.5°C and 25°C. (b) Cycling performance of Li||NCM811 coin cells with different anodes at 0.5°C and −25°C. (c) Comparison of the performance of LMBs prepared with conventional low‐temperature electrolyte strategies [50, 51, 52, 53, 54, 55, 56, 57, 58]. (d) Cycling performance of Li (40 µm)||NCM811 coin cells with different anodes at 0.5 C and −25°C. (e) EIS plots after different cycles of Bare‐Li and cPSiA₃F₇‐Li. (f) Cycling performance of cPSiA₃F₇‐Li||NCM811 pouch cell at −25°C. The inset is the optical image of the corresponding pouch cell lighting LEDs after cycling.

It was found that the role of double‐layer SEI derived from the elastomeric solvent‐phobic coating was more significant in low‐temperature performance than at room temperature. At 0.5 C, cPSiA₃F₇‐Li||NCM811 coin cell exhibited remarkable cycling stability with a capacity retention of 99.3% and stable CE over 300 cycles at −25°C (Figure 5b). In contrast, the Bare‐Li||NCM811 coin cell demonstrated inferior cycling performance with only 51.8% of capacity retention after 300 cycles. Meanwhile, its CE fluctuated sharply after 100 cycles, implying the continuous side reactions on the Bare‐Li surface at −25°C. When compared to the recently reported low‐temperature electrolyte strategies, the solvent‐phobic elastomeric coating demonstrated significant advantages in terms of rate performance, cycle life, and capacity retention (Figure 5c). Meanwhile, with a thin Li anode (40 µm), the cPSiA₃F₇‐Li||NCM811 coin cell presented a stable cycling over 200 cycles with a capacity retention of 80.2%, while that of the Bare‐Li‐based coin cell was only 44.2% (Figure 5d). It can be explained that the double‐layer SEI effectively improved the interfacial stability and significantly diminished the consumption of liquid electrolyte. On the contrary, at low temperature, the chemically/mechanically unstable interface of Bare‐Li resulted in the continuous growth and exposure of Li dendrites with fast consumption of lithium inventory. The changes in the electrode/electrolyte interface impedance of Bare‐Li and cPSiA₃F₇‐Li after long‐term cycling were evaluated by EIS in combination with a simulation circuit (Figures 5e and S44). It showed that the SEI resistance (R_SEI ) of BareLi exhibited a continuous increase with cycling progression, while that of cPSiA₃F₇‐Li stabilized after the 20th cycle, demonstrating the high electrochemical stability of such a double‐layer SEI at low temperature. SEM images also provided clear evidence of the mechanical stability of such SEI at low temperatures, where cPSiA₃F₇‐Li exhibited a smooth surface morphology, yet Bare‐Li showed an extremely irregular anode surface (Figure S45).

Moreover, a simple spraying technique was employed to fabricate the cPSiA₃F₇‐Li (40 µm)||NCM811 pouch cells to evaluate the scalability of such elastomeric solvent‐phobic coating. The cPSiA₃F₇‐Li (40 µm)||NCM811 (3 mAh cm⁻²) pouch cell exhibited a high specific capacity of 186.2 mAh g⁻¹ at 0.5 C with capacity retention of 90% over 60 cycles (Figure S46). Meanwhile, the pouch cell also showed stable cycling performance at −25°C, with a capacity retention over 80% for 100 cycles at 0.5 C. The pouch cell could consistently illuminate a light‐emitting diode after low‐temperature cycling (Figure 5f).

3. Conclusion

In this study, we proposed a solvent‐phobic elastomeric coating with high chemo‐mechanical stability, which enabled the stable cycling of LMBs under low temperatures. After optimizing the intrinsic properties of solvent‐phobic coating, that is, surface energy, mechanical resilience, and dielectric constants, a double‐layer SEI structure comprising an organic polymer outer layer and an inorganic‐rich inner layer was obtained. The multi‐physical modeling and experimental results show that the double‐layer SEI constructed by such solvent‐phobic elastomeric coating can (1) effectively suppress the side reactions between organic solvents and Li anode by selectively transporting ions yet repelling organic solvents, (2) alleviate the electrode polarization on the anode surface with homogenous Li⁺ deposition, (3) and mitigate the stress concentration between the Li dendrite and inorganic‐rich SEI with reduced risk of mechanical failure. With a synergistic effect between organic and inorganic phases, the derived double‐layer SEI enabled the LMB to stably cycle for 300 cycles with a capacity retention of 99.3% at low temperatures. The significantly improved cycling performance of such cell systems, coupled with other commercial liquid electrolytes, also manifested the universality of cPSiA₃F₇ as a protective layer. In addition, the cPSiA₃F₇‐based elastomeric coating demonstrated remarkable electrochemical performance, which enabled a Li||NCM811 pouch cell with a high specific capacity of 184.7 mAh g⁻¹ at 25°C and exceptional cycling stability at −25°C. With a facile fabrication approach, multi‐physical insight, stable cycling performance, wide applicability, and scalability, the delicate design of solvent‐phobic elastomeric coating should shed light on large‐scale development of safe and long‐cycling LMBs and related energy‐storage systems.

Author Contributions

Xinyuan Shan: Writing – review and editing. Linming Bai: Data curation and writing – review and editing. Ming Tian: Writing – review and editing, funding acquisition, supervision, and resources. Jingren Gou: Writing – review and editing, supervision, and writing – original draft. Han Qin: Writing – review and editing, and software. Hang Ding: Conceptualization, methodology, writing – original draft, writing – review and editing, data curation, visualization, validation, investigation, and formal analysis. Yue Li: Methodology, writing – review and editing. Peng‐Fei Cao: Writing – review and editing, resources, funding acquisition, supervision, project administration, and conceptualization. Sijin Jin: Writing – review and editing and validation. Jiamin Gao: Writing – review and editing and visualization.

Conflicts of Interest

None of the authors has conflicts of interest to disclose

Funding

This research was supported by the National Natural Science Foundation of China (grant no. 52373275 received from P. F. Cao and grant no. 52221006 received from M. Tian).

Supporting information

Supporting File: adma73917‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.