Cost-effective interfacial high-concentration electrolyte for stable lithium metal batteries

Abstract

High-concentration electrolytes are promising candidates for high-energy-density lithium metal batteries, yet their practical application is hindered by sluggish ionic transfer and high costs. Here, we show an interfacial high-concentration electrolyte that overcomes these limitations by localizing 5 M LiTFSI electrolyte at the Li anode surface, while maintaining a 1 M electrolyte in the bulk. This approach is facilitated by a bi-continuous phase-separated polymer layer that provides physical confinement and leverages hydrogen-bonding and ion-dipole interactions to prevent salt diffusion from the interfacial layer into the bulk electrolyte. This design preserves the high ionic conductivity of the bulk electrolyte while ensuring sufficient Li⁺ for rapid interfacial charge transfer. It also cuts lithium salt usage and associated costs by up to 70%, while retaining the benefits of high-concentration electrolytes, including the formation of a LiF-rich solid-electrolyte interphase. As a result, a 6.8 Ah Li | |NCM811 pouch cell delivers a specific energy of 506 Wh kg^-1 at 0.1 C and maintains stable cycling over 200 cycles with 75.8% capacity retention at 0.5 C. This work demonstrates an effective electrolyte design that offers a cost-effective and sustainable pathway for high-energy-density lithium metal batteries.

Introduction

Lithium (Li) metal stands out as an ideal negative electrode material for high energy density batteries due to its low potential (−3.04 V vs. standard hydrogen electrode), and impressive theoretical specific capacity (3860 mAh g^-1 and 2061 mAh cm^-3)^1,2,3,4,5. However, the practical application of Li metal is hampered by issues including low Coulombic efficiency (CE), limited cycle life and safety concerns^6,7. These challenges stem from the inevitable growth of Li dendrites^8,9,10, and uncontrollable parasitic reactions between Li and electrolyte¹¹.

Electrolyte engineering is acknowledged as a pivotal and cost-effective strategy to stabilize solid electrolyte interphase (SEI) and regulate Li morphology by tuning electrolyte formulations and salt concentrations^1,12,13,14. High-concentration electrolytes (usually ≥ 3 M, HCEs) have aroused tremendous interest because of their unusual functionalities in simultaneously improving interfacial stability and the electrochemical properties of Li metal batteries (LMBs)^15,16,17. As shown in Fig. 1a, traditional “standard” concentration electrolytes (~1 M, SCEs) have high ionic conductivity and are cost-effective, making them suitable for most commercial applications. Yet, their low Li⁺ transference number causes a great concentration polarization and insufficient Li⁺ mass transfer to the electrode interface, leading to uneven Li deposition. Moreover, the SEI formed in SCEs, typically derived from organic-rich components, is fragile and prone to promoting Li dendrite growth, ultimately leading to degradation of cycling performance. In contrast, HCEs with elevated salt concentrations exhibit reduced concentration polarization and more homogeneous ionic mass transfer. Furthermore, HCEs, dominated by anion-rich solvation structures^18,19,20,21, tend to form an inorganic-rich SEI with a high Young’s modulus²², which is mechanically robust enough to inhibit Li dendrite growth, thereby enabling the long-term cycling stability of LMBs²³ (Fig. 1b).

Fig. 1: Schematic for Li⁺ distribution and drawbacks/advantages in different electrolytes.

a “Standard” concentration electrolyte (SCE): high ionic conductivity, fragile SEI, and cost-effective. b High-concentration electrolyte (HCE): low ionic conductivity, robust SEI, and high cost. c Interfacial high-concentration electrolyte (IHCE): high ionic conductivity in bulk electrolyte, improved interfacial charge transfer kinetics, a robust SEI with an additional polymer protective layer, and cost-effective.

Although HCEs perform many unique functions not shared with SCEs, their commercialization was impeded by challenges such as high cost, low ionic conductivity and poor wettability^20,24. Taking the traditional electrolyte composed of 1 M lithium hexafluorophosphate in ethylene carbonate/diethyl carbonate (1: 1, v/v) as an example, increasing the LiPF₆ concentration from 1 M to 4 M leads to a significant rise in both lithium salt consumption and cost. Specifically, lithium salt usage jumps from 11% to 33%, while the cost escalates from 32% to 65% (Supplementary Note 12). This substantial increase in lithium salt usage and associated costs poses challenges to resource conservation and hinders progress toward sustainable development. To further advance HCEs, a family of inert diluents was introduced into HCEs to form localized high-concentration electrolytes (LHCEs)^25,26,27. LHCEs can partially preserve the solvation structures of HCEs at low salt concentrations, which not only inherits the merits of HCEs, but also compensates for its shortcomings of high viscosity and poor wettability^28,29. Nonetheless, commonly used diluents, such as highly fluorinated ethers^29,30, fluorinated orthoformate³¹, fluorobenzene³² and fluorinated alkanes³³, are highly toxic and environmentally harmful³⁴. Moreover, fluorinated solvents are expensive³⁵, posing challenges for practical applications and sustainable development.

Herein, we propose an interfacial high-concentration electrolyte (IHCE) design by pairing a high salt concentration interfacial layer with SCE to stabilize the solid-electrolyte interphase (SEI), enhance Li⁺ transfer kinetics, and decrease the electrolyte costs (Fig. 1c). The high salt concentration interfacial layer is constructed through a densely bi-continuous phase-separated network, which imposes spatial confinement, along with hydrogen bonding and ion-dipole interactions, effectively preventing salt diffusion from the interfacial layer into SCE. We demonstrate that the gradient-concentration electrolyte design of IHCE plays a crucial role in improving Li⁺ flux in the interfacial layer, retaining high ionic conductivities of the SCE, and reducing lithium salt usage and associated costs by ~70% compared to the HCE. The IHCE strategy also induced the formation of salt-derived SEI on the Li electrode. Consequently, under a high charge/discharge rate of 2 C, the IHCE@Li | |NCM811 full cells delivered a discharge capacity (~164 mAh g^-1, 82% of the capacity at 0.1 C). Furthermore, we achieved a 6.8 Ah IHCE@Li | |NCM811 pouch cell with a high specific energy of 506 Wh kg^-1 and sustained stable cycling over 200 cycles with a capacity retention of 75.8% in the SCE at 0.5 C.

Results

Design and characterization of IHCE

In this work, we constructed a salt-concentrated interfacial layer based on a polymer matrix of poly(vinylidene fluoride-co-hexafluoropropylene) (PVDF-HFP) and cross-linked polyethylene glycol diacrylate (CPEGDA) (Fig. 2a, b). This interfacial layer was integrated with standard-concentration bulk electrolyte, consisting of 1 M LiFSI in DME, for use in half-cell configurations. As shown in Supplementary Table 1, PVDF-HFP and CPEGDA exhibit a large Flory-Huggins interaction parameter (χ = 6.5 k), determined through contact angle measurements (Supplementary Fig. 1 and Note 1), which facilitates the formation of a bicontinuous phase microstructure³⁶. The PVDF-HFP with rich fluorine (F) atoms was chosen for its excellent chemical stability, film-forming integrity, and mechanical properties³⁷. Notably, C-F bonds in PVDF-HFP are anticipated to offer hydrogen-bonding and ion-dipole interactions among PVDF-HFP, solvents, and TFSI^- anions, thereby trapping LiTFSI in the interfacial layer and preventing its diffusion into the bulk electrolyte. The detailed salt-retention mechanism of IHCE will be discussed in the next section. CPEGDA, enriched with oxygen-containing groups (-C-O-C-, C = O), was selected for its favorable ionic conductivity and ability to absorb electrolyte solvents. The cross-linked structure was designed to enhance the mechanical properties of the flexible PEGDA chain segments and strengthen the overall structural integrity of the PVDF-HFP/CPEGDA polymer matrix. Details regarding the synthesis of the IHCE are provided in the Methods section and Supplementary Figs. 2, 3. High ionic conductivity of 1.4 × 10^-5 S cm^-1, as well as tensile strength of 91 MPa, were obtained at a PVDF-HFP/CPEGDA weight ratio of 8: 2 with 50 wt% LiTFSI, as demonstrated by Nyquist plots (Supplementary Fig. 5a and Table 2) and stress-strain curves (Supplementary Fig. 5b). The solvent uptake of the polymer matrix for dimethyl ether (DME) solvent was controlled at ~30 wt% (Supplementary Table 3). This limited solvent uptake was beneficial for constructing high-concentration electrolytes.

Fig. 2: Design and characterization of IHCE.

a, b Schematic illustration of design and chemical structure for IHCE. The chemical structure of IHCE shows a bicontinuous phase that comprises interpenetrating PVDF-HFP (blue color) and CPEGDA (pink color). c Cross-sectional SEM image of dry-IHCE coated on Cu. d, e SEM images d and corresponding electron energy dispersive X-ray spectroscopy mapping e of dry-IHCE.

As shown in Fig. 2c, a dense thin film (~2.0 μm) was formed on the surface of Cu foil by using a facile doctor-blade coating method. As expected, this film exhibits a distinct phase-separation structure with interconnected island-shaped domains surrounded by a cross-linked matrix, as shown in Fig. 2d, e and Supplementary Fig. 6. The characteristic dispersive regional cluster distribution of F and O elements, corresponding to scanning electron microscopy (SEM) images, confirmed that the F-rich PVDF-HFP phase is embedded in the O-rich CPEGDA phase. This observation was further supported by polarizing optical microscopy (POM), which provided macroscopic evidence consistent with SEM observations (Supplementary Fig. 7). Additionally, 3D laser scanning confocal microscope images (Supplementary Fig. 8) show that PVDF-HFP, selectively labeled with Nile Red, is uniformly distributed throughout the IHCE membrane. This uniform phase distribution ensures consistent polymer blending without large-scale phase separation, contributing to the structural stability and ionic transport properties of IHCE.

Salt-retention mechanism of IHCE

To quantify the salt-retention ability of IHCE, we submerged IHCE film in DME for 50 days and continuously monitored the absorbance of the soaking solution at an absorption peak of ~210 nm, corresponding to the n→π* transition of -SO₂ in LiTFSI. According to Lambert-Beer law, the absorbance of LiTFSI is directly proportional to its concentration. We used this linear relationship to establish the UV standard curve³⁸ (Supplementary Fig. 9), determine the concentration of LiTFSI in the soaking solution, and calculate the residual LiTFSI concentration in the IHCE. The details are described in Supplementary Tables 4-6 and Note 4. As shown in Fig. 3a and Supplementary Fig. 10, after soaking in DME for 50 days, the residual concentration of LiTFSI in IHCE remained at 3.53 M. Compared to the initial 5 M concentration of LiTFSI in IHCE, the daily dissolution rate was less than 0.88% during 2–50 days, demonstrating the good salt-retention ability of IHCE in electrolyte solvents. Moreover, IHCE also exhibited similarly strong salt-retention performance in carbonate-based solvents such as FEC/DEC, as detailed in Supplementary Figs. 11, 12, Tables 7, 8, and Note 5.

Fig. 3: Salt-retention mechanism of IHCE.

a Dissolution and residual concentration profiles of LiTFSI in IHCE versus immersion time. b Fractional free volume (FFV) models of IHCE by molecular dynamics simulations. The grey surface indicates the van der Waals surface, and the blue surface is the Connolly surface with a probe radius of 1.3 Å. Purple, lithium; red, oxygen; blue, nitrogen; ice-blue, fluorine; yellow, sulfur; grey, carbon; white, hydrogen. c Pore size distribution obtained from FFV analysis. d Diffusion coefficients of TFSI^- in SCE (1 M LiTFSI-DME) and IHCE layer. e Binding energies of TFSI^- with PVDF-HFP, CPEGDA and DME. f FTIR spectra of dry-IHCE and IHCE. g ¹H NMR spectra of DME, PVDF-HFP-DME and PVDF-HFP-DME-LiTFSI. h Schematic of the IHCE in LMBs.

To investigate the salt-retention mechanism, molecular dynamics (MD) simulations were conducted to analyze the physical structure of IHCE. As shown in Figs. 3b, Supplementary Figs. 13, 14 and Note 6, the introduction of DME increases the free volume of the PVDF-HFP/CPEGDA network from 7275 ų to 10319 ų, primarily attributed to an expanded polymer network and enhanced polymer chain mobility. Upon incorporation of LiTFSI, the total free volume decreases to 9530 ų, indicating partial occupation of the cavities by TFSI^- anions. More importantly, pore size distribution analysis (Fig. 3c) shows that the dominant pore sizes in IHCE are below 2.5 Å, much smaller than the solvation diameter of TFSI^-39, thereby imposing a strong physical barrier to anion migration. This sub-nanometer confinement effect is further validated by the measured diffusion coefficient of TFSI^- in IHCE (Fig. 3d and Supplementary Fig. 15), which is 0.04 × 10^-2 Ų ps^-1, representing a ~ 98.4% reduction compared to 2.47 × 10^-2 Ų ps^-1 in SCE (1 M LiTFS-DME). These results confirm that the polymer matrix effectively restricts anion mobility through sub-nanometer confinement. Beyond spatial confinement, molecular-level interactions also contribute to TFSI^- immobilization. Density functional theory (DFT) calculations (Fig. 3e and Supplementary Fig. 16) show that TFSI^- exhibits stronger binding to PVDF-HFP (−15.03 kcal mol⁻¹) and CPEGDA (−10.22 kcal mol⁻¹) than to DME (-7.91 kcal mol⁻¹), indicating that the polymer chains provide favorable anchoring sites for anions. In the presence of DME, ternary complexes such as PVDF-HFP-DME-TFSI^- (−19.26 kcal mol^-1) and CPEGDA-DME-TFSI^- (-22.60 kcal mol^-1) show enhanced binding energies (Supplementary Fig. 17), suggesting that DME serves as a coordination bridge that strengthens polymer-anion anchoring. This is further supported by electrostatic potential (ESP) analysis (Supplementary Fig. 18), which shows that DME in polymer-DME complexes exhibits higher local ESP maximum (+15.3 and +10.4 kcal mol^-1) and less negative ESP minimum (−31.7 and −25.6 kcal mol⁻¹) compared to pure DME ( + 12.4 and −42.1 kcal mol⁻¹). The increased electropositivity of DME facilitates stronger electrostatic attraction with negatively charged TFSI^-, thereby promoting anion immobilization within the IHCE framework. These computational findings are corroborated by spectroscopic evidence. As shown in Fig. 3f, FTIR spectra reveal redshifts in the -CF₂ vibrational bands of PVDF-HFP from 1063 cm^-1 and 1147 cm^-1 (dry IHCE) to 1056 cm^-1 and 1133 cm^-1 in DME-swollen IHCE, indicating hydrogen-bond-like interactions among PVDF-HFP, DME, and TFSI^−40,41. Consistently, ¹H NMR (Fig. 3g) resonance of DME shifted to the downfield direction, going from 3.30 and 3.33 ppm in PVDF-HFP-DME to 3.48 and 3.51 ppm in PVDF-HFP-DME-TFSI^-, further validating H···F and H···O interactions among DME, TFSI^- and PVDF-HFP. These results demonstrate that the salt-retention ability of IHCE is primarily governed by sub-nanometer confinement of solvated anions, and solvent-mediated interactions between TFSI^- and the polymer matrix.

To more clearly identify the dominant mechanism responsible for salt retention, we conducted comparative experiments using a control polymer matrix that possesses hydrogen-bonding and ion-dipole functionalities but lacks a sub-nanoconfined structure. As shown in Supplementary Fig. 19, Table 9, and Note 7, this control sample exhibited rapid salt leaching, indicating that weak molecular interactions alone are insufficient for effective anion retention. Additionally, we also conducted temperature-dependent salt-retention tests at 40 °C, which is above typical ambient conditions (25 °C) but remains below the glass transition temperature of our polymer matrix. The residual salt content remained nearly unchanged compared to that at 25 °C (Supplementary Fig. 20 and Table 10), confirming that salt retention in IHCE is insensitive to moderate temperature. These results collectively confirm that sub-nanometer physical confinement is the primary mechanism governing salt retention in the IHCE, while hydrogen-bonding and ion-dipole interactions act as supportive factors.

Raman spectroscopy was employed to characterize the coordination environment of TFSI^- and elucidate the solvation structure of IHCE. Typically, the deconvoluted peaks between 740–749 cm⁻¹ are divided into three distinct TFSI^- dissociation modes: solvent-separated ion pairs (SSIPs) at ~740 cm⁻¹, contact ion pairs (CIPs) at ~744 cm⁻¹, and ion aggregates (AGGs) at ~749 cm^{−1 37}. As shown in Supplementary Fig. 21, the relative intensities of CIPs and AGGs are higher in IHCE (5 M LiTFSI-DME) than in bulk SCE (1 M LiTFSI-DME), indicating strong Li⁺-TFSI^- interactions and fewer free solvent molecules, consistent with previously observed TFSI^- dissociation modes in traditional HCEs²⁰. In general, a larger number of free solvent molecules in SSIPs contributes to the higher ionic conductivity, but is typically associated with lower Li⁺ transference numbers (t_Li⁺)⁴². In contrast, improved Li⁺-TFSI^- interactions within AGGs enhance t_Li⁺, but often reduce ionic conductivity due to restricted ion mobility²⁴. Compared to traditional SCE, the IHCE with saturated DME ( ~ 30 wt%) exhibits a considerable proportion of CIPs (47.5%), which form smaller ion clusters than AGGs and are favorable for increasing the t_Li⁺ and promoting Li⁺ mobility. This solvation environment results in a t_Li⁺ of 0.57 (Supplementary Fig. 30), significantly higher than those of SCE (0.33) and HCE (0.42), indicating suppressed anion migration and more efficient Li⁺ transport. In addition, ionic conductivity measurements show that IHCE layer (Supplementary Fig. 22 and Table 2) exhibits a conductivity of 1.12 × 10⁻³ S cm⁻¹, which is significantly higher than that of conventional artificial protective layers (~10⁻⁵ to 10⁻⁴ S cm⁻¹)^43,44. Thereby, these combined advantages of a high t_Li⁺ and moderate ionic conductivity enable efficient Li⁺ transport within the IHCE layer. The schematic of the IHCE on the Li anode surface is depicted in Fig. 3h.

Morphology and interfacial chemistry of Li electrode

SEM analysis was conducted to observe the morphological changes of the Li electrode at a high current density of 4 mA cm^-2. For the SCE, the resulting morphology (Fig. 4a, b) exhibits loose and porous rod-like structures with a thickness of ~8.3 μm after the initial Li deposition. After 50 cycles, more pronounced cracks appeared (Fig. 4c), and the thickness of deposited Li in the SCE increased to ~13.9 μm (Fig. 4d), reflecting a ~ 67.5% increase compared to the initial Li deposition. By contrast, regular Li growth with larger granules was observed in HCE after initial Li deposition. However, the thickness of deposited Li increased from ~5.5 μm to ~8.9 μm after 50 cycles (Supplementary Fig. 23), indicating a ~ 61.8% increase relative to the initial Li deposition. In sharp contrast, IHCE@Cu showed large granular sizes and smooth scale-like Li deposition with a thickness of ~5.1 μm (Fig. 4e-h), closely resembling the theoretical thickness (5 μm) for the same deposition capacity of 1 mAh cm^-2. Even after 50 cycles, a flat and dense Li morphology was maintained, with a slight thickness increase to ~6.7 μm. 36.5% and 31.4% decreases in the thickness relative to SCE and HCE were achieved, respectively. This suppression of Li dendrite growth and volume expansion is also evident under a lower current density of 1 mA cm^-2 (Supplementary Figs. 24, 25). In the SCE system, the deposited Li expands from ~7.8 μm to ~13.2 μm after 50 cycles, corresponding to a ~ 69.2% increase. In the HCE system, the thickness increases from ~5.4 μm to ~7.9 μm (~46.3% increase). In sharp contrast, IHCE@Cu maintains a compact morphology, with only a slight thickness increase from ~5.1 μm to ~5.9 μm (~15.7% increase).

Fig. 4: Characterization of Li morphology and interfacial chemistry.

SEM images of deposited Li at a current density of 4 mA cm^-2. a, c, e, g Top view and b, d, f, h cross-sectional SEM images of deposited Li on bare Cu a–d and IHCE@Cu e–h in SCE after 1 cycle and 50 cycles. FEM simulations showing the evolution of concentration distribution and electric field strength at t = 0 and 300 s. i-l The Li⁺ concentration distributions on the bare Cu i, j and IHCE@Cu k, l. m, p The electric field strength evolution on the bare Cu m, n and IHCE@Cu o, p. q The normalized curves of Li⁺ concentration versus time at the Li electrode interface. Interfacial chemistry of IHCE. r, s TOF-SIMS depth profiles r and 3D reconstructed sputtering images s of SEI on the cycled IHCE@Cu surface. The red color represents LiF₂^-, the blue color represents C₂H₃O^- and the grey color represents Cu₂^-.

Finite-element method (FEM) simulations based on cubic current distribution were employed to investigate the temporal evolution of interfacial Li⁺ concentration and electric field strength at the IHCE@Li electrode. According to Chazalviel’s “space-charge model”⁴⁵, rapid Li⁺ deposition in SCE causes a sharp drop in anion concentration at the electrode surface, leading to significant space charge formation and generating an electric field that induces dendritic Li growth. Under high current densities, insufficient ion transfer exacerbates concentration polarization, further intensifying dendrite nucleation and growth. As shown in Fig. 4k, l, IHCE maintains a high and uniform Li⁺ distribution within the IHCE layer during prolonged simulation, effectively minimizing concentration polarization at the IHCE@Li electrode surface. Additionally, a relatively low and consistent electric field strength is obtained at the IHCE@Li electrode (Fig. 4o, p). In contrast, the bare Li exhibits a significant drop in ion concentration (Fig. 4i, j) and an uneven electric field distribution (Fig. 4m, n), which accelerates the growth of Li dendrites. Correspondingly, the normalized Li⁺ concentration profiles over 500 s further highlight the high Li⁺ concentration at the IHCE@Li surface (Fig. 4q). In comparison, the bare Li in SCE experiences a rapid decline in Li⁺ concentration, dropping to near-zero levels within 100 s. These results suggest that IHCE effectively compensates for Li⁺ depletion and mitigates surface concentration variations by enriching the average Li⁺ concentration in the interfacial layer of Li electrode, thereby creating a uniform and low electric field strength that inhibits dendritic Li growth.

To explore the interfacial chemistry derived from IHCE, we employed cryogenic transmission electron microscopy (Cryo-TEM) to visualize the nanostructure of the SEI. As shown in Supplementary Fig. 26, the SEI presents a uniform and gradient structure, consisting of an outer organic amorphous layer (~4 nm) and an inner organic/inorganic composite layer (~3 nm). The predominant crystalline LiF embedded in the inner layer, displaying a mosaic-type structure, is confirmed by ordered lattices (2.00 Å) with lattice planes (200) in corresponding fast Fourier transform images⁴⁶. Additionally, time-of-flight secondary ion mass spectrometry (TOF-SIMS) depth profiling was conducted to analyze the relative proportions and three-dimensional (3D) distribution of components in the SEI. As seen in Fig. 4r, s, the SEI species, including C₂H₃O^- (corresponding to organic components) and LiF₂^- (corresponding to LiF), exhibit an obvious gradient distribution throughout the spatial region, consistent with the Cryo-TEM observations. The pronounced C₂H₃O^- signal and no distinctive LiF₂^- signals in the initial sputtering indicate a higher concentration of organic components in the top part of the SEI. In addition, the signal of LiF₂^- segments gradually increases during 90 s of sputtering, confirming the formation of inorganic-rich fractions in the inner SEI layer.

Fast Li⁺ transfer kinetics

Under high current densities, the hysteretic Li⁺ transport leads to severe concentration polarization. This, in turn, induces dendrite growth and dead Li formation, while also exacerbating SEI fracture and reformation, ultimately resulting in a sharp decline in cycling stability^47,48. The Li | |Li symmetric cells were first fabricated to evaluate the overall ionic migration and overpotential under high current densities. As shown in Fig. 5a-c and Supplementary Fig. 27, IHCE exhibits better rate capability and flatter voltage profiles compared to both HCE and SCE, with low overpotential values of 27.0, 57.1, and 69.2 mV at current densities of 1, 4, and 6 mA cm^-2, respectively. IHCE also achieves a prolonged lifespan of 300 h with a stable overpotential of only 43 mV at a high current density of 4 mA cm^-2, whereas both HCE and SCE exhibit overpotentials exceeding 200 mV after only 20 h (Supplementary Fig. 28). Additionally, IHCE@Li | |Cu asymmetric cells display a high CE of 98.3% and an extended lifespan of 370 cycles under the same conditions (Supplementary Fig. 29).

Fig. 5: Fast Li⁺ transfer kinetics.

a Rate performance of Li | |Li symmetric cells at different current densities. b, c Voltage profiles of Li | |Li half cells at different current densities for HCE b and IHCE c. d The Li⁺ transference number of the SCE, HCE and IHCE. e Tafel plots for Li plating/stripping in SCE, HCE and IHCE. The inset shows the values of the corresponding exchange current densities. f Activation energies derived from Nyquist plots of cycled Li | |Cu cells with SCE, HCE and IHCE, respectively.

Fast Li⁺ transport is indispensable for good rate performance. As shown in Fig. 5d and Supplementary Fig. 30, the as-fabricated IHCE interfacial layer exhibits a higher Li⁺ transference number of 0.57, compared to HCE (0.42) and SCE (0.33), endowing reduced concentration polarization and facile interfacial Li⁺ transfer kinetics. Tafel plots in Fig. 5e display that the exchange current density in IHCE (0.84 mA cm^-2) is nearly twice that of HCE (0.49 mA cm^-2) and three times that of SCE (0.30 mA cm^-2), indicating IHCE significantly enhances the interfacial transfer kinetics. We further calculated the activation energy (E_a) for Li⁺ migration across the SEI using Arrhenius equation by fitting the temperature-dependent EIS profiles (Supplementary Fig. 31). As illustrated in Fig. 5f, the E_a (54.3 kJ mol^-1) of IHCE is lower than that of SCE (58.9 kJ mol^-1) and HCE (80.6 kJ mol^-1), indicating a smaller Li⁺ transport barrier and enhanced diffusion of Li⁺ through SEI. This is further evidenced by the Nyquist plots of the IHCE@Li | |Cu asymmetric cells, which display small and stable SEI resistances (R_SEI) (Supplementary Fig. 32), attributed to the high ionic conductivities of the SEI. Thus, compared to the conventional HCE strategies, the enhanced Li⁺ transport in IHCE is attributed to the gradient-concentration electrolyte design: (i) the bulk SCE, with its high ionic conductivity (~2 times higher than that of HCE, Supplementary Fig. 33 and Table 13) and low viscosity, ensures rapid transport of solvated Li⁺ through the electrolyte; (ii) the IHCE layer, enriched with high-concentration electrolyte, provides a sustainable and stable Li⁺ supply, reducing concentration polarization at the interface, which enhances the kinetics of interfacial charge transfer and Li deposition/stripping process; and (iii) IHCE-assisted SEI exhibits high ionic conductivity, thereby lowering Li⁺ transport barrier through SEI.

Electrochemical performance of Li | |NCM811 full cells

The effectiveness of IHCE was first evaluated in IHCE@Li | |NCM811coin-type full cells, utilizing thin Li metal electrodes (50 μm-thick) and high-loading positive electrodes (~22 mg cm^-2). Figure 6a and Supplementary Fig. 34 show that the performance of IHCE@Li | |NCM811 full cells in SCE outperforms bare Li | |NCM811 full cells in both SCE and HCE. Specifically, IHCE@Li | |NCM811 cells achieved 80% capacity retention over 328 cycles with a high Coulombic efficiency of 99.84%. In sharp contrast, bare Li | |NCM811 cells in SCE and HCE display fast capacity fading and severe voltage polarization after only 50 cycles (Supplementary Fig. 35). Our IHCE design represents a significant improvement in cell cycle life at high-loading positive electrodes compared to other state-of-the-art strategies with area-capacity > 3 mAh cm^-2 (Fig. 6b and Supplementary Table 14). Additionally, the improved rate performance of IHCE@Li | |NCM811 cells is demonstrated in Supplementary Fig. 36, which delivers higher discharge capacity (~164 mAh g^-1) than that of bare Li | |NCM811 cells in HCE ( ~ 158 mAh g^-1) and SCE ( ~ 72 mAh g^-1) at 2 C (1 C = 180 mA g^-1). The corresponding galvanostatic charge/discharge voltage profiles for bare Li | |NCM811 cells also reveal a drastic polarization increase at 2 C, especially in SCE. These results demonstrate that IHCE can facilitate the electrochemical stability of the cells at a high charge/discharge rate of 2 C, due to its fast Li⁺ transfer kinetics, as shown in Fig. 5.

Fig. 6: Electrochemical performance of Li | |NCM811 full cells.

a Long-term cycling of the coin-type full cells with SCE, HCE and IHCE at 0.5 C discharge (1 C = 180 mA g^-1). b Comparison of Li-metal coin-type full cells in previous reports and this work (criteria: area capacity ≥3 mAh cm^-2 and capacity retention ≥80%). c Voltage profiles of 506 Wh kg^-1 IHCE@Li | |NCM811 pouch cells during the activation processes at 0.1 C (1 C = 6600 mA g^-1). d Weight ratio of cell components in the 506 Wh kg^-1 IHCE@Li | |NCM811 pouch cells; detailed information is provided in Supplementary Table 15. e Comparison of lithium salt usage and electrolyte costs for Li-metal pouch cells with HCE and IHCE at an E/C ratio of 1.29 g Ah^-1. f, Cycling performance of 506 Wh kg^-1 IHCE@Li | |NCM811 pouch cells at 0.5 C discharge.

To meet the long-term objective of developing rechargeable batteries with a specific energy exceeding 500 Wh kg^-1, it is essential to adopt aggressive battery parameters, such as lean electrolyte (electrolyte mass/cell capacity (E/C) ratio <1.3 g Ah^-1), low negative/positive capacity ratio (N/P ratio <2), and high-capacity (> 4 mAh cm^-2) positive electrodes^49,50 and so on. Increasing the areal loading of positive electrode and reducing the E/C ratio are recognized as effective strategies to enhance specific energy. However, these approaches often pose challenges, as they can compromise cycle life. High loading typically leads to poor electrolyte wetting and increased internal polarization resistance, resulting in capacity fade and accelerated cycle degradation. Additionally, lean electrolyte conditions exacerbate electrolyte depletion, further accelerating cell degradation and shortening cycle life. In this study, we fabricated and evaluated an industrial-level IHCE@Li | |NCM811 pouch cell with a low N/P ratio of 1.51, comprised of 0.1 mm Li negative electrode and 30 mg cm^-2 (capacity: ~6.8 mAh cm^-2) NCM811 positive electrode, under lean electrolyte conditions (E/C: ~1.29 g Ah^-1) (Supplementary Fig. 37). As shown in Fig. 6c, the IHCE@Li | |NCM811 pouch cell delivers a high initial discharge capacity of 6.8 Ah during the activation processes and achieved a specific energy of 506 Wh kg^-1 at 0.1 C. The weight ratio and cell components are shown in Fig. 6d and Supplementary Table 15. Notably, the 506 Wh kg^-1 IHCE@Li | |NCM811 pouch cell delivers 75.8% capacity retention after 200 cycles at a discharge rate of 0.5 C (1 C = 6600 mA g^-1, Fig. 6f and Supplementary Fig. 38). Moreover, the performances of our IHCE@Li | |NCM811 pouch cells, in terms of specific energy and cycle number, shows improvement with respect to previously reported LMBs based on the criteria of specific energy ≥ 400 Wh kg^-1 (Supplementary Table 16). The primary factor contributing to the achievement of this high reversible 506 Wh kg^-1 LMBs is gradient-concentration electrolyte design. The bulk SCE, with its low viscosity and high ionic conductivity, endows good electrolyte wetting and decreases internal polarization resistance under high-loading (~30 mg cm^-2) positive electrode. Simultaneously, the high salt concentration interfacial layer promotes robust SEI formation on the Li electrode and compensates for electrolyte depletion under lean electrolyte conditions (E/C ratio ~1.29 g Ah^-1). These advancements enable long-lasting practical 500 Wh kg^-1 cells over 200 cycles.

The IHCE strategy has demonstrated advantages in improving both specific energy and long-term cycling stability of Li-metal pouch cells, which are crucial for sustainable development. Equally important is the cost control of battery manufacturing, especially in the context of sustainability. Reducing production costs can boost market competitiveness, encourage the adoption of electric transportation and renewable energy, and ultimately maximize environmental benefits. We calculated and compared lithium salt usage and associated costs for SCE, HCE, and IHCE per ampere-hour (Ah) in pouch cells, based on the E/C ratio (1.29 g Ah^-1) established in this work. As shown in Fig. 6e and Supplementary Tables 18-21, the lithium salt usage and associated costs of IHCE are reduced by ~70% compared to HCE and are closely aligned with SCE. Furthermore, the total cost of IHCE was reduced by ~23% relative to that of HCE. Thus, beyond meeting the escalating energy-density demands, the cost advantages of IHCE also contribute to promoting economic sustainability in practical applications.

Discussion

In summary, we have developed an interfacial high-concentration electrolyte (IHCE) that localizes a high-concentration electrolyte (5 M LiTFSI-DME) at the Li interface, while the bulk electrolyte remains at 1 M. This design was achieved through a densely bi-continuous phase-separated polymer layer, wherein a mechanically robust PVDF-HFP domain is embedded in a continuous ionic-conductivity CPEGDA matrix. This structure provides physical confinement of high-concentration electrolyte while leveraging hydrogen-bonding and ion-dipole interactions to prevent salt diffusion from the interfacial layer into the bulk electrolyte. We demonstrated that the IHCE strategy combines the advantages of HCE and SCE, offering fast Li⁺ transfer kinetics, reduced lithium salt usage, an ideal salt-derived SEI, and minimized electrolyte costs. The as-fabricated IHCE@Li | |NCM811 full cells achieved a high discharge capacity of 164 mAh g^-1 under a high charge/discharge rate of 2 C. Notably, a 6.8 Ah practical pouch-type cell enabled a specific energy of ~506 Wh kg^-1 and exhibited long cycling stability over 200 cycles under 0.5 C.

More importantly, the development of IHCE marks a significant advancement in energy storage, aligning with the Sustainable Development Goals. By improving specific energy and extending battery lifespan, IHCE contributes to more efficient and durable LMBs, which are crucial for renewable energy systems and electric vehicles. Additionally, reduced lithium salt usage, lower electrolyte costs, and enhanced performance at higher charge/discharge rates make IHCE economically and environmentally sustainable. In the long term, IHCE will reduce reliance on non-renewable resources and decrease the carbon footprint of energy storage systems, paving the way for a sustainable, low-carbon future.

Methods

Materials

Poly(vinylidene fluoride-co-hexafluoropropylene) (PVDF-HFP, 99%, average Mw: ~40,000, pellets) were purchased from Shanghai Macklin Biochemical Co., Ltd. Polyethylene glycol diacrylate (PEGDA, 99%, average Mw: ~400), 2,2’-Azbis(2-methylpropionitrile) (AIBN), anhydrous tetrahydrofuran (THF, ≥ 99.9%) and anhydrous N, N-Dimethylformamide (DMF, 99.8%, AR) were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. Pentaerythritol tetra(3-mercaptopropionate) (4-SH, 95%) was purchased from Shanghai Myriad Chemical Technology Co., Ltd. Lithium bis(trifluoromethane sulfonyl)amine (LiTFSI, 99.9%) was provided from the Suzhou Dodo Chem Co., Ltd. The electrolyte components of lithium tetrafluoroborate (LiBF₄, 99.9%), lithium difluoro(oxalato)borate (LiDOFB, 99.9%), lithium bis(fluorosulfonyl)imide (LiFSI, 99.9%), fluoroethylene carbonate (FEC, ≥ 99.9%), diethyl carbonate (DEC, ≥ 99.9%), 1,2-dimethoxyethane (DME, ≥ 99.9%) were purchased from Suzhou Dodo Chem Co., Ltd. All purchased salts were dried under vacuum at 80 °C for 12 h before use, and all solvents were dried with 4 Å molecular sieves (Aladdin) before use.

Synthesis of IHCE

PVDF-HFP (0.74 g), LiTFSI (0.46 g), and DMF (8.34 g) were added to a round-bottom flask and stirred at 50 °C until a homogeneous solution was obtained. After cooling to 25 ± 1 ^oC, PEGDA (0.15 g), 4-SH crosslinker (0.035 g), and AIBN (0.0018 g) were introduced into the mixture and dispersed thoroughly. The resulting solution was deoxygenated by purging with nitrogen for 1 h. Polymerization was then carried out by heating the mixture in an oil bath at 65 °C for 1 h, yielding a crosslinked CPEGDA network embedded with PVDF-HFP and LiTFSI. The weight ratios of PVDF-HFP and CPEGDA were controlled at 6: 4, 4: 6, 2: 8, and 8: 2 to optimize the mechanical properties and ionic conductivity of the PVDF-HFP/CPEGDA polymers.

Preparation of IHCE@Cu and IHCE@Li electrodes

The IHCE solution (0.2 wt%) was prepared by dissolving IHCE in THF with magnetic stirring overnight. The polymer matrix of IHCE has a weight-average molecular weight (M_w) of 297,281 g mol^-1 (Supplementary Fig. 4). The resulting homogeneous IHCE-THF solution was then coated on Cu foil using the doctor-blading method. Finally, the Cu foil coated with IHCE was dried under vacuum at 60 °C for 2 h to get IHCE@Cu electrodes. The IHCE@Li electrodes were prepared using the same method. 0.2 wt% IHCE solution was applied at a controlled volume of ~2 μl cm^-2 relative to the surface area of the electrode for coating. IHCE@Cu and IHCE@Li foils were prepared in a dry room (dew point temperature: -40) using an automatic coater (BEVS1811, Guangzhou, Shenghua). The copper current collector had a thickness of ~8 μm and a purity of ≥ 99.8%. IHCE@Cu foils were punched into 19 mm disks using a punching machine (T07, Kejing, Hefei). For coin-cell assembly, IHCE@Li foils were manually punched into 16 mm disks. For pouch-cell assembly, IHCE@Li foils were cut into rectangular pieces (8.4 cm × 6.2 cm) using a steel ruler and surgical blade.

Characterizations

The scanning electron microscopy (SEM, 2.0 kV, 25 pA) images and energy dispersive spectrometer (EDS, 10.0 kV, 0.2 nA) elemental mapping were acquired using a field-emission scanning electron microscope (FEI Verios 460). Electrodes for SEM characterization were harvested after cycling, rinsed with anhydrous DME to remove residual salts, dried under vacuum in an argon-filled glovebox (O₂ < 0.1 ppm, H₂O < 0.1 ppm) at 25 ± 1 °C, and subsequently transferred to the SEM chamber using an airtight transfer vessel to minimize air exposure. Fourier-transform infrared spectroscopy (FTIR) was performed on a Bruker Alpha П spectrometer with a diamond ATR crystal for each spectrum within 500–4000 cm^-1. Raman spectroscopy was performed using a Thermo Scientific DXR3xi Raman microscope (laser wavelength: 532 nm). A polarized optical microscope (POM, ECLIPSE LV100N POL) and a confocal laser scanning microscope (LSCM, Leica TCS SP8 STED) were used to characterize the phase structure of the dry-IHCE membrane. The POM was operated without polarizers, functioning as a high-performance optical microscope to provide macroscopic structural insights that complement the high-resolution SEM observations. Time-of-flight secondary ion mass spectrometry (TOF-SIMS, M6) was utilized to measure the chemical composition and distribution of the SEI. Measurements were carried out in the negative mode using a 2 keV Cs⁺ ion beam (~70 nA) for sputtering and a 30 keV Bi⁺ ion beam (~4 pA) for secondary ion analysis. Depth profiling was conducted by raster scanning over a 100 × 100 μm² area inside the regressed Cs-sputtered area (250 × 250 μm²). Samples for TOF-SIMS were harvested in an argon-filled glovebox, gently rinsed with anhydrous DME, dried under vacuum, and transferred using a sealed vessel to avoid ambient contamination. An ultraviolet spectrophotometer (UV-3600) was used to track the dissolution of LiTFSI from IHCE films in DME over 50 days. The dissolved salt concentration was determined by calibration with standard UV absorption spectra. IHCE films for this characterization were prepared by casting on the polytetrafluoroethylene (PTFE) molds and drying at 70 °C under high vacuum (-0.1 MPa) for 12 h.

Cryo-TEM characterization was conducted using 400-mesh Cu TEM grids coated with ultra-thin carbon film (3–5 nm). IHCE solution (0.2 wt%) was dissolved in THF and stirred overnight. The IHCE solution was drop-coated onto the Cu grid and dried in a vacuum oven at 60 ^oC for 2 h to obtain the IHCE@Cu TEM grid. For cryo-TEM experiments, a coin-type cell was assembled with the IHCE@Cu TEM grid as the working electrode and Li foil (diameter: 16 mm) as the counter electrode. Li was first plated onto the IHCE@Cu TEM grid at a current density of 0.5 mA cm^-2 with an area capacity of 0.25 mAh cm^-2. After plating, the cell was deconstructed, and the TEM grid was washed with anhydrous DME to remove the residual salts and dried under vacuum in an argon-filled glove box. The sample was loaded onto a cryo-transfer holder and inserted into the TEM column. Liquid nitrogen was used to maintain a cryogenic environment around -170 °C to minimize beam damage to the sample. Cryo-TEM characterization was performed on a field-emission transmission electron microscope (Talos-F), operated at an accelerating voltage of 200 kV.

Electrochemical measurements

Electrochemical measurements were performed in 2032-type coin cells. All Li | |Cu, Li | |Li, and Li | |NCM811 coin cells were tested on the Neware battery system (Shenzhen, China) at 25 °C. For the Li | |Cu half cells, the first 10 cycles were operated between 0.01-2 V at 0.08 mA cm^-2 to form a stable solid electrolyte interphase (SEI). Subsequently, 1 mAh cm^-2 of Li was plated on Cu and then stripped to 1 V at 4 mA cm^-2. The rate performance of the Li | |Li symmetric cells was tested with a fixed areal capacity of 1 mAh cm^-2 at current densities of 0.5, 1, 2, 4 and 6 mA cm^-2 for every 10 cycles. The long-term cycling performance of Li | |Li half cells was cycled at 4 mA cm^-2. Li | |NCM811 full cells were cycled between 3.0 V and 4.3 V at 0.1 C charge/0.5 C discharge, with three activation cycles at 0.1 C before long-term cycling. Rate performance of Li | |NCM811 full cells cycled at rates ranging from 0.1 C to 2 C (1 C = 180 mA g^-1). Electrochemical impedance spectroscopy (EIS) measurements were performed using an Autolab electrochemical workstation on Li | |Li asymmetric cells by applying a potentiostatic signal with an amplitude of 10 mV over a frequency range of 1 MHz to 100 mHz. 20 data points per decade of frequency were collected, and the measurements were conducted around the cell’s open-circuit voltage without additional operating conditions. The exchange current density was measured using a CHI 600 C electrochemical workstation, with a voltage ranging from -0.15 to 0.15 V. Both EIS and DC polarization techniques were employed on Li | |Li cells to determine the Li⁺ transference number. The initial and steady-state polarization currents were recorded via DC polarization, applying a polarization potential of 10 mV. Correspondingly, the initial and steady-state bulk resistances were evaluated using EIS before and after the DC polarization. Ionic conductivities were measured by EIS with a frequency range of 1 MHz to 1 Hz, using an IHCE film between two stainless steel (SS) blocking electrodes. The activation energy (E_a) was determined via EIS measurements under different temperatures, ranging from 15 to 35 °C, with 5 °C intervals. The values of activation energies (E_a) were calculated using T/R_SEI=Aexp(-E_a/RT), where T is the absolute temperature, R is the gas constant, R_SEI is the interfacial Li⁺ transfer resistance, and A is the pre-exponential factor.

Fabrication of Li-metal pouch cell

The pouch cells were fabricated in a dry room (dew point temperature: -40). 8.2 cm × 6.0 cm double-sided NCM811was used as the positive electrode. The active mass loading of NCM811 was controlled at around 30 mg cm^-2 per side, yielding a high capacity (~6.6 mAh cm^-2 on each side), ideal for building Li-metal pouch cells with a specific energy beyond 500 Wh kg^-1. The as-prepared IHCE@Li (dimensions: 8.4 cm × 6.2 cm) served as the negative electrode with Li foil thickness of 100 μm (~10 mAh cm^-2 on each side). A wet single-sided ceramic polyethylene (16 μm, Shenzhen Numi Technology Co., Ltd) was used as the separator. The positive electrode was prepared by casting the slurry with a weight ratio of 95.5 wt% NCM811, 2.2 wt% polyvinylidene fluoride (PVDF, Mw ≈ ~1,2000,000, ≥ 99.9%) binder, 1.8 wt% Super-P, and 0.5 wt% carbon nanotube (CNT) electrically conductive additives, using N, N-dimethylformamide (DMF, ≥99.8%) as the solvent. The solvent-to-solid ratio in the electrode slurry was ~60%. The slurry was coated on both sides of an aluminum foil current collector (thickness ≈ 4.6 μm, purity ≥99.9%) using an automatic coating machine (Model ZY-TSF435080, Jiangsu Jinfan Zhanyu New Energy Technology Co., Ltd.). The positive electrodes were then cut into 8.2 cm × 6.0 cm sheets using a semi-automatic electrode cutting machine (Model 450, Shenzhen MiChe Precision Machinery Automation Co., Ltd.) for further use. The electrolyte was prepared by dissolving 0.6 M Lithium difluoro(oxalato)borate (LiDFOB) and 0.4 M lithium tetrafluoroborate (LiBF₄) in fluoroethylene carbonate (FEC) and diethyl carbonate (DEC) (1: 2, v/v). For IHCE@Li | |NCM811 pouch cell, the negative/positive capacity ratio (N/P ratio) was 1.51. The electrolyte to capacity (E/C) ratio was controlled at approximately 1.29 g Ah^-1. Galvanostatic cycling of the pouch cells was conducted within a voltage window of 3-4.3 V at 25 °C. The cells were first charged/discharged (0.1 C, 1 C = 6600 mA g^-1) for two cycles as activation processes, then performed 0.1 C charge/0.5 C discharge in the subsequent cycles. The details and supplier information are shown in Supplementary Table 17.

Coin cell assembly

All 2032-type coin cells were assembled with a wet single-sided ceramic polyethylene (PE) separator and 70 μl of electrolytes in an argon-filled glovebox (H₂O < 0.1 ppm, O₂ < 0.1 ppm). The electrolytes, consisting of 1 M (denoted as SCE) or 4 M (denoted as HCE) lithium bis(fluorosulfonyl)imide (LiFSI) in DME, were used for asymmetric Li | |Cu and symmetric Li | |Li half cells. For asymmetric Li | |Cu half cells, IHCE@Cu (or bare Cu) foil was used as the working electrode, and IHCE@Li (or bare Li) foil was employed as the counter electrode. For symmetric Li | |Li half cells, IHCE@Li (or bare Li) foils were used for both working and counter electrodes. The above ether-based electrolytes, including SCE and HCE, were used in bare Li | |Cu (or Li | |Li) half cells as control samples, and SCE was utilized in IHCE@Li | |Cu (or IHCE@Li | |Li) cells. For Li | |NCM811 full cells, high mass loading NCM811 (22 mg cm^-2, corresponding to 4 mAh cm^-2) was used as the positive electrode, while bare Li foil (~50 μm thick) or IHCE@Li foil was used as the negative electrode, respectively. Two types of carbonate-based electrolytes, denoted as SCE and HCE, were used to accommodate the high voltage of NCM811. SCE and HCE were used in bare Li | |NCM811 full cells as control samples. While SCE was used for IHCE@Li | |NCM811 full cells. Specifically, SCE was prepared by dissolving 0.6 M lithium difluoro(oxalato)borate (LiDFOB) and 0.4 M lithium tetrafluoroborate (LiBF₄) in fluoroethylene carbonate (FEC) and diethyl carbonate (DEC) (1: 2, v/v). HCE was prepared using 3 M LiDFOB and 2 M LiBF₄ in the same solvent mixture.

Finite element methods

Finite element method (FEM) simulations were conducted to investigate the Li concentration distribution and electric field strength. In the finite elements model, the IHCE layer, with a thickness of 1 μm, was placed on the top of Li electrode with a thickness of 2 μm. The initial Li⁺ concentration in the IHCE layer was set at 5 M, with a diffusion coefficient of 2.933 × 10^-13 m² s^-1. Additionally, the Li⁺ transference number was set at 0.57.

The Nernst-Planck Eq. (1) was employed to describe Li⁺ transportation in the electrolyte and IHCE layer:

$$J=-D\nabla c-z\frac{D}{{RT}}{Fc}\nabla \varphi$$

(1)

Where J represents Li⁺ flux, D is the diffusion coefficient, c is the concentration, z is the charge number, F is Faraday’s constant, and φ is the potential. The electric field strength is then calculated as ∇φ.

The local current density of Li⁺ plating on IHCE@Li interface is described by the Butler-Volmer Eq. (2):

$${i}_{{\mbox{local}}}={i}_{0}\left({e}^{\frac{{\alpha }_{a}F\eta }{{RT}}}-{e}^{-\frac{{\alpha }_{c}F\eta }{{RT}}}\right)$$

(2)

Where i₀ represents the exchange current density, F is the Faraday’s constant, η is the overpotential, and T is the temperature.

The Li⁺ concentration near the IHCE@Li interface was determined under the assumption of electrical neutrality. Periodic boundary conditions were applied to both sides of the electrolyte and SEI. A current density of 10 A m^-2 was applied to the positive electrode while the negative electrode was grounded.

Molecular dynamics simulations

Molecular dynamics (MD) simulations were employed to model IHCE membrane structures, followed by pore landscape distribution analysis using the Zeo + + code. The initial system consisted of 20 PVDF-HFP chains, each containing two repeat units made up of nine vinylidene fluoride (VDF) monomers and one hexafluoropropylene (HFP) monomer. In addition, the system included 5 CPEGDA chains and 123 DME molecules. These components were randomly packed into a cubic simulation box with PVDF-HFP and CPEGDA physically mixed. The molecular structures of PVDF-HFP, CPEGDA, LiTFSI, and DME used in the simulation are shown in Supplementary Fig. 13. Simulations were performed using the COMPASS Ⅲ force field in Materials Studio 2020. Van der Waals and Coulomb interactions were calculated using atom-based and Ewald methods, respectively, with a cut-off value of 12.5 Å. The model was initially subjected to structural optimization. To achieve a thermodynamically stable configuration, we performed an annealing process of 20 cycles, ramping the system temperature between 298 K and 500 K. Simulations were carried out under the NPT ensemble, using the Nose thermostat for temperature control and the Berendsen barostat for pressure regulation, respectively. A time step of 1 fs was used, and the total annealing time was 100 ps. Subsequently, the lowest-energy configuration obtained after annealing was fully relaxed under periodic boundary conditions for 800 ps in the NPT ensemble (P = 1 atmosphere, T = 298.0 K). A 5 ns production run was finally conducted in the canonical (constant NVT) ensemble under the Nose-Hoover thermostat. This simulation time was sufficient for the system to reach equilibrium of temperature, potential and total energy. Finally, the equilibrium configuration was analyzed by Zeo + + (version 0.3) and atom volumes & surfaces tool in MS. The initial and final configurations of MD simulations are provided in Supplementary Data 1.

DFT calculations

All DFT calculations were performed using Gaussian 16. Geometry optimizations and vibrational frequency analyses were conducted at the M062X hybrid functional⁵¹ with the 6-311 + G(d,p) basis set. The solvent effects were incorporated via the SMD solvation model⁵² to simulate an implicit solvent environment. Frequency calculations confirmed the absence of imaginary frequencies, ensuring all optimized structures corresponded to local minima on the potential energy surface.

Subsequent single-point energy calculations were carried out at the M06-2X functional with the Def2-TZVP basis set⁵³. For M06-2X, the default DFT-D3(0) dispersion correction with zero-damping was applied, as the functional inherently includes medium-range correlation effects⁵⁴. To account for Basis Set Superposition Error (BSSE) in interaction energy calculations, the counterpoise correction method⁵⁵ was employed. The binding energy was calculated (E_BE) as:

$${E}_{{BE}}={E}_{{AB}}+{E}_{{BSSE}}-{E}_{A}-{E}_{B}$$

(3)

where “E_AB” is the energy of the dimer system, “E_A” and “E_B” are the energies of the isolated monomers, and “E_BSSE” denotes the BSSE correction term derived from the counterpoise correction method.

The optimized molecular structures are provided in Supplementary Data 2.