Boosting ultra-fast charging in lithium metal batteries through enhanced solvent–anion interaction via conjugation effect

Abstract

In pursuit of higher energy density, adopting a lithium metal anode holds promise for the evolving battery technology. Nevertheless, practical obstacles persist, including explosion hazards, restricted fast charging (>5C), and lithium metal compatibility issues. To tackle these challenges, we devised a non-flammable deep eutectic electrolyte (DEE) comprising lithium bisfluorosulfonimide (LiFSI) and prop-1-ene-1,3-sultone (PES). This DEE demonstrates exceptional attributes, including near 98.76% Coulombic efficiency and a high ionic conductivity of 1.96 mS cm⁻¹. We found that the electron-absorbing effect of the double-bonded structure enhances the interactions between PES and FSI⁻, releasing more free lithium ions and boosting the Li⁺ transference number of 0.78. Our engineered DEE fosters the formation of a robust organic–inorganic gradient solid electrolyte interphase (SEI) with a surface layer rich in organic species and an inner layer rich in inorganic species. The high Li⁺ transfer number and stable SEI together enable ultra-fast charging and sustained cycling, with 81.32% capacity retention after 1000 cycles at 10C in the LiFePO₄‖DEE‖Li battery. Meanwhile, the mechanistic reasons behind fast charging performance are elaborated by theoretical calculations, and its practical applicability is underscored through successful implementation in pouch cells with high loading and 30 μm of Li metal.

---

Broader context

The increasing demand for high density, improved safety, and ultra-fast charging capabilities is driving advancements in the new energy industry. Lithium metal is regarded as the “holy grail” of anodes for high energy density batteries because of its high specific capacity and low electrochemical potential. However, carbonate electrolytes fall short of enabling ultra-fast charging for lithium–metal batteries, posing flammable safety hazards. Here, we pioneer LPES2.5, a novel DEE composed of LiFSI and PES in a molar ratio of 2 : 5, possessing high Li⁺ mobility, non-flammability, and a wide electrochemical window. The double-bonded structure with the conjugate effect of PES boosts the solvent–anion interaction and effectively promotes the dissociation of Li⁺ from FSI⁻, with a Li⁺ transfer number of up to 0.78. Benefiting from the high Li⁺ transfer number, it exhibits exceptional ultra-fast charging performance, as demonstrated by an LFP‖Li battery achieving an impressive capacity retention of 81.32% over 1000 cycles at 10C. Its successful use in practical lithium metal pouch cells highlights immediate applicability and scalability.

Introduction

In light of global carbon-neutral initiatives, the burgeoning new energy industry is experiencing significant growth driven by the increasing demand for batteries with high energy density, rapid charging capabilities, and enhanced safety.^1–3 Lithium-ion batteries, the current workhorse in the industry, face limitations in meeting the heightened performance requirements of this evolving market.⁴ Consequently, a compelling candidate for the next generation of ideal electrode materials has emerged: the lithium–metal anode.⁴ It is distinguished by its ultra-high theoretical specific capacity (3860 mA h g⁻¹), low redox potential (−3.040 V vs. standard hydrogen electrodes), and low weight density (0.534 g cm⁻³).⁵ However, the widespread use of carbonate electrolytes in lithium metal batteries (LMBs) presents formidable challenges, including poor compatibility with lithium metal,⁶ low Coulombic efficiency,⁷ flammability,⁸ and the potential for short circuits stemming from dendrite formation, especially at high charge rates.^9,10 Maintaining a favorable cycling performance at high current densities is particularly challenging for carbonate electrolytes because of their inherently low Li⁺ transference number and solid electrolyte interphase (SEI), characterized by low ionic conductivity and high interface impedance.¹¹ In addition, Li precipitation and structural degradation of the cathode material may occur under rapid charging and discharging conditions, and the use of flammable commercial carbonate electrolytes increases the safety risks of batteries.¹² Consequently, there is an urgent need for alternative ultra-fast charging and non-flammable electrolytes for LMBs.¹³

Deep eutectic electrolytes (DEEs), a recent focus of increasing interest and attention, draw inspiration from the concept of deep eutectic solvents.¹⁴ Composed of intermolecular forces, DEEs are synthesized without generating by-products, rendering them environmentally friendly and cost-effective.¹⁵ Notably, DEEs share many favorable characteristics with ionic liquids, including low vapor pressure, excellent thermal stability, high chemical stability, non-flammability, and compositional versatility.¹⁶ Geiculescu et al. prepared DEEs based on lithium salts and an alkyl sulfonamide solvent that showed promise for lithium-ion battery applications with low residual current densities and reasonable capacities for various electrode materials.¹⁷ Moreover, DEEs with succinonitrile (SN) and lithium bis(fluorosulfonyl)imide (LiFSI) as the main components were synthesized by Hou et al.¹⁸ The DEE exhibited high ionic conductivity, good thermal stability, and better cycling stability at high temperatures. However, DEEs still face challenges stemming from their high viscosities, low conductivities, and substantial polarizations, making them unsuitable for ultra-fast charging in LMBs.

Prop-1-ene-1,3-sultone (PES) has been used as an SEI-forming additive in lithium-ion batteries.¹⁹ It serves to stabilize both the cathode and anode electrodes at high voltages and effectively improves the long-term cycling performance of batteries.^20,21 However, no previous studies have reported the use of PES as a single solvent. In this study, we introduce a novel DEE composed of LiFSI and PES in varying molar ratios, possessing high Coulombic efficiency, non-flammability, and a wide electrochemical window. It is considered that PES interacts with FSI⁻ by the conjugation effect and contributes to the dissociation of lithium ions. The strong solvent–anion interactions were confirmed by spectral characterization, molecular dynamics, and density functional theory. As a result, it exhibits exceptional ultra-fast charging performance, as demonstrated by an LFP‖Li battery achieving capacity retention of 81.32% over 1000 cycles at 10C. This remarkable performance can be attributed to the high Li⁺ transference number and the formation of a robust organic–inorganic gradient SEI. Compared to carbonate electrolytes or other DEEs, this DEE may offer distinct advantages in cycling performance under high-current-density conditions.

Results and discussion

In this study, the DEE was prepared by combining LiFSI and PES at molar ratios of 1 : 1 to 1 : 2.5, which resulted in the formation of a homogeneous and clear liquid upon heating (Fig. 1a). As shown in Fig. S1 (ESI†), the molar ratio of 1 : 2.5 DEE with 5 wt% FEC as a film-forming additive was selected for its favorable characteristics, including low viscosity (191.33 mPa s) and high ionic conductivity (1.96 mS cm⁻¹). This ratio is referred to as LPES2.5, and the molar ratio of 1 : 2 is denoted as LPES2. Importantly, a molar ratio of 1 : 2.5, demonstrates stability without precipitation, whereas DEE could not be prepared at a 1 : 3 molar ratio (Fig. S3, ESI†). Therefore, a 1 : 2.5 molar ratio stands out as the most suitable choice for the DEE as an electrolyte in LMBs.

Fig. 1 (a) Synthesis pathway illustrating the preparation of LPES2.5 electrolytes. (b) Thermogravimetric analysis of different electrolytes. (c) Images of LPES2.5 and the carbonate electrolyte 15 seconds after ignition. (d) FTIR spectra of LiFSI, PES, LPES2 and LPES2.5. (e) and (f) Raman spectra of LiFSI, PES, LPES2 and LPES2.5. (g) Raman spectra of LPES2 and LPES2.5 in the range of 1200 cm⁻¹ and 1250 cm⁻¹ for FSI⁻. (h) MD simulation snapshots of LPES2.5 at 298 K, using ball and stick models for LiFSI and wireframe representation for PES (green). (i) RDF of different ion pairs within LPES2.5, calculated from MD simulation. (j) DFT geometry optimization of LPES2.5 (Li in purple, N in blue, O in red, C in cyan, H in white, S in yellow, and F in pink).

Thermal stability, as evidenced by thermogravimetric analysis (Fig. 1b), demonstrates minimal weight loss up to 150 °C in LPES2.5. Importantly, LPES2.5 surpasses the thermal stability of PES due to the strong interactions between LiFSI and PES. In addition, differential scanning calorimetry results show the observation of the glass transition temperature (T_g) after the formation of DEE (Fig. S4, ESI†). As the content of PES increased, the T_g decreases to −74.3 °C.

Given the inherent risks associated with lithium dendrite growth, ensuring the non-flammability of the electrolyte is crucial for the safety of LMBs.²²Fig. 1c illustrates that LPES2.5 did not ignite even after 15 seconds, whereas the carbonate electrolyte ignited instantly and continued to burn for the same duration. Therefore, the non-flammable nature of LPES2.5 contributes significantly to the development of high-safety LMBs. Additionally, LPES2.5 has an expansive electrochemical window of nearly 5.2 V, as demonstrated in Fig. S5 (ESI†). A wide electrochemical window can contribute to improved safety and stability, as it provides a buffer against overcharging or discharging, thereby reducing the risk of thermal runaway and cell failure.²³

To elucidate the mechanism of DEE formation, we characterized LPES2 and LPES2.5 using Fourier-transform infrared spectroscopy (FTIR), Raman spectroscopy, and ¹H nuclear magnetic resonance spectroscopy (¹H NMR). Fig. 1d illustrates the shifts in the C–C–H stretching vibration peak of PES from 1612.2 cm⁻¹ to 1609 cm⁻¹ upon DEE formation. A similar shift is observed for the C=C–H stretching vibration peak of PES. Additionally, the –SO₂ stretching vibration peak of LiFSI at 1171 cm⁻¹ shifts to 1169 cm⁻¹, indicating an interaction between the oxygen atoms in S=O of LiFSI and the hydrogen atoms of PES.²⁴ These findings highlight the presence of hydrogen bonding interactions between PES and LiFSI. The Raman spectra (Fig. 1e and f) confirm these results, showing shifts in the peaks associated with the S=O stretching of PES and LiFSI. Furthermore, peak splitting analysis (Fig. 1g) in the wavenumber range of 1200 cm⁻¹ to 1250 cm⁻¹ reveals that the proportion of solvent-separated ion pair (SSIP) and contact ion pairs (CIP) in DEE increases with higher PES content, signifying the promotion of Li⁺ and FSI⁻ dissociation with the introduction of PES. Besides, ¹H NMR test was employed to strengthen the key role of hydrogen bonding in DEE formation and support the conclusions drawn from FTIR and Raman. These results demonstrate that the chemical shifts of all H atoms shift to higher fields after the formation of the DEE (Fig. S8, ESI†). Specifically, the chemical shift of H2 and H3 near the double bond experiences shifts from 7.38 to 6.58 ppm and 7.42 to 6.95 ppm, respectively. These chemical shifts indicate the existence of hydrogen bonding between LiFSI and PES, corroborating the formation of the DEE.^25,26

To further validate the mechanism of DEE formation between LiFSI and PES, we conducted a molecular dynamics (MD) simulation, shown as a snapshot (Fig. 1h), radial distribution function (RDF) of the simulation (Fig. 1i), and Li⁺ coordination number (Fig. 1j). Fig. 1h shows two strong peaks at 1.98 Å, corresponding to Li–O_FSI⁻ and Li–O_PES interactions, indicating that PES and FSI⁻ dominated the initial solvated structure of Li⁺. The coordination number of Li⁺ with O_FSI⁻ and O_PES are calculated as 2.98 and 2.37, respectively, implying the formation of an anion-rich solvated structure with three FSI⁻ ions. Moreover, the lower involvement of PES in the coordination with lithium ions indicates further promotion of Li⁺ dissociation. Subsequently, the prevalence of this anion-rich solvation structure leads to preemption of the electrode surface reduction by an increased abundance of FSI⁻, thus facilitating the generation of an inorganic-rich SEI layer. This phenomenon also plays a pivotal role in forming a compositional gradient within the SEI, as observed by the X-ray photoelectron spectroscopy (XPS) analysis. From the RDF of the simulation in Fig. S7 (ESI†), the first Li⁺ coordination shell (within 2.80 Å) of LPES2.5 is mainly composed of O_FSI⁻–H_PES at 2.80 Å with a coordination number of 2.14, revealing a preference to form hydrogen bonds between O_FSI− and H_PES. This is consistent with the FTIR spectroscopy results. Hence, the enhanced interaction between PES and FSI⁻ can furtherweaken the coordination of FSI⁻ with Li⁺, contributing to the increase in the number of free Li⁺ and high Li⁺ transference number.

The geometry optimization structures of LiFSI–xPES (x = 0, 1, 2, 2.5, and 3) based on the DFT calculations are shown in Fig. S10 (ESI†). As the molar ratio of PES increases from 1 to 3, the coordination distance between Li⁺ and O_FSI⁻ increases from 1.87 Å to 2.08 Å (the coordination distance of the other oxygen atom changes from 1.88 Å to 2.09 Å), indicating a weakening of the Li⁺–FSI⁻ interaction. Interestingly, when the molar ratio of LiFSI to PES is 2 : 5, one oxygen atom on the FSI⁻ still coordinates with the Li⁺ within a range of coordination distance from 1.93 Å to 2.64 Å. However, the additional oxygen atom escapes the range of coordination forces, with a coordination distance to Li⁺ exceeding 3.00 Å, indicating that Li⁺ exhibits strong coordination with oxygen atoms in PES, with coordination distances ranging from 1.93 Å to 2.09 Å. This supports the role of PES in the formation of DEEs. Fig. S11 (ESI†) shows the electron density of LiFSI–xPES determined by electrostatic potential (ESP) analysis. The negatively charged regions around the oxygen atoms on the PES molecule exhibit strong mutual attraction with the positively charged regions around Li⁺, inducing the formation of a stable structure in LPES2.5. The inter- and intramolecular interactions were determined using the interaction region indicator (IRI) of LiFSI–xPES (Fig. S12, ESI†). The presence of a large green zone between the FSI⁻ and the PES molecule suggests a stronger hydrogen bonding interaction between FSI⁻ and PES. When the molar ratio reaches 2 : 5 of LiFSI:PES, a blue zone surrounds the Li⁺, indicating a stronger coordination interaction between the ion and the oxygen atom of PES.

To assess the migration ability of specific ions, considering the Li⁺ transference number is essential, particularly in fast-charging scenarios.²⁷ Electrolytes with high Li⁺ transference number are known to reduce polarization during charge and discharge processes. Remarkably, compared with ionic liquids, highly concentrated electrolytes, and other DEEs (Table S3, ESI†), LPES2.5 achieves a Li⁺ transference number of 0.78 (Fig. S13, ESI†), greatly benefiting fast-charging performance. It is worth noting that the lithium salt concentration of LPES2.5 is only 2.1 mol kg⁻¹, effectively reducing costs compared to high-concentration electrolytes, while providing safety levels similar to ionic liquid electrolytes. Besides, the heightened Li⁺ transference number observed in LPES2.5 (0.78) can also lead to the suppression of lithium dendrite, because ion concentration gradients within the electrolyte, the drivers of dendritic growth in lithium metal batteries, are effectively mitigated.^28–31

To investigate lithium-ion migration mechanism in LPES2.5, we employed pulsed-field gradient (PFG) NMR to study the self-diffusion coefficients of each component in LPES2.5, including PES, FSI⁻, and Li⁺. The results reveal that the diffusion coefficient of Li⁺ (7.467 × 10⁻⁸ cm² S⁻¹) is larger than those of PES (5.753 × 10⁻⁸ cm² S⁻¹) and FSI⁻ (4.918 × 10⁻⁸ cm² S⁻¹), indicating faster diffusion of Li⁺ compared to the solvent and anions. This supports the presence of a Li⁺ hopping conduction mechanism in our DEE, which contributes to its high Li⁺ transference number.^32–35

Moreover, in further exploration of the reasons behind LPES2.5's high Li⁺ transference number, we substituted PES with PS while maintaining LiFSI to create a DEE for comparison (referred to as LPS2.5). The structure of PS closely resembles that of PES, differing solely in the double-bond structure. Initially, our findings reveal that LPS2.5 exhibits less favorable lithium-ion mobility, featuring a Li⁺ transference number of only 0.58 and an ionic conductivity of 1.658 mS cm⁻¹ (Fig. S13, ESI†).

In electrolyte systems, Li⁺–anion interaction and Li⁺–solvent interaction are two primary intermolecular forces that drive the dissolution of lithium salts and the formation of solvation structures. Strong solvent–anion interaction could exert a diminishing effect on Li⁺–anion interaction, contributing more free Li⁺. A range of computational simulation tests, such as MD, DFT, and ESP, can assist in analyzing and determining intermolecular forces. The interaction between Li⁺ and FSI⁻ in LPES2.5 appears weaker compared to that in LPS2.5, as evident in MD and DFT simulation results (Fig. 2a and b). The RDF peak of Li⁺–O_FSI⁻ at 1.98 Å in LPS2.5 surpasses that of Li⁺–O_FSI⁻ in LPES2.5, as shown in Fig. 2a, suggesting a weakened interaction between Li⁺ and FSI⁻ in LPES2.5. This weakening is beneficial as it allows lithium ions to detach from the solvated structure and engage in electrochemical processes at the electrode surface. The bond critical point (BCP) in atoms in molecules (AIM) Theory pinpoints key locations along atom interaction paths, revealing bond characteristics and strengths.^36,37 Higher electron density (ρ(r)) and more negative potential energy density (V(r)) at BCPs indicate stronger bonds, with ρ(r) reflecting electron concentration and V(r) representing bond stabilization energy. Plotting V(r) against the horizontal and vertical coordinate (i.e., ρ(r)) (Fig. 2b) reveal that the point representing the Li⁺–FSI⁻ chemical bond in LPS1 (LiFSI : PS = 1 : 1) exhibits a more negative V(r) with a larger (r), suggesting a stronger Li⁺–FSI⁻ interaction in LPS1 than in LPES1 (LiFSI : PES = 1 : 1).³⁸ Furthermore, we calculated the binding energies of PES–Li⁺, PS–Li⁺, PES–FSI⁻ and PS–FSI⁻ to further characterize the Li⁺–anion interaction and Li⁺–solvent interaction (Fig. 2c and Fig. S15, ESI†). These results show that the binding energy of PES–FSI⁻ (−18.47 kcal mol⁻¹) surpasses that of PS–FSI⁻ (–17.71 kcal mol⁻¹), signifying a stronger interaction force between PES and FSI⁻ and weakened Li⁺–anion interaction. However, the binding energy of PES–Li⁺ (–57.26 kcal mol⁻¹) is lower than that of PS–Li⁺ (−58.04 kcal mol⁻¹). The lower binding energy of Li ions facilitates their detachment from the solvent before insertion into the electrode material, a crucial aspect for enabling rapid charging and discharging capabilities.

Fig. 2 (a) RDF of Li–O_FSI⁻ in LPES2.5 and LPS2.5 calculated from MD simulation. (b) Diagram illustrating the potential energy density, V(r), as a function of electron density, ρ(r), at the bond critical points (BCPs) of Li–O_FSI− in LPES2.5 and LPS2.5. (c) Calculated binding energy of PES–FSI⁻ and PS–FSI⁻ obtained from DFT simulation. (d) ¹⁹F NMR spectra of LiFSI, LPES2.5 and LPS2.5. (e) and (f) Schematic diagrams illustrating the competitive nature of solvent–anion interaction in LPES2.5 and LPS2.5.

To highlight the differences in anion–solvent interactions between LPES2.5 and LPS2.5, the ¹⁹F NMR spectra of LiFSI, LPES2.5, and LPS2.5 were conducted. The F atom is only present in the FSI⁻, and the peak shift illustrates the anion–solvent interaction.³⁹ As shown in Fig. 2d, the F-peak signals of both LPES2.5 and LPS2.5 exhibited high-field shifts. Notably, the F-peak signals of LPES2.5 showed higher shifts compared to those of LPS2.5, indicating a stronger interaction force between PES and FSI⁻. These findings provide additional experimental confirmation of the anion–solvent interactions, complementing our theoretical and MD simulation results.

From the aforementioned analysis, it becomes evident that there exist more robust hydrogen bonding interactions between PES and FSI⁻ in comparison to PS. Moreover, we delved deeper into investigating the influence of PES's double-bond structure on the strength of these hydrogen bonds by electrostatic potential analysis. The electrostatic potential graph illustrates the distribution of electrostatic potential values on a surface equivalent to electron density, with positive values indicating the energy needed to move a positive charge and negative values representing the opposite. Higher electrostatic potential near hydrogen atoms suggests greater electron deficiency, potentially increasing the likelihood of dissociation due to their electron-deficient nature. According to the results of the electrostatic potential analysis of PES and PS (Fig. S16, ESI†), the electrostatic potential energies near the H atoms connected to C=C in PES are higher than that of the corresponding H atoms in PS, indicating higher electron deficiency in the H atoms of PES, making them more prone to dissociation.

In general, the presence of a double bond structure enhances the electron deficiency of hydrogen atoms nearby, favoring their engagement in hydrogen bonding with highly electronegative atoms. Additionally, the stronger binding energy between PES and FSI⁻ in LPES2.5 corroborates intensified solvent–anion interactions. However, solvated FSI⁻ encounters greater steric hindrance during electrolyte diffusion, leading to a weakened interaction between lithium ions and solvents, as demonstrated by MD simulations and BCP analyses. Consequently, LPES2.5 exhibits a higher Li⁺ transference number and faster Li-ion migration rate, as confirmed by MSD analysis (Fig. S17, ESI†). The diffusion coefficients, derived from the Einstein equation, for LPES2.5 and LPS2.5 are 2.95 × 10⁻⁷ cm² S⁻¹ and 1.52 × 10⁻⁷ cm² S⁻¹, respectively. This notable disparity underscores a substantial improvement in the diffusion kinetics of Li⁺ within LPES2.5. Hence, the competitive nature of solvent–anion interaction is illustrated in Fig. 2e and f, where H atoms in PES exhibit stronger hydrogen bonding with FSI⁻, signifying strong solvent–anion interactions, consequently releasing more free lithium ions. In contrast, LPS2.5 displays diminished Li⁺ migration ability due to the weaker solvent–anion interactions.

Comprehensive tests were conducted to assess the electrochemical performance of LPES2.5. As shown in Fig. 3a, LPES2.5 achieves a high Coulombic efficiency of 98.76% after 20 cycles at 0.5 mA cm⁻², surpassing the Coulombic efficiency of carbonate electrolytes (76.08%). High Coulombic efficiency typically indicates that lithium deposition occurs at the previous lithium active sites, with minimal dead lithium generated during cycling.⁴⁰ Such efficiencies are associated with a favorable high-rate performance, potentially owing to the increased number of active Li sites on the anode surface.⁴¹ Furthermore, cyclic voltammetry (CV) was used to evaluate the lithium plating/stripping processes in a Li‖Cu cell using LPES2.5, and carbonate electrolytes (Fig. 3b). The higher current response of LPES2.5 during the lithium plating and stripping process signified improves charge transfer and more rapid reaction kinetics.⁴²

Fig. 3 (a) Coulombic efficiency of Li‖Cu cells with LPES2.5 and carbonate electrolyte at 0.5 mA cm⁻². (b) First-turn CV curves of LPES2.5 and carbonate electrolytes at 1 mV s⁻¹ in Li‖Cu cells. Galvanostatic cycling performance of Li‖Li symmetric cells with LPES2.5 and carbonate electrolyte (c) at gradient current densities of 0.2 mA cm⁻², 0.5 mA cm⁻², 1 mA cm⁻², 2 mA cm⁻², and 3 mA cm⁻² and (d) 1 mA cm⁻². (e and f) Visual documentation of lithium metal flakes immersed in different electrolytes for different days.

To assess compatibility with lithium metal, we used symmetric Li‖Li cell tests under various current densities, ranging from 0.2 mA cm⁻² to 3 mA cm⁻². As shown in Fig. 3c, while the carbonate electrolyte experiences cell failure at 2 mA cm⁻² during the 17th cycle, the cell using LPES2.5 maintains stable cycling without short-circuiting at 3 mA cm⁻². This outcome suggests enhanced compatibility between LPES2.5, and Li metal, likely due to the formation of a robust SEI layer during cycling. The long-term cycling performance of the Li‖Li cells is shown in Fig. 3d. Employing LPES2.5 results in stable cycling for 700 h at 1 mA cm⁻² with a polarization potential comparable to that of the carbonate electrolyte. In contrast, the carbonate electrolyte exhibits an abrupt increase in the polarization potential after 300 h and short circuits after 570 h. Further validation of compatibility with Li metal was conducted by immersing lithium metal in the electrolyte and monitoring the changes in the lithium surface and electrolyte (Fig. 3e and f). After 63 days, the lithium metal remains unchanged in LPES2.5, showing no signs of precipitation or surface alteration. Conversely, in the carbonate electrolyte, the lithium metal displays discoloration and oxidation after only 14 days, leading to the formation of a noticeable black oxide layer after 63 days. These results confirm the superior compatibility of LPES2.5 with lithium metal compared with the carbonate electrolyte.

To elucidate the compatibility and stability of the previously analyzed LPES2.5 with lithium metal, we conducted XPS depth profiling on the lithium metal surface after cycling to analyze the components of SEI. In the C 1s spectrum (Fig. 4a), three primary peaks are identified on the SEI surface: C–C (284.8 eV), O–C=O (286.6 eV), and C–F (290.1 eV), with an additional CO₃²⁻ peak emerging within the inner layer of the SEI. In the F 1s spectrum (Fig. 4b), the peak at 686.9 eV is attributed to C–F, whose intensity notably diminishes with increasing etching depth, while the LiF peak at 684.8 eV exhibits an opposite trend and primarily resides at the bottom of the SEI.⁴³ This trend is further supported by the data in Fig. 4f, which illustrates the C–F-rich surface (atomic ratio: 70%) and LiF-rich bottom of the SEI (atomic ratio: 83%). In the N 1s spectrum (Fig. 4c), Li₃N (398.8 eV) is prominently observed on the inner side of the SEI. This compound primarily derives from the reductive decomposition of FSI⁻ and facilitates rapid Li⁺ conduction between SEI layers due to its high ionic conductivity.⁴⁴ Additionally, the S 2p spectrum (Fig. 4d) reveals various inorganic sulfide components, including Li₂SO₃, Li₂SO₄, and LiSO₂. Notably, Li₂S is exclusively found in the inner layer of the SEI, promoting uniform lithium deposition and stabilizing the lithium metal anodes.⁴⁵ The presence of LiF and Li₃N peaks is also confirmed in the Li 1s spectra. The contents of C and O are related to the organic species, whereas N, F, and S pertain to the inorganic species. As the etching time increases, the C and O contents decrease from 64.04% to 47.69%, whereas the N, F, and S contents increase from 12.93% to 15.99% (Fig. 4e). In summary, these results indicate the formation of an organic–inorganic gradient SEI layer with different compositions and contents on the lithium metal after cycling with LPES2.5 (Fig. 4f and g).

Fig. 4 (a)–(d) XPS spectra of C 1s, F 1s, N 1s, and S 2p in LFP‖Li cells with LPES2.5 after 10 cycles. (e) Quantified atomic ratios of the SEI formed in LPES2.5 at varying etching depths. (f) Relative contents of C–F and LiF in SEI layers at different etching depths. (g) Schematic representation of the SEI concentration in LPES2.5.

The SEI surface layer primarily comprises C–F and O–C=O bonds, while inorganic species, such as LiF, Li₃N, and Li₂S, are predominantly situated in the inner layer of the SEI. The highly polar C–F bonds serve as effective Lewis base sites for Li⁺ adsorption onto the SEI surface.⁴⁶ The outer organic layer, enriched with PES decomposition products, provides a flexible matrix that accommodates volume changes during cycling.⁴⁷ This layer helps to reduce continuous solvent decomposition by passivating the anode surface, maintaining a stable SEI over extended cycles.⁴⁸ The presence of a significant amount of LiF in the inner layer enhances the mechanical strength and stability of the SEI.⁴⁹ Additionally, the inorganic-rich inner layer, containing Li₃N and Li₂S, facilitates rapid Li⁺ transport to the electrode surface.^50,51 Consequently, the SEI formed in the LPES2.5 electrolyte exhibits a unique gradient structure with a higher concentration of inorganic compounds near the Li metal and more organic compounds towards the electrolyte. This design enhances mechanical stability, reduces continuous solvent decomposition, and improves ionic conductivity, all contributing to the high Coulombic efficiency (CE) and fast-charging performance observed.

To evaluate the electrochemical performance of LPES2.5, we assembled full cells using LFP cathodes and Li–metal anodes. Remarkably, LPES2.5 demonstrates superior rate performance under high current densities of 5C and 10C compared to the carbonate electrolyte and LPS2.5, yielding discharge capacities of 119.50 mA h g⁻¹ and 85.31 mA h g⁻¹, respectively, while performing similarly at low current densities (Fig. 5a and Fig. S21, ESI†). Notably, LPES2.5 exhibits a lower polarization potential and higher capacity above 2C compared to carbonate electrolytes, indicating reduced internal resistance to polarization (Fig. 5b and c). After 250 cycles, the LPES2.5 cell maintains a capacity retention of 92.76% at 2C, while the carbonate electrolyte experiences a substantial capacity decrease after 200 cycles (Fig. S22, ESI†). At 5C, the carbonate electrolyte shows a drastic performance decline after 280 cycles, exhibiting unstable Coulombic efficiency (Fig. 5d). Conversely, LPES2.5 maintains a capacity retention of 89.90% after 600 cycles, with an exceptionally stable average Coulombic efficiency close to 100.00%. The minor capacity drops at the 450th cycle are influenced by a temperature decrease. Surprisingly, during ultra-fast 10C charging with a mere 3-min duration, the LFP‖Li cell with LPES2.5 delivers an initial capacity of 85.34 mA h g⁻¹ and retains 81.32% capacity after 1000 cycles, achieving a high Coulombic efficiency of 99.98% (Fig. 5e). This demonstrates the suitability of LPES2.5 for ultra-fast charging applications. In comparison, the battery with the carbonate electrolyte is struggling to cycle at a high rate of 10C, with capacity dropping to 0 mA h g⁻¹ at 100th cycles. Additionally, cycling performance at a high temperature of 50 °C. in LFP‖Li cell (Fig. S24, ESI†) reveals that cells using LPES2.5 exhibit superior cycling performance with a capacity retention of 92.89% after 200 cycles at 2C. This may be attributed to the reduced cell polarization owing to the lower viscosity of LPES2.5, and its high thermal stability at elevated temperatures.

Fig. 5 Cycling performance comparison of LFP‖Li cells utilizing LPES2.5 and carbonate electrolytes at 25 °C. (a) Rate performance of LFP‖Li cells with LPES2.5 and carbonate electrolytes. Voltage profiles with (b) LPES2.5 and (c) carbonate electrolytes. Cycle performance of LFP‖Li cells (d) at 5C and (e) 10C. (f) Cycle performance of LFP‖Li cells with a highly loaded LFP electrode (15 mg cm⁻²) and thin Li cathode (30 μm) at 0.465mA cm⁻² (0.2C). (f) Cycle performance of LFP‖Li cells with a highly loaded LFP electrode (15 mg cm⁻²) and thin Li cathode (30 μm) at 8.79 mA cm⁻² (5C) at 60 °C. (i) Overview of reported DEEs,^17,52–58 ionic liquids,^59–64 and highly concentrated electrolytes,^65–69 concerning cycle rate on the horizontal coordinate and the number of cycles on the vertical axis. (h) Cycle performance of a pouch cell with 11.5 mg cm⁻² LFP electrode and 30 μm Li cathode. (j) Comparative performance of carbonate electrolytes with LPES2.5 in this study.

To simulate practical applications, thick LFP cathodes with a high loading of approximately 15 mg cm⁻² and a 30 μm lithium metal anode was applied in a full cell, whose N/P ratio was 2.5 (Fig. 5f). When LPES2.5 is used as the electrolyte, the cell demonstrates satisfactory performance, with a capacity retention of 97.00% after 75 cycles at 0.465 mA cm⁻² (0.2C). Furthermore, at an elevated temperature of 60 °C, the viscosity of LPES2.5 decreases. The cell with an 11.5 mg cm⁻² LFP cathode and a 30 μm lithium metal anode using LPES2.5 exhibited an initial capacity of 152.9 mA h g⁻¹ with an average Coulombic efficiency of 99.31% over 20 cycles at current densities up to 8.79 mA cm⁻² (5C), as shown in Fig. 5g.

Pouch cells differ from coin cells in several aspects, including their high cathode loading, low Li content, and battery impedance. Emphasizing the pouch cell performance can broaden electrolyte practicality in diverse application scenarios.^70,71 The electrochemical performance of pouch cells featuring high-loading LFP cathodes and 30 μm lithium anodes is presented in Fig. 5h. Employing LPES2.5, the discharge capacities during three formation cycles at a low rate of 0.3 mA cm⁻² show slight decay over 100 cycles of charging and discharging at 0.75 mA cm⁻², resulting in capacity retention of 85.95%. Furthermore, the cycle life of an LFP‖Gr full cell using LPES2.5 can be extended (Fig. S29, ESI†). These findings demonstrate that LPES2.5 can serve as a safe electrolyte in both lithium-ion batteries and high-energy LMBs. Besides, a float charging test in LFP‖Li battery was conducted to assess stability after deep charging, LPES2.5 demonstrates greater capacity retention compared to the carbonate electrolyte (Table. S4, ESI†).

Based on the cycling performance of LFP‖Li cells using LPES2.5 as the electrolyte, it is evident that our meticulously designed DEE outperforms previously reported DEEs,^17,52–58 ionic liquids,^59–64 and highly concentrated electrolyte^65–69 systems in both high current densities and extended cycling capabilities (Fig. 5i). Overall, LPES2.5 offers significant advantages over carbonate electrolytes regarding Coulombic efficiency, oxidation voltage, Li⁺ transfer number, fast-charging capability, and prolonged cycling performance, as detailed in the radar chart shown in Fig. 5j. Importantly, LPES2.5 showcases the immense potential for fulfilling the fast-charging demands in future energy systems.

Scanning electron microscopy (SEM) images depicting the surface and cross-section of Li anode, and LFP cathode surface after varying cycles at 1C are displayed in Fig. 6a and b and Fig. S30–S32 (ESI†). Notably, after 200 cycles with carbonate electrolytes, the Li anode surface appears rough with accumulated by-products. Conversely, when utilizing LPES2.5, the lithium metal electrode maintains a flat and smooth surface. Furthermore, the LFP cathode surface displays fewer signs of by-product adhesion when using LPES2.5. LPES2.5, as an LMB electrolyte, fosters uniform lithium deposition and suppresses side reactions between the electrode and electrolyte, thereby improving Coulombic efficiency and cycling performance.

Fig. 6 Comparative SEM images depicting the (a) surface and (b) LFP cathode after 200 cycles of LFP‖Li cells using LPES2.5 and carbonate electrolytes. (c) TEM images of LFP cathodes after 60 cycles in LPES2.5 and carbonate electrolytes. The electrochemical impedance spectra for LFP‖Li cells at different cycles with (d) LPES2.5 and (e) carbonate electrolytes.

Transmission electron microscopy (TEM) images capturing the cathode–electrolyte interface (CEI) on the LFP particles after cycling (Fig. 6c) correlate with the SEM results. The CEI layer thickness often signifies electrolyte degradation, with a thicker CEI layer potentially fostering more side reactions within the battery.⁷² Interestingly, a noticeable discrepancy in CEI layer thickness on the surface of the LFP cathode is observed. LPES2.5 results in a thinner CEI layer, measuring only 13.04 nm, notably thinner than the 20.23 nm thickness observed with the carbonate electrolyte. Consequently, LPES2.5 facilitates the formation of a thinner and more uniform CEI layer during cycling, ultimately reducing charge transfer impedance. This, in turn, promotes rapid lithium-ion migration, thereby facilitating cell cycling performance at elevated current densities. The optical microscope was also employed to compare the lithium deposition morphology in situ using different electrolytes (Fig. S33, ESI†). After 30 minutes of deposition at 0.5 mA cm⁻², lithium deposition with carbonate electrolyte exhibits obvious lithium dendrites, whereas no obvious lithium dendrites are observed in the Li‖LPES2.5‖Cu cell.

Assessing the long-cycle performance of batteries, particularly at high rate, relies significantly on evaluating SEI layer resistance variations post-cycling. Cycling the battery at 1C for 50, 100, and 300 cycles reveals relatively stable interphase impedance when LPES2.5 is used (Fig. 6d). In contrast, utilizing the carbonate electrolyte leads to a significant increase in the SEI impedance (Fig. 6e).

Conclusions

In conclusion, we successfully developed a DEE comprising LiFSI and PES for application in LMBs. LPES2.5 exhibits non-flammability and excellent compatibility with lithium metal anodes, rendering it a promising option for constructing highly safe LMBs. The conjugate effect of PES's double-bonded structure greatly enhances solvent–anion interactions and effectively promotes the dissociation of Li⁺ from FSI⁻, with a Li⁺ transfer number of up to 0.78. Moreover, LPES2.5 forms a robust organic–inorganic gradient SEI, with an organic species-rich surface layer and an inorganic species-rich inner layer. This configuration enables ultra-fast 10C charge/discharge capabilities while maintaining smooth lithium metal surfaces and achieving an impressive Coulombic efficiency of 98.76%. Notably, it achieves a remarkable capacity retention of 81.32% and a Coulombic efficiency of 99.98% after 1000 cycles at 10C. More importantly, our successful application of LPES2.5 in lithium–metal pouch cells underscores its potential for future practical use and further expanded its promising applications. Our work not only introduces a novel DEE design concept ensuring high safety and lithium compatibility but also highlights the crucial role of a high Li⁺ transference number, a conductive and stable SEI layer, and superior Coulombic efficiency in enabling super-fast charging performance.

Data availability

All data generated or analysed during this study are included in this article and its ESI.† The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the Shenzhen Science and Technology Program (JCYJ20180507183907224, JCYJ20210324121411031, JSGG202108021253804014, RCBS20210706092218040, GXWD20221030205923001, and KQTD20170809110344233).

Notes and references

1. Z. P. Cano, D. Banham, S. Ye, A. Hintennach, J. Lu, M. Fowler and Z. Chen, Nat. Energy, 2018, 3, 279–289.
2. K. An, M. J. Joo, Y. H. T. Tran, S. Kwak, H. G. Kim, C. S. Jin, J. Suk, Y. Kang, Y. J. Park and S. W. Song, Adv. Funct. Mater., 2023, 33, 2301755.
3. X.-B. Cheng, R. Zhang, C.-Z. Zhao and Q. Zhang, Chem. Rev., 2017, 117, 10403–10473.
4. J. Liu, Z. Bao, Y. Cui, E. J. Dufek, J. B. Goodenough, P. Khalifah, Q. Li, B. Y. Liaw, P. Liu, A. Manthiram, Y. S. Meng, V. R. Subramanian, M. F. Toney, V. V. Viswanathan, M. S. Whittingham, J. Xiao, W. Xu, J. Yang, X.-Q. Yang and J.-G. Zhang, Nat. Energy, 2019, 4, 180–186.
5. M. Winter, B. Barnett and K. Xu, Chem. Rev., 2018, 118, 11433–11456.
6. K. Xu, Chem. Rev., 2014, 114, 11503–11618.
7. F. Ding, W. Xu, G. L. Graff, J. Zhang, M. L. Sushko, X. Chen, Y. Shao, M. H. Engelhard, Z. Nie, J. Xiao, X. Liu, P. V. Sushko, J. Liu and J.-G. Zhang, J. Am. Chem. Soc., 2013, 135, 4450–4456.
8. Y. Liang, W. Wu, D. Li, H. Wu, C. Gao, Z. Chen, L. Ci and J. Zhang, Adv. Energy Mater., 2022, 12, 2202493.
9. K. Xu, Chem. Rev., 2004, 104, 4303–4418.
10. G. Zheng, S. W. Lee, Z. Liang, H.-W. Lee, K. Yan, H. Yao, H. Wang, W. Li, S. Chu and Y. Cui, Nat. Nanotechnol., 2014, 9, 618–623.
11. N. Piao, X. Ji, H. Xu, X. Fan, L. Chen, S. Liu, M. N. Garaga, S. G. Greenbaum, L. Wang, C. Wang and X. He, Adv. Energy Mater., 2020, 10, 1903568.
12. P. Zhang, T. Yuan, Y. Pang, C. Peng, J. Yang, Z.-F. Ma and S. Zheng, J. Electrochem. Soc., 2019, 166, A5489.
13. F. Li and G. Li, Energy Technol., 2023, 11, 2201397.
14. H. Vanda, Y. Dai, E. G. Wilson, R. Verpoorte and Y. H. Choi, C. R. Chim, 2018, 21, 628–638.
15. F. M. Perna, P. Vitale and V. Capriati, Curr. Opin. Green Sustainable Chem., 2020, 21, 27–33.
16. J. Wu, Q. Liang, X. Yu, Q. F. Lü, L. Ma, X. Qin, G. Chen and B. Li, Adv. Funct. Mater., 2021, 31, 2011102.
17. O. E. Geiculescu, D. D. DesMarteau, S. E. Creager, O. Haik, D. Hirshberg, Y. Shilina, E. Zinigrad, M. D. Levi, D. Aurbach and I. C. Halalay, J. Power Sources, 2016, 307, 519–525.
18. Q. Hou, P. Li, Y. Qi, Y. Wang, M. Huang, C. Shen, H. Xiang, N. Li and K. Xie, ACS Energy Lett., 2023, 8, 3649–3657.
19. J. Self, D. S. Hall, L. Madec and J. R. Dahn, J. Power Sources, 2015, 298, 369–378.
20. R. Petibon, L. Madec, L. M. Rotermund and J. R. Dahn, J. Power Sources, 2016, 313, 152–163.
21. J. Xia, L. Ma and J. R. Dahn, J. Power Sources, 2015, 287, 377–385.
22. W. Wu, Y. Bo, D. Li, Y. Liang, J. Zhang, M. Cao, R. Guo, Z. Zhu, L. Ci, M. Li and J. Zhang, Nano-Micro Lett., 2022, 14, 44.
23. S. Murugesan, O. A. Quintero, B. P. Chou, P. Xiao, K. Park, J. W. Hall, R. A. Jones, G. Henkelman, J. B. Goodenough and K. J. Stevenson, J. Mater. Chem. A, 2014, 2, 2194–2201.
24. L. Li, S. Zhou, H. Han, H. Li, J. Nie, M. Armand, Z. Zhou and X. Huang, J. Electrochem. Soc., 2011, 158, A74.
25. X. Zhang, H. Dai, H. Yan, W. Zou and D. Cremer, J. Am. Chem. Soc., 2016, 138, 4334–4337.
26. H. Hong, J. Zhu, Y. Wang, Z. Wei, X. Guo, S. Yang, R. Zhang, H. Cui, Q. Li, D. Zhang and C. Zhi, Adv. Mater., 2024, 36, 2308210.
27. K. M. Diederichsen, E. J. McShane and B. D. McCloskey, ACS Energy Lett., 2017, 2, 2563–2575.
28. Y. Chen, X. Dou, K. Wang and Y. Han, Adv. Energy Mater., 2019, 9, 1900019.
29. D. H. Wong, J. L. Thelen, Y. Fu, D. Devaux, A. A. Pandya, V. S. Battaglia, N. P. Balsara and J. M. DeSimone, Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 3327–3331.
30. Y. Lu, M. Tikekar, R. Mohanty, K. Hendrickson, L. Ma and L. A. Archer, Adv. Energy Mater., 2015, 5, 1–7.
31. T. Zeng, Y. Yan, M. He, R. Zheng, D. Du, L. Ren, B. Zhou and C. Shu, Chem. Commun., 2021, 57, 12687–12690.
32. K. Dokko, D. Watanabe, Y. Ugata, M. L. Thomas, S. Tsuzuki, W. Shinoda, K. Hashimoto, K. Ueno, Y. Umebayashi and M. Watanabe, J. Phys. Chem. B, 2018, 122, 10736–10745.
33. K. Shigenobu, K. Dokko, M. Watanabe and K. Ueno, Phys. Chem. Chem. Phys., 2020, 22, 15214–15221.
34. A. Nakanishi, K. Ueno, D. Watanabe, Y. Ugata, Y. Matsumae, J. Liu, M. L. Thomas, K. Dokko and M. Watanabe, J. Phys. Chem. C, 2019, 123, 14229–14238.
35. C. Zhang, K. Ueno, A. Yamazaki, K. Yoshida, H. Moon, T. Mandai, Y. Umebayashi, K. Dokko and M. Watanabe, J. Phys. Chem. B, 2014, 118, 5144–5153.
36. M. Shakourian-Fard, G. Kamath and S. K. R. S. Sankaranarayanan, ChemPhysChem, 2016, 17, 2916–2930.
37. T. Lu and Q. Chen, J. Comput. Chem., 2022, 43, 539–555.
38. K. Chen, X. Shen, L. Luo, H. Chen, R. Cao, X. Feng, W. Chen, Y. Fang and Y. Cao, Angew. Chem., Int. Ed., 2023, 62, e202312373.
39. J. Xu, V. Koverga, A. Phan, A. m Li, N. Zhang, M. Baek, C. Jayawardana, B. L. Lucht, A. T. Ngo and C. Wang, Adv. Mater., 2024, 36, 2306462.
40. J. Xiao, Q. Li, Y. Bi, M. Cai, B. Dunn, T. Glossmann, J. Liu, T. Osaka, R. Sugiura, B. Wu, J. Yang, J.-G. Zhang and M. S. Whittingham, Nat. Energy, 2020, 5, 561–568.
41. G. M. Hobold, K.-H. Kim and B. M. Gallant, Energy Environ. Sci., 2023, 16, 2247–2261.
42. X. Ma, J. Yu, X. Zou, Y. Hu, M. Yang, F. Zhang and F. Yan, Cell Rep. Phys. Sci., 2023, 4, 101379.
43. Y. Yu, G. Huang, J.-Z. Wang, K. Li, J.-L. Ma and X.-B. Zhang, Adv. Mater., 2020, 32, 2004157.
44. T. Zhang, H. Lu, J. Yang, Z. Xu, J. Wang, S.-I. Hirano, Y. Guo and C. Liang, ACS Nano, 2020, 14, 5618–5627.
45. T. Li, X. Q. Zhang, N. Yao, Y. X. Yao, L. P. Hou, X. Chen, M. Y. Zhou, J. Q. Huang and Q. Zhang, Angew. Chem., Int. Ed., 2021, 60, 22683–22687.
46. F. Li, J. He, J. Liu, M. Wu, Y. Hou, H. Wang, S. Qi, Q. Liu, J. Hu and J. Ma, Angew. Chem., Int. Ed., 2021, 60, 6600–6608.
47. D. W. Kang, J. Moon, H.-Y. Choi, H.-C. Shin and B. G. Kim, J. Power Sources, 2021, 490, 229504.
48. E. Peled and S. Menkin, J. Electrochem. Soc., 2017, 164, A1703.
49. T. D. Pham, A. Bin Faheem and K.-K. Lee, Small, 2021, 17, 2103375.
50. S. Liu, X. Ji, J. Yue, S. Hou, P. Wang, C. Cui, J. Chen, B. Shao, J. Li, F. Han, J. Tu and C. Wang, J. Am. Chem. Soc., 2020, 142, 2438–2447.
51. W. Wu, D. Li, C. Gao, H. Wu, Y. Bo, J. Zhang, L. Ci and J. Zhang, Adv. Sci., 2024, 2310136.
52. H. Wu, B. Tang, X. Du, J. Zhang, X. Yu, Y. Wang, J. Ma, Q. Zhou, J. Zhao, S. Dong, G. Xu, J. Zhang, H. Xu, G. Cui and L. Chen, Adv. Sci., 2020, 7, 2003370.
53. J.-H. Baik, S. Kim, D. G. Hong and J.-C. Lee, ACS Appl. Mater. Interfaces, 2019, 11, 29718–29724.
54. B. Joos, T. Vranken, W. Marchal, M. Safari, M. K. Van Bael and A. T. Hardy, Chem. Mater., 2018, 30, 655–662.
55. B. Joos, J. Volders, R. R. da Cruz, E. Baeten, M. Safari, M. K. Van Bael and A. T. Hardy, Chem. Mater., 2020, 32, 3783–3793.
56. H. Wang, J. Song, K. Zhang, Q. Fang, Y. Zuo, T. Yang, Y. Yang, C. Gao, X. Wang, Q. Pang and D. Xia, Energy Environ. Sci., 2022, 15, 5149–5158.
57. G. Chen, Y. Zhang, C. Zhang, W. Ye, J. Wang and Z. Xue, ChemSusChem, 2022, 15, e202201361.
58. X. Pei, Y. Li, T. Ou, X. Liang, Y. Yang, E. Jia, Y. Tan and S. Guo, Angew. Chem., Int. Ed., 2022, 61, e202205075.
59. H. Zhang, W. Qu, N. Chen, Y. Huang, L. Li, F. Wu and R. Chen, Electrochim. Acta, 2018, 285, 78–85.
60. B. Tong, X. Chen, L. Chen, Z. Zhou and Z. Peng, ACS Appl. Energy Mater., 2018, 1, 4426–4431.
61. N. Chen, Y. Guan, J. Shen, C. Guo, W. Qu, Y. Li, F. Wu and R. Chen, ACS Appl. Mater. Interfaces, 2019, 11, 12154–12160.
62. L. Liu and C. Sun, ChemElectroChem, 2020, 7, 707–715.
63. Z. Hu, J. Chen, Y. Guo, J. Zhu, X. Qu, W. Niu and X. Liu, J. Membr. Sci., 2020, 599, 117827.
64. M. D. Widstrom, K. B. Ludwig, J. E. Matthews, A. Jarry, M. Erdi, A. V. Cresce, G. Rubloff and P. Kofinas, Electrochim. Acta, 2020, 345, 136156.
65. S. Chen, J. Zheng, D. Mei, K. S. Han, M. H. Engelhard, W. Zhao, W. Xu, J. Liu and J. G. Zhang, Adv. Mater., 2018, 30, 1706102.
66. P. Liu, Q. Ma, Z. Fang, J. Ma, Y.-S. Hu, Z.-B. Zhou, H. Li, X.-J. Huang and L.-Q. Chen, Chin. Phys. B, 2016, 25, 078203.
67. X.-Q. Zhang, X. Chen, L.-P. Hou, B.-Q. Li, X.-B. Cheng, J.-Q. Huang and Q. Zhang, ACS Energy Lett., 2019, 4, 411–416.
68. F. Qiu, X. Li, H. Deng, D. Wang, X. Mu, P. He and H. Zhou, Adv. Energy Mater., 2018, 9, 1803372.
69. Q. Ma, Z. Fang, P. Liu, J. Ma, X. Qi, W. Feng, J. Nie, Y. S. Hu, H. Li, X. Huang, L. Chen and Z. Zhou, ChemElectroChem, 2016, 3, 531–536.
70. H. Liu, X. Sun, X.-B. Cheng, C. Guo, F. Yu, W. Bao, T. Wang, J. Li and Q. Zhang, Adv. Energy Mater., 2022, 12, 2202518.
71. Z.-X. Chen, M. Zhao, L.-P. Hou, X.-Q. Zhang, B.-Q. Li and J.-Q. Huang, Adv. Mater., 2022, 34, 2201555.
72. J. Zhao, Y. Liang, X. Zhang, Z. Zhang, E. Wang, S. He, B. Wang, Z. Han, J. Lu, K. Amine and H. Yu, Adv. Funct. Mater., 2021, 31, 2009192.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ee00391h

‡ J. W. and L. X. contributed equally to this work.

---