Unlocking high-performance lithium metal batteries through a unique solvation structure engineered using an ether solvent

Abstract

Lithium metal batteries (LMBs) offer exceptional theoretical energy density and an ultra-low reduction potential, making them a leading candidate for next-generation energy storage. However, challenges such as dendritic lithium growth and electrolyte instability hinder their commercial viability by causing capacity decline and safety risks. This study presents an electrolyte formulation based on a single-salt, single-solvent system of lithium bis(fluorosulfonyl)imide (LiFSI) in diethylene glycol diethyl ether (DEGDEE). The key advantage of this system stems from a unique, anion-participating solvation structure, engineered through the molecular design of the DEGDEE solvent. This structure, particularly at an optimized concentration of 1.75 M LiFSI in DEGDEE, facilitates the formation of protective layers on both the anode and cathode that effectively stabilize interfacial side-reactions, leading to a significant enhancement in cycle life. The resulting Li‖Cu cells exhibit an average Coulombic efficiency of ∼98% at both 25 °C and 60 °C, and Li‖Li symmetric cells demonstrate ultra-stable cycling for over 1500 h with a minimal polarization of ∼0.02 V. When paired with practical LiFePO₄ cathodes, the full cell achieves a specific capacity of 147 mAh g⁻¹ attaining 85.4% capacity retention over 1000 cycles at 25 °C and 163 mAh g⁻¹ with 95.7% over 200 cycles at 60 °C, all while maintaining a high efficiency (99.8%) at 1.0C. This work demonstrates that engineering the Li⁺ solvation structure through rational solvent design provides a powerful strategy for creating highly stable interfaces, advancing LMBs toward practical, high-performance energy storage.

---

New concepts

The commercial adoption of high-energy-density lithium metal batteries (LMBs) is obstructed by the instability of the solid electrolyte interphase (SEI) layer and the persistent growth of dendrites, arising from the high reactivity of lithium metal. Herein, we introduce a distinctive electrolyte design based on diethylene glycol diethyl ether (DEGDEE) that enables a unique solvation structure to stabilize the Li metal interface. With its engineered structure, the DEGDEE molecule combines a glyme-derived backbone with two ethylene oxide units, providing sufficient chelating sites for Li⁺ coordination for the long cycling stability while maintaining low viscosity for efficient ion transport. Importantly, the terminal ethyl groups introduce steric hindrance, weakening excessively strong solvent–Li⁺ interactions and promoting anion participation in the solvation sheath. This anion-enriched coordination environment facilitates the formation of a robust, uniform SEI, effectively suppressing dendrite growth and unwanted side reactions. Compared with conventional glymes, this molecular design offers a new pathway to control solvation structure via simultaneous modulation of backbone coordination sites and terminal group sterics, and delivers long lifespan batteries. This work provides fresh insight into electrolyte molecular engineering for high-energy-density LMBs, with potential applicability across a range of advanced battery chemistries.

1. Introduction

With its widespread adoption and significant impact on modern life, lithium-ion battery (LIB) technology has become a cornerstone of rechargeable energy storage worldwide.^1–3 The rapid rise of electric vehicles (EVs) and the global push toward carbon neutrality have driven technological advancements, emphasizing the urgent need for batteries with greater energy storage capacity, faster charging capabilities, and enhanced durability. However, LIBs are approaching their theoretical energy capacity limit of 372 mAh g⁻¹, which poses challenges in meeting the demands of next-generation EVs and other high-energy applications.³ This has spurred extensive research into alternative storage systems, with metal batteries emerging as a promising solution due to their significantly higher theoretical energy densities compared to conventional technologies.^3–5 Among these, lithium metal batteries (LMBs) stand out for their exceptionally high theoretical energy density (∼3860 mAh g⁻¹) and lowest reduction potential (−3.04 V vs. standard hydrogen electrode), making them a strong candidate for the next generation of high-energy-density battery systems designed to power the transition toward sustainable transportation and beyond.^6,7

Despite over a century of research on metallic lithium (Li) in batteries, major challenges continue to impede its commercial viability, with electrolyte–electrode incompatibility being one of the most critical issues.² Due to its status as the most electropositive alkali metal, Li exhibits high reactivity with organic electrolyte components, leading to pronounced chemical and electrochemical instability.⁸ A solid electrolyte interphase (SEI) forms at the interface, serving as a passivation layer that prevents direct contact between lithium and the electrolyte. This nanometer-scale layer inhibits further electrolyte decomposition while remaining permeable to lithium ions (Li⁺), enabling battery operation.⁹ However, the SEI generated by the uncontrolled reaction typically exhibits structural nonuniformity and Li-dendritic formation.¹⁰ When the SEI layer is primarily composed of organic components, it lacks mechanical stability, leading to frequent breakdown and increased electrolyte exposure to the Li. This promotes the uncontrolled growth of lithium dendrites, which can cause internal short circuits, limiting cycle life and raising safety concerns.^11,12 Enhancing the homogeneity and stability of SEI is essential for preserving Li metal anodes and minimizing side reactions that degrade battery performance.¹⁰ To extend battery lifespan and improve performance, most strategies have focused on increasing SEI stability by modifying its chemical composition. Various approaches have been proposed to achieve a stable SEI, including (1) applying artificial protective layers on the electrode and separator surfaces,¹³ (2) incorporating functional additives,¹⁴ and (3) adjusting the electrolyte solvation structure using (localized) high concentration electrolytes (LHCEs and HCEs), binary salt systems, or new solvents.¹⁵ Although LHCEs and HCEs (>3.0 M) can significantly enhance battery performance, their high cost has limited practical implementation, underscoring the need for cost-effective organic electrolytes with strong stability toward Li.

Electrolyte solvents underpin the chemistry of both LIBs and LMBs. While carbonate solvents drive the success of LIBs, their high reactivity with lithium makes them incompatible for LMBs.^12,16 Instead of forming a protective SEI, they decompose into fragile lithium alkyl carbonates, leaving the metal surface unstable and prone to dendrites, low efficiency, and short circuits.⁸ These limitations have shifted attention toward ethers, whose greater reductive stability offers a more compatible pathway for pairing with lithium metal.^8,17 Among them, linear ethers often outperform their cyclic counterparts, delivering enhanced interfacial stability and electrochemical performance.¹⁷ Among linear ethers, glymes, which are oligoethers with the formula R–O–(CH₂CH₂O)_n–R′, have emerged as molecularly versatile and tunable in electrolyte design.^18–21 Variation in the number of ethylene oxide units (n) enables precise control over the physical and electrochemical properties of glymes. Yet this molecular flexibility faces a critical trade-off: increasing the chain length also increases viscosity and reduces ionic conductivity.²² In this study, we focused on considering a backbone glyme with n = 2, as this chain length represents an optimal balance; it is sufficiently long to provide multiple chelating sites for Li⁺ ions while remaining short enough to maintain a relatively low viscosity and favorable ionic conductivity. However, the tridentate chelating nature of a glyme with n = 2, which allows it to strongly bind to Li⁺, can also be a disadvantage. This strong solvating power can lead to a high population of solvent-separated ion pairs (SSIPs), where “free” anions are abundant.^23,24 While this is beneficial for ionic conductivity, it can inhibit the formation of a robust, anion-derived SEI, which is critical for stabilizing the lithium metal anode. This highlights the need for additional molecular design considerations to achieve a solvent with truly balanced performance. A promising strategy, therefore, involves fixing the oligoether chain length while modifying the terminal alkyl groups (R and R′). Specifically, extending the terminal methyl groups to longer ethyl groups, as in DEGDEE, introduces significant steric hindrance around the terminal oxygen atoms. This steric hindrance weakens the Li⁺–solvent coordination, thereby altering the solvation structure to favor the formation of contact ion pairs (CIPs) and aggregates (AGGs), even at moderate concentrations. This shift is crucial because it promotes the participation of anions in the formation of the SEI. The resulting anion-derived SEI is typically more inorganic, mechanically robust, and stable compared to the organic-rich, unstable interphases formed from solvent decomposition. Therefore, our approach focuses on leveraging the unique solvation structure engineered through the rational design of the DEGDEE molecule to overcome the inherent limitations of conventional ether-based electrolytes.

In this study, a single-salt single-solvent electrolyte system is proposed, utilizing diethylene glycol diethyl ether (DEGDEE) as the solvent and lithium bis(fluorosulfonyl)imide (LiFSI) as the salt. This formulation leverages the unique properties of DEGDEE to create a beneficial solvation environment, enhancing the cycling performance of LMBs while maintaining a favorable balance of physicochemical properties suitable for elevated-temperature operation. Among the various concentrations of LiFSI in DEGDEE, 1.75 M LiFSI–DEGDEE exhibited the best performance with a high Li plating/stripping CE of ∼99% and formed smooth dendrite-free Li deposits. Furthermore, when a LiFePO₄ (LFP) cathode is paired with a Li metal anode, the full cells show a specific capacity of 147.0 mAh g⁻¹ and a capacity retention of 85.4% is achieved after 1000 cycles at 25 °C (163.1 mAh g⁻¹ and 95.7% after 200 cycles at 60 °C), along with a high efficiency of 99.8% at 1.0C. Raman spectroscopy was performed to understand the unique solvation structures in the DEGDEE-based electrolyte at different concentrations. The morphologies and components of the SEI layer were analyzed using modern methods such as scanning electron microscopy (SEM) and X-ray photoelectron spectroscopy (XPS).

2. Experimental section

2.1. Electrolyte preparation

Lithium bis(fluorosulfonyl)imide (LiFSI, battery grade) was obtained from Chunbo Co. Ltd, and diethylene glycol diethyl ether (DEGDEE; >98%) was purchased from TCI Chemicals. Before preparing the electrolytes, DEGDEE was dried using 4 Å molecular sieves (Sigma Aldrich). The solubility of LiFSI in DEGDEE at room temperature was determined by dissolving the salt in a precisely measured volume of solvent until saturation, from which the saturated concentration was obtained. The resulting saturated electrolyte solution was then diluted to prepare solutions with lower electrolyte concentrations. The reference electrolyte was 1.0 M LiPF₆ in EC : EMC : DMC (1 : 1 : 1, volume ratio) which was ordered from Dongwha Electrolyte Co., Ltd. All electrolytes were prepared and stored inside an Ar-filled glovebox (O₂ and H₂O < 1.0 ppm).

2.2. Electrode preparation

Li metal anodes (16 mm diameter, 250 µm thickness) were ordered from MTI Corporation and used as received. Cu electrodes (Wellcos Corporation) were punched into disks with a diameter of 19 mm. The cathodes were prepared as follows: LiFePO₄ (LFP, battery grade, MTI Corporation), super-P (conductive agent, 99+% (metals basis), Alfa Aesar), and poly(vinylidene difluoride) (PVDF, M_w = 534 000, Sigma Aldrich) were uniformly dispersed in N-methyl-2-pyrrolidone (NMP, anhydrous, 99.5%, Sigma Aldrich) at a weight ratio of 80 : 10 : 10. The mass loading of the LFP cathode electrodes was controlled at approximately 6.24 mg cm⁻². The resulting slurry was cast onto aluminum foil (MTI Corporation) by using the doctor blade technique, and then dried at 80 °C for 12 h. Next, the electrodes were punched into 15 mm disks and dried again at 80 °C under vacuum for 12 h. All electrodes were stored in an Ar-filled glove box before fabrication.

2.3. Physicochemical properties

Ionic conductivity measurements were performed with a conductivity bench meter (S30, Mettler Toledo) and a Cond Probe Inlab710 (Mettler Toledo). Prior to each measurement, the apparatus was calibrated using an aqueous KCl (0.1 M) solution. Viscosity measurements were performed using a DV3T shear viscosity rheometer (Brookfield), calibrated with deionized water before testing. All measurements were carried out at 25 °C. A contact angle tester (DSA100E, KRÜSS Scientific) was utilized to measure the angle between solvent/electrolyte and the surface of the polypropylene (PP) separator. Differential scanning calorimetry (DSC) analysis was performed using a DSC 250 (TA Instrument) from 30 °C to −60 °C at a heating rate of 5 °C min⁻¹ under an air atmosphere. The sealed pan with electrolyte was cooled to −60 °C at a rate of 5 °C min⁻¹ using an RCS90 (TA Instrument) and then scanned from −60 °C to 30 °C at 5 °C min⁻¹.

2.4. Solvation structure

Laser Raman spectroscopy was performed at room temperature using a RamanTouch system (Nanophoton) with a laser wavelength of 532 nm. The samples were sealed in the capillary tube and maintained throughout the measurements. The Raman intensity was normalized based on the CH₂ bending/scissoring and CH₃ bending modes within the 1400–1500 cm⁻¹ region. Spectral deconvolution was performed by fitting the measured spectra with a Gaussian–Lorentz mixed function.²⁵

Nuclear magnetic resonance (NMR) spectra were performed using an ECZL400S (JEOL) for ⁷Li-NMR and a Bruker Avance III 600 spectrometer (600 MHz) for ¹⁷O-NMR. The external standard of 1.0 M LiCl in D₂O (δ = 0.00) was used to quote the chemical shift signals of NMR spectra. A capillary tube with D₂O solution was sealed and afterward placed in the NMR tube alongside the solution samples. All NMR measurements were conducted at room temperature following the preparation of the samples within the glove box.

2.5. Electrochemical measurements

A BioLogic SP50e instrument was used at experimental room temperature to measure the lithium transference number (t₊), linear sweep voltammetry (LSV), and electrochemical impedance spectroscopy (EIS). Symmetrical lithium cells with alternating current (AC) impedance and direct current (DC) polarization were used to obtain t₊.^26–28 Electrode resistances (R₀, R_ss) were obtained by the AC impedance measurement, while initial currents (I₀) and stable currents (I_ss) were determined by using DC polarization. Impedances were measured in a frequency range from 10 mHz to 0.1 MHz; chronoamperometry experiments were done with an applied voltage of 10 mV for 1 h. The t₊ was calculated according to the following equation:

A three-electrode configuration was employed for LSV to evaluate the electrochemical stability window of the electrolytes. A platinum (Pt) disk was used as the working electrode, with lithium metal serving as both the reference and counter electrodes. The LSV measurements were conducted at a scan rate of 2 mV s⁻¹. Cell resistance was measured by EIS. The experiments were carried out across a wide range of frequencies, from 10 mHz to 1.0 MHz, with an amplitude of 10 mV.

Galvanostatic experiments were carried out using CR2032-type coin cells on a WBCS3000L battery cycler system (Wonatech, Korea) at constant temperatures of 25 °C and 60 °C. The compatibility of the investigated electrolytes with Li was examined using Li‖Cu cells with PP (Celgard® 2400) as a separator. During each cycle, 1.0 mAh cm⁻² of Li metal was deposited onto the Cu substrate at a current density of 1.0 mA cm⁻², and then stripped until the potential reached 1.0 V versus Li/Li⁺.

Average CE (CE_avg) was determined by a modified method of Aurbach (Method 3)²⁹ using Li‖Cu cells, Cu served as the substrate for Li metal deposition. The Cu surface was conditioned by plating 4 mAh cm⁻² of Li and stripping to 1 V at 0.4 mA cm⁻². Afterward, a Li reservoir of 5 mAh cm⁻² was plated on the Cu, followed by cycling (0.5 mAh cm⁻²) for 10 cycles at 0.4 mA cm⁻². Again, the final Li on Cu was stripped to 1 V at 0.4 mA cm⁻². The CE_avg over 10 cycles can be calculated using eqn (2):

The long-term cycling of Li‖Cu cells (method 1) was performed at 1.0 mA cm⁻²/1.0 mAh cm⁻² with a cut-off voltage of 1.0 V vs. Li/Li⁺. The CE can be calculated for each cycle according to eqn (3):

The plating/stripping of the Li‖Li symmetric cell was carried out at different current densities from 0.5 mA cm⁻² to 5.0 mA cm⁻², with a capacity of 1.0 mAh cm⁻². The Li‖LFP cells were tested to compare the role of the investigated solvents in the performance of LMBs. The Li‖LFP coin cells were tested galvanostatically between 2.5 and 4.0 V, applying a charge/discharge current density of 1.0C (1.0C = 170 mA g⁻¹). This was conducted following two formation cycles at a charge/discharge current density of 0.1C. Each coin cell was filled with 70 µL of the electrolyte. The fabrication of all coin cells was performed in an Ar-filled glovebox (O₂ < 1.0 ppm, H₂O < 1.0 ppm).

2.6. Characterization of the electrode

The morphological and surface characteristics of the electrode were comprehensively examined using scanning electron microscopy (SEM) and X-ray photoelectron spectroscopy (XPS). Prior to characterization, Li deposition on Cu substrates was soaked in pure 1,2-dimethoxyethane (DME) for 24 h to remove loosely bound surface species. To further eliminate residual salt, the substrates were subjected to multiple sequential rinses with fresh DME. The samples were then vacuum-dried in the antechamber of a glovebox for 24 h. For cross-sectional imaging, vertical cuts were made with a precision razor blade to reveal the internal structure. SEM imaging was conducted using a Hitachi S4800 operated at an accelerating voltage of 5 kV and an emission current of 10 µA, enabling detailed surface topography analysis at high resolution. XPS measurements were carried out using a Thermo Fisher K-Alpha system equipped with a monochromatic Al Kα X-ray source (photon energy = 1486.7 eV). Spectral calibration was performed using the C 1s peak at 285.0 eV, which corresponds to the C–C bonding, as an internal reference to ensure data consistency across samples. All sample handling, including preparation and transfer, was conducted entirely under an inert argon atmosphere. Samples were transported to the SEM and XPS facilities in sealed, argon-filled glass containers to prevent contamination from ambient air.

3. Results and discussion

3.1. Physical properties

The physical properties of several ether solvents widely used in LMBs are compared in Table S1 (SI), and their corresponding molecular structures are shown in Fig. S1 (SI). Among these ether solvents, 1,2-dimethoxyethane (DME) and 1,2-diethoxyethane (DEE) are widely studied for their high ionic conductivity and low viscosity; however, their low boiling points pose challenges for high-temperature operation.^8,30 In contrast, longer-chain glymes like diethylene glycol dimethyl ether (DEGDME) and DEGDEE offer improved thermal stability. Specifically, DEGDEE exhibits a compelling combination of properties: a high boiling point (188 °C), a high flash point (70 °C), and low vapor pressure (0.5 mmHg), which collectively enhance operational safety, particularly under thermal-abuse conditions. Furthermore, its favorable dipole moment (1.96 D) ensures adequate salt dissociation, allowing for enhanced ion transport.³¹ These factors collectively position it as a promising and safer solvent candidate for robust, high-temperature LMB applications.

The saturation concentration of a lithium salt in a solvent is a critical parameter in electrolyte design, as it defines the maximum salt content that can be dissolved without precipitation. A higher saturation concentration offers more flexibility in tailoring key electrolyte properties such as ionic conductivity, viscosity, and solvation structure. In this study, the saturation concentration of LiFSI in DEGDEE at room temperature was determined to be 1.75 M. Although 1.75 M is a moderate concentration, it meets the essential requirements for use as a practical electrolyte in LMBs, offering a good balance between performance and cost.

A key challenge for electrolytes at low temperatures is maintaining salt solubility, as a decrease can lead to salt crystallization and performance degradation. To evaluate this behavior, LiFSI–DEGDEE solutions with various concentrations were stored at temperatures ranging from 25 °C and −40 °C (Fig. S2, SI). While salt crystallization was observed in the higher concentration solutions (1.25–1.75 M), the 1.0 M LiFSI–DEGDEE electrolyte remained completely free of visible precipitation even at a harsh temperature of −40 °C, demonstrating its excellent thermal stability for low-temperature applications. To investigate the low-temperature phase behavior of the DEGDEE-based electrolytes, differential scanning calorimetry (DSC) was performed on pure DEGDEE and the 1.0 M LiFSI–DEGDEE electrolyte from −60 to 30 °C (Fig. S3, SI). Surprisingly, no distinct phase transition peaks were detected in either sample, which deviates from previously reported data suggesting a melting point of −44 °C for DEGDEE.³² This result confirms that the 1.0 M LiFSI–DEGDEE electrolyte remains liquid even at extremely low temperatures, making it a promising candidate for reliable ultra-low-temperature LMB applications.

An electrolyte system can enable the tuning of physical and electrochemical properties by varying salt concentrations. Accordingly, the macroscopic properties, including ionic conductivity, shear viscosity, Li-ion transference number, and contact angle, of LiFSI in DEGDEE electrolyte systems with different salt concentrations were measured. As illustrated in Fig. 1, the ionic conductivity follows a well-known trend for non-aqueous electrolytes: it initially increases with increasing LiFSI concentration, reaches a maximum, and then decreases.^33,34 This behavior typically results from the interplay between increasing charge carriers and counteracting effects such as higher viscosity and greater ion pairing.³⁵ Specifically, the ionic conductivity peaked at ∼5.54 mS cm⁻¹ at a concentration of 1.25 M and remained sufficiently high at 5.06 mS cm⁻¹ even at the saturated concentration of 1.75 M. These values indicate that the DEGDEE-based electrolyte system exhibits adequate ionic conductivity for practical applications in LMBs.³⁴ The viscosity of an electrolyte is also fundamentally governed by the intermolecular interactions within the solution. These forces, encompassing ion–solvent attractions, ion–ion electrostatic forces, and solvent–solvent interactions, collectively determine its resistance to flow.³⁵ As the salt concentration increases, the relative contents of charge carriers increase, intensifying these intermolecular forces. This leads to an increase in overall internal friction and thus results in higher macroscopic viscosity. In other words, at a constant temperature, viscosity increases with salt concentration due to intensified ion pairing and the restricted mobility of solvent molecules.^8,35 At 25 °C, viscosity increases from 1.2 to 12.1 cP as the LiFSI concentration rises from 0.25 to 1.75 M. While DEGDEE-based electrolytes exhibit higher viscosity than their shorter-chain counterparts (Table S2, SI), this value is within the acceptable range for practical LMBs.

Fig. 1 Concentration-dependent ionic conductivity and viscosity of LiFSI in DEGDEE at 25 °C.

The lithium-ion transference number (t₊) of an electrolyte is a critical property governing ion transport, representing the fraction of total current carried by lithium ions.³⁶ A higher t₊ is essential for efficient battery operation. Crucially, as the transference number is determined by the size and mobility of solvated ions, t₊ serves as a valuable indicator of the Li⁺ solvation structure within the electrolyte. In this context, we measured t₊ for LiFSI in DEGDEE electrolyte systems across a concentration range (1.0 M to 1.75 M), as detailed in Fig. S4 and Table S3 (SI). Remarkably, the t₊ values remained consistently high, between 0.43 and 0.45. This exceptional stability in t₊ strongly suggests that the Li⁺ solvation structure in the DEGDEE system is uniquely maintained, regardless of the salt concentration. This finding is particularly noteworthy because, in most conventional electrolytes, increasing the salt concentration significantly alters the solvation environment and consequently changes the t₊. For example, the LiFSI–DME electrolyte shows a significant change in its transference number, with a difference of approximately 0.26 between 0.5 M and 2.0 M concentrations.³⁷ Similarly, other linear ether solvents exhibit considerable variation (0.6–0.9) in transference numbers with changing salt concentrations.³⁸ This remarkable stability of the Li⁺ transference number in the DEGDEE system is a distinctive feature compared to other ether-based electrolytes. This behavior stems directly from the unique solvation environment engineered by the DEGDEE molecular structure, and it provides significant electrochemical advantages, the details of which will be discussed in the following sections.

To assess the wettability of the investigated solvents and electrolytes, the contact angle (θ) between the solvent/electrolyte and polypropylene (PP) separator was measured (Fig. S5, SI). Good wettability of the electrolyte on the separator promotes uniform lithium-ion flux, which is crucial for achieving even lithium deposition and suppressing dendrite formation.³⁹ The pure solvents such as ethyl methyl carbonate (EMC), dimethyl carbonate (DMC), and DEGDEE were measured to investigate the solvent effect on separator wettability toward the electrolytes (refer to Fig. S5a–c). The contact angle values of EMC, DMC, and DEGDEE were 27.4°, 33.6°, and 25.1°, respectively. The observed variation in contact angle with different solvents can be influenced by multiple factors such as solvent polarity, viscosity, and surface tension, as well as solvent/separator molecular interactions.^39,40 Table S1 (SI) shows that DEGDEE has higher polarity/viscosity (5.7/1.12 cP, respectively) compared to DMC (3.0/0.59 cP, respectively) and EMC (2.96/0.65 cP, respectively); however, DEGDEE exhibits a small contact angle, indicating enhanced wettability due to lower surface tension.⁴⁰ This is attributed to its flexible molecular geometry; as a relatively large molecule, it exhibits greater degrees of freedom, allowing it to adopt conformations that maximize interaction with the surface. This adaptability contributes to a reduced surface tension, making it easier for the liquid to spread and further enhancing its ability to wet the separator, which promotes better interfacial compatibility.⁴¹ As a salt dissolves in a solvent, it is noted that the contact angle generally increases with higher salt concentration.³⁹ This trend is largely attributed to the increase in the overall polarity of the electrolyte solution. As more salt dissolves, the concentration of ionic species rises, making the electrolyte more polar. Since the PP separator is non-polar, the growing mismatch in polarity between the electrolyte and the separator surface leads to reduced wettability and a higher contact angle. A similar trend is observed in Fig. S5d–g (SI), where increasing salt concentrations from 1.0 to 1.75 M LiFSI in DEGDEE results in contact angles ranging from 34.7° to 59.1°. Notably, compared to 1.0 M LiPF₆ in EC : EMC : DMC (1 : 1 : 1, volume ratio) (hereafter denoted as RE) displayed a larger contact angle of 75° (Fig. S5h, SI), the 1.75 M LiFSI–DEGDEE demonstrated a lower contact angle. As a result, DEGDEE-based electrolytes facilitate superior separator wetting and promote a more uniform Li-ion flux, which has a positive impact on suppressing the growth of Li dendrites and improving overall cell performance.

3.2. Solvation structure

To establish a clear link between the electrolyte's molecular-level interactions and its macroscopic electrochemical behavior, the solvation structure of the DEGDEE-based system was investigated. This local environment around the Li⁺ ion is of paramount importance, as it directly governs the composition of the SEI layer and the kinetics of ion transport, which are the primary factors controlling overall battery performance.^42,43 In liquid electrolytes, both solvents and anions interact with Li⁺ via ion–dipole or ion–ion interactions, shaping its solvation structure based on the competition between Li⁺–solvent and Li⁺–anion interactions. To investigate the effect of LiFSI concentration on ionic speciation in DEGDEE, Raman spectroscopy was performed in the 680–780 cm⁻¹ range, corresponding to the S–N–S symmetric stretching vibration of the FSI⁻ anion (refer to Fig. 2).^17,44 It is generally observed that as the salt concentration increases, the population of free anions decreases while the proportion of associated ionic complexes, such as CIPs and AGGs, increases. This shift in ionic association is typically reflected in the Raman spectrum as a blue shift of the anion's characteristic vibrational modes. Indeed, in the conventional DME-based electrolytes, a distinct blue shift is evident with increasing concentration, signifying a clear change in the Li⁺ coordination environment.⁴⁵ Strikingly, the DEGDEE-based electrolyte exhibits a much less pronounced spectral change, suggesting an unusually stable solvation structure. To gain a more quantitative understanding of the ionic speciation, the Raman spectra were deconvoluted (Fig. 2a). The peaks at ≈720 cm⁻¹, ≈730 cm⁻¹, and ≈740 cm⁻¹ are assigned to free anions, CIPs, and AGGs, respectively.⁴⁵ The analysis (Fig. 2b) reveals a remarkably stable distribution of ionic species in the DEGDEE system across all tested concentrations, uniquely dominated by a combination of free anions (∼63.3 ± 1.4%), CIPs (∼21.1 ± 0.7%), and AGGs (∼15.6 ± 0.6%). Furthermore, the nearly identical normalized spectra (Fig. 2c) confirm that the solvation structure and ionic speciation are exceptionally stable across different concentrations, supporting the deconvoluted results in Fig. 2a and b. While the relative ratios of these species remain remarkably constant, the absolute number of associated complexes (CIPs and AGGs) naturally increases with concentration. This increased population of anion-involved ionic species in electrolytes at higher concentration is particularly significant, as they serve as the direct precursors for building a stable, anion-derived SEI. This stable, anion-rich solvation sheath is a core advantage of the DEGDEE system, as it ensures that anions are readily available to participate in the formation of a stable and inorganic-rich SEI.^8,15

Fig. 2 Raman analysis. (a) Raman spectra of LiFSI–DEGDEE electrolytes in the region from 700 to 800 cm⁻¹ (S–N–S symmetric stretching mode of the FSI⁻ anion). Black solid line and gray solid lines represent the original spectra and the fitting results, respectively. (b) Distribution of the Li⁺ solvates for different concentrations of LiFSI in DEGDEE electrolytes from Raman spectra. (c) Raman spectra of LiFSI–DEGDEE with different concentrations after normalization.

To further corroborate the Raman spectra and probe the solvation environment, nuclear magnetic resonance (NMR) measurements were conducted.^46,47 Because ¹⁷O-NMR measures the average chemical shift of all oxygen atoms in both FSI⁻ ions and DEGDEE solvent molecules, the observed shift is a direct indicator of the average electron density on these oxygens. As shown in Fig. S6a (SI), the spectra of the pure solvent exhibit two ¹⁷O-NMR resonances corresponding to chain oxygen (R–O–R) at 34.0 ppm and terminal oxygen (–O–R′) at 40.8 ppm.⁴⁷ Upon the addition of salt, an upfield displacement of the ethereal oxygen resonances occurs, with slight shifts in both oxygen types, resulting in spectral broadening, which intensifies with increasing salt concentration. This broadening of the solvent signal strongly suggests a decrease in the population of free solvent molecules, confirming their integral role in the Li⁺ solvation complexes. The spectra of the sulfonyl oxygen from FSI⁻ anions appear with only negligible chemical shifts as the salt concentrations increase at around 205.3 to 205.6 ppm (Δδ = 0.3 ppm), unlike the increased chemical shifts (Δδ = 1.7 ppm) of the LiFSI–DME electrolyte at concentration ranges of 0–2.0 M.⁴⁶

To gain deeper insight into the local electronic environment of the lithium cation, the ⁷Li-NMR chemical shift was also analyzed, as it is highly sensitive to interactions with neighboring anions and solvent molecules.^46,47 As shown in Fig. S6b (SI), the total change in the ⁷Li chemical shift across the entire concentration range from 0.25 M to 1.75 M is merely ∼0.3 ppm. This variation is remarkably small, especially when compared to the large concentration-dependent shifts typically observed in conventional ether-based electrolyte systems; for instance, the chemical shift changes were approximately ∼2.0 ppm for the LiFSI–DME system.^46,47 This minimal change strongly corroborates the Raman spectroscopy results and almost constant lithium transference number, confirming that the fundamental Li⁺ solvation structure in the DEGDEE-based electrolyte remains exceptionally stable regardless of the salt concentration. This characteristic solvation behavior of DEGDEE is a key advantage. The resulting stability of the solvation structure is highly advantageous for battery performance because it establishes a favorable equilibrium between “free” anions, which ensure sufficient ionic conductivity, and associated ion complexes that contribute to forming a robust, anion-derived SEI layer.^15,43 This structural stability is key to enabling effective ion transport and homogeneous lithium deposition, as it effectively mitigates local concentration fluctuations during cycling.

3.3. Electrochemical stability of Li metal anodes

The electrochemical anodic stability of the investigated electrolytes was evaluated using linear sweep voltammetry (LSV) in a three-electrode system, where platinum served as the working electrode and lithium metal was used as both the counter and reference electrodes. The LSV measurements were conducted at room temperature (25 °C) with a scan rate of 2 mV s⁻¹ (refer to Fig. S7, SI). The measured anodic and cathodic stability limits are summarized in Table S4 (SI). Notably, the LiFSI–DEGDEE electrolytes exhibit a wide and stable electrochemical stability window (ESW). The oxidation potential is approximately 4.89 V vs. Li/Li⁺, while the reduction potential is approximately −0.08 V vs. Li/Li⁺ across the entire concentration range from 1.0 M to 1.75 M. Consequently, the overall ESW is approximately 4.97 V, with negligible variation observed with changing salt concentration. This remarkable consistency in electrochemical stability strongly corroborates the previous Raman and NMR results, which further demonstrated that the fundamental solvation structure remains structurally invariant regardless of salt concentration. Such a wide and stable ESW is highly advantageous, as it confirms the electrolyte's suitability for applications with both low-potential lithium metal anodes and high-voltage cathodes.⁴⁸

To evaluate the compatibility with Li metal, the Coulombic efficiency (CE) of Li plating and stripping was quantified in Li‖Cu cells. Following the methodology described by Adams et al.,²⁹ cells were cycled at 1.0 mA cm⁻² with a fixed capacity of 1.0 mAh cm⁻², and the results are presented in Fig. 3a and b. At 25 °C, the DEGDEE-based electrolytes demonstrated a remarkable improvement in both CE and cycling stability compared to the conventional RE electrolyte. Notably, the CE steadily improved as the LiFSI concentration increased, reaching a high and stable average of ∼98% in the 1.75 M electrolyte (Fig. 3a). This trend strongly suggests that the formation of a more uniform and robust SEI, facilitated by the anion-rich solvation structure at higher concentrations, is critical for minimizing the irreversible loss of active lithium during cycling. This enhanced stability was even more pronounced at an elevated temperature of 60 °C. While the CE in the RE system showed unstable behavior and rapidly degraded after just a few cycles, the 1.75 M LiFSI–DEGDEE electrolyte maintained an exceptionally stable performance for over 250 cycles (Fig. 3b). Taken together, these results provide compelling evidence that the engineered DEGDEE-based electrolyte fosters a highly stable Li metal interface across a broad range of operating temperatures. Effect of the chain lengths on the electrochemical performance was evaluated by cycling test asymmetric Li‖Cu cells with a LiFSI concentration of 1.75 M in various linear ether solvents (Fig. S1, SI). Under cycling with plating/stripping of 1.0 mAh cm⁻² at 25 °C, all electrolytes except 1.75 M LiFSI–DME displayed stable CEs exceeding 97% with delayed activation around 150 cycles (Fig. S8a, SI). Subsequently, the cycle number at which substantial fluctuation began varied, ranking as DEGDEE > DEGDME > DEE > DME. A similar trend was observed at 60 °C (Fig. S8b, SI), where 1.75 M LiFSI–DEGDEE again exhibited the most delayed fluctuation. These results are consistent with our hypothesis that solvent chain length is an important factor impacting performance, as the alkyl and methoxy chains in ether solvents are expected to enhance the stability of lithium metal anodes.

Fig. 3 Reversibility of the Li metal anode. CE at a current density of 1 mA cm⁻² and a capacity of 1 mAh cm⁻² for Li plating/stripping in Li‖Cu cells in the electrolytes at (a) 25 and (b) 60 °C. (c) Voltage profiles of Li‖Li symmetric cells cycled in the electrolytes at 25 °C.

The initial (ICE) and average CE (CE_avg) were investigated by Li‖Cu cells using a modified Aurbach test, excluding contributions from the formation cycle.²⁹ The ICE reflects the amount of active lithium retained after the first cycle. As shown in Fig. S9 (SI), DEGDEE-based electrolytes delivered remarkable improvements with increasing salt concentration, reaching ICE values of 93%–96% compared to 35.94% for RE. A higher ICE indicates fewer irreversible side reactions and more active lithium preserved for long-term reversibility. Following that, the reaction loss associated with the substrate surface is eliminated in the initial cycle due to its impact on the accuracy of CE measurement. The CE_avg value is then determined after subsequent cycles, providing a continuous view of the ongoing efficiency and stability of the battery, which serves as a critical predictor of its overall lifespan and performance.²⁹ The cell applied with RE achieved a CE_avg of 62.1%, and the DEGDEE-based electrolyte sustained ≈96%–98%. Importantly, the saturated 1.75 M LiFSI–DEGDEE system achieved the highest stability with a CE_avg of 98.5%. These results underscore that its low reactivity with lithium effectively suppresses side reactions, enabling highly efficient cycling. Both ICE and CE_avg indicated the potential of 1.75 M LiFSI–DEGDEE electrolyte minimizes the side reactions between the electrodes and electrolytes, highlighting its ability to suppress side reactions for highly efficient cycling, as widely recognized in the field.

To further investigate the interfacial stability and Li deposition/stripping behavior, galvanostatic cycling of Li‖Li symmetric cells was performed. Fig. 3c shows the long-term cycling stability of Li‖Li symmetric cells with RE and DEGDEE-based electrolytes at a current density of 1.0 mA cm⁻² and a capacity of 1.0 mAh cm⁻². The cycling stability of the Li‖Li cell in RE electrolyte was significantly lower than that observed in the LiFSI–DEGDEE electrolytes. The overpotential of a symmetric cell with RE electrolyte started to increase after ∼40 h of cycling. The voltage hysteresis increased to 560 mV after 97 h, and the cell eventually succumbed to a short circuit, indicated by a sudden rise in potential. In stark contrast, the DEGDEE-based electrolytes demonstrated superior performance, enabled by their unique solvation structure. For instance, the DEGDEE-based electrolytes with concentrations of 1.0 M, 1.25 M, and 1.5 M maintained stable voltage hysteresis for 210, 400, and 630 h, respectively. Notably, the 1.75 M electrolytes achieved a stable polarization for up to 1500 h, with the voltage hysteresis remaining at 350 mV. This outstanding durability directly reflects the effectiveness of the engineered interface in preventing dendrite formation and parasitic side reactions. The Li‖Li symmetric cells were further evaluated at 60 °C to validate the reversibility of the Li anode in these electrolytes, as shown in Fig. S10 (SI). Temperature significantly affects the shape of the voltage profile, with voltage polarization decreasing across all electrolytes due to enhanced Li⁺ transport (ionic conductivity) at elevated temperatures.^49–51 The cells applying RE, 1.0 M, and 1.25 M LiFSI–DEGDEE electrolytes exhibited a gradual increase in voltage polarization after approximately 100, 200, and 300 h, respectively. In sharp contrast, the cells with 1.5 M LiFSI–DEGDEE maintained a stable plating/stripping profile for about 600 h, and this stability was further extended to beyond 1000 h when the concentration was increased to 1.75 M. Typically, the voltage profiles display a gradual rise in polarization prior to the onset of a latent internal short circuit, characterized by a sudden voltage drop followed by an artificially stable potential plateau.⁵² This phenomenon can be ascribed to the depletion of electrochemically active lithium and the accumulation of interfacial resistance.⁵⁰ Fig. S10b (SI) presents a comparison of the voltage profiles of symmetric cells with those of DEGDEE-based electrolytes. By applying 1.75 M LiFSI–DEGDEE, lithium symmetric cells also demonstrate long-term cycling stability at 60 °C. The results reveal that the reversibility of Li deposition and dissolution at high temperatures is improved by increasing the salt concentration of DEGDEE-derived electrolyte, which stabilizes the interface and suppresses parasitic processes. Furthermore, the rate capability of the symmetric cells was evaluated by subjecting them to a range of current densities ranging from 0.5 to 5.0 mA cm⁻² and an area capacity of 1.0 mAh cm⁻² (Fig. S11, SI). The RE-based cell exhibited substantially higher polarization at all current densities, particularly at the demanding rate of 5.0 mA cm⁻², signifying sluggish kinetics. In contrast, the DEGDEE-based electrolytes, especially the 1.75 M formulation, maintained a low and stable overpotential even at high rates. This indicates rapid Li-ion transport across a robust interface, highlighting the system's potential for high-power applications.

To probe the interfacial resistance and its evolution over time, electrochemical impedance spectroscopy (EIS) was conducted on Li‖Li symmetric cells during standing time and on Li‖Cu cells after cycling (Fig. 4). The resulting Nyquist plots, composed of the ohmic resistance (R_e), SEI resistance (R_SEI), and charge-transfer resistance (R_ct), provide critical insights into the interfacial stability.⁵³ First, the chemical stability was assessed by monitoring Li‖Li cells during a rest period (0, 12, 24, 48, and 72 h), as shown in Fig. 4a–c. The R_SEI of the Li‖Li symmetric cell in RE electrolyte grew continuously, reaching 910 Ω after 72 h, which signifies ongoing parasitic reactions and the formation of a defective interface. In sharp contrast, the R_SEI of cells in DEGDEE-based electrolytes remained stable and low around ∼180 Ω and ∼160 Ω for 1.0 M and 1.75 M concentrations, respectively, throughout the resting period, confirming the formation of an effective and electronically passivating SEI layer that suppresses further electrolyte decomposition.⁵³ Next, the dynamic stability of the interface was evaluated by performing EIS on Li‖Cu cells after repeated plating/stripping cycles (Fig. 4d–f). The surface resistance (R_s = R_SEI + R_ct) in the RE electrolyte increased dramatically with cycling, reaching 165 Ω after 50 cycles.^14,54 This behavior is characteristic of the continuous cracking and reformation of a brittle SEI, which leads to significant polarization and the poor CE observed in Fig. 3a. In contrast, the 1.0 M LiFSI–DEGDEE electrolyte demonstrated a slight increase in R_s as the cycle number increased; however, R_s remained lower compared to RE electrolyte. Furthermore, the 1.75 M LiFSI–DEGDEE electrolyte demonstrated the most remarkable performance, with only a slight increase in resistance. The R_s in the 1.75 M LiFSI–DEGDEE electrolyte was approximately 28 Ω, which is approximately 6 times lower than that in RE electrolyte. These results highlight the outstanding performance of 1.75 M LiFSI–DEGDEE, emphasizing its exceptional ability to reduce cell resistance, suppress dendrite formation, and enhance electrochemical kinetics. The formation of a highly stable SEI layer with 1.75 M LiFSI–DEGDEE plays a critical role in significantly enhancing both CE and cycling performances. These outstanding performances stem from the unique solvation structure which preferentially promotes anion decomposition over solvent decomposition. This leads to the formation of an inorganic-rich SEI, such as LiF, which enhances mechanical strength and effectively suppresses dendrite growth. Concurrently, a stable SEI provides high lithium-ion conductivity, lowering the interfacial resistance and enabling uniform lithium deposition/stripping. As a result, side reactions are minimized, leading to high CE and exceptional long-term cycling stability.

Fig. 4 (a)–(c) Nyquist plots of Li‖Li cells upon standing time. (d)–(f) Nyquist plots of Li‖Cu cells after various cycles with 1.0 M LiPF₆-EC : EMC : DMC (1 : 1 : 1, vol.) (RE), 1.0 M LiFSI–DEGDEE, and 1.75 M LiFSI–DEGDEE.

3.4. Characterization of the SEI layer

To visually and chemically connect the enhanced electrochemical performance to the properties of the electrode–electrolyte interface, the morphology and composition of the SEI were investigated. Li metal was deposited on Cu substrates at a current density of 1.0 mA cm⁻² for 1 h in various electrolytes, and the resulting surfaces were analyzed by SEM and XPS. As shown in the SEM images in Fig. 5a and d, the Li deposit formed in the conventional RE electrolyte exhibited a non-uniform, dendritic morphology. Consistent with numerous reports for carbonate-based systems these high-aspect-ratio, needle-like structures are prone to penetrating the separator, which can lead to internal short circuits and critical safety failures.⁸ This phenomenon occurred because the SEI layer was not strong enough to accommodate the morphological evolution of the lithium surface, which was consistent with the low CEs reported in Fig. 3a. In sharp contrast, the Li deposited in DEGDEE-based electrolytes exhibited remarkably dense, uniform, and compact, nodule-like grains (Fig. 5b, c, e and f). This morphology is highly desirable as it effectively suppresses dendrite growth, thereby minimizing the risk of short circuits and significantly improving the safety and reliability of the cell. The formation of this stable morphology is a direct result of the unique, ionically conductive SEI formed from the DEGDEE-based electrolyte, which in turn leads to the high CE and low polarization observed during cycling (refer to Fig. 3 and 4d–f).

Fig. 5 Morphological evolutions of deposited Li on a Cu substrate at 1 mA cm⁻² for 1 h in (a) and (d) RE, (b) and (e) 1.0 M LiFSI–DEGDEE, and (c) and (f) 1.75 M LiFSI–DEGDEE electrolytes.

To understand the chemical origins of these morphological differences, the composition of the SEI layers was analyzed by XPS. The spectrum and its deconvoluted peaks, reflecting the composition of the SEI layers formed on the electrodes, are presented in Fig. 6. The hydrocarbon C 1s peak (285.0 eV) served as a reference to accurately calibrate the spectral energy scale.^48,55 The C 1s spectra in Fig. 6a reveals that the SEI layers formed in RE and 1.0 M LiFSI–DEGDEE electrolytes exhibit a higher intensity of C–C bonding (285.0 eV), indicating a greater tendency to form SEI layers with higher organic content compared to the 1.75 M LiFSI–DEGDEE electrolyte. Organic species are mechanically soft and often partially soluble in the electrolyte, destabilizing the SEI layer. This exposes the fresh Li electrode to continuous electrolyte interactions, leading to repeated SEI formation and hindering the achievement of high CE. Furthermore, the presence of Li₂CO₃, typically formed from the reduction of carbonate-based solvents during lithium deposition, leads to the formation of a brittle, low-conductivity SEI layer.⁴⁹ This layer is easily fractured and fails to adhere effectively to the metal surface. The F 1s spectra observed in Fig. 6b and summarized in Table S5 (SI) reveal the extremely low intensity of the S–F peak at 687.9 eV in the LiFSI–DEGDEE electrolytes emphasizing that FSI⁻ anion has completely decomposed, contributing to an elevated level of LiF in the SEI. This behavior was further corroborated by the considerably higher intensities of LiF (∼685 eV) identified in DEGDEE-based electrolytes compared to RE electrolyte. The ability to form a LiF-rich SEI is a key advantage of this system. This robust inorganic layer enhances mechanical durability and chemical stability, and offers high Li-ion diffusivity, thereby improving cell kinetics and suppressing lithium dendrite formation.^42,48,55 It is reasonable to state that RE and 1.0 M LiFSI–DEGDEE have similar intensities of C–C peaks; however, 1.0 M LiFSI–DEGDEE, thanks to the larger amount of LiF helps enhance the strength of the SEI layer resulting in being dendrite-free (refer to Fig. 5b and e). Accordingly, the formation of LiF-rich layers in 1.75 M LiFSI–DEGDEE can be regarded as a critical factor for inhibiting Li dendrites (refer to Fig. 5c and f) and reaching high CEs (refer to Fig. 3a).^48,49

Fig. 6 The surface properties of Li deposited on the Cu substrate for 1 h at 1.0 mA cm⁻². (a) C 1s and (b) F 1s of XPS spectra for the SEI layers formed in RE, 1.0 M LiFSI–DEGDEE, and 1.75 M LiFSI–DEGDEE electrolytes.

3.5. Lithium metal anode paired with the LFP cathode

To demonstrate the practical viability of our electrolyte, we integrated it into a full-cell configuration using an LFP cathode, a material widely recognized for its exceptional thermal stability, non-toxicity, safety, and low cost.⁵⁶ To evaluate the effectiveness of the RE and LiFSI–DEGDEE electrolytes for LMBs, coin cells with a Li metal anode and LFP cathode were subjected to rigorous cycling tests. Two formation cycles were operated at 0.1C to ensure the formation of stable passivation layers on electrodes. Following this, subsequent cycles were carried out at charge/discharge current densities of 1.0C/1.0C (1.0C = 170 mA g⁻¹). The long-term cycling performance of these Li‖LFP coin cells, including efficiency and discharge capacity with RE and varying LiFSI concentrations in DEGDEE at 25 °C and 60 °C, is presented in Fig. 7. After the formation cycles, the cells with RE electrolyte showed an initial capacity of approximately 146 mAh g⁻¹ at 25 °C, however, the capacity declined rapidly over subsequent cycles (refer to Fig. 7a). The Li‖LFP cell with RE electrolyte retained less than 80% of its initial capacity after only 140 cycles at 25 °C. According to industry standards, a battery is considered to have reached its end-of-life when its capacity falls to 70–80% of its initial value.⁵⁷ The substantial capacity loss in the RE system is unequivocally attributed to undesirable parasitic reactions between the lithium metal electrode and the electrolyte. Specifically, the brittle, organic-rich SEI layer formed in the RE electrolyte failed to effectively protect the Li electrodes over extended cycles, leading to diminished electrochemical performance. In stark contrast, the DEGDEE-based electrolytes demonstrated superior electrochemical performance, as shown in Fig. 7a, due to their ability to form robust protective layers derived from their unique solvation structure. Among them, the Li‖LFP cells with 1.75 M LiFSI–DEGDEE exhibited the best performances, maintaining an impressive capacity retention of 97% over 750 cycles (with a stable discharge capacity of approximately 147 mAh g⁻¹), and ultimately retaining 85.4% of its initial capacity after 1000 cycles. As can be seen in Table S2 (SI), in comparison, the cell performance of our investigated electrolytes is higher than that of other electrolytes. These results highlight the exceptional cycling stability of the 1.75 M LiFSI–DEGDEE electrolyte. This strongly suggests that the stable, inorganic-rich SEI layer formed in this electrolyte effectively prevents unwanted reactions between Li and the electrolyte, thereby enhancing electrochemical performance and enabling practical long-term operation. To further investigate the stability of Li‖LFP cells in different electrolytes at high temperatures, charge/discharge performance at 60 °C was also evaluated (Fig. 7b). Elevated temperatures exacerbate parasitic reactions, making this a critical test of interfacial stability. Unstable passivation layers lead to additional electrolyte decomposition during galvanostatic cycles at elevated temperatures, which increase cell resistance and ultimately cause capacity loss.^58,59 Li‖LFP cells using DEGDEE-based electrolytes demonstrated significantly longer cycling stability compared to those using the RE electrolyte at 60 °C. Although the Li‖LFP cells with RE electrolyte initially delivered a discharge capacity of ∼138 mAh g⁻¹, the capacity rapidly faded to ∼124 mAh g⁻¹ within only 5 cycles due to the instability of the passivation layers at high temperature. Typically, cells with RE electrolyte (a carbonate-based electrolyte containing LiPF₆) become unstable at high temperatures because the decomposition of LiPF₆ is accelerated, producing large amounts of corrosive species like PF₅, HF, and POF₃. This leads to the decomposition of carbonate solvents (e.g., ring-opening polymerization of EC and disproportionation of EMC) and corrosion of the passivation layers on both the Li anode and LFP cathode, causing catastrophic capacity fading.^60,61 In contrast, after formation cycles, the Li‖LFP cell with 1.75 M LiFSI–DEGDEE maintained a high initial discharge capacity of ∼163 mAh g⁻¹ and a capacity retention of 95.7% after 200 cycles. The superior performance of the Li‖LFP cell with the electrolyte at high temperatures is attributed to the formation of robust, LiF-rich SEI, which is a direct outcome of its unique solvation environment. These LiF-rich layers exhibit significantly higher thermal stability than the organic-rich SEI, which notably enhances the high-temperature stability of LMBs. The voltage-capacity profiles of the Li‖LFP cells in the different electrolytes at 25 °C and 60 °C are presented in Fig. S12 and S13 (SI), respectively. The representative curves are displayed at the plotted cycles from 1st to 200th (RE), 700th (1.0 M, 1.25 M, and 1.5 M), and 1000th (1.75 M), allowing evaluation of polarization growth and plateau stability. The voltage profiles reveal that the voltage hysteresis (voltage difference between charge and discharge) increases continuously, and there is a gradual fading of the discharge plateau, indicative of increased polarization during cycling.⁶² As shown in Fig. S12 (SI), RE exhibits high degradation until the 200th cycle with the large voltage hysteresis, whereas DEGDEE-based electrolytes display more stable and overlapping profiles, with narrower voltage gaps between charge and discharge plateaus. After the 700th cycle, the DEGDEE-based electrolytes exhibited polarization changes and plateau stability with concentrations of 1.0 M, 1.25 M, and 1.5 M, respectively. Conversely, the cell with 1.75 M LiFSI–DEGDEE maintains negligible polarization changes and impressive plateau stability in 1000 cycles (Fig. S12e, SI). Furthermore, the voltage profiles for the Li‖LFP cells at high temperature have also shown the ultimate gradual capacity loss due to elevated temperatures exacerbating parasitic reactions.⁵⁸ Note that there is a noticeable voltage fluctuation near 3.8–3.9 V that appears after prolonged cycling (600–700th) during the charge process of the Li‖LFP cells at 25 °C with 1.0 M and 1.5 M electrolytes (Fig. S12b and d, SI).⁶³ This behavior is mainly caused by the inhomogeneous Li-ion content, which likely arises from transient inhomogeneity in the lithiation state of the LFP electrode, where sluggish Li⁺ transport and uneven desolvation cause transient electrolyte oxidation or surface-film restructuring.⁶⁴ Such inhomogeneous reactions are more pronounced at lower temperatures, where Li⁺ mobility is restricted. Faster ion and charge transfer kinetics at elevated temperature suppress this behavior, yielding smooth charge curves and explaining the absence of the fluctuation at 60 °C (Fig. S13b and d, SI).^63,64 DEGDEE-based electrolytes demonstrated significantly longer cycling stability compared to those using the RE electrolyte at 60 °C (as shown in Fig. S13, SI). Additionally, 1.75 M LiFSI–DEGDEE electrolyte maintains a high capacity while the average discharge voltage decay is slower, exhibiting low polarization changes and excellent plateau stability. These results demonstrate that the DEGDEE-based electrolytes effectively suppress polarization growth and maintain stable redox plateaus of the LFP cathode.

Fig. 7 Electrochemical performances. The Li‖LFP cells were tested with various electrolytes at a charge/discharge current density of 1.0C/1.0C at (a) 25 °C and (b) 60 °C.

The rate capability, a critical metric for practical applications, is inherently linked to the charging current density, which significantly influences both overall electrochemical performance and the morphology of Li deposition. To assess their potential for high-power operation, the rate capabilities of the full cells were systematically investigated by cycling them at various C-rates, ranging from 0.1C to 5.0C. Fig. S14 (SI) compares the specific capacities delivered by the cells with different electrolytes across the tested range of current densities. The cell employing the RE electrolyte exhibited a dramatic decline in rate capability as the current density increased, delivering specific capacities of only ∼77 mAh g⁻¹ at 3.0C and a negligible ∼5 mAh g⁻¹ at 4.0C. In contrast, the rate capability of the DEGDEE-based electrolytes improved with increasing LiFSI concentration, with the 1.75 M electrolyte consistently delivering the highest specific capacities across all tested C-rates. This stark difference in performance is attributed to the nature of the SEI: the superior rate capability in the 1.75 M LiFSI–DEGDEE electrolyte stems from a stable, low-impedance SEI that facilitates rapid Li-ion kinetics.

Post-mortem analyses were conducted to investigate the morphological evolution of electrodes after cycling. Li‖LFP cells were disassembled after 30 cycles to harvest the Li metal anodes for SEM analysis. As illustrated in Fig. S15 (SI), the morphology and integrity of the cycled anodes were profoundly dependent on the electrolyte used. In the RE electrolyte, the Li anode surface was rough, loose, and cracked, indicative of a non-uniform and unstable SEI (Fig. S15a and b, SI). This led to significant volumetric expansion, with the anode thickness increasing to a porous layer of ∼158 µm (Fig. S15c, SI). Such SEI failure allows for irregular Li deposition, resulting in the formation of dendrites and electrically isolated “dead” Li, which in turn accelerates electrolyte decomposition due to the increased surface area.^6,65 In stark contrast, the anodes cycled in the LiFSI–DEGDEE electrolytes displayed a dense, uniform, and globular nodule-like morphology, which effectively suppresses dendrite growth (Fig. S15d, e, g and h, SI). Consequently, volumetric expansion was minimal, with the anode surface remaining compact, cohesive, and largely free of corrosion after 30 cycles (Fig. S15f and i, SI). These observations underscore the superior protective qualities of the SEI formed in the LiFSI–DEGDEE system. Li metal anodes harvested from Li‖LFP cells were further analyzed to investigate the SEI composition (Fig. S16 and Table S6, SI). The C 1s spectra of the RE reveal dominant C–C signals, which can be attributed to extensive solvent decomposition. This observation indicates that the SEI formed in RE is primarily composed of organic species, resulting in a relatively fragile and unstable interphase (Fig. S16a, SI). In contrast, for the 1.75 M LiFSI–DEGDEE electrolyte, the intensity of C–C bonding is significantly reduced compared with the other investigated electrolytes. This finding suggests that fewer organic species are incorporated into the SEI, leading to a denser and more stable interphase. Analysis of the F 1s spectra further highlights the differences in SEI chemistry. In the case of LiPF₆-based electrolytes, incomplete salt decomposition results in the accumulation of Li_xPOF_y species. These inorganic byproducts produce a heterogeneous and mechanically fragile SEI layer, and the amount of LiF formed is relatively low (Fig. S16b, SI). By contrast, DEGDEE-based electrolytes promote more complete decomposition of LiFSI, leading to an SEI enriched in inorganic species and reduced organic residues. This compositional shift is particularly evident at higher salt concentrations. Strikingly, in the 1.75 M LiFSI–DEGDEE electrolyte, both the F 1s and N 1s spectra confirm a substantial increase in LiF and the clear presence of Li₃N, surpassing the levels observed in 1.0 M LiFSI–DEGDEE (refer to Fig. S16c, SI). The emergence of Li₃N is especially noteworthy. Previous studies have demonstrated that Li₃N, together with LiF, plays a critical role in stabilizing the SEI, facilitating fast Li⁺ ion transport, and suppressing parasitic electrolyte decomposition at the electrode surface.^66,67 The combined enrichment of LiF and Li₃N therefore provides a plausible explanation for the improved electrochemical stability and dendrite suppression observed in the 1.75 M LiFSI–DEGDEE electrolyte.

Furthermore, the stability of the cathode–electrolyte interphase (CEI) is also critical for long-term performance, as it must prevent parasitic side reactions and the dissolution of transition metal ions. To examine the CEI composition, LFP cathodes were analyzed using XPS after 30 cycles (Fig. S17 and Table S7, SI). The analysis revealed distinct differences in the CEI composition, directly reflecting the foundational role of the electrolyte's solvation structure. On the LFP cathodes cycled in the RE electrolyte, a thick CEI layer rich in solvent decomposition products was formed. This layer was so substantial that it obscured the signals from the underlying binder (CH₂–CF₂) and conductive carbon, a finding consistent with the cell's poor electrochemical performance (Fig. S17a, SI).³⁰ The high intensity of LiF in the F 1s spectra, along with the presence of Li_xPF_y and Li_xPO_yF_z species in the P 2p spectrum (Fig. S17e, SI), confirms that the degradation originates from the continuous, incomplete decomposition of the LiPF₆ salt, likely exacerbated by trace water to produce corrosive HF.⁶⁸ In stark contrast, the CEI formed in the DEGDEE-based electrolyte is primarily derived from anion decomposition, creating a more stable, inorganic-rich protective layer. Although the investigated electrolytes exhibit a higher intrinsic oxidation potential than that of the LFP materials (as shown in Fig. S7, SI), electrolyte decomposition can still occur due to catalytic oxidation reactions on the cathode surface.¹⁶ In the DEGDEE-based system, the unique solvation structure promotes the preferential decomposition of the FSI⁻ anion over the solvent to form the CEI layer (Fig. S17b and c, SI). This is evidenced by the significantly lower intensity of organic byproducts (O=C–O and C–O/O–H) in the O 1s spectra compared to the RE electrolyte (refer to Fig. S17d, SI), indicating that the DEGDEE-based system effectively minimizes solvent decomposition on both the cathode and the anode. Instead, the FSI⁻ anion decomposition, confirmed by signals in the S 2p, F 1s, and N 1s spectra, constructs a beneficial composite CEI. The N 1s spectrum (Fig. S17f, SI) is particularly revealing, confirming the formation of Li₃N. This component is highly advantageous due to its high ionic conductivity (σ = 6 × 10⁻³ S cm⁻¹ for the single-crystal structure at 25 °C)⁶⁹ and superior stability against lithium metal, which effectively compensates for the sluggish kinetics of LiF (σ = 10⁻³¹ S cm⁻¹).⁷⁰ The creation of a dual-component LiF–Li₃N protective layer is known to stabilize the electrode–electrolyte interface and enhance cycling performance.⁷¹

XPS analysis further reveals that at saturated concentrations, the formation of highly conductive Li₃N is enhanced while the LiF content remains consistent. Therefore, the exceptional efficiency and extended capacity retention of the full cell in the DEGDEE-based electrolyte can be directly correlated with the formation of a stable and highly conductive inorganic SEI/CEI on both the Li anode and LFP cathode.

Conclusion

In this study, we have systematically demonstrated that diethylene glycol diethyl ether (DEGDEE) is a highly effective solvent for practical lithium metal batteries, offering a compelling combination of superior performances. The core advantage of this system lies in the unique Li⁺ solvation structure engineered by the DEGDEE solvent, which was thoroughly investigated across various concentrations. The optimal formulation, 1.75 M LiFSI in DEGDEE, delivered a broad electrochemical stability window and low interfacial resistance. Surface analyses via SEM and XPS revealed that DEGDEE-based electrolytes promote the formation of inorganic-rich SEI/CEI layers primarily driven by salt decomposition preceding solvent reduction, an effect attributed to the high proportion of associated species in the solvation structure. This stabilization of electrode–electrolyte interfaces significantly extends battery lifespan, particularly under high-temperature conditions. Harnessing the advantages of a robust SEI layer and comprehensive electrochemical optimization, the optimal DEGDEE-based electrolyte unlocks remarkable cycling stability. Li plating/stripping reversibility persists over 300 cycles at 25 °C and remains resilient for 250 cycles even under the demanding conditions of 60 °C. With outstanding lithium compatibility, Li symmetric cells achieve ultra-stable cycling for over 1500 h. When coupled with LFP cathodes, Li–metal anodes deliver an exceptional lifespan, exhibiting enhanced self-discharge and superior performance. As a result, Li‖LFP cells sustain stable operation while maintaining a high efficiency (>99.0%) and a capacity retention of 85.4% for over 1000 cycles at 25 °C. Even at 60 °C, the cycling lifespan extends by more than 200 cycles, preserving an outstanding efficiency of 99.8% and a capacity retention of 95.7%. This study underscores DEGDEE-based electrolytes as a breakthrough in stable and high-performance LMBs. Further experimental and theoretical studies are currently underway to generalize this solvent effect and pave the way for the next generation of energy storage solutions.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting the findings of this study are available within the article and its supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d5nh00585j.

Acknowledgements

This research was supported by the National Research Foundation of Korea (NRF) (RS-2024-00348663), and by the Korea Innovation Foundation (INNOPOLIS) (RS-2025-25444020), both funded by the Korean government (MSIT). Additionally, we also acknowledge the support that was provided by Gunsan City, Korea, through the Human Resources Program for the EV industrial cluster.

References

1. J. B. Goodenough and K. S. Park, J. Am. Chem. Soc., 2013, 135, 1167–1176.
2. J. Xie and Y. C. Lu, Nat. Commun., 2020, 11, 1–4.
3. J. M. Tarascon and M. Armand, Nature, 2001, 414(6861), 359–367.
4. H. Kim, G. Jeong, Y. U. Kim, J. H. Kim, C. M. Park and H. J. Sohn, Chem. Soc. Rev., 2013, 42, 9011–9034.
5. M. A. Rahman, X. Wang and C. Wen, J. Electrochem. Soc., 2013, 160, A1759–A1771.
6. B. Liu, J. G. Zhang and W. Xu, Joule, 2018, 2, 833–845.
7. S. Kim, G. Park, J. Lee, S. Seo, K. Ryu, C. H. Kim, J. W. Choi, S. Kim, G. Park, J. W. Choi, S. J. Lee, S. Seo, K. Ryu and C. H. Kim, Adv. Mater., 2023, 35, 2206625.
8. J. Qian, W. A. Henderson, W. Xu, P. Bhattacharya, M. Engelhard, O. Borodin and J. G. Zhang, Nat. Commun., 2015, 6, 1–9.
9. E. Peled, J. Electrochem. Soc., 1979, 126, 2047–2051.
10. C. Fang, X. Wang and Y. S. Meng, Trends Chem., 2019, 1, 152–158.
11. M. D. Tikekar, S. Choudhury, Z. Tu and L. A. Archer, Nat. Energy, 2016, 1, 1–7.
12. D. Lin, Y. Liu and Y. Cui, Nat. Nanotechnol., 2017, 12(3), 194–206.
13. Z. Yu, Y. Cui and Z. Bao, Cell Rep. Phys. Sci., 2020, 1, 100119.
14. G. Zheng, Y. Xiang, S. Chen, S. Ganapathy, T. W. Verhallen, M. Liu, G. Zhong, J. Zhu, X. Han, W. Wang, W. Zhao, M. Wagemaker and Y. Yang, Energy Storage Mater., 2020, 29, 377–385.
15. J. Xiang and Y. C. Lu, ACS Nano, 2024, 18, 10726–10737.
16. Y. Jie, X. Ren, R. Cao, W. Cai, S. Jiao, Y. L. Jie, X. D. Ren, R. G. Cao, W. B. Cai and S. Jiao, Adv. Funct. Mater., 2020, 30, 1910777.
17. D. Ruan, Z. Cui, J. Fan, D. Wang, Y. Wu and X. Ren, Chem. Sci., 2024, 15, 4238–4274.
18. K. Yoshida, M. Nakamura, Y. Kazue, N. Tachikawa, S. Tsuzuki, S. Seki, K. Dokko and M. Watanabe, J. Am. Chem. Soc., 2011, 133, 13121–13129.
19. S. Tobishima, H. Morimoto, M. Aoki, Y. Saito, T. Inose, T. Fukumoto and T. Kuryu, Electrochim. Acta, 2004, 49, 979–987.
20. S. Wei, Z. Li, K. Kimura, S. Inoue, L. Pandini, D. Di Lecce, Y. Tominaga and J. Hassoun, Electrochim. Acta, 2019, 306, 85–95.
21. D. Di Lecce, V. Marangon, H. G. Jung, Y. Tominaga, S. Greenbaum and J. Hassoun, Green Chem., 2022, 24, 1021–1048.
22. A. Dutta, K. Matsushita, Y. Kubo, A. Dutta, Y. Kubo and K. Matsushita, Adv. Sci., 2024, 11, 2404245.
23. Z. Li, H. Rao, R. Atwi, B. M. Sivakumar, B. Gwalani, S. Gray, K. S. Han, T. A. Everett, T. A. Ajantiwalay, V. Murugesan, N. N. Rajput and V. G. Pol, Nat. Commun., 2023, 14(1), 1–13.
24. G. Cai, J. Holoubek, M. Li, H. Gao, Y. Yin, S. Yu, H. Liu, T. A. Pascal, P. Liu and Z. Chen, Proc. Natl. Acad. Sci. U. S. A., 2022, 119, e2200392119.
25. Z. Zeng, V. Murugesan, K. S. Han, X. Jiang, Y. Cao, L. Xiao, X. Ai, H. Yang, J.-G. Zhang, M. L. Sushko and J. Liu, Nat. Energy, 2018, 3, 674–681.
26. J. Evans, C. A. Vincent and P. G. Bruce, Polymer, 1987, 28, 2324–2328.
27. S. Zugmann, M. Fleischmann, M. Amereller, R. M. Gschwind, H. D. Wiemhöfer and H. J. Gores, Electrochim. Acta, 2011, 56, 3926–3933.
28. L. Y. Lin and C. C. Chen, J. Power Sources, 2024, 603, 234236.
29. B. D. Adams, J. Zheng, X. Ren, W. Xu, J.-G. Zhang, B. D. Adams, J.-G. Zhang, J. Zheng, X. Ren and W. Xu, Adv. Energy Mater., 2018, 8, 1702097.
30. T. D. Pham, K.-K. Lee, T. D. K. Pham and K. Lee, Small, 2021, 17, 2100133.
31. P. C. Li, Z. Q. Zhang, Z. W. Zhao, J. Q. Li, Z. X. Xu, H. Zhang and G. Li, Adv. Mater., 2024, 36, 2406359.
32. B. W. Schick, M. Uhl, M. Al-Shakran, J. Bansmann, S. Riedel, Z. Zhao-Karger and T. Jacob, ChemElectroChem, 2025, 2500046.
33. M. Amiri and D. Bélanger, ChemSusChem, 2021, 14, 2487–2500.
34. D. Reber, R. Figi, R. S. Kühnel and C. Battaglia, Electrochim. Acta, 2019, 321, 134644.
35. L. Li, S. Zhou, H. Han, H. Li, J. Nie, M. Armand, Z. Zhou and X. Huang, J. Electrochem. Soc., 2011, 158, A74.
36. P. Kurzweil and K. Brandt, Electrochemical Power Sources: Fundamentals, Systems, and Applications Li-Battery Safety, 2019, pp. 47–82.
37. P. Zhang, H. Jin, T. Wang and M. Wang, Electrochim. Acta, 2018, 277, 116–126.
38. T. Ishikawa, S. Haga, K. Shigenobu, T. Sudoh, S. Tsuzuki, W. Shinoda, K. Dokko, M. Watanabe and K. Ueno, Faraday Discuss., 2024, 253, 385–406.
39. Y. Xie, H. Zou, H. Xiang, R. Xia, D. Liang, P. Shi, S. Dai and H. Wang, J. Membr. Sci., 2016, 503, 25–30.
40. Y. Sun, C. J. Radke, B. D. McCloskey and J. M. Prausnitz, J. Colloid Interface Sci., 2018, 529, 582–587.
41. Y. Xie, H. Xiang, P. Shi, J. Guo and H. Wang, J. Membr. Sci., 2017, 524, 315–320.
42. P. Shi, L. Zhang, H. Xiang, X. Liang, Y. Sun and W. Xu, ACS Appl. Mater. Interfaces, 2018, 10, 22201–22209.
43. K. Lee, S. H. Kwon, J. Kim, E. Park, I. Kim, H. C. Ahn, A. Coskun and J. W. Choi, ACS Energy Lett., 2024, 9, 2201–2211.
44. X. Liu, J. Zhang, X. Yun, J. Li, H. Yu, L. Peng, Z. Xi, R. Wang, L. Yang, W. Xie, J. Chen and Q. Zhao, Angew. Chem., Int. Ed., 2024, e202406596.
45. T. D. Pham, A. Bin Faheem, J. Kim, H. M. Oh, K.-K. Lee, T. D. Pham, A. Bin Faheem, J. Kim, K.-K. Lee and H. M. Oh, Small, 2022, 18, 2107492.
46. C. Wan, M. Y. Hu, O. Borodin, J. Qian, Z. Qin, J. G. Zhang and J. Z. Hu, J. Power Sources, 2016, 307, 231–243.
47. J. Peng, L. Carbone, M. Gobet, J. Hassoun, M. Devany and S. Greenbaum, Electrochim. Acta, 2016, 213, 606–612.
48. R. Miao, J. Yang, Z. Xu, J. Wang, Y. Nuli and L. Sun, Sci. Rep., 2016, 6(1), 1–9.
49. Z. Wang, C. Chen, D. Wang, Y. Zhu and B. Zhang, Angew. Chem., Int. Ed., 2023, 62(28), e202303950.
50. D. Yu, X. Pan, J. E. Bostwick, C. J. Zanelotti, L. Mu, R. H. Colby, F. Lin and L. A. Madsen, Adv. Energy Mater., 2021, 11, 2003559.
51. J. Wang, W. Huang, A. Pei, Y. Li, F. Shi, X. Yu and Y. Cui, Nat. Energy, 2019, 4(8), 664–670.
52. Y. Lu, Z. Tu and L. A. Archer, Nat. Mater., 2014, 13(10), 961–969.
53. W. Hu, Y. Peng, Y. Wei and Y. Yang, J. Phys. Chem. C, 2023, 127, 4495.
54. S. Li, S. Fang, H. Dou and X. Zhang, ACS Appl. Mater. Interfaces, 2019, 11, 20804–20811.
55. T. Duong Pham, A. Bin Faheem, S. Yeon Chun, J.-R. Rho, K. Kwak, K.-K. Lee, T. D. Pham, A. Bin Faheem, K. Lee, S. Y. Chun, K. Kwak and J. Rho, Adv. Energy Mater., 2021, 11, 2003520.
56. S. Kaushik, T. Mehta, P. Chand, S. Sharma and G. Kumar, J. Energy Storage, 2024, 97, 112818.
57. F. Xia, K. Wang and J. Chen, J. Energy Storage, 2023, 64, 107161.
58. J. Jeong, J. N. Lee, J. K. Park, M. H. Ryou and Y. M. Lee, Electrochim. Acta, 2015, 170, 353–359.
59. J. Hou, M. Yang, D. Wang and J. Zhang, Adv. Energy Mater., 2020, 10, 2070079.
60. E. W. C. Spotte-Smith, T. B. Petrocelli, H. D. Patel, S. M. Blau and K. A. Persson, ACS Energy Lett., 2023, 8, 347–355.
61. P. J. Buitrago Botero, A. W. Ells, A. Svirinovsky-Arbeli, M. Juelsholt and L. E. Marbella, J. Am. Chem. Soc., 2025, 147(11), 9159–9174.
62. P. P. Prosini, J. Electrochem. Soc., 2005, 152, A1925.
63. L. Wang, J. Qiu, X. Wang, L. Chen, G. Cao, J. Wang, H. Zhang and X. He, eScience, 2022, 2, 125–137.
64. Y. Kobayashi, H. Miyashiro, A. Yamazaki and Y. Mita, J. Power Sources, 2020, 449, 227502.
65. 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.
66. S. Xia, Z. Lin, B. Peng, X. Yuan, J. Du, X. Yuan, L. Liu, L. Fu, R. Holze and Y. Wu, Energy Environ. Sci., 2024, 17, 5461–5467.
67. S. Wang, J. Zhang, L. Zhang, X. Li, R. Zhao, Y. Liu, Z. Wang, X. Lu, Y. Xin, H. Tian, F. Kang and B. Li, J. Mater. Chem. A, 2024, 12, 5815–5823.
68. T. D. Pham, A. Bin Faheem, H. D. Nguyen, H. M. Oh and K. K. Lee, J. Mater. Chem. A, 2022, 10, 12035–12046.
69. T. Lapp, S. Skaarup and A. Hooper, Solid State Ionics, 1983, 11, 97–103.
70. Q. Zhang, J. Pan, P. Lu, Z. Liu, M. W. Verbrugge, B. W. Sheldon, Y. T. Cheng, Y. Qi and X. Xiao, Nano Lett., 2016, 16, 2011–2016.
71. Y. Lei, X. Xu, J. Yin, K. Xi, L. Wei, J. Zheng, Y. Wang, H. Wu, S. Jiang and Y. Gao, Small, 2024, 20, 2400365.

---