Novel heterocyclic solvent for high-performance lithium–metal and lithium–sulfur batteries

Abstract

Lithium–sulfur (Li–S) batteries are considered promising candidates for next-generation rechargeable batteries owing to their high theoretical energy density of 2600 Wh kg⁻¹ and the natural abundance of sulfur. However, their practical viability is limited by polysulfide dissolution, parasitic side reactions at the lithium anode, and sluggish sulfur redox kinetics, all of which result in rapid capacity fading and poor cycling stability. To address these challenges, we introduce 2-methylfuran (2MeF), a novel heterocyclic solvent with intrinsically weak solvating ability, into conventional ether-based electrolytes to rationally tailor solvation structures and interfacial chemistry. Spectroscopic and computational analyses reveal that the incorporation of 2MeF decreases the fraction of solvent-separated ion pairs while promoting contact ion pairs and aggregates, thereby strengthening Li⁺–solvent interactions. Electrochemical and interfacial characterization studies demonstrate that the optimized DOL:2MeF:DME formulation facilitates the formation of a thin and uniform LiF-rich solid electrolyte interphase on the Li anode and mitigates cathode passivation, resulting in homogeneous Li deposition and more reversible sulfur redox processes. Consequently, Li–S cells exhibit improved cycling and enhanced rate capability. This study underscores the critical role of solvent design in governing solvation chemistry and electrode interfacial stability, offering new insights into the development of sustainable, high-performance Li–S batteries.

---

Introduction

Lithium–sulfur (Li–S) batteries are regarded as one of the most promising next-generation energy storage systems due to the high theoretical specific capacity of sulfur (1675 mAh g⁻¹) and a high theoretical energy density (∼2600 Wh kg⁻¹),¹ far exceeding that of conventional lithium-ion batteries (∼250–300 Wh kg⁻¹).^2,3 Combined with the natural abundance, low cost, and environmental benignity of sulfur, Li–S batteries are promising candidates for next-generation electric vehicles, large-scale energy storage, and urban air mobility (UAM) systems.

Despite these advantages, the practical viability of the Li–S battery system remains unclear due to several challenges. At the cathode, the stepwise conversion of sulfur to soluble lithium polysulfides (LiPSs) and ultimately to Li₂S is often accompanied by sluggish redox kinetics,^4,5 uncontrolled precipitation of insulating Li₂S, and passivation of the cathode surface.^6,7 As a result, limited ion/electron transport across the interface leads to poor reversibility of sulfur redox reactions,⁸ reduced active material utilization, and rapid capacity fading.⁹ On the anode side, dissolved LiPS intermediates readily migrate to the Li–metal anode, where they undergo parasitic side reactions,¹⁰ consuming active lithium and generating an unstable solid electrolyte interphase (SEI) layer.^10,11 This parasitic process gives rise to the polysulfide shuttle effect, leading to low coulombic efficiency and poor long-term cycling stability.^12,13

Electrolyte design has emerged as a key strategy to simultaneously regulate interfacial stability and redox processes of Li–S batteries. Beyond serving as simple Li⁺ ion conductors, electrolytes directly control LiPS reactivity, ion transport, and interfacial chemistry, thereby critically influencing polysulfide shuttle, SEI characteristics, and the reversibility of sulfur redox reactions. Ether-based solvents, particularly mixtures of 1,2-dimethoxyethane (DME) and 1,3-dioxolane (DOL), remain the state-of-the-art electrolyte systems for Li–S batteries. DME, a linear ether, provides high ionic conductivity and high Li salt solubility, but aggravates shuttle behaviour due to excessive LiPS dissolution.¹⁴ In contrast, DOL, a cyclic ether, partially suppresses the shuttle and promotes SEI formation,^15,16 but often results in passivating interphases with limited ionic conductivity and stability.¹⁷ Although binary mixtures of DOL and DME have been widely adopted, their inherent limitations—including insufficient polysulfide control and unstable SEI formation—remain unresolved, highlighting the necessity for novel electrolyte formulations.

Recent efforts have focused on introducing alternative solvents or co-solvents to tailor solvation structures and SEI chemistry in Li–S batteries. Moderately or weakly solvating electrolytes have been reported to mitigate polysulfide shuttling and extend cycle life.^18–24 For example, Kim et al. demonstrated that incorporating fluorinated ether co-solvents into DME-based electrolytes enables the formation of moderately solvating electrolytes that suppress polysulfide dissolution while preserving favorable redox kinetics, thereby extending the cycle life of lean-electrolyte Li–S batteries.²⁵ Despite their effectiveness, these strategies rely primarily on fluorinated solvents, whose environmental persistence and emerging regulatory restrictions raise concerns over their long-term viability. This underscores the need for non-fluorinated alternatives capable of tuning solvating power. In this study, we identify 2-methylfuran (2MeF), a heterocyclic ether with weaker solvating ability, as a promising candidate for such electrolyte design.

Previous studies have shown that 2MeF, when used as an electrolyte additive, can promote SEI formation on Li–metal anodes.^26,27 Although its application in Li–S batteries has not been reported, the delocalization of oxygen lone pairs in the furan ring suggests a potential role in suppressing polysulfide dissolution. In this work, we introduce 2MeF as a non-fluorinated cyclic ether co-solvent and formulate three electrolyte systems—DOL : DME (3 : 7), DOL : 2MeF : DME (2 : 1 : 7), and 2MeF : DME (3 : 7). Their solvation structures were investigated using density functional theory (DFT) calculations, Raman spectroscopy, and nuclear magnetic resonance (NMR) spectroscopy, while interfacial properties and Li deposition behavior were probed by scanning electron microscopy (SEM), galvanostatic electrochemical impedance spectroscopy (GEIS), and X-ray photoelectron spectroscopy (XPS). This comprehensive study demonstrates that 2MeF-based electrolytes enable stable Li deposition and reversible sulfur redox chemistry, highlighting the promise of non-fluorinated cyclic ethers as viable electrolyte components for high-performance Li–S batteries.

Results and discussion

Electrochemical performance of Li–metal cells

This work introduces a novel heterocyclic ether solvent, 2MeF, in the electrolyte formulation. The electrolyte composition was fixed with 0.75 M LiFSI, 70 vol% of the linear ether DME, and 5 wt% LiNO₃ additive, while the amounts of cyclic solvents (DOL and 2MeF) were varied. Three electrolyte compositions are compared: DOL : DME (3 : 7 vol%), DOL : 2MeF : DME (2 : 1 : 7 vol%), and 2MeF : DME (3 : 7 vol%). First, to investigate the effect of each electrolyte composition on Li plating/stripping efficiency and cycling stability, Li–Cu half cells and Li–Li symmetric cells were assembled and evaluated through galvanostatic cycling experiments (Fig. 1).

Fig. 1 Electrochemical measurements of Li–Cu half cells and Li–Li symmetric cells. (a) Coulombic efficiency (CE) of Li–Cu half cells during Li plating/stripping at a current density of 1 mA cm⁻² and an areal capacity of 1 mAh cm⁻². (b) CE of Li–Cu half cells at a current density of 3 mA cm⁻² and an areal capacity of 1 mAh cm⁻². Voltage profile curves are shown in Fig. S1. (c) Polarization overpotential (η_polarization) extracted from the cycling data shown in (b). (d) Voltage profile during initial Li plating onto Cu foil at 0.1 mA cm⁻². (e) Average coulombic efficiencies for Li plating/stripping, evaluated using the modified Aurbach method with Li–Cu half cells. (f) Cycling of Li–Li symmetric cells at a current density of 2 mA cm⁻² with a half-cycle areal capacity of 4 mAh cm⁻².

Lithium was plated onto the Cu electrode at a current density of 1 mA cm⁻² with an areal capacity of 1 mAh cm⁻², followed by a stripping cycle up to 1 V. The voltage profiles of Li plating/stripping are provided in the SI (Fig. S1). The typical voltage profile observed in the Li–Cu half-cell shows an initial voltage dip at the beginning of the Li plating process, corresponding to Li nucleation, followed by a plateau region attributed to Li⁺ mass transfer. During the stripping process, a relatively stable voltage plateau is observed, with a gradual increase in overpotential toward the end, eventually reaching the cutoff voltage of 1 V. The coulombic efficiency (CE) for each cycle is shown in Fig. 1a. Among the tested electrolytes, the Li–Cu cell with DOL:2MeF:DME exhibited the highest average CE of 96.4% over 300 cycles. The cells prepared with DOL:DME and 2MeF:DME electrolytes showed average CEs of 89.3% and 95.4%, respectively. Notably, the DOL:2MeF:DME composition demonstrated the smallest fluctuations of CE throughout the cycle, suggesting the most stable Li plating/stripping behavior.

As the applied current density is increased to 3 mA cm⁻², all cells exhibited greater CE fluctuation; however, the DOL:2MeF:DME system still showed the smallest fluctuation, indicating relatively stable Li cycling behavior (Fig. 1b). The average CE values for DOL:DME, DOL:2MeF:DME, and 2MeF:DME were 92.7%, 97.3%, and 94.8%, respectively, with DOL:2MeF:DME exhibiting the highest CE throughout the cycle. The variations of polarization overpotential (η_polarization) of Li–Cu cells over 300 cycles are presented in Fig. 1c. The η_polarization was calculated as the voltage difference between the plating and stripping plateaus, with the plateau voltage determined at an areal capacity of 0.5 mAh cm⁻² in each cycle. The initial cycle shows a relatively large η_polarization with DOL:DME, DOL:2MeF:DME, and 2MeF:DME exhibiting polarization values of 145, 136, and 238 mV, respectively. This result suggests that the DOL:2MeF:DME electrolyte supports more efficient ion transport and the formation of a relatively stable interfacial layer.^28,29 While a gradual decrease in η_polarization was observed during the initial cycles for the cell with 2MeF:DME, it exhibited the highest overall overpotential among the three cells. In the early cycles, the Li–Cu cell with DOL:DME showed lower polarization; however, the polarization of the DOL:2MeF:DME cell decreased gradually as cycling proceeds, resulting in an η_polarization of 44 mV for DOL:DME and 28 mV for DOL:2MeF:DME at the 300th cycle. This indicates that the use of the DOL:2MeF:DME electrolyte leads to gradual SEI stabilization and improved interfacial reactivity over extended cycling, which is consistent with the improved long-term cyclability observed in Fig. 1a and b.

Fig. 1d shows the voltage profile of the initial Li plating cycle on a Cu substrate at a current density of 0.1 mA cm⁻². The nucleation overpotential (η_nucleation) is defined as the voltage difference between the minimum of the initial voltage dip and the subsequent plateau region, reflecting the lithiophilicity of the substrate.^30,31 The η_nucleation values for DOL:DME, DOL:2MeF:DME, and 2MeF:DME were measured to be −47, −38, and −57 mV, respectively, with the DOL:2MeF:DME system exhibiting the lowest overpotential. Since nucleating Li on the lithiophobic Cu surface requires an overpotential to overcome the energy barrier for initial deposition, the smaller η_nucleation observed with the DOL:2MeF:DME electrolyte indicates a lower nucleation barrier for Li plating. This suggests the formation of a more favorable SEI for Li deposition on Cu in this electrolyte formulation.

Next, a modified Aurbach test³² was employed to more precisely evaluate the average CE of Li–metal cycling with varying electrolyte compositions (Fig. 1e). The 2MeF:DME system exhibited a relatively greater overpotential during the plating/stripping process and exhibited a CE of 97.2%. In contrast, the DOL:DME cell showed a CE of 97.8%, while the cell with DOL:2MeF:DME exhibited the highest CE of 98.8%. The high CE observed with DOL:2MeF:DME in Aurbach cycling indicates improved reversibility of Li plating/stripping and the formation of a stable SEI that minimizes Li loss and electrolyte decomposition.^33–35

To evaluate the long-term cycling stability as a function of electrolyte composition, galvanostatic cycling tests of the Li–Li symmetric cell were performed at a current density of 2 mA cm⁻² with an areal capacity of 4 mAh cm⁻² (Fig. 1f). The Li–Li cell with 2MeF:DME exhibited large overpotential from earlier cycles and short-circuited after 210 h. The DOL:DME-based cell initially exhibited the lowest overpotential; however, signs of micro short-circuiting emerged after 250 h, followed by complete short-circuiting at approximately 310 h. In contrast, the Li–Li cell with DOL:2MeF:DME exhibited a gradual decrease in overpotential during cycling, and maintained stable operation for over 400 h without short-circuiting. One notable feature in the voltage profile is the increasingly pronounced asymmetric peaking at the end of each plating and stripping cycle with higher 2MeF content in the electrolyte. This suggests that 2MeF influences SEI characteristics and spatially varying surface kinetics, leading to a more pronounced peaking feature in the voltage trace.³⁶ To summarize the electrochemical cycling results of Li–metal cells, the DOL:2MeF:DME electrolyte demonstrates superior performance in terms of cycle life, Li plating/stripping reversibility, and Li nucleation behavior.

Morphology of deposited lithium

To evaluate the Li deposition behavior under varying electrolyte compositions, Li was deposited onto Cu foil at a current density of 0.3 mA cm⁻² and areal capacities of 0.9, 1.8, and 3.0 mAh cm⁻² (Fig. 2). At a low deposition capacity of 0.9 mAh cm⁻², distinct morphological differences were observed depending on the electrolyte composition. In the DOL:DME system, Li deposits appeared as rod-like particles (Fig. 2a), while the 2MeF:DME system produced granular particles that further assembled into aggregated structures (Fig. 2g).

Fig. 2 SEM images of Li deposits on Cu foil, obtained after plating Li at a current density of 0.3 mA cm⁻² and areal capacities of 0.9 mAh cm⁻² (left panel), 1.8 mAh cm⁻² (middle panel), and 3.0 mAh cm⁻² (right panel). The electrolyte solvent compositions are: (a–c) DOL : DME (3 : 7), (d–f) DOL : 2MeF : DME (2 : 1 : 7), and (g–i) 2MeF : DME (3 : 7).

In both electrolyte compositions, some regions of the Cu foil remained uncovered by Li, indicating incomplete surface coverage. In contrast, the DOL:2MeF:DME electrolyte resulted in the formation of a smooth and compact lithium layer, with no detectable exposure of the underlying copper substrate (Fig. 2d). This uniform morphology aligns well with the favorable early-stage electrochemical response shown in Fig. 1c and d—specifically the lower nucleation overpotential and smaller polarization observed with the DOL:2MeF:DME electrolyte—demonstrating the significant influence of electrolyte composition on initial Li nucleation and growth behavior.

Upon further deposition to 1.8 mAh cm⁻², morphological divergence among the electrolytes became more evident. The DOL:DME and 2MeF:DME electrolytes yielded larger and more irregular Li structures, either through the growth of existing granules or the formation of larger rod-like deposits (Fig. 2b and h). The uncovered regions of the Cu foil remain, indicating non-uniform growth of Li. In the case of the DOL:2MeF:DME electrolyte, small quasi-spherical Li particles (∼3 μm in diameter) were deposited at the electrode–electrolyte interface, forming a conformal and uniform base layer that fully covered the Cu substrate, which facilitated the subsequent growth of larger Li deposits (Fig. 2e). This growth behavior is likely attributed to the initial SEI formed on the Cu surface in the DOL:2MeF:DME system, where the favorable SEI characteristics promote uniform lithium nucleation and deposition.

After depositing 3 mAh cm⁻² capacity of Li, the DOL:DME system produced enlarged, curved rod-like Li deposits (Fig. 2c), whereas the 2MeF:DME electrolyte yielded an entangled structure consisting of large, rounded particles and rod-like features, accompanied by a rough surface with fine particulate aggregation (Fig. 2i). In contrast, Li deposited from the DOL:2MeF:DME electrolyte maintained a uniform and densely packed morphology, with no apparent voids (Fig. 2f). The DOL:2MeF:DME electrolyte consistently exhibited superior Li deposition behavior across the entire range of deposited capacities. This stable deposition behavior is in line with the enhanced electrochemical cycling performance of Li–metal batteries observed in Fig. 1.

Electrolyte solution structure

To investigate the correlation between the electrolyte structure and electrochemical performance, the solvation structure was analyzed using DFT calculations, Raman spectroscopy, and NMR spectroscopy (Fig. 3). The electrostatic potential (ESP) map visualizes the spatial distribution of electrostatic potential across a molecule, providing critical insights into ion interactions, solvation behavior, and chemical reactivity.^37–39 As shown in Fig. 3a, the minimum ESP values for DME, DOL, and 2MeF are located on the oxygen atom, with values of −154.95 kJ mol⁻¹, −152.43 kJ mol⁻¹, and −114.54 kJ mol⁻¹, respectively. The relatively less negative ESP value of 2MeF indicates a lower electron-donating ability toward Li⁺ ions, suggesting that 2MeF acts as a weaker coordinating solvent compared to DME and DOL. The lower electron-donating ability of 2MeF is likely attributed to the delocalization of the oxygen lone pair into the π-conjugated furan ring, which reduces its ability to coordinate with Li⁺ ions.

Fig. 3 (a) Electrostatic potential maps of DME, DOL, and 2MeF. Color coding ranges from −160.00 kJ mol⁻¹ (red, electron-rich areas) to +160.00 kJ mol⁻¹ (blue, electron-deficient areas). Gray, white, and red spheres represent carbon, hydrogen, and oxygen atoms, respectively. (b–d) Raman measurements of 0.75 M LiFSI in DOL : DME (3 : 7), DOL : 2MeF : DME (2 : 1 : 7), and 2MeF : DME (3 : 7) electrolytes. (b) Raman modes corresponding to the FSI⁻ anion. (c) Deconvolution analysis of the FSI⁻ Raman mode for each electrolyte. (d) Raman modes associated with DME. (e) ⁷Li NMR spectra of each electrolyte. The LiFSI salt concentration was maintained at 0.75 M for all electrolyte compositions.

Next, Raman spectroscopy was performed to probe the solvation structures of Li⁺ ions and identify the coordination states in each electrolyte. Specifically, the S–N–S stretching vibration of the FSI⁻ anion, appearing in the 700–780 cm⁻¹ region, provides a sensitive spectroscopic probe for elucidating the solvation clusters.⁴⁰ The Raman spectra of FSI⁻ mode and deconvolution results are presented in Fig. 3b and c, respectively. The DOL : DME (3 : 7) electrolyte exhibits a solvation structure dominated by solvent-separated ion pairs (SSIPs, ∼69%), along with smaller contributions from contact ion pairs (CIPs, ∼26%) and aggregates (AGGs, ∼5%). Upon increasing the 2MeF content, the SSIP fraction decreases to 66% in DOL : 2MeF : DME (2 : 1 : 7) and further to 63% in 2MeF : DME (3 : 7), while the combined fractions of CIPs and AGGs increase to 34% and 37%, respectively. These results indicate that increasing the 2MeF content drives the solvation structure toward a lower SSIP fraction and a higher proportion of CIPs and AGGs.

In addition to the FSI⁻ mode, the Raman mode corresponding to the DME solvent was analyzed to further investigate changes in the solvation environment (Fig. 3d). The peaks observed at 821 cm⁻¹ and 849 cm⁻¹ originate from free (noncoordinated) DME,⁴¹ while coordination of DME to Li⁺ shifts the vibration to a higher wavenumber, resulting in a Li⁺–DME peak at around 877 cm⁻¹.⁴² In the DOL : DME (3 : 7) electrolyte, the Li⁺–DME coordination peak is located at 877 cm⁻¹, whereas in the 2MeF : DME (3 : 7) electrolyte, it is observed at a higher wavenumber of 881 cm⁻¹. This blue shift is likely associated with the weaker solvating ability of 2MeF, which promotes Li⁺–FSI⁻ association and reduces the number of accessible DME molecules within the solvation shell, thereby enhancing the interaction between Li⁺ and DME molecules. Consistent with the results shown in Fig. 3b and c, increasing the 2MeF content induces a transition of the solvation structure from SSIP-dominant to relatively CIP/AGG-rich configurations, with Li⁺ ions increasingly coordinated by FSI⁻ anions and concurrently strengthening the interaction with DME molecules.

This structural transition is directly reflected in the physicochemical properties of the electrolytes (Table S1). Although the ionic conductivity remains comparable or even slightly enhanced in the 2MeF-containing systems, the Li⁺ transference number decreases significantly from 0.498 (DOL:DME) to 0.363 (DOL:2MeF:DME) and 0.360 (2MeF:DME), as the fraction of CIP/AGG increases. The stronger Li⁺–anion coordination in these configurations restricts the mobility of Li⁺ as a free charge carrier. Instead, Li⁺ ions move together with anions, which contribute little to the overall ionic current. Consequently, the increase in ion-paired species leads to reduced Li⁺ mobility and a lower transference number, even under conditions of similar total ionic conductivity.

To further validate the solvation structure, ⁷Li NMR spectroscopy was performed (Fig. 3e). The ⁷Li nucleus is highly sensitive to changes in the local solvation environment, providing valuable insights into the coordination of Li⁺ ions in different electrolyte systems. In the ⁷Li NMR spectra, the Li⁺ peak progressively shifts upfield from −1.23 ppm in the DOL : DME (3 : 7) electrolyte to −1.25 ppm in the DOL : 2MeF : DME (2 : 1 : 7) electrolyte and −1.31 ppm in the 2MeF : DME (3 : 7) electrolyte with increasing 2MeF content. The observed upfield shift reflects an increased electron density around the Li⁺ ions, which can be attributed to increased ion pairing with FSI⁻ anions. As the 2MeF content increases, the weaker solvating ability of 2MeF leads to a solvation structure where Li⁺ ions are highly associated with FSI⁻ anions, thereby shielding the Li⁺ nuclei and causing the chemical shift to move upfield. These structural characteristics are expected to strongly influence SEI formation at the Li–metal anode and interfacial processes in Li–S cells, which will be further discussed.

Electrochemical performance of Li–S cells

The electrochemical performance of Li–S batteries employing 0.75 M LiFSI in DOL : DME (3 : 7), DOL : 2MeF : DME (2 : 1 : 7), and 2MeF : DME (3 : 7) as electrolytes was evaluated by galvanostatic charge–discharge experiments at room temperature (Fig. 4). Cycling was performed over 50 cycles at 0.1C with a sulfur loading of ∼2 mg cm⁻² (Fig. 4a). The initial discharge cycle capacities were 995, 1125, and 1023 mAh g⁻¹ for the DOL:DME, DOL:2MeF:DME, and 2MeF:DME electrolytes, respectively, with the DOL:2MeF:DME formulation exhibiting the highest initial discharge capacity. Notably, the DOL:2MeF:DME electrolyte retained 73% of its initial capacity after 50 cycles, demonstrating the most stable cycling performance among the three. In contrast, both DOL:DME and 2MeF:DME showed pronounced capacity fading, with DOL:DME exhibiting rapid degradation after 40 cycles, and 2MeF:DME retaining only 46% of its initial capacity at the end of cycle 50. Additionally, the DOL:2MeF:DME cell maintained nearly 100% coulombic efficiency throughout the test, whereas the other two electrolytes exhibited CE fluctuations, suggesting unstable redox reversibility. This superior cycling stability can be correlated with the lowest shuttle current observed in the DOL:2MeF:DME electrolyte (Fig. S2), which effectively suppresses parasitic polysulfide shuttling. These results demonstrate that the optimized electrolyte enables improved reversibility of Li–S chemistry and supports long-term cycling stability.

Fig. 4 Electrochemical cycling performance of Li–S cells assembled with a Li–metal anode, a sulfur cathode, and electrolytes composed of 0.75 M LiFSI in DOL : DME (3 : 7), DOL : 2MeF : DME (2 : 1 : 7), and 2MeF : DME (3 : 7). Cells were cycled at 0.1C between 1.2 and 3.0 V vs. Li/Li⁺. (a) Discharge capacity and coulombic efficiency over 50 cycles. (b) Rate capability at various C-rates. Voltage profiles recorded at (c) cycle 1, (d) cycle 2, and (e) cycle 30, respectively.

Next, C-rate-dependent cycling tests were performed to evaluate the rate capability of Li–S cells with different electrolytes. Li–S cells were cycled at 0.1C, 0.2C, 0.5C, and 1C, each for 10 cycles, followed by recovery at 0.1C (Fig. 4b).

Increasing the C-rate to 1C results in a progressive decrease in discharge capacity for all three cells; however, the capacity fading was less pronounced in the cell with the DOL:2MeF:DME electrolyte. The DOL:2MeF:DME formulation exhibited the highest discharge capacities across all C-rates and demonstrated superior capacity recovery upon returning to 0.1C. When comparing DOL:DME and 2MeF:DME electrolytes, the 2MeF:DME-based cell initially delivered a higher capacity at 0.1C; however, it exhibited a markedly pronounced performance degradation at higher current densities. This noticeable decline in capacity at higher current densities in the 2MeF:DME-based cell is likely associated with the greater cell resistance (Fig. S3) and the resistive nature of its SEI, which will be further discussed.

Voltage profiles at the 1st, 2nd, and 30th cycles were compared to evaluate the electrochemical cycling behavior (Fig. 4c–e). In the 1st cycle, all three cells exhibited typical discharge–charge features of Li–S batteries, with two discharge plateaus near 2.3 V and 2.0 V and a charge plateau at around 2.4 V. The DOL:2MeF:DME-based cell exhibited the most stable discharge and charge plateaus, resulting in the highest capacity and the lowest overpotential among the tested electrolytes.

As cycling progressed, the DOL:DME- and 2MeF:DME-based cells showed noticeable changes in their voltage profiles. In particular, the length of the lower discharge plateau and the charge plateau became significantly shortened, as clearly observed in the 2nd cycle (Fig. 4d). This degradation behavior is associated with the surface passivation by Li₂S on the sulfur cathode, likely caused by uncontrolled growth of Li₂S, which impedes charge transfer and suppresses the further reduction of polysulfides. Although all cells exhibited increased polarization over extended cycling, the DOL:2MeF:DME cell demonstrated the smallest increase in overpotential, maintaining a well-defined voltage profile even after 30 cycles. This result suggests that the optimized solvation structure in DOL:2MeF:DME effectively mitigates cathode passivation and facilitates more reversible sulfur redox reactions, thereby sustaining more stable voltage profiles and promoting enhanced cycling stability in Li–S batteries.

To investigate the influence of electrolyte composition on Li morphology in Li–S cells, the Li anodes retrieved after one galvanostatic cycle were analyzed by SEM (Fig. S4). The Li morphology strongly depended on the electrolyte composition: the DOL:DME and 2MeF:DME systems exhibited rough and non-uniform surfaces, whereas the DOL:2MeF:DME system showed a smooth, compact morphology, consistent with its superior electrochemical performance.

In situ impedance analysis during cycling

To gain further insight into interfacial evolution during cycling, in situ galvanostatic electrochemical impedance spectroscopy (GEIS) measurements were conducted to monitor resistance changes in real time under operating conditions. To investigate the evolution of cell resistance during discharge, GEIS spectra were collected at 20-min intervals throughout the discharge process of Li–S cells. Impedance measurement points were marked along the voltage profiles and numbered accordingly (Fig. 5a–d). The discharge process was segmented into three regions based on the voltage profile: the initial sloping region prior to the first plateau (red shaded), the plateau region (green shaded), and the region following the plateau (blue shaded). For each segment, corresponding Nyquist plots were constructed to capture the dynamic changes in interfacial impedance (Fig. 5b and e). The temporal evolution of impedance was visualized using a color gradient, with deeper hues representing later time points.

Fig. 5 In situ GEIS results of Li–S cells cycled in (a–c) DOL:DME, and (d–f) DOL:2MeF:DME electrolytes. Li–S cells were discharged at 0.1C, with intermittent EIS spectra collected every 20 min during the discharging process, with the dots representing the point where EIS was performed. The voltage profile (left panel), overlaid Nyquist plot (middle panel), and corresponding DRT plot (right panel) are shown. In situ GEIS analysis results of 2MeF:DME electrolyte are displayed in Fig. S5.

To deconvolute the impedance responses, the corresponding distribution of relaxation times (DRT) analysis was carried out, as presented in Fig. 5c and f. The DRT method enables the deconvolution of overlapping electrochemical processes and facilitates the identification of equivalent circuit components by resolving distinct relaxation time constants (τ) and their corresponding resistances.⁴³ The DRT results of both electrolytes show four characteristic peaks, each representing a specific electrochemical process. The first peak at τ ≈ 10⁻⁶ s (P1) corresponds to the interparticle ohmic resistance arising from ionic and electronic conduction within the electrode and electrolyte. The second peak at τ ≈ 10⁻⁴–10⁻⁵ s (P2) corresponds to the resistance associated with SEI formation at the Li–metal anode. The third peak at τ ≈ 10^−0.8 s (P3) reflects the charge-transfer resistance and diffusion resistance of polysulfides near the cathode surface. The slowest relaxation process at τ ≈ 10^0.7 s (P4) represents bulk ion transport and long-range polysulfide diffusion across the cell. This peak assignment is guided by previously reported interpretations of DRT spectra in Li–S systems, as described by Miller and co-workers.⁴⁴

First, we discuss the initial sloping region before the discharge plateau, indicated by the red-shaded area in the voltage profile (data points #1 to #5 in Fig. 5a and d). This region corresponds to the stepwise reduction of elemental sulfur (S₈) into long-chain polysulfides and subsequently short-chain polysulfides, during which the continuous formation of soluble polysulfides leads to a gradual decrease in cell voltage.⁴⁵ Although both DOL:DME and DOL:2MeF:DME electrolytes exhibit similar voltage profiles in terms of shape and potential range, their impedance evolution reveals significant differences. In the DOL:DME-based cell, the diameter of the high-frequency semicircle in the Nyquist plots increases substantially with discharge, from 54 Ω cm² at point #1 to 160 Ω cm² at point #5 (Fig. 5b, top panel). In contrast, the DOL:2MeF:DME-based cell maintains a relatively stable interfacial resistance (∼80 Ω cm²) over the same interval (Fig. 5e, top panel), suggesting suppressed resistance growth. This trend is further corroborated by the DRT spectra. In the DOL:DME system, the P2 peak (τ ≈ 10^−4.4 s), which is attributed to Li⁺ transport across the SEI at the Li–metal surface, increases in magnitude and shifts to longer relaxation times as discharge progresses (Fig. 5c, top panel). Conversely, the DOL:2MeF:DME cell shows minimal change in the P2 peak (Fig. 5f, top panel), indicating a more stable interfacial environment.

These observations suggest that the SEI formed in the DOL:DME system becomes increasingly resistive during early discharge, likely due to uncontrolled reactions between soluble lithium polysulfides and the lithium–metal surface. Such interactions are known to exacerbate interfacial passivation and impede ion transport.^7,46,47 In contrast, the DOL:2MeF:DME formulation appears to effectively suppress LiPS-induced SEI degradation, thereby mitigating resistance buildup and preserving interfacial conductivity during this stage. As a minor yet meaningful observation, the electrolyte ohmic resistance (R_s), corresponding to the intercept of the semicircle at high frequency in the Nyquist plot, increased from 5 Ω cm² to 19 Ω cm² during the initial discharge of the DOL:DME cell (Fig. 5b, top panel). This increase indicates a growing resistance to ionic conduction, likely associated with elevated viscosity due to enhanced dissolution of lithium polysulfides.^44,48 In contrast, the DOL:2MeF:DME cell maintained a nearly constant R_s value throughout the same discharge interval (Fig. 5e, top panel), suggesting that polysulfide dissolution was more effectively suppressed in this system, thereby mitigating the increase in bulk electrolyte resistance. The P4 peak, located at approximately τ ≈ 10^0.7 s and attributed to the bulk diffusion resistance of lithium ions and polysulfide species, progressively increased during the early stages of discharge in both electrolyte systems. However, the magnitude of this increase was more pronounced in the DOL:DME cell, indicating a greater degree of mass transport limitation across the electrolyte and electrode interfaces in this system.

Next, we discuss the plateau region of the discharge curve, indicated by the green-shaded area corresponding to data points #6 through #15. This region represents the electrochemical conversion of short-chain polysulfides into insoluble species such as Li₂S₂ and Li₂S. In the DOL:DME-based cell, the Nyquist plots exhibit the emergence of a second semicircle in the low-frequency region, which progressively grows in size as discharge proceeds (Fig. 5b, middle panel). Correspondingly, the DRT spectra show an increasing P3 peak (τ ≈ 10^−0.8 s), associated with the charge transfer and diffusion resistance of polysulfides, and a P4 peak (τ ≈ 10^0.7 s), attributed to bulk ion or polysulfide diffusion across the cell (Fig. 5c, middle panel). Notably, these peaks gradually grow in intensity and merge into a single broad feature, indicating that both interfacial and bulk transport resistances increase during discharge. This impedance buildup is primarily attributed to the progressive accumulation of insoluble Li₂S₂ and Li₂S species on the cathode surface, which forms an electronically and ionically insulating layer, thereby slowing the conversion of the remaining soluble polysulfides.⁴⁴ Simultaneously, the continued generation of these solid species increases the viscosity of the electrolyte, further hindering the long-range transport of ionic species, including Li⁺ and unreacted polysulfides near the cathode–electrolyte interface.⁴⁴ As a result, both the charge transfer at the cathode and the bulk ionic diffusion across the cell become increasingly restricted, leading to pronounced electrochemical polarization.

In contrast, the DOL:2MeF:DME cell exhibits a similar trend in impedance evolution, but with significantly smaller overall resistance. The Nyquist plots reveal smaller semicircles, and the DRT spectra show a much less pronounced growth in the P3 and P4 peaks (Fig. 5e and f, middle panel). These observations indicate that the resistance increase associated with the formation of solid Li₂S₂/Li₂S species is effectively mitigated in the DOL:2MeF:DME system. This suggests that the electrolyte formulation plays a crucial role in modulating the deposition behavior of insoluble discharge products. As a result, both polysulfide reduction and diffusion processes are better maintained, supporting more efficient sulfur redox kinetics during discharge.

Finally, in the post-plateau region (blue shade, data points #16 onward), corresponding to the steep voltage drop near the end of discharge, the impedance growth becomes more pronounced. In the DOL:DME cell, the Nyquist plots show an enlargement of the low-frequency semicircles, while the DRT spectra reveal significant increases in the merged P3/P4 peaks, indicating severe charge-transfer and diffusion limitations. This behavior is consistent with extensive cathode passivation by insulating Li₂S, which blocks active sites and hinders polysulfide transport across the electrolyte.

In contrast, the DOL:2MeF:DME cell exhibits a much more sluggish impedance rise, with P3/P4 peaks growing only moderately. This suggests that the optimized solvation structure in DOL:2MeF:DME mitigates excessive Li₂S accumulation and preserves ion transport pathways even in the deep-discharge regime. Such resistance mitigation likely arises from reduced polysulfide dissolution—limiting parasitic SEI thickening—and the formation of a more stable interfacial layer on the Li–metal anode.

The GEIS measurement results for the 2MeF:DME cell are presented in Fig. S5. Unlike the other electrolyte systems, the 2MeF:DME cell exhibited the appearance of a second semicircle already in the pre-plateau stage (red shade), indicating the early onset of charge-transfer resistance associated with polysulfide reduction. Furthermore, in the plateau stage (green shade), the P4 peak corresponding to bulk diffusion resistance was already observed, suggesting that cathode surface passivation by prematurely formed Li₂S₂ and Li₂S occurred at an earlier stage compared to the other electrolytes. This accelerated emergence of both charge-transfer and diffusion resistances highlights the inherently faster degradation dynamics in the 2MeF:DME system.

Interfacial chemistry revealed by XPS depth profiling

To further investigate the interfacial characteristics underlying these impedance trends, XPS depth profile analysis was performed on Li deposits obtained after repeated plating/stripping cycles in each electrolyte. As shown in the F 1s spectra (Fig. 6a–c), the outermost SEI surface (i.e., before etching) in all three electrolyte systems consists of two components: C–F (∼688 eV) and Li–F (∼685 eV). Upon surface etching, only LiF was detected in all three electrolyte compositions. LiF is widely recognized as a mechanically robust and ionically conductive SEI component that can suppress dendrite growth and stabilize Li plating/stripping.⁴⁹

Fig. 6 F 1s XPS depth profiles of the SEI formed after the 5th Li plating in Li–Cu cells using (a) DOL : DME (3 : 7), (b) DOL : 2MeF : DME (2 : 1 : 7), and (c) 2MeF : DME (3 : 7) electrolytes. Li stripping/plating was conducted at a current density of 1 mA cm⁻² with an areal capacity of 1 mAh cm⁻². Depth profiling was performed by Ar⁺ ion etching at 1 keV in 10 s intervals to investigate SEI composition and LiF distribution. (d) Variations of the LiF peak area as a function of etching cycle number.

As shown in Fig. 6, the LiF distribution as a function of SEI depth exhibited markedly different behaviors among the electrolytes. For the DOL:DME electrolyte, the LiF peak area increased sharply during the initial etching cycles, reaching a maximum at the second cycle before gradually decreasing. This trend suggests that LiF is concentrated in the upper regions of the SEI, farther from the electrode interface. The 2MeF:DME cell exhibited the opposite trend, showing a continuous growth of the LiF signal with depth, indicating the formation of a thick and resistive interphase in the inner SEI layer. In both electrolytes, the large amounts of LiF and its steep depth-dependent variation suggest the formation of non-uniform SEI layers that hinder ion transport. In contrast, the DOL:2MeF:DME electrolyte exhibits a smaller overall LiF content that is uniformly distributed throughout the SEI, suggesting the formation of a thin, stable interphase capable of maintaining low interfacial resistance and suppressing early passivation during cycling. These findings are in strong agreement with the impedance and DRT analysis in Fig. 5 and Fig. S5, where the DOL:2MeF:DME cell exhibited the smallest increase in SEI-associated resistance and stable interfacial transport properties. The more uniform LiF incorporation in this electrolyte likely enhances the mechanical robustness and ionic conductivity of the SEI/CEI, thereby facilitating uniform lithium deposition and supporting more favorable Li–S redox kinetics during cycling.

Taken together, the XPS depth profile confirms that DOL:2MeF:DME promotes the formation of a thinner and more uniformly distributed LiF layer, while DOL:DME and 2MeF:DME generate thicker and spatially heterogeneous LiF-rich interphases that hinder ion transport. Such differences strongly imply that the solvation environment of the electrolyte governs the reduction pathway of FSI⁻ anions, ultimately dictating the interfacial chemistry.

Correlation between the solvation structure, interfacial stability, and cell performance

The results presented above collectively highlight the strong correlation between the electrolyte solvation structure, interfacial chemistry, and the overall electrochemical performance of Li–metal and Li–S cells. We discuss how the interplay between the solvation structure and interfacial stability governs the electrochemical behaviors of the three electrolyte formulations.

(1) Influence of the solution structure on SEI formation. Raman and NMR analyses revealed that the solvation environment of Li⁺, particularly the coordination state of FSI⁻ anions, varies strongly with the solvent composition. In DOL:2MeF:DME, the solvation structure is dominated by CIPs, where FSI⁻ directly coordinates with Li⁺. This configuration undergoes moderate reduction, favoring the formation of a thin and uniform LiF-based SEI.⁵⁰ Such a structure effectively suppresses uncontrolled Li dendrite growth and ensures stable Li plating/stripping behavior. In contrast, the 2MeF:DME electrolyte promotes AGG-type solvation, in which multiple FSI⁻ anions cluster around Li⁺. This state is more susceptible to deep reduction, resulting in excessive LiF generation and the formation of a thick, resistive SEI.⁴⁰ The DOL:DME system exhibits an intermediate behavior, with a smaller CIP fraction and relatively unstable SEI evolution, as evidenced by the accumulation of LiF in the upper SEI region observed in XPS depth profiling. Collectively, these results demonstrate that the solvation structure directly governs the reduction pathway of FSI⁻ and thereby dictates the thickness, uniformity, and ion transport properties of the SEI formed on Li–metal surfaces.

(2) Influence of SEI characteristics on Li–metal electrochemistry and deposition morphology. The properties of the SEI strongly affect the electrochemical response of the Li–metal anode. In the DOL:2MeF:DME electrolyte, the optimized solvation structure produces a thin, uniform, and ionically conductive SEI, resulting in the lowest nucleation overpotential, the highest coulombic efficiency, extended symmetric-cell lifetime, and reduced overpotential compared to the other systems. SEM analysis further confirms uniform and compact Li deposits, in contrast to the porous and dendritic morphologies observed in other electrolytes. These results indicate that a mechanically stable and chemically homogeneous SEI is essential to regulate Li nucleation and growth, enabling reversible Li plating/stripping.

(3) Implications for Li–S cell performance. The impact of electrolyte-induced interfacial chemistry extends beyond the Li–metal to critically shape the electrochemical performance of Li–S cells. Among the tested systems, the DOL:2MeF:DME electrolyte delivered the most favorable cycling performance, exhibiting superior capacity retention and markedly improved rate capability. This enhancement originates not only from the stabilized Li anode interface but also from suppressed cathodic impedance growth, as evidenced by the delayed emergence of charge-transfer (P3) and bulk-diffusion (P4) resistances in GEIS/DRT analysis. In sharp contrast, the 2MeF:DME cell exhibited early development of these resistive features, indicative of premature passivation of the sulfur cathode by insulating Li₂S species, ultimately leading to accelerated performance degradation. The DOL:DME electrolyte exhibited an intermediate behavior, showing a progressive increase in interfacial resistance over cycling. Taken together, these findings highlight that the optimized solvation structure in DOL:2MeF:DME electrolyte simultaneously stabilizes the Li anode interface and mitigates cathode passivation, thereby facilitating more reversible sulfur redox kinetics and significantly enhanced cycling stability and rate performance in Li–S batteries.

Conclusions

In this study, we systematically investigated the impact of the electrolyte solvation structure on interfacial chemistry and electrochemical performance in Li–metal and Li–S cells. Raman and NMR analyses revealed that tuning the solvent composition with 2MeF modulates the Li⁺–FSI⁻ coordination environment, shifting the balance between SSIP, CIP, and AGG states. The DOL:2MeF:DME electrolyte exhibited the highest fraction of CIPs, which promoted moderate anion reduction and the formation of a thin, uniform LiF-rich SEI.

Electrochemical evaluation of Li–metal anodes demonstrated that the uniform and stable SEI formed in DOL:2MeF:DME resulted in lower nucleation overpotential, higher coulombic efficiency, and improved cycling stability in Li–metal cells, accompanied by more uniform Li deposition morphology. Furthermore, GEIS/DRT analysis highlighted that DOL:2MeF:DME effectively suppressed the growth of interfacial and diffusion resistances on both anode and cathode surfaces, in contrast to the early passivation observed with 2MeF:DME. These favorable interfacial properties translated into superior Li–S cell performance, with the DOL:2MeF:DME electrolyte delivering the highest capacity retention, improved rate capability, and stable voltage profiles over prolonged cycling.

Overall, 2-MeF plays a dual mechanistic role in improving the performance of Li–metal and Li–S cells. (i) At the Li anode, it promotes the formation of a thin and uniform LiF-rich SEI by enabling a CIP-dominated solvation structure that undergoes controlled reduction. (ii) At the sulfur cathode, its participation in the Li⁺ solvation sheath moderates the interfacial charge-transfer process, thereby mitigating Li₂S passivation and enabling stable sulfur redox kinetics during cycling.

These findings emphasize the importance of tailoring solvent coordination environments to optimize SEI/CEI chemistry but also provide guiding principles for designing advanced electrolytes for high-energy-density Li–S and Li–metal batteries. Looking forward, tuning solvation environments through molecular-level electrolyte engineering may provide versatile design principles for next-generation rechargeable batteries.

Experimental

Electrolyte preparation

Lithium bis(fluorosulfonyl)imide (LiFSI, >99.0%, TCI Sejin CI), 1,2-dimethoxyethane (DME, 99.5%, Sigma-Aldrich), 1,3-dioxolane (DOL, 99.8%, Sigma-Aldrich), 2-methylfuran (2MeF, 99%, Sigma-Aldrich), and lithium nitrate (LiNO₃, 99.98%, Thermo Fisher) were purchased and used without further purification. All electrolyte formulations contained 0.75 M LiFSI with 5 wt% LiNO₃ as an additive, while the solvent composition was varied as follows: (i) DOL : DME (3 : 7 vol%), (ii) DOL : 2MeF : DME (2 : 1 : 7 vol%), and (iii) 2MeF : DME (3 : 7 vol%). Electrolytes were prepared in an argon-filled glovebox with O₂ and H₂O levels maintained below 1 ppm. Physicochemical properties of these electrolytes are summarized in Table S1.

Electrode preparation

Sulfur powder (S₈, 99.98%, Sigma Ald.) and conductive carbon (Super P Li, Timcal Inc.) were dried in an oven at 100 °C for 4 h prior to use. The dried powders were mixed at a weight ratio of 5 : 4 (S₈ : C) and ball-milled using a planetary mill (XQM-0.4A, Changsha Tianchuang, China) at 250 rpm. The milling procedure consisted of 30 min milling followed by a 10 min rest, repeated for 20 cycles. Poly(vinylidene fluoride) (PVDF, Solef 5130, Solvay) was dissolved in N-methyl-2-pyrrolidinone (NMP, 99.5%, Sigma Ald.) under magnetic stirring to prepare a stock solution (PVDF/NMP = 40 mg mL⁻¹). The S₈/C composite was mixed with the PVDF/NMP solution at a weight ratio of 5 : 4 : 1 (S₈ : C : PVDF), and additional NMP was added to adjust slurry viscosity. The resulting slurry was first mixed using a paste mixer (HPM-S1.5H, HanTech, Korea) for 15 min, followed by further mixing with a vortex mixer (Vortex-Genie 2, Scientific Industries, Inc., USA) to ensure homogeneity. The sulfur slurry was cast onto aluminum foil (35 μm thickness) and dried at room temperature for 24 h, followed by vacuum drying at 55 °C for 8 h to obtain the sulfur cathode.

Cell assembly

CR2032-type coin cells were assembled using a hydraulic crimper (MTI Corporation, USA). For Li–Cu half cells, Cu foil discs (14 mm diameter, 0.025 mm thickness, 99.8%, Alfa Aesar) were used as substrates, paired with Li foil (15 mm diameter, 0.15 mm thickness, 99.9%, HONJO Metal) and a polypropylene separator (Celgard 2400, 19 mm diameter). Li–S cells were prepared using sulfur cathodes (14 mm diameter, ∼2 mg cm⁻² sulfur loading) and Li–metal anodes. The electrolyte volume was fixed at 60 μL for all cell configurations, corresponding to an electrolyte-to-sulfur (E/S) ratio of ∼20 μL mg⁻¹ for the Li–S cells. All assembly processes were conducted in an Ar-filled glovebox with O₂ and H₂O levels maintained below 1 ppm.

Electrochemical measurements

Galvanostatic cycling was carried out using a multichannel battery tester (CT-4008Tn-5V10mA-164, Neware, China). All measurements were conducted at 25 °C in a temperature-controlled chamber. For Li–Cu half cells, Li was deposited onto Cu substrates at a current density of 0.3 mA cm⁻² for 20 min or 0.1 mA cm⁻² for 1 h, at an areal capacity of 1 mAh cm⁻². The deposited Li was subsequently stripped at the same current density until reaching a cutoff voltage of 1.0 V vs. Li/Li⁺. The Aurbach cycling test was conducted to evaluate the coulombic efficiency. First, 5 mAh cm⁻² of Li was deposited onto a Cu substrate at 0.5 mA cm⁻² and subsequently stripped to 1.0 V vs. Li/Li⁺, after which the same amount of Li was redeposited under identical conditions. This was followed by 10 cycles of Li plating/stripping at 1 mAh cm⁻² at a current density of 0.5 mA cm⁻². Finally, all remaining Li was completely stripped to 1.0 V vs. Li/Li⁺ at a current density of 0.5 mA cm⁻². Li–Li symmetric cells were cycled at 2 mA cm⁻² with an areal capacity of 4 mAh cm⁻².

Galvanostatic charge–discharge tests of Li–S cells were performed at 0.1C within a voltage window of 1.2–3.0 V vs. Li/Li⁺. For rate capability evaluation, the cells were cycled sequentially at 0.1C, 0.2C, 0.5C, and 1C for 10 cycles each, followed by a return to 0.1C. During charging, a constant voltage hold at 3.0 V vs. Li/Li⁺ for 30 min was applied after the constant current step. Electrochemical impedance spectroscopy (EIS) was conducted using a BioLogic SP-150 potentiostat. Galvanostatic electrochemical impedance spectroscopy (GEIS) measurements of Li–S cells were carried out intermittently every 20 min during the charge–discharge process by applying a current perturbation equal to 10% of the charge/discharge current. The applied current corresponded to the 0.1C rate based on the sulfur cathode, and the EIS spectra were recorded over the frequency range of 1 MHz to 0.1 Hz. Distribution of relaxation times (DRT) analysis was subsequently carried out using RelaxIS 3 software (Rhd Instruments) to deconvolute overlapping electrochemical processes and identify characteristic time constants.

Potentiostatic electrochemical impedance spectra were collected in the as-prepared state within the frequency range of 1 MHz to 0.1 Hz with a voltage amplitude of 10 mV. Shuttle current measurements were conducted in Li–S cells by first discharging the cell galvanostatically at 0.1C to 1.2 V, charging to 3.0 V, and subsequently discharging potentiostatically to 2.35, 2.34, 2.25, 2.20, 2.15, and 2.1 V vs. Li/Li⁺. At each voltage step, the potential was held for 1 h.

The ionic conductivity of electrolyte solutions was measured by EIS. A blocking cell configuration was employed, where a polytetrafluoroethylene (PTFE) flat gasket was placed between two stainless steel blocking electrodes and filled with 200 μL of electrolyte. EIS spectra were recorded using a BioLogic SP-300 potentiostat over the frequency range of 7 MHz to 0.1 Hz at 25 °C. The bulk resistance (R_b) was determined from the high-frequency intercept of the Nyquist plot, and the ionic conductivity (σ) was calculated using σ = L/(R_b·A), where L is the gasket thickness, and A is the effective electrode area (cm²).

The Li⁺ transference number (t₊) was determined using the Bruce–Vincent–Evans method.⁵¹ Li–Li symmetric cells were assembled using glass fiber separators soaked with 300 μL of electrolyte. A potential bias of 10 mV was applied for 10 h until the current reached a steady state, and EIS spectra were recorded before and after the DC bias over the frequency range of 1 MHz to 0.1 Hz at 25 °C. Both chronoamperometry and EIS measurements were conducted using a BioLogic SP-300 potentiostat. The transference number was calculated according to the following equation:

where ΔV is the applied potential (10 mV), I₀ and I_ss are initial and steady-state currents, and R₀ and R₁ are the interfacial resistances before and after polarization, respectively.

Characterization

SEM. The surface morphology of Li deposits was examined using a field-emission scanning electron microscope (FE-SEM, JSM-7500F, JEOL Ltd., Japan). For the Li–Cu half cells, Li was plated onto Cu foil in Li–Cu half cells at a current density of 0.3 mA cm⁻² for plating durations of 3, 6, and 10 h. For the Li–S cells, galvanostatic discharge was performed at 0.1C within a voltage window of 1.2–3.0 V vs. Li/Li⁺. After the respective tests, the cells were disassembled using a hydraulic crimping/disassembling machine (MTI Corporation, USA). The retrieved Cu and Li electrodes were rinsed with DME to remove residual electrolyte and subsequently dried under vacuum at room temperature for 24–48 h prior to SEM measurements.

DFT calculation. Electrostatic potential (ESP) maps of the electrolyte solvents were obtained by density functional theory (DFT) calculations at the B3LYP/6-311+G** level of theory using the Spartan'14 software package.

Raman spectroscopy. Raman spectra were collected using a DXR Raman Microscope (Thermo Scientific) equipped with a 532 nm excitation laser operating at 10 mW. Electrolyte samples were prepared in an argon-filled glovebox and sealed in 2.5 mm glass capillaries. Spectra were acquired with an acquisition time of 2 s and 180 accumulations. Raman spectra were deconvoluted using Voigt functions to determine the peak areas.

NMR. NMR spectra were acquired at room temperature on a JEOL JNM-ECZ500R/S1 spectrometer operating at 500 MHz. Electrolyte samples were prepared in an Ar-filled glovebox and transferred into 5 mm screw-cap NMR tubes to avoid exposure to air and moisture. ⁷Li chemical shifts were externally referenced to 1 M LiCl (99.9%, ThermoFisher) in D₂O (99.9%, Sigma Ald.) using a coaxial sealed capillary.

XPS. XPS measurements were conducted using a Thermo Scientific K-Alpha spectrometer equipped with a monochromatic Al Kα X-ray source (hν = 1486.6 eV). Depth profiling was performed with a monatomic Ar⁺ ion source operated at 1000 eV, with the surface etched in 10 s intervals. All binding energies were calibrated to the C 1s peak at 284.8 eV. Background subtraction was applied using the Shirley method, and the spectra were fitted with a GL(30) line shape to quantify the peak areas.

Author contributions

Conceptualization (Y. H. and M. S.); data curation (Y. H.); formal analysis (Y. H. and M. S.); funding acquisition (M. S.); investigation (Y. H. and M. S.); methodology (Y. H. and M. S.); project administration (M. S.); supervision (M. S.); validation (Y. H. and M. S.); visualization (Y. H. and M. S.); writing – original draft (Y. H.); writing – review & editing (M. S.)

Conflicts of interest

There are no conflicts to declare.

Data availability

The authors declare that the data supporting the findings of this study are available within the paper and its supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d5ta07058a.

Acknowledgements

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (Grant No. RS-2022-NR071972 and Grant No. RS-2025-24535363). This research has been supported by the POSCO Science Fellowship of the POSCO TJ Park Foundation.

References

1. P. Zuo, J. Hua, M. He, H. Zhang, Z. Qian, Y. Ma, C. Du, X. Cheng, Y. Gao and G. Yin, J. Mater. Chem. A, 2017, 5, 10936–10945.
2. J. Li, Z. Du, R. E. Ruther, S. J. An, L. A. David, K. Hays, M. Wood, N. D. Phillip, Y. Sheng and C. Mao, JOM, 2017, 69, 1484–1496.
3. K. Zhu, C. Wang, Z. Chi, F. Ke, Y. Yang, A. Wang, W. Wang and L. Miao, Front. energy res., 2019, 7, 123.
4. X. Jiao, X. Tang, J. Li, Y. Xiang, C. Li, C. Tong, M. Shao and Z. Wei, Chem. Sci., 2024, 15, 7949–7964.
5. S. Lang, S.-H. Yu, X. Feng, M. R. Krumov and H. D. Abruña, Nat. Commun., 2022, 13, 4811.
6. G. Zhou, A. Yang, G. Gao, X. Yu, J. Xu, C. Liu, Y. Ye, A. Pei, Y. Wu and Y. Peng, Sci. Adv., 2020, 6, eaay5098.
7. H. Chu, H. Noh, Y.-J. Kim, S. Yuk, J.-H. Lee, J. Lee, H. Kwack, Y. Kim, D.-K. Yang and H.-T. Kim, Nat. Commun., 2019, 10, 188.
8. J. T. Kim, A. Rao, H.-Y. Nie, Y. Hu, W. Li, F. Zhao, S. Deng, X. Hao, J. Fu and J. Luo, Nat. Commun., 2023, 14, 6404.
9. H. Pan, J. Chen, R. Cao, V. Murugesan, N. N. Rajput, K. S. Han, K. Persson, L. Estevez, M. H. Engelhard and J.-G. Zhang, Nat. Energy, 2017, 2, 813–820.
10. X. Q. Zhang, Q. Jin, Y. L. Nan, L. P. Hou, B. Q. Li, X. Chen, Z. H. Jin, X. T. Zhang, J. Q. Huang and Q. Zhang, Angew. Chem., Int. Ed., 2021, 60, 15503–15509.
11. J. Li, L. Gao, F. Pan, C. Gong, L. Sun, H. Gao, J. Zhang, Y. Zhao, G. Wang and H. Liu, Nano–Micro Lett., 2024, 16, 12.
12. W. Lu, Y. Xu, C. Dazon, F. Macaulay, V. Pimenta and C. Serre, Energy Mater., 2025, 5, 500075.
13. X.-Y. Li, S. Feng, Y.-W. Song, C.-X. Zhao, Z. Li, Z.-X. Chen, Q. Cheng, X. Chen, X.-Q. Zhang and B.-Q. Li, J. Am. Chem. Soc., 2024, 146, 14754–14764.
14. R. Atwi and N. N. Rajput, Patterns, 2023, 4, 100799.
15. S. S. Zhang, J. Power Sources, 2013, 231, 153–162.
16. M. Xie, Y. Wu, Y. Liu, P. Yu, R. Jia, W. A. Goddard III and T. Cheng, Mater. Today Energy, 2021, 21, 100730.
17. S. Song, W. Gao, G. Yang, Y. Zhai, J. Yao, L. Lin, W. Tang, N. Hu and L. Lu, Mater. Today Energy, 2022, 23, 100893.
18. S. C. Kim, X. Gao, S.-L. Liao, H. Su, Y. Chen, W. Zhang, L. C. Greenburg, J.-A. Pan, X. Zheng and Y. Ye, Nat. Commun., 2024, 15, 1268.
19. Y. Liu, Y. Elias, J. Meng, D. Aurbach, R. Zou, D. Xia and Q. Pang, Joule, 2021, 5, 2323–2364.
20. G. Liu, Q. Sun, Q. Li, J. Zhang and J. Ming, Energy Fuels, 2021, 35, 10405–10427.
21. H. Ye and Y. Li, Nano Res. Energy, 2022, 1, e9120012.
22. M. Shin, H.-L. Wu, B. Narayanan, K. A. See, R. S. Assary, L. Zhu, R. T. Haasch, S. Zhang, Z. Zhang and L. A. Curtiss, ACS Appl. Mater. Interfaces, 2017, 9, 39357–39370.
23. W. Shin, L. Zhu, H. Jiang, W. F. Stickle, C. Fang, C. Liu, J. Lu and X. Ji, Mater. Today, 2020, 40, 63–71.
24. J. Wu, M. Li, L. Ma, X. Li, X. Chen, J. Long, Y. Wang, X. Li, J. Liu and Z. Guo, ACS Nano, 2024, 18, 32984–32994.
25. I. Kim, S. Kim, H. Cho, J. Jung, H. Kwon, D. Kim, Y. Shin and H. T. Kim, Adv. Energy Mater., 2025, 15, 2403828.
26. K. Abraham, J. Foos and J. Goldman, J. Electrochem. Soc., 1984, 131, 2197.
27. L. Dominey and J. Goldman, Proceedings of the 34th International Power Sources Symposium, 1990, pp. 82–86.
28. Z. Yu, H. Wang, X. Kong, W. Huang, Y. Tsao, D. G. Mackanic, K. Wang, X. Wang, W. Huang and S. Choudhury, Nat. Energy, 2020, 5, 526–533.
29. Y. Li, Y. An, Y. Tian, H. Fei, S. Xiong, Y. Qian and J. Feng, J. Electrochem. Soc., 2019, 166, A2736–A2740.
30. X. Chen, X.-R. Chen, T.-Z. Hou, B.-Q. Li, X.-B. Cheng, R. Zhang and Q. Zhang, Sci. Adv., 2019, 5, eaau7728.
31. A. Mohammadi, L. Monconduit, L. Stievano and R. Younesi, J. Electrochem. Soc., 2022, 169, 070509.
32. B. D. Adams, J. Zheng, X. Ren, W. Xu and J. G. Zhang, Adv. Energy Mater., 2018, 8, 1702097.
33. H. Kwon, H. Kim, J. Hwang, W. Oh, Y. Roh, D. Shin and H.-T. Kim, Nat. Energy, 2024, 9, 57–69.
34. R. Li, X. Huang, H. Zhang, J. Wang, Y. Fan, Y. Huang, J. Liu, M. Yang, Y. Yu and X. Xiao, Nat. Commun., 2025, 16, 4672.
35. G. Wang, M. Zhu, Y. Zhang, C. Song, X. Zhu, Z. Huang, Y. Zhang, F. Yu, G. Xu and M. Wu, InfoMat, 2022, 4, e12293.
36. K.-H. Chen, K. N. Wood, E. Kazyak, W. S. LePage, A. L. Davis, A. J. Sanchez and N. P. Dasgupta, J. Mater. Chem. A, 2017, 5, 11671–11681.
37. M. Qin, Z. Zeng, Q. Wu, X. Liu, Q. Liu, S. Cheng and J. Xie, J. Energy Chem., 2023, 85, 49–57.
38. Y. Zhao, T. Zhou, M. Mensi, J. W. Choi and A. Coskun, Nat. Commun., 2023, 14, 299.
39. L. Dong, S. Zhong, B. Yuan, Y. Ji, J. Liu, Y. Liu, C. Yang, J. Han and W. He, Research, 2022, 9837586.
40. S. Kim, B. Seo, H. V. Ramasamy, Z. Shang, H. Wang, B. M. Savoie and V. G. Pol, ACS Appl. Mater. Interfaces, 2022, 14, 41934–41944.
41. Z. Cui, Z. Jia, D. Ruan, Q. Nian, J. Fan, S. Chen, Z. He, D. Wang, J. Jiang and J. Ma, Nat. Commun., 2024, 15, 2033.
42. Y. Zhao, T. Zhou, T. Ashirov, M. E. Kazzi, C. Cancellieri, L. P. Jeurgens, J. W. Choi and A. Coskun, Nat. Commun., 2022, 13, 2575.
43. Y. Lu, C.-Z. Zhao, J.-Q. Huang and Q. Zhang, Joule, 2022, 6, 1172–1198.
44. R. Soni, J. B. Robinson, P. R. Shearing, D. J. Brett, A. J. Rettie and T. S. Miller, Energy Stor. Mater., 2022, 51, 97–107.
45. Y. Huang, L. Lin, C. Zhang, L. Liu, Y. Li, Z. Qiao, J. Lin, Q. Wei, L. Wang and Q. Xie, Adv. Sci., 2022, 9, 2106004.
46. É. A. Santos, M. M. Amaral, B. S. Damasceno, L. M. Da Silva, H. G. Zanin, J. N. Weker and C. B. Rodella, Nano Energy, 2024, 110098.
47. I. Kim, J. Jung, S. Kim, H. Cho, H. Chu, W. Jo, D. Shin, H. Kwon and H. T. Kim, EcoMat, 2024, 6, e12483.
48. S. Risse, N. A. Cañas, N. Wagner, E. Härk, M. Ballauff and K. A. Friedrich, J. Power Sources, 2016, 323, 107–114.
49. S. Zaman, M. Narayanasamy, S. M. Naqvi, T. Hassan, A. Iqbal, U. Zafar, N. Hussain, S. Jeong, S. Y. Cho and S. Jung, Energy Mater., 2025, 5, 500028.
50. T. Hou, K. D. Fong, J. Wang and K. A. Persson, Chem. Sci., 2021, 12, 14740–14751.
51. J. Evans, C. A. Vincent and P. G. Bruce, Polymer, 1987, 28, 2324–2328.

---