Introduction

The demand for energy sources is continuously increasing1, driving the need for battery systems with higher energy density and longer cycle life than conventional lithium-ion batteries2. Li−S batteries have garnered significant attention as next-generation energy storage systems because they utilize high energy density S (theoretical capacity = 1672 mAh g−1)3,4,5 cathodes and Li metal (theoretical capacity = 3860 mAh g−1, potential = −3.04 V vs. SHE)6 anodes. This combination of high-energy density electrodes yields an exceptionally high theoretical energy density of over 2600 Wh kg−1. Moreover, S, employed as the cathode material, is abundant in the Earth’s crust and environmentally benign, providing notable benefits in terms of cost-effectiveness and sustainability7,8,9.

Despite these advantages, several critical challenges must be addressed to enable the reliable operation of Li−S batteries. One major issue in Li−S batteries is the dissolution of LiPS and the accompanying parasitic reactions. During the discharge process, the conversion of sulfur to lithium sulfide proceeds via a solid−liquid−solid reaction mechanism10. This phase change involves the generation of LiPS (Li2Sx, 4 ≤ x ≤ 8) as intermediates11, which dissolve into ether-based electrolytes and trigger the LiPS shuttle effect12,13. The shuttle effect of LiPS leads to the loss of active sulfur and undesirable interactions with the Li metal anodes, ultimately resulting in reduced coulombic efficiency and capacity fading. Furthermore, the Li metal anodes present additional limitations, such as dendrite formation and interfacial instability, both of which cause rapid cycling fading14,15,16. Since these simultaneous changes at the cathode and anode mutually accelerate each other, in most cases, the performance degradation of Li−S batteries is often more severe than that of conventional lithium-ion counterparts.

A number of approaches have been introduced to mitigate these challenges in Li−S batteries. Electrolyte additives have been one of the primary remedies, as the given interfacial phenomena at both the cathode and anode are strongly influenced by the structure of the electrolyte. A representative additive, lithium nitrate (LiNO3), forms a stable solid electrolyte interphase (SEI) on the Li metal anode surface, effectively suppressing side reactions with LiPS17,18,19. Following this, other nitrate-based additives, such as CsNO320, KNO321, ZrO(NO3)222, LaNO323, and NaNO324, have been investigated. Additional additives, including SOCl2 (forming LiCl-enriched SEI)25, KPF6 (forming LiF-enriched SEI)26, and CuPc (containing copper sulfide)27, have also been developed to further improve the cycling performances. Despite significant progress, the development of additives for Li−S batteries remains limited by challenges such as complicated mechanistic interactions and long-term stability issues.

In this study, we propose a closo-type complex hydride, Li(CB11H12), as an additive for Li−S batteries. In ether-based electrolytes, Li(CB11H12) dissociates into Li+ and disordered [CB11H12] complex anions. The dissociated [CB11H12] modifies the solvation structure around Li+, promoting Li+−solvent coordination while driving the anions outward. This modified electrolyte weakens LiPS solvation by reducing their solvation free energy, thereby suppressing the shuttle effect, while simultaneously forming a complex hydride-based interfacial layer that mitigates side reactions at the Li metal anodes.

Complex hydrides have been intensively investigated as solid-state electrolytes due to their high lithium−ion conductivity, low material density, and outstanding chemical/electrochemical stability against the Li metal anodes28,29,30,31,32,33,34,35,36,37,38,39. Based on these conspicuous properties, various types of all-solid-state batteries employing complex hydride solid electrolytes such as Li−S40,41,42, Li−TiS243,44, Li−LiCoO245, Li−NMC have been studied46,47. These characteristics of complex hydrides can be applied to a wide range of battery electrolyte systems, including aqueous and organic liquid electrolytes. Therefore, to the best of our knowledge, the utilization of complex hydrides as additives in liquid-electrolyte lithium batteries, such as those demonstrated in this study, represents an approach distinct from their conventional role in all−solid−state batteries.

Results and discussion

Electrolyte synthesis and solvation structure

Complex hydrides, generally denoted as Mx(MyHz), where M and MyHz represent a metal cation and a complex anion, respectively, exhibit excellent chemical and electrochemical stability toward the Li metal anodes because of their strong reducing capability30,42. In particular, the closo-type complex hydride Li(CB11H12) includes a large cage−like [CB11H12] polyatomic complex anion formed by strong covalent bonds among the central atoms (boron (B) and carbon (C)) and multiple hydrogen (H) atoms (Fig. 1a). The structure of the [CB11H12] complex anion is characterized by delocalized electron-deficient bonding, which leads to weak electrostatic interactions with lithium−ions48,49. The X-ray diffraction (XRD) profile of Li(CB11H12) that is used as an electrolyte additive in this study is indexed by the orthorhombic unit cell, which is consistent with that of its low-temperature phase (space group Pca21 (Z = 4))32 (Supplementary Fig. 1).

Fig. 1: Synthesis and characterization of the electrolyte.
figure 1

a Crystal structure of Li(CB11H12), showing ionic bonding between Li+ and [CB11H12] and covalent bonding within the [CB11H12] complex anion. b Schematic illustration of the Li+ solvation structure in the pristine electrolyte and 0.05 M Li(CB11H12) electrolyte. c Raman spectra of the pristine and 0.05 M Li(CB11H12) electrolytes. d NMR (left: 7Li, middle: 17O, and right: 19F) spectra of the pristine electrolyte and 0.05 M Li(CB11H12) electrolytes.

The electrolytes were prepared by simply mixing LiTFSI salt, LiNO3 additive, DOL:DME solvent, and Li(CB11H12) additive. From the testing of various contents of Li(CB11H12) (Supplementary Fig. 2), 0.05 M was chosen as the main modified electrolyte in this investigation. Hereafter, the electrolytes prepared without and with 0.05 M Li(CB11H12) are denoted as pristine and 0.05 M Li(CB11H12), respectively. Figure 1b illustrates a schematic representation of Li+ solvation structures in the pristine and 0.05 M Li(CB11H12) electrolytes. The Li(CB11H12) additive increases the number of solvent molecules surrounding Li+ compared to the pristine electrolyte, as clarified by the subsequent Raman and nuclear magnetic resonance (NMR) analyses.

Vibrating modes of the constituent ions and molecules in the pristine and 0.05 M electrolytes were examined by Raman spectroscopy measurements (Fig. 1c and Supplementary Fig. 5). In the range of 700−800 cm−1, the Raman profile of the pristine electrolyte exhibits various CF3 bending and S−N stretching modes of the TFSI anion, including free TFSI (solvent-separated ion pairs) at 739 cm−1, contact ion pairs (CIP; a TFSI anion interacting with one Li+) at 744 cm−1, and aggregates (AGG; a TFSI anion interacting with two or more Li+ ions) at 749 cm−1, respectively50,51. For the 0.05 M Li(CB11H12) electrolyte, the peak intensity of free TFSI increases, whereas those of CIP and AGG decrease, indicating that the Li(CB11H12) additive reduces the involvement of TFSI in Li+−anion pairing. In addition, the Raman spectra of the 830−890 cm−1 and 1000−1060 cm−1 regions of the pristine electrolyte present the Raman peaks from the DOL:DME solvents at 870 cm−1 and 1024 cm−1, both of which correspond to Li+−coordinated solvent52. The intensity of the Li+−coordinated solvent peak is enhanced in the 0.05 M Li(CB11H12) electrolyte, which indicates strengthened Li+−solvent coordination. A similar weak interaction between Li+ and anions was observed from the NO stretching modes of NO3 anions. The Raman peaks at 1037, 1042, and 1052 cm−1 represent free NO3, CIP (one Li+ interacting with NO3), and AGG (NO3 interacting with multiple Li+), respectively53. In the 0.05 M Li(CB11H12) electrolyte, the intensity of the free NO3 peak increases, whereas that of the CIP and AGG peaks decreases compared to the pristine electrolyte, confirming a higher portion of NO3 remaining in the free-ion state. Also, the enhanced peak at 1024 cm−1 in the 0.05 M Li(CB11H12) electrolyte supports this interpretation, indicating enhanced Li+−solvent coordination.

These changes in the Li+ solvation structure are accompanied by a significant modification in the symmetric structure of the [CB11H12] complex anions, compared to that in solid-state Li(CB11H12). Solid-state Li(CB11H12) features multiple B−H stretching modes associated with anisotropic bonding, characterized by variations in their covalent bond lengths49,54, as reflected in its Raman profile that exhibits multiple B−H peaks. In contrast, in the 0.05 M Li(CB11H12) electrolyte, these modes are consolidated into two peaks, which imply that the B−H bonds become isotropic. The equalized bond lengths demonstrate a decrease in the symmetry of the [CB11H12] complex anions (and thus an increase in disorder), as observed in the high−temperature phases of complex hydrides32,55 of solid state Li(CB11H12). In the same context, the shift toward a higher Raman shift of the C–H stretching mode in the 0.05 M Li(CB11H12) electrolyte indicates a decrease in bond length, which can also be interpreted as a result of bond alignment in the [CB11H12] complex anions. The 7Li NMR spectrum of the 0.05 M Li(CB11H12) electrolyte exhibits a reduced chemical shift caused by the elevated electron density around Li+, reconfirming enhanced coordination by solvent molecules (Fig. 1d). The 17O NMR peaks of the solvents also show a reduced chemical shift, which can be attributed to an increased coordination number around Li+ that lowers the electron donation from each individual solvent molecule56,57. Moreover, the weakly coordinating anion [CB11H12]48,49, with its high reducing ability, preferentially interacts with solvent molecules, thereby enhancing the shielding of oxygen nuclei58,59. Consistently, the 19F NMR spectrum of the 0.05 M Li(CB11H12) electrolyte shows a reduced chemical shift, indicating weakened Li+−TFSI coordination. The increased Li diffusivity and the lowered H diffusivity also support the dominant interaction between Li+ and solvent (Supplementary Fig. 3)60,61.

Overall, these findings reveal that the Li(CB11H12) additive forms a unique solvation structure in which solvent-solvated Li+ ions and free anions coexist. This structural change can be understood by the low coordinating nature of the [CB11H12] complex anions. It is well known that the disordering of complex anions in the solid-state complex hydrides significantly reduces interactions with surrounding ions42,62. Therefore, the disordered structure of the [CB11H12] complex anions in the 0.05 M Li(CB11H12) electrolyte, as demonstrated by Raman analyses, directly reflects its weak coordinating ability. The redshift of the Raman peak associated with the symmetric vibrational mode of the closo-type anion in the 0.05 M Li(CB11H12) electrolyte (Supplementary Fig. 4) indicates weakened electrostatic interaction with Li+, supporting its role as a weakly coordinating anion. The weak coordinating characteristic of disordered [CB11H12] complex anions weakens the interaction with anions such as TFSI and NO3, thereby stabilizing primarily Li+−solvent coordination in the first solvation shell (Fig. 1b). This solvent-dominant solvation structure can suppress the stabilization and dissolution of LiPS, thereby potentially suppressing their solubility in the electrolyte. Importantly, as Li(CB11H12) contains multiple H atoms in its structure, even a small amount of this additive can effectively reconfigure the local solvation environment around Li+, potentially improving various aspects of the performance of Li−S batteries, as verified by the electrochemical characterization in the following section.

Electrochemical properties

The electrochemical performance of both electrolytes was evaluated using Li−S cells. Detailed cell preparation and measurement conditions are described in the Methods section. The 0.05 M Li(CB11H12) electrolyte exhibited good electrochemical performance in various aspects, such as specific capacity, rate capability, and cycle life. Figure 2a shows the first discharge−charge profiles of cells with the 0.05 M Li(CB11H12) electrolyte and pristine electrolyte under a current density of 0.1 C (1 C = 935.2 mAh g−1). When galvanostatically tested in the voltage range of 1.8−2.8 V vs. Li+/Li at 0.1 C, the 0.05 M Li(CB11H12) electrolyte cell presented higher first discharge and charge capacities than the pristine electrolyte. In the first discharge, the 0.05 M Li(CB11H12) electrolyte and pristine electrolyte cells delivered 895.9 and 843.1 mAh g−1, respectively, whereas in the first charge, they showed 896.7 and 852.9 mAh g−1, respectively.

Fig. 2: Electrochemical performance of Li−S full cells.
figure 2

a First discharge−charge voltage profiles of the pristine electrolyte and 0.05 M Li(CB11H12) electrolyte cells at 0.1 C. b Rate performance of the pristine electrolyte and 0.05 M Li(CB11H12) electrolyte cells at various C-rates. The E/S ratio was 35 μL mg−1. Discharge capacities after 1st and 100th cycles of c the pristine electrolyte and d 0.05 M Li(CB11H12) electrolyte cells at various E/S ratios (12, 25, and 35 μL mg−1). The capacity retention between the 1st and 100th cycles was described as a percentage. e Cycling performances of discharge capacity and coulombic efficiency of the pristine electrolyte and 0.05 M Li(CB11H12) electrolyte cells at 0.5 C. The E/S ratio was 12 μL mg−1.

The 0.05 M Li(CB11H12) cell also displayed improved rate capability (Fig. 2b and Supplementary Fig. 2). As the C-rate increased to 2, 5, 10, and 20 times from 0.1 C, the 0.05 M Li(CB11H12) cell retained 79.8%, 68.7%, 59.2%, and 47.6% of the capacity (895.9 mAh g−1) in the first cycle. In contrast, with the same C-rate variations, the pristine electrolyte cell retained 79.8%, 68.8%, 57.9%, and 41.7%, even though its capacity (843.1 mAh g−1) in the first cycle was lower.

The enhanced cycling stability of the 0.05 M electrolyte cell was more pronounced under low E/S conditions (Fig. 2c, d, and Supplementary Figs. 68). The 0.05 M Li(CB11H12) electrolyte cells showed superior cycling stability during 100 cycles under all E/S ratios (12, 25, and 35 μL mg−1) used (Fig. 2c). Indeed, the capacity retention (60.1%) after 100 cycles at 12 μL mg−1 was almost the same as that (62.6%) at 35 μL mg−1. The discharge capacity after 100 cycles of the 0.05 M Li(CB11H12) electrolyte cell at 12 μL mg−1 was 628.5 mAh g−1 (Supplementary Fig. 8). However, the pristine electrolyte cell at 12 μL mg−1 retained only 11.5% after 100 cycles. The enhanced cycling stability of the 0.05 M electrolyte was also reflected in the dQ/dV results (Supplementary Fig. 9). When measured at higher rates of 0.5 C, 1 C, and 2 C (after an initial 3 activation cycles at 0.05 C), the 0.05 M Li(CB11H12) electrolyte cell showed superior cycle retention compared to the pristine electrolyte counterpart (Fig. 2e and Supplementary Fig. 10). Moreover, under high sulfur loading (4 mg cm−2), the 0.05 M Li(CB11H12) electrolyte exhibited more stable cycling performance than the pristine electrolyte (Supplementary Fig. 11). The beneficial effect of the Li(CB11H12) additive was observed even in the absence of LiNO3 (Supplementary Figs. 12 and 13).

The Li(CB11H12) additive resulted in improvements in kinetic behavior and stability. Based on these results, the electrochemical evaluation was expanded to measurements intended to separately assess the reactions of the S cathodes and the Li metal anodes. First, the potential difference method was employed to examine the relationship between LiPS dissolution from the S cathodes and the solvation structure of the electrolyte prepared. Figure 3a shows the discharge voltage profiles of the pristine electrolyte and 0.05 M Li(CB11H12) electrolyte cells during the first cycle at a current density of 0.1 C and an E/S ratio of 12 μL mg−1. Both cells exhibit two distinct voltage plateaus; the 0.05 M Li(CB11H12) electrolyte cell displays a lower first plateau voltage (V1st) and a higher second plateau voltage (V2nd) compared to the pristine electrolyte counterpart.

Fig. 3: Electrochemical reactions at the S cathodes and the Li metal anodes.
figure 3

a First discharge voltage profiles of cells with the pristine electrolyte and 0.05 M Li(CB11H12) electrolyte at 0.1 C. b Galvanostatic cycling profiles of Li||Li symmetric cells at a current density of 1 mA cm−2. The areal capacity was 1 mAh cm−2.

In Li−S batteries, the discharge process proceeds through a two-step consecutive reaction, and because both the reactants (S, Li) and the final product (Li2S) are solid-phase species10,12, the overall Gibbs free energy change (∆G_total) remains constant regardless of the electrolyte composition. V1st observed originates from the formation of LiPS as an intermediate phase during the discharge reaction Li (s) + S (s) ↔ LiPS (sol)63, and thus this value is determined by the solvation free energy (∆G1, sol.) of the generated LiPS. According to the relevant equation, ∆G1, sol = −nFV1, where n and F represent the number of moles of electrons and Faraday’s constant, respectively, a weak solvation free energy during discharge results in a low V1st (Supplementary Fig. 14)64,65. The high V2nd detected can also be understood in the same context. Since the discharge process proceeds through a sequential two-step reaction, the overall free energy change (∆G_total) remains constant. Thus, a decrease in ∆G1, sol leads to a corresponding increase in ∆G2.

Our structural analyses revealed that the Li(CB11H12) additive induces primary interactions between Li+ and solvent molecules. Additionally, the [CB11H12] complex anions exhibited a disordered state with very low coordination ability. It is therefore concluded that the decreased V1st and the increased V2nd result from the reduced ∆G1, sol. due to the Li(CB11H12) additive. Moreover, this behavior can be applied to interpret the solubility of LiPS, as the ∆G1, sol is proportional to −ln(Ksp), the solubility product constant64. Consequently, the lowered solvation free energy implies a reduced solubility of LiPS during cycling.

The solubility of LiPS was further examined experimentally using ultraviolet-visible (UV-vis) absorption spectroscopy (Supplementary Fig. 15). The Li2S6 solution without any added electrolyte (referred to as the blank) exhibits characteristic absorption peaks of S62− (470 nm) and S42− (420 nm) species. After 12 h, the 0.05 M Li(CB11H12) electrolyte exhibits lower absorbance peaks of S62− and S42− compared to the pristine counterpart, supporting that the Li(CB11H12) additive effectively reduces the solubility of LiPS. Free solvent molecules are known to promote the formation of lithium polysulfides (LiPS)64,66,67. In contrast, our solvation structure analysis reveals that Li(CB11H12) induces enhanced Li+−solvent coordination. Consequently, the 0.05 M Li(CB11H12) electrolyte mitigates the effect of free solvent molecules, leading to a reduced LiPS solubility.

Next, Li plating/stripping experiments were carried out to assess lithium-ion transfer capability across the Li metal anode interface. When galvanostatically cycled at a current density of 1 mA cm−2 with a constant capacity of 1.0 mAh cm−2, the 0.05 M Li(CB11H12) electrolyte cell exhibited superior cycle performance during 1000 h (Fig. 3b). In contrast, the pristine electrolyte cell indicates large voltage fluctuations, confirming unstable interfacial reactions. The slightly elevated overpotential observed during the initial cycles in the 0.05 M Li(CB11H12) electrolyte is attributed to the high interfacial resistance prior to cell saturation (Supplementary Fig. 16). The stable Li plating/stripping reactions induced by the Li(CB11H12) additive were also confirmed through measurements at various current densities (Supplementary Figs. 17 and 18), SEM analyses after cycling (Supplementary Fig. 19), and evaluations in Li||Cu cells (Supplementary Fig. 20). The enhanced stability against the Li metal anodes is ascribed to the reversible reaction by the strong reducing ability35,68 of the Li(CB11H12) additive.

Analysis of the electrode interface and morphology

The electrochemical processes of the 0.05 M Li(CB11H12) electrolyte cell were further investigated by various analyses. The SEM and XPS analyses on the S cathodes after 10 cycles present no significant changes in morphology and chemical state (Supplementary Figs. 21 and 22). These results indicate that the Li(CB11H12) additive mainly affects the dissolution of LiPS identified in the potential difference analyses (Fig. 3a).

Therefore, our additional characterizations focus on the reaction of the lithium metal anodes. After cycling with the 0.05 M Li(CB11H12) electrolyte at 0.1 C, SEM measurements reveal that the Li metal surface remains highly smooth and uniform (Fig. 4a). In contrast, the pristine electrolyte cells exhibit a rough and porous Li metal surface, which becomes progressively more damaged with ongoing cycles, likely due to continuous side reactions with the electrolyte (Fig. 4b).

Fig. 4: Structural characterizations of the Li metal anodes during cycling.
figure 4

SEM images of the Li metal anodes after 1 and 10 cycles at 0.1 C: a the 0.05 M Li(CB11H12) and b pristine electrolytes. Scale bars, 100 μm. XPS spectra of c S 2p, d Li 1s, e F 1s, and f C 1s of the Li metal anodes after 10 cycles. g Schematic illustration of the roles of the Li(CB11H12) additive in Li−S batteries.

To further investigate the reactions at the Li metal interfaces, XPS analyses were conducted. Importantly, the S 2p spectra indicate that in the 0.05 M Li(CB11H12) electrolyte cell, the formation of the highly oxidized LixSOy species is suppressed (Fig. 4c). It is known that the LixSOy species are formed from the oxidation of LiPS and/or their side reactions with decomposition byproducts of TFSI and NO3 in the electrolyte17,69. In addition, the 0.05 M Li(CB11H12) electrolyte cell exhibits a significantly higher ratio of LiPS among the surface compounds compared to the pristine cell, clarifying that their oxidation reactions are mitigated. Importantly, for the Li 1s spectrum, a strong new peak, which is assigned to the complex hydride-based layer on the Li metal surface70, is observed at 57 eV in the 0.05 M Li(CB11H12) electrolyte cell (Fig. 4d). These results suggest that the Li(CB11H12) additive not only suppresses the dissolution of LiPS from the S cathode but also inhibits their subsequent oxidation at the lithium metal interface. The F 1s and C 1s spectra further confirm that the Li(CB11H12) additive mitigates the formation of electrolyte decomposition byproducts, including LiF, CF3, ROLi, HCO2Li, ROCO2Li, and RCH2OCO2Li (Fig. 4e, f)71,72. Our Li plating/stripping experiments (Fig. 3b) and SEM analyses (Supplementary Fig. 19) demonstrated that the intrinsic reducing ability of Li(CB11H12) additive stabilizes the interfacial reactions of the lithium metal anode, independently of its reactions with LiPS (Fig. 3b, Supplementary Figs. 17 and 18). In addition, the relative atomic concentrations of surface elements were quantitatively determined based on XPS survey scans (Supplementary Fig. 23 and Supplementary Table 1).

Therefore, the improved performances of the Li−S batteries are attributed to the synergistic effects of suppressed LiPS dissolution, mitigated LiPS oxidation, and stabilized Li metal plating/stripping reactions. The roles of the Li(CB11H12) additive are graphically summarized in Fig. 4g.

To the best of our knowledge, this study represents the first application of a complex hydride additive in Li−S batteries, providing an important starting point for the further development of this material system. Our results offer a strategy for modulating the Li+ coordination environment and suggest the potential of a complex hydride additive to address fundamental limitations of the Li−S batteries. Together with the extensive research on Li−S batteries, prior studies have explored various additive chemistries23,73,74,75. In particular, ionic liquid-based additives designed to improve long-term cycling performance have shown substantial improvements through continuous developments of early-stage formulations76,77,78,79,80. In this work, although only Li(CB11H12) was studied among various complex hydride candidates, it highlights the possibility of designing a broader range of complex hydride additives (e.g., other complex hydrides and mixed complex hydrides)42,43,55, which can enable diverse molecular and structural design.

Conclusions

By adopting the closo-type complex hydride Li(CB11H12) as an electrolyte additive, we have effectively enhanced the performance of Li−S batteries. In the Li−S battery liquid electrolyte, the [CB11H12] complex anions exhibit a weakly coordinating disordered state that induces enhanced interaction between Li+ and solvent molecules, thereby suppressing LiPS dissolution and mitigating the shuttle effect. Additionally, the Li(CB11H12) additive forms a highly reducing complex hydride-based interfacial layer on the lithium metal anode, which not only reduces parasitic oxidation reactions of LiPS and electrolyte decompositions but also stabilizes the lithium metal interface. From a broader perspective, the insights provided in this work could be applied to a variety of liquid electrolyte systems that require weakly coordinating solvation structures and/or reducing environments.

Methods

Preparation of CNT-based S cathodes

Sulfur powder (Sigma Aldrich, 99.98%) and multi-walled carbon nanotubes (MWCNTs, Nanocyl, diameter = 9.5 nm) were mixed in ethanol at a weight ratio of S:MWCNT = 7:3 wt.% using ball milling (450 rpm, 30 min, 10 cycles) to ensure uniform dispersion. Ethanol was subsequently removed through vacuum filtration, and the obtained MWCNT@S mixture was dried in a vacuum oven at 60 °C for 12 h to eliminate residual solvents. The dried mixture was then heated at 155 °C for 2 h to facilitate the melt-diffusion of sulfur into the CNT network. The resulting MWCNT@S material was combined with Super P carbon black (SPB, Alfa Aesar) and carboxymethyl cellulose/styrene-butadiene rubber (CMC/SBR) in a weight ratio of 8:1:1. The slurry was cast on an Al current collector using the doctor blade technique and dried in a vacuum oven at 60 °C for 12 h. The loading density of final CNT-based sulfur cathodes was ~1.3 mg cm−2.

Preparation of carbon-coated separator

The carbon mixture (YP@SPB) was prepared using a mechanical mixing method with a weight ratio of 8:2. Carbon powder was dispersed in 1-Methyl-2-pyrrolidinone (Sigma Aldrich, 99%) solvent and subsequently mixed with a polyvinylidene fluoride (PVDF, Solvay Corp)-based polymeric binder at an 8:2 weight ratio. The resulting Carbon@PVDF slurry was cast onto a polyethylene membrane (W-Scope, pore volume: 43%, pore size: 60 nm, thickness: 16 µm) with a film thickness of approximately 30 µm. The coated membrane was then vacuum dried at 50 °C for 16 h. The completed carbon-coated separator was punched into 19 mm diameter discs for further use.

Preparation of electrolyte

The pristine electrolyte was fabricated by dissolving 1 M lithium bis(trifluoromethanesulfonyl)imide (LiTFSI, Sigma Aldrich, 99.95%) and 1.0 wt.% lithium nitrate (LiNO3, Sigma Aldrich, 99.99%) in 1,3-dioxolane (DOL, Sigma Aldrich, 99.95%) and 1,2-dimethoxyethane (DME, Sigma Aldrich, 99.5%) (1:1, volume ratio). The LiTFSI and LiNO3 were dried under vacuum at 80 °C for 12 h prior to use. The Li(CB11H12) was obtained by drying the hydrated compounds under vacuum (<5 × 10−2 Pa); Li(CB11H12)∙xH2O (Katchem) at 160 °C for 12 h. For comparison, various amounts of Li(CB11H12) were added to the pristine electrolyte. The Li(CB11H12) concentrations were adjusted to 0.05, 0.1, 0.3, and 0.5 M. These electrolytes were stirred for 12 h in a glove box filled with argon gas.

Electrochemical measurements

All CR2032 coin cells (Wellcos Co.) were assembled in an argon-filled glove box (O2 < 1 ppm and H2O < 1 ppm). A polyethylene membrane was used as the separator. The Li−S full cell tests were performed in a two-electrode configuration with a carbon-coated separator, using lithium foil (Honjo, thickness of 100 μm). The Li−S full cells were filled with 25, 50, and 70 μL of electrolyte. To evaluate electrochemical performance, a battery cycler (WBCS 3000, WonAtech) was used for galvanostatic discharge-charge measurements. The tests were cycled at 0.1 C (1 C = 1672 mA g−1) and 25 °C from 1.8 to 2.8 V (vs. Li+/Li). For the rate performance evaluation, various currents ranging from 0.1 C to 2 C were used to measure the rate performance in the potential range of 1.8−2.8 V (vs. Li+/Li). To study the Li stripping/plating process, the Li||Li symmetric cells were conducted at current densities of 0.5, 1.0, and 2.0 mA cm−2. The symmetric cells were filled with 70 μL of electrolyte using a lithium foil (Honjo, thickness of 200 μm). After assembly, all the cells were kept at 25 °C for 12 h to allow electrolyte penetration. The cyclic voltammetry (CV) measurements were performed in the potential range of 1.8−2.8 V (vs. Li+/Li) for full cells at a scanning rate of 0.1 mV s−1. All electrochemical measurements were conducted in a 25 °C thermostatic chamber.

Material characterization

The phase of the Li(CB11H12) was performed with XRD (SmartLab, Rigaku) measurements with Cu radiation (wavelength λ = 1.5406 Å for Kα1 and 1.5444 Å for Kα2) in the 2θ range of 5°−50°. For the powder XRD measurement, the sample was loaded into a thin-walled glass capillary under an Ar atmosphere and sealed with paraffin liquid. The bonding states of the electrolytes were characterized using Raman spectroscopy (LabRAM HR Evolution, Horiba). Nuclear Magnetic Resonance spectroscopy (NMR, JEOL−400 MHz spectrometer) was performed to confirm the enhanced coordination of Li⁺ by solvent molecules. NMR measurements were conducted in a coaxial NMR tube to avoid mixing with D2O81. The dissolution behavior of Li2S6 in DOL:DME was analyzed using UV-visible-near infrared (UV-Vis-NIR) absorption spectroscopy with a LAMBDA 950 spectrophotometer (PerkinElmer). Electrochemical impedance spectroscopy (EIS) measurements were performed in the 1 MHz−10 mHz frequency range at a voltage amplitude of 10 mV using a potentiostat (SP-300, Biologic). All electrochemical measurements were conducted in a 25 °C thermostatic chamber. The morphology and microstructure of the samples were examined using field-emission scanning electron microscopy (FE-SEM, GEMINI500, ZEISS) operated at 2 kV. The surface elemental composition of the S cathodes and the Li metal anodes before and after the plating process was confirmed using X-ray photoelectron spectroscopy (XPS). XPS measurements were performed using a Nexsa system (ThermoFisher Scientific) with a 100 µm X-ray spot size.