INTRODUCTION
The rapid expansion of the electric vehicle market has significantly increased the demand for advanced battery technologies that can replace fossil fuels and reduce greenhouse gas emissions [1–3]. Among these, lithium-ion batteries (LIBs) have emerged as one of the most mature and widely utilized energy storage technologies [4–6]. However, critical limitations such as insufficient energy density and significant safety concerns impede their ability to meet future demands. Commercial LIBs depend on flammable liquid electrolytes, which represent substantial safety risks, including fire and explosion during overcharging or mechanical impact [7–10]. Furthermore, the uncontrolled growth of lithium dendrites can breach the separator, inducing short circuits and exacerbating safety concerns [11–14]. Addressing these challenges necessitates the development of next-generation batteries with enhanced energy density and improved safety.
All-solid-state lithium batteries (ASSLBs) have garnered significant attention as a promising alternative to conventional LIBs [15–19]. ASSLBs provide inherent safety, high energy density, and a broad operating temperature range, which renders them ideally suited for high-performance applications. The employment of solid electrolytes (SEs), which are non-flammable and thermally stable, addresses the safety concerns associated with liquid electrolytes [20–25]. Furthermore, their mechanical properties inhibit lithium dendrite growth, allowing the application of high-energy-density lithium metal anodes [26–32]. These characteristics establish ASSLBs as a transformative technology for future energy storage systems.
Among various SEs, sulfide-based solid electrolytes (SSEs) are distinguished by their superior lithium-ion conductivity and ductile mechanical properties, which enable processing at room temperature. Several SSEs exhibit ionic conductivities (~10–2 S/cm) that rival those of liquid electrolytes [19,33–38]. Nonetheless, a significant limitation of SSEs is their poor stability when exposed to moisture, resulting in hydrolysis and the emission of hazardous hydrogen sulfide (H2S) gas [16,19,39–42]. For example, in lithium argyrodites such as Li6PS5Cl (LPSCl), water attacks P−S bonds, leading to the formation of P−O bonds and the degradation of the PS43– units. This reduction in ionic conductivity necessitates synthesis and handling in moisture-free conditions, greatly elevating production costs and complicating commercialization efforts [43–46]. LPSCl is a promising solid electrolyte for all-solid-state lithium batteries (ASSLBs) due to its high room-temperature ionic conductivity (~10–3 S cm–1), which originates from its cubic argyrodite crystal structure with a highly disordered anion sublattice, enabling fast three-dimensional Li+ transport with low activation energy. Its relatively soft mechanical nature facilitates dense pellet formation and intimate solid–solid contact with electrode materials under moderate pressure, thereby reducing interfacial resistance. Furthermore, LPSCl exhibits low grain-boundary resistance, good compatibility with cathode materials, and compositional tunability, making it attractive for practical solid-state battery applications.
To address the atmospheric instability of SSEs, several strategies have been suggested. A prevalent method includes doping with metal oxides like ZnO, Fe2O3, or Bi2O3, which inhibit the generation of H2S gas by forming robust metal sulfides [47,48]. Another approach entails the partial substitution of sulfur or phosphorus with more stable elements, such as oxygen, leading to the formation of oxy-sulfide compounds. Such alterations enhance air stability and expand the electrochemical stability window [49–51]. Furthermore, according to the Hard-Soft Acid-Base (HSAB) theory, substituting hard acids like P5+ with soft acids such as As5+ or Sb5+ enhances bonding with S2–, thus diminishing the rate of hydrolysis. Nonetheless, while these strategies improve chemical stability, they frequently reduce ionic conductivity or interfacial stability with lithium metal, thereby limiting their overall effectiveness in ASSLBs [52,53].
In this study, we pioneered an oxy-sulfide-coated sulfide solid electrolyte on Li6PS5Cl (LPSCl) under benign conditions, including ambient temperature and pressure. This strategy markedly enhanced the moisture stability of the electrolyte, curtailed H2S gas production, and preserved high ionic conductivity. To confirm its practical viability, we constructed and tested all-solid-state batteries utilizing this modified electrolyte. The findings underscore the potential of this advanced electrolyte to surmount the stability hurdles of SSEs, thereby facilitating the development of more dependable and cost-efficient ASSLBs.
EXPERIMENTAL
Sample preparation
Initially, Li6PS5Cl (LPSCl) was synthesized via a mechanochemical method. Li2S (99.9%, Sigma-Aldrich), P2S5 (99%, Sigma-Aldrich), and LiCl (≥ 99%, Sigma-Aldrich) were blended using a mortar and pestle within an argon-purged glove box. The homogenized mixture was then transferred to an 80-ml zirconia jar containing both 3 mm and 5 mm zirconia balls (total 250 g) and subjected to reactive milling in a high-energy ball mill (Fritsch, Pulverisette 6) at 400 rpm for 15 hours. Subsequently, the mixture was re-ground for 15 minutes using a mortar and pestle, compacted under a pressure of 20 kN, and sintered at 550°C for 10 hours. The product was further refined by an additional grinding for 15 minutes to achieve the final form.
The synthesized LPSCl solid electrolyte was placed in a Schlenk flask within an argon-filled glove box and hermetically sealed. Oxygen gas was introduced at a constant pressure of 0.5 bar under ambient conditions to conduct the oxygen treatment on the solid electrolyte. The treatment was carried out for durations of 0, 1, 3, 6, and 12 hours (O2 x hours, where x = 0, 1, 3, 6, and 12), with continuous stirring using a magnetic bar throughout the process.
Characterization
To examine the crystal structure of the synthesized samples, powder X-ray diffraction (XRD) measurements were performed using a Rigaku-Ultima IV diffractometer equipped with Cu-Kα radiation (λ = 1.5418 Å) across a 2θ range of 10° to 80°, with a step size of 0.01 s–1. All samples were secured in air-sensitive sample holders within an argon-filled glove box to prevent moisture exposure. The morphology and distribution of the prepared solid electrolytes were investigated using field-emission scanning electron microscopy (FE-SEM, JEOL JSM-7610F) and energy-dispersive X-ray spectroscopy (EDS) at an accelerating voltage of 10 kV. Additionally, X-ray photoelectron spectroscopy (XPS, VG Scientific, ESCALAB 250) was employed to delineate the oxidation states of the elements within the solid electrolytes.
First principles calculations
Quantum ESPRESSO software was employed to investigate the material properties of the Li6PS5Cl crystal structure [54–57]. Ultra-soft pseudopotentials were utilized for Li, P, S, Cl, and O elements [58,59]. The initial structural of Li6PS5Cl was obtained from the Materials Project (MP-ID: mp-985592, https://next-gen.materialsproject.org/) [60]. The source code was further modified to account for oxygen incorporation during the computational treatment.
Electrochemical performance evaluation
The ionic conductivity of the synthesized solid electrolyte was determined using electrochemical impedance spectroscopy (EIS). Measurements were conducted at room temperature employing a Biologic SP-300, covering a frequency range from 7 MHz to 1 Hz at an amplitude of 50 mV. An In/SE/In symmetric cell configuration was utilized for these measurements. The cell comprised a mold with a diameter of 10 mm, containing a 0.05 mm indium foil layer, 250 mg of solid electrolyte, and an additional 0.05 mm indium foil layer. The assembly was compressed under a load of 20 kN to form a compact pellet. The ionic conductivity was calculated using the equation:
where L denotes the pellet thickness, Rtotal represents the total resistance, and S is the pellet’s surface area. EIS measurements were also carried out over a temperature range from 25°C to 100°C, and the activation energy (Ea) for conduction was computed using the Arrhenius equation:
here Ea signifies the activation energy for conduction, T denotes the absolute temperature, k symbolizes the Boltzmann constant, and σ₀ represents the pre-exponential factor.
To evaluate the electrochemical stability of the solid electrolyte, cyclic voltammetry (CV) measurements were conducted. A 2032-type coin cell was assembled, and CV scans were performed at a scan rate of 1 mV/s over a voltage range of 0–5 V (vs. Li/Li+). To ascertain the compatibility with lithium metal, DC polarization analysis was conducted. A symmetrical Li/SE/Li cell was assembled, and a constant current density of 0.1 mA/cm2 was applied at room temperature.
A 2032-type coin cell was assembled to evaluate the galvanostatic charge-discharge performance. Comprising a cathode composite (LiNbO₃-coated NCM811, solid electrolyte, and VGCF), a solid electrolyte layer, and a lithium metal anode, the cell was assembled in an argon-filled glove box. The LiNbO3 coating on the NCM811 surface was prepared according to a previously reported procedure [61]. For cell preparation, 200 mg of solid electrolyte was cold-pressed at 20 kN to form a pellet with a diameter of 16 mm. The cathode composite was uniformly applied onto the solid electrolyte pellet, and the assembly underwent another pressing. Indium foil and a 0.5T spacer were positioned on the cathode side and pressed at 60 kN. On the opposite side, lithium foil was coupled with a 0.5 T spacer as the anode, and the assembly was pressed at 10 kN. This complete coin cell was subjected to galvanostatic charge-discharge testing using a WonATech WBCS 3000 system across a voltage range of 2.5–4.25 V at a current rate of 0.1 C.
Air stability measurements
To investigate the reaction stability of the solid electrolyte under atmospheric moisture, the generation of H2S gas was monitored. The experiment was carried out in a glove box filled with dry air (77% N2, 23% O2) at 10% relative humidity. A 500 mg sample of solid electrolyte powder was positioned in a sealed desiccator with a volume of 32,940 cm3, and the generation of H2S gas was monitored using a Minimax-X4 sensor. Following exposure to air, the ionic conductivity of the solid electrolyte was measured, and XRD analysis was performed to assess structural changes.
RESULTS AND DISCUSSION
Structural analysis of the Li6PS5Cl solid electrolyte
To explore the structural changes in the Li6PS5Cl (LPSCl) solid electrolyte following oxygen treatment, X-ray diffraction (XRD) analysis was performed, revealing the results in Fig. 1 (a) displays the XRD patterns of LPSCl samples exposed to oxygen for various durations (x = 0, 1, 3, 6, and 12 hours). In the pristine sample (x = 0), no characteristic peaks corresponding to the starting materials (Li2S, P2S5, and LiCl) were observed, verifying the successful synthesis of the cubic argyrodite structure of LPSCl (ICSD 2592000) [62,63].
As Fig. 1(b) and (c) demonstrate, the XRD peak positions shifted slightly with prolonged oxygen exposure. Following 1 hour of exposure (x = 1), primary diffraction peaks shifted to lower angles, indicative of possible lattice expansion. However, after durations of 3 (x = 3) and 6 (x = 6) hours, the peaks reverted to their original configurations. After 12 hours of exposure (x = 12), a similar shift to lower angles was observed.
The observed shifts in peak positions can be interpreted as structural transformations within the lattice, consequent to the cyclical absorption and release of oxygen during the treatment process. Initially, during the first hour of exposure, the lattice possibly absorbed oxygen or underwent partial oxidation of certain components (e.g., PS43-), resulting in lattice expansion. Between 3 and 6 hours of exposure, the structure likely reverted to its original form as oxygen was released or redistributed. Prolonged exposure to 12 hours possibly led to further oxygen absorption or the formation of defects, subsequently causing more substantial lattice expansion [64,65].
It is significant to note that the XRD patterns did not display new peaks following the incorporation of oxygen. This suggests that oxygen did not induce a new phase within the LPSCl structure, but rather facilitated structural adjustments through the absorption of oxygen and reorganization of existing components. The resultant lattice expansion was likely due to localized chemical alterations (e.g., bond reorganization within PS43-) and internal stresses arising from oxygen integration [66].
To further understand the local crystal distortion, we simulated the Li6PS5Cl and oxygen-incorporated Li6PS5Cl solid electrolytes (SEs) using input data obtained from the Materials Project Database (https://next-gen.materialsproject.org/). Fig. 2 illustrates the crystal structures of both Li6PS5Cl (supercell 2×2×2), and oxygen-incorporated Li6PS5Cl (supercell 2×2×2) SEs, while the optimized crystal structures are provided in the Supporting Information (Videos S1 and S2). A total of 416 atoms were considered for the optimization of the Li192P32S160Cl32 structure (supercell 2×2×2), and 417 atoms were used for the optimization of the oxygen-incorporated Li192P32S160Cl32O structure (supercell 2×2×2). The optimized results reveal that oxygen incorporation causes slight distortions in the atomic positions of the crystal lattice. Moreover, the lattice parameter increased marginally from 20.559 Å to 21.299 Å upon oxygen incorporation, confirming that the expanded lattice facilitates faster Li-ion transport within the solid electrolyte.
Fig. S1 presents the O 1s, P 2p, and S 2p XPS spectra of LPSCl electrolytes subjected to oxygen treatment for different durations (x = 1, 3, 6, and 12 h). The S 2p spectra reveal the presence of S-O bonding, with characteristic peaks appearing at approximately 166-168 eV. Although the initial intensity of these peaks is low due to the trace concentration of oxysulfide species in the LPSCl electrolyte, the S-O peak intensity increases systematically with increasing oxygen treatment time [49,78]. Furthermore, density functional theory (DFT) calculations confirm that oxygen preferentially forms S-O bonds rather than P-O bonds. Based on these experimental and computational results, we confirm the successful formation of oxysulfide surface species in the prepared electrolytes.
The ionic conductivity of each solid electrolyte composition was quantitatively assessed using In/SE/In symmetric cells and electrochemical impedance spectroscopy (EIS) measurements. Fig. 3(a) presents the Nyquist plot and provides comprehensive data for the oxygen-treated Li6PS5Cl solid electrolyte, whereas Fig. 3(b) delineates alterations in ionic conductivity correlated with the duration of oxygen treatment (x = 0, 1, 3, 6, and 12 hours). Notably, samples treated with oxygen for 1 hour and 3 hours exhibited enhanced ionic conductivity relative to the pristine solid electrolyte, with the sample treated for 1 hour displaying the highest ionic conductivity measured at 4.15 mS/cm. Conversely, samples exposed for extended periods of 6 and 12 hours manifested substantial reductions in ionic conductivity.
The increase in ionic conductivity in the sample treated for 1 hour can be attributed to lattice expansion in the argyrodite structure, as evidenced by the XRD results, which likely facilitated improved lithium-ion pathways. For the sample treated for 3 hours, the ionic conductivity and XRD peak positions closely resembled those of the pristine electrolyte, implying that the lattice parameters reverted to their original state. Conversely, for samples treated for 6 and 12 hours, additional shifts in XRD peaks were noted, accompanied by a significant reduction in ionic conductivity. This suggests that prolonged oxygen treatment leads to oxygen accumulation within the solid electrolyte, obstructing lithium-ion conduction pathways despite the lattice expansion [67].
Furthermore, temperature-dependent ionic conductivity analysis was conducted on both the pristine sample and the 3-hour oxygen-treated sample (x = 3) to evaluate the activation energy for lithium-ion conduction. These results, displayed in Fig. 4, indicate that the activation energy for both samples remained constant at 0.316 eV, demonstrating that the 3-hour oxygen treatment did not appreciably alter the mobility of Li+ ions, given the similar ionic conductivities between the two samples.
The air stability test was performed in a dry air setting with 10% humidity at 25°C for 60 minutes. The experimental setup is depicted in Fig. 5. Fig. 6 illustrates the amount of H2S gas produced during the test.
The results reveal that the oxygen-treated samples generated substantially less H2S compared to the pristine sample. Particularly, the sample exposed to oxygen for 3 hours showed the most significant reduction in H2S generation, with a measurement of 3.49 cm³/g after 60 minutes - a decrease of about 66% compared to the 10.15 cm³/g measured for the pristine sample (Table 1). These results confirm that oxygen treatment markedly enhances the moisture resistance of the sulfide-based solid electrolytes [62,65].
Fig. S2 presents the H2S gas generation behavior of LPSCl electrolytes subjected to oxygen treatment for different durations (x = 0 and 3 h), along with Li6PS4.9O0.1Cl, under 30% relative humidity (RH) conditions. All samples were evaluated under identical 30% RH conditions to assess their moisture stability. As expected, the P2O5-doped LPSCl electrolyte exhibits superior moisture stability, as evidenced by lower H2S gas generation, compared with the electrolytes investigated in this study. Both the pristine and oxygen-treated LPSCl samples generated relatively high amount of H2S upon moisture exposure; however, the oxygen-treated sample consistently exhibits reduced H2S evolution compared to the pristine LPSCl electrolyte. The reduction indicates that oxygen treatment partially suppresses moisture-induced degradation, although its stabilizing effect remains inferior to that of P2O5 doping.
XRD analysis was conducted to explore the structural modifications post-air exposure (Fig. 7 (a)). The results verified the consistency of the crystal structure with the argyrodite phase. However, as demonstrated in Fig. 7 (b), impurity peaks corresponding to LiCl at approximately 35o were detected in all samples, except the one exposed to oxygen for 3 hours. This suggests that the decomposition of the solid electrolyte due to reaction with atmospheric moisture was effectively inhibited in the 3-hour oxygen-treated sample [44,68].
Further, EIS analysis confirmed that the sample treated with oxygen for 3 hours exhibited superior air stability. As indicated in Fig. 8 (a) and (b), EIS measurements following air exposure showed that the ionic conductivity of the oxygen-treated LPSCl was 1.03 mS/cm, significantly higher than 0.58 mS/cm observed in the pristine sample. This underscores the enhanced stability and performance of the oxygen-treated solid electrolyte.
To assess the electrochemical stability of oxygen-treated Li6PS5Cl, cyclic voltammetry (CV) experiments were performed over the voltage range of 0–5 V. Fig. 9 (a) and (b) display the CV results of the first and second cycles for both untreated Li6PS5Cl and Li6PS5Cl treated with oxygen for 3 hours. The cell configuration used for these experiments consisted of Li/solid electrolyte/solid electrolyte + VGCF.
In Fig. 9 (a), distinct oxidation and reduction peaks were observed during the first cycle due to electrochemical activation. The oxidation and reduction peaks appeared at 0.3 V, 2 V, 3.5 V during the first cycle and shifted to 0.3 V, 2.8 V, 4.2 V during the second cycle. In Fig. 9 (b), the peaks were observed at 0.3 V, 3.7 V, 1 V during the first cycle and at 0.3 V and 2.8 V during the second cycle. The observed reactions in the CV measurements are attributed to the redox processes of sulfur species (S → S2–) and phosphorus species (P5+ → P/Pn–), which are decomposition products of Li6PS5Cl. These reactions yield products such as PS43–, Li2S, and Li3P [45,69]. During oxidation, sulfur (S) and lithium phosphide (Li3P) are generated, while these products are reduced during discharge. An oxygen treatment stabilizes the surface of Li6PS5Cl, thereby suppressing the formation of decomposition products and enhancing its electrochemical reversibility.
To further investigate the electrochemical stability with lithium metal and assess the ability to suppress lithium dendrite formation, a Li/SE/Li symmetric cell was assembled, and long-term DC cycling tests were conducted at 25°C. Fig. 10 (a) and (b) depict the results of DC cycling performed at 0.1 mA/cm2 for 400 hours, revealing the cycling stability during the charging and discharging processes. The oxygen-treated Li6PS5Cl exhibited stable cycling for more than 400 hours without any short-circuiting events, whereas the untreated Li6PS5Cl encountered an internal short circuit within 30 hours. This demonstrates that the oxygen-treated electrolyte significantly improved cycling performance compared to the untreated sample.
The electrochemical performance of the all-solid-state battery was assessed through galvanostatic charge-discharge measurements after assembling the cell with a cathode composite, solid electrolyte, and lithium metal. The cathode composite comprised NCM811 particles coated with LiNbO₃, a solid electrolyte, and a conductive material (VGCF).
Fig. 11 (a) illustrates the first cycle charge-discharge profiles of cells prepared with bare Li6PS5Cl and oxygen-treated Li6PS5Cl (treated for 3 hours) within the voltage range of 2.5–4.25 V at a rate of 0.1 C and a temperature of 25°C. The electrolyte treated with oxygen exhibited a slightly higher discharge capacity (162.2 mAh/g) compared to the bare electrolyte (157.6 mAh/g). Fig. 11 (b) displays the long-term cycling stability and coulombic efficiency of both cells. The oxygen-treated Li6PS5Cl demonstrated a higher capacity retention rate of 88% after 50 cycles, compared to 86% for the bare Li6PS5Cl. This data suggests that oxygen treatment effectively preserves the stability of the electrolyte over extended cycling periods. Table 2 summarizes a comparison of the electrochemical performance of the synthesized electrolytes with previously reported results. The electrolytes developed in this study demonstrate competitive and reasonable performance relative to the literature [69–77]. The enhanced electrochemical performance of the oxygen-treated Li6PS5Cl is attributed to its increased air stability, reduced degradation during cycling, and mitigated side reactions at the cathode-electrolyte interface, which are crucial for the longevity and reliability of all-solid-state batteries.
Fig. S3 (a) presents the first-cycle galvanostatic charge-discharge profiles of cells assembled with oxygentreated Li6PS5Cl (3 h treatment) within the voltage windows of 2.0-3.6 V and 2.5-4.25 V at 0.5 C rate and 25°C. The cell capacity at higher C rates remains relatively low, indicating kinetic limitations under fast-charging conditions. Adjusting the operating potential using different anode configurations did not result in noticeable improvement in high-rate performance. Ongoing efforts are therefore focused on optimizing both cathode architecture and anode design to enhance interfacial kinetics and overall electrochemical performance at elevated C rates.
CONCLUSIONS
Oxygen-treated Li6PS5Cl (LPSCl) sulfide-based solid electrolytes were developed to enhance their moisture stability, and their electrochemical performance was evaluated in all-solid-state batteries (ASSLBs). The oxygen treatment was performed under ambient temperature and pressure, resulting in lattice expansion and stabilization of the PS43– structural units, thereby enhancing the air stability of the electrolyte. In the H2S gas generation test, the oxygen-treated LPSCl samples showed a reduction of up to 66% in gas emission. Electrochemical impedance spectroscopy (EIS) analysis confirmed that the samples treated for 1 and 3 hours exhibited enhanced ionic conductivity compared to the pristine sample. In galvanostatic charge-discharge testing, the oxygen-treated electrolyte demonstrated improved initial discharge capacity (162.2mAh/g) and capacity retention (88% after 50 cycles). Furthermore, DC cycling tests showed stable operation for over 400 hours without short-circuiting, highlighting enhanced interfacial stability with lithium metal. This study demonstrates that a simple oxygen treatment process can simultaneously improve the moisture stability and electrochemical performance of sulfide-based solid electrolytes, providing a promising pathway for the development of commercially viable ASSLBs. Future research will aim to optimize the oxygen treatment conditions and to explore compatibility with various materials.
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation
Supplement1
