Mechanistic insights into electrolyte decomposition induced by polysulfide migration in large-scale Li–S battery cells

Abstract

Despite their high theoretical energy density, Li–S batteries suffer from polysulfide (PS) migration and electrolyte degradation. Here, we present one of the first integrated operando and ex situ studies in 18650-format Li–S cells, revealing that a sulfur@activated carbon cathode suppresses PS crossover, reduces gas evolution by 35.7-fold, and stabilizes the solid electrolyte interphase (SEI). Multi-technique analysis confirms that cathode architecture is crucial for enabling durable, and scalable Li–S batteries.

---

Despite their high theoretical capacity (1675 mA h g⁻¹) and energy density (2600 W h kg⁻¹), lithium–sulfur (Li–S) batteries face persistent challenges that hinder their practical application, most notably the dissolution and migration of polysulfide (PS) intermediates.^1,2 During discharge, elemental sulfur (S₈) undergoes stepwise reduction to soluble long-chain polysulfides (Li₂S_x, 4 ≤ x ≤ 8) around 2.3 V, followed by further reduction to insoluble short-chain species (Li₂S₂ and Li₂S) near 2.1 V.³ However, the reverse oxidation of these discharge products is incomplete, allowing unreacted PS species to accumulate in the electrolyte over repeated cycles. Driven by concentration gradients, these species migrate toward the lithium metal anode, where they are reduced and deposited, leading to uneven SEI formation.⁴ This ongoing process consumes electrolyte, destabilizes the interface, and generates degradation products—including gases and soluble organics—that contribute to internal pressure buildup, increased safety risks, and ultimately, premature cell failure.⁵ Our previous studies on 18650-format Li–S cells have linked poor cycling stability to a combination of PS shuttling, electrode corrosion, and severe electrolyte decomposition.⁶ To address these limitations, we present one of the first studies to integrate operando gas evolution analysis with ex situ chemical diagnostics in practical, full-format 18650-type jelly-roll Li–S batteries. By combining differential electrochemical mass spectrometry (DEMS), gas chromatography coupled with flame ionization and thermal conductivity detectors (GC-FID/TCD), nuclear magnetic resonance (NMR), and depth-resolved X-ray photoelectron spectroscopy (XPS), we systematically investigate how PS migration governs electrolyte degradation under realistic cycling conditions. Furthermore, we demonstrate that employing a sulfur@activated carbon (S@AC) composite cathode significantly suppresses polysulfide crossover, reduces gas evolution by up to 35.7-fold, and stabilizes the SEI on the lithium anode. These insights highlight the critical role of cathode architecture in controlling interfacial reactions and provide a practical design strategy for advancing durable, and scalable Li–S battery systems.

The cells were assembled with a conventional electrolyte formulation of 1 M LiTFSI in DOL/DME (1 : 1 v/v) with 0.1 M LiNO₃, and an electrolyte-to-sulfur ratio (E/S) of 15 μL mg⁻¹ to simulate practical configurations. The S@AC composite was synthesized via a straightforward melt-diffusion technique, in which elemental sulfur was infused into porous activated carbon at 155 °C for 6 h, as illustrated in Fig. S1 (ESI†).⁷ Thermogravimetric analysis (Fig. S2a, ESI†) revealed a distinct weight loss between 200 and 550 °C, corresponding to a sulfur content of 56.2 wt% in the composite. The successful infiltration of sulfur into the porous carbon framework was further confirmed by a substantial decrease in both specific surface area—from 2302.07 m² g⁻¹ to 1.73 m² g⁻¹—and total pore volume—from 1.204 cm³ g⁻¹ to 0.004 cm³ g⁻¹ (Fig. S2b, ESI†). Transmission electron microscopy (TEM) and energy-dispersive X-ray (EDX) analysis (Fig. S3, ESI†) demonstrated that the carbon framework retained its structural integrity while sulfur was uniformly distributed throughout the nanopore network (Fig. S4 and S5, ESI†). These confined sulfur domains within the porous carbon matrix promote intimate electrical contact and serve as physical barriers to polysulfide diffusion, thereby mitigating polysulfide migration during cycling.⁸ For cathode fabrication, the S@AC composite was homogenized with carbon black and polyvinylidene fluoride (PVDF) binder in a weight ratio of 7 : 2 : 1 using N-methyl-2-pyrrolidone (NMP) as the solvent. The resulting slurry was processed via roll-to-roll casting onto aluminum foil current collectors, as shown in Fig. S6 (ESI†). The fabricated electrodes achieved a sulfur loading of 3.68 mg cm⁻² with a total electrode thickness of 137 μm. For comparative analysis, control cathodes comprising bare sulfur (without the activated carbon host) were prepared using an identical formulation and processing method. Detailed specifications of both electrode types are provided in Table S1 (ESI†). The electrochemical performance of the Li–S batteries was evaluated within a voltage window of 1.6–3.0 V. Galvanostatic charge–discharge (GCD) profiles at the formation stage (0.05C, where 1C = 1675 mA h g⁻¹) are shown in Fig. 1a and b, where n referred to the 3 duplicated cell (Fig. S7 and Table S2, ESI†). Both the bare sulfur and S@AC cathodes exhibit the characteristic two-step discharge plateaus at approximately 2.3 V and 2.1 V, corresponding to the reduction of S₈ to soluble long-chain polysulfides (Li₂S₆–Li₂S₄) and subsequently to insoluble short-chain species (Li₂S₂ and Li₂S), respectively. The S@AC composite cathode delivers an initial discharge capacity of 1272.9 ± 15.8 mA h g⁻¹ and an initial coulombic efficiency (ICE) of 93.2%, which is significantly higher than the 825.8 ± 52.5 mA h g⁻¹ capacity and 82.3% ICE obtained from the bare sulfur cathode. This ∼54% improvement in capacity underscores the role of the activated carbon host in enhancing sulfur utilization and mitigating polysulfide dissolution.⁹

Fig. 1 GCD profiles of large-scale cylindrical Li–S batteries during formation cycles at 0.05C (1C = 1675 mA h g⁻¹) for (a) bare sulfur cathode and (b) S@AC composite cathode. (c) Capacity retention and (d) coulombic efficiency over 50 cycles at 0.1C, highlighting the improved cycling stability and suppressed polysulfide shuttle behavior in the S@AC-based cell.

To assess long-term cycling stability, the cells were further tested at a current density of 0.1C for 50 cycles. As shown in Fig. 1c, both cathodes exhibited capacity decay, reflecting the progressive loss of electrochemically active material due to polysulfide shuttling. After 50 cycles, the S@AC cathode retained 580.6 mA h g⁻¹ (70.0% capacity retention), whereas the bare sulfur cathode declined to 393.5 mA h g⁻¹ (60.6% retention). The coulombic efficiencies (CEs), plotted in Fig. 1d, also declined with cycling. In the first 15 cycles, the S@AC cathode showed the inferior CEs due to the presence of oxygen-containing functional groups on the activated carbon surface including hydroxyl, carbonyl, and carboxylic which are suggested by FTIR spectra shown in Fig. S8 (ESI†). These oxygen species can weaken the polysulfides adsorption towards the carbon host as well as reacts with the reactive lithium anode to form lithium oxide and lithium hydroxide on the passivation layer, resulting in the loss of active lithium and sulfur species and decrease of CEs.^10,11 However, the bare sulfur cathode exhibited a more pronounced drop and greater fluctuated CEs in the subsequent cycles, indicative of the more severe dissolution and migration of unconfined polysulfides.¹² These results confirm that the S@AC architecture effectively stabilizes the redox environment and prolongs battery lifespan by suppressing the progressive interfacial degradation. While a higher C-rate of 0.5C led to a significant capacity reduction in both bare sulfur and S@AC cathodes, primarily due to kinetic hindrance, their retention and coulombic efficiencies followed the same patterns seen at 0.1C (Fig. S9, ESI†). To investigate electrolyte decomposition induced by polysulfide migration, in situ gas evolution analysis was performed during the initial two formation cycles using DEMS. Fig. 2 presents the real-time partial pressure profiles of gaseous species as a function of cell voltage. Among the detected products, ethane (C₂H₆) emerged as the predominant gas evolved during cycling. Additionally, trace amounts of oxygen (O₂), carbon monoxide (CO), and carbon dioxide (CO₂) were observed. According to ab initio calculations by Chen et al.,¹³ ethane formation is attributed to the recombination of methyl radicals produced by bond cleavage in the DME solvent. In contrast, the oxygenated gas species are unlikely to originate from electrolyte degradation and may instead result from oxidative processes involving carbon black or activated carbon with oxygen-containing functional groups in the cathode structure or the reaction of lithium metal anode with some remaining moisture.¹⁴ Notably, the bare sulfur cathode exhibited more pronounced pressure spikes compared to the S@AC cathode, indicating significantly greater gas evolution. These results are consistent with the visual observations of bubble formation recorded during operation (Videos S1 and S2, ESI†), further supporting the role of polysulfide-induced electrolyte degradation in driving gas generation. To quantitatively assess electrolyte decomposition resulting from polysulfide migration, ex situ gas analysis was performed using GC-FID/TCD. Following both the formation (two cycles at 0.05C) and prolonged cycling (50 cycles at 0.1C), the gas accumulated in sealed glass tubes containing the Li–S cells was sampled and analyzed using argon as the carrier gas (Fig. S10, ESI†). The resulting chromatograms, shown in Fig. 3a and b, identified three primary hydrocarbon species—methane (CH₄), C₂H₆, and ethylene (C₂H₄)—with retention times of 1.8, 2.3, and 3.4 minutes, respectively. Notably, CH₄ and C₂H₄ were not detected in the in situ DEMS results, suggesting that these lighter gases may form via post-operando breakdown of ethane or delayed decomposition of solvent residues after electrochemical cycling. This hypothesis, while plausible, remains to be fully validated. These hydrocarbons are consistent with known degradation products of DOL and DME-based electrolytes.¹³ In addition to hydrocarbon gases, trace amounts of hydrogen (H₂) and CO₂ were detected by TCD. Previous studies have reported that H₂ may be generated from the reaction between the electrolyte and freshly exposed lithium metal before the establishment of a stable SEI.¹⁵ The source of the CO₂ gas was still inconclusive. The oxidation of the oxygen-containing species in cathode structure or reaction of lithium metal with the remaining moisture may be possible to produce this trace CO₂. Importantly, comparative quantification (Table S3, ESI†) reveals that the total gas evolution from the bare sulfur cathode is 7.6-fold higher after formation and 35.7-fold higher after extended cycling compared to the S@AC composite cathode. These findings underscore the substantial suppression of electrolyte decomposition in the S@AC-based system and further validate the role of polysulfide confinement in improving interfacial stability and safety in practical Li–S battery configurations. To further elucidate the chemical degradation pathways occurring in the liquid phase, post-formation and post-retention electrolytes were analyzed by NMR spectroscopy. The ¹H NMR spectrum of the fresh electrolyte revealed distinct peaks at 3.24 and 3.42 ppm corresponding to the DME solvent, and at 3.68 and 4.67 ppm corresponding to DOL (Fig. S11, ESI†). In the post-cycling electrolyte samples, new trace peaks emerged, attributed to unidentified decomposition byproducts. Notably, a reduction in the signal intensity of DOL—more pronounced than that of DME—was observed in both the ¹H NMR (Fig. S13, ESI†) and ¹³C NMR (Fig. S16, ESI†) spectra, suggesting that DOL is more susceptible to chemical degradation. This observation is consistent with previously reported findings indicating that DOL has a lower energy barrier for decomposition when activated by lithium cations or sulfur radical anions.¹⁶ In addition, a subtle upfield chemical shift was detected in the ¹H NMR spectra of the post-retention electrolyte, which is indicative of increased electron shielding caused by the presence of soluble electron-rich polysulfide species.¹⁷ This upfield shift was more pronounced in electrolytes from cells with the bare sulfur cathode, implying a higher concentration of dissolved polysulfides after extended cycling compared to the S@AC-based system. The solid-phase decomposition products and the composition of the passivation layer formed on the lithium metal anode were further analyzed via XPS with argon beam etching at five depth levels. Note that each argon etching level (500 eV for 30 seconds) is approximately 25 nm in depth relative to LiNbO₃ standard. The depth profiles (Fig. S17–S22, ESI†) revealed a stratified interphase structure. The outermost layers were dominated by organic components such as lithium formate (HCO₂Li; 530 eV O1s, 54 eV Li1s), while the inner regions primarily contained inorganic species such as lithium fluoride (LiF; 685 eV F1s) and lithium oxide (Li₂O; 528 eV O1s).^18–20 Additional peaks corresponding to decomposition products of LiTFSI and LiNO₃ were also observed, including SO₃²⁻, SO₄²⁻, and NSO₂CF₃ species (167–170 eV, S2p), as well as LiN_xO_y, Li₃N, C=N, LiNO₂, and residual LiNO₃ (397–408 eV, N1s).

Fig. 2 In situ gas evolution profiles of large-scale cylindrical Li–S batteries equipped with (a) bare sulfur cathode and (b) S@AC composite cathode during formation cycles at 0.05C (1C = 1675 mA h g⁻¹), measured using DEMS. The results highlight suppressed gas generation in the S@AC cell, indicative of reduced electrolyte decomposition.

Fig. 3 Ex situ gas analysis of large-scale cylindrical Li–S batteries with bare sulfur (black) and S@AC (red) cathodes conducted using GC-FID/TCD after (a) two formation cycles at 0.05C and (b) prolonged cycling for 50 cycles at 0.1C. The results reveal substantially reduced gas evolution in the S@AC cell, indicating mitigated electrolyte decomposition.

Furthermore, S2p spectra exhibited characteristic peaks at 161 and 163 eV associated with Li₂S and Li₂S_x species, particularly on lithium from cells with bare sulfur cathodes, indicating the deposition of polysulfide reduction products. Based on the integrated intensity of the Li1s signal (Table S4, ESI†), the passivation layer formed in the bare sulfur cathode cell was notably thicker than that formed with the S@AC composite, supporting the conclusion that uncontrolled polysulfide migration leads to more extensive electrolyte decomposition and interfacial instability.^21,22 Additionally, the UV-Vis spectroscopy measurements of the post-retention electrolytes revealed that there were 1.75 M and 0.38 M of the polysulfides dissolving from the bare sulfur and S@AC cathode, respectively, confirming the previous observations from NMR shift and XPS analysis (Fig. S23 and S24, ESI†).

In conclusion, this study provides mechanistic insight into how polysulfide migration drives electrolyte decomposition in large-scale cylindrical Li–S batteries. By employing a S@AC composite cathode, polysulfide dissolution and shuttling were effectively suppressed, resulting in enhanced sulfur utilization, improved capacity retention, and more stable coulombic efficiency compared to the bare sulfur cathode. Multi-technique analysis—including in situ DEMS, ex situ GC-FID/TCD, NMR spectroscopy, and XPS depth profiling—revealed that the bare sulfur cathode induced significantly greater gas evolution and liquid-phase degradation of the DOL solvent. The S@AC cathode reduced total gas generation by up to 35.7-fold after extended cycling and minimized the formation of a thick, heterogeneous passivation layer on the lithium anode. These findings underscore the critical role of cathode architecture in regulating polysulfide behavior and stabilizing the electrolyte–electrode interface. Additionally, the employment of other carbon hosts such as carbon nanotubes or graphene is the promising option to the cathode structure (Table S5, ESI†).^23,24 This work contributes to the fundamental understanding of degradation mechanisms in Li–S systems and offers a viable design strategy for developing safer and longer-lasting Li–S batteries for practical applications.

This work was financially supported by the Program Management Unit for National Competitiveness Enhancement (PMU-C), PTT Public Company Limited, and IRPC Public Company Limited; the Fundamental Fund from Thailand Science Research and Innovation (TSRI) and VISTEC (Grant No. FRB680014/0457); and the Energy Policy and Planning Office (EPPO), Ministry of Energy, Thailand. Additional support was provided by the Frontier Research Centre at VISTEC. N. B. gratefully acknowledges VISTEC for her scholarship.

Conflicts of interest

The authors declare that they have no known competing financial interest or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The data supporting this article have been included as part of the ESI.†

References

1. Q. Shao, Z.-S. Wu and J. Chen, Energy Storage Mater., 2019, 22, 284–310.
2. M. Sawangphruk, Chem. Commun., 2025, 61, 7770–7794.
3. J. Liu, Y. Zhou, T. Yan and X.-P. Gao, Adv. Funct. Mater., 2024, 34, 2309625.
4. F. Huang, S. Wang, Y. Jie, E. Hansen, S. Wang, Z. Lei, J. Liu, R. Cao, G. Zhang and S. Jiao, J. Energy Chem., 2020, 49, 257–261.
5. L. Huang, T. Lu, G. Xu, X. Zhang, Z. Jiang, Z. Zhang, Y. Wang, P. Han, G. Cui and L. Chen, Joule, 2022, 6, 906–922.
6. S. Kaenket, S. Duangdangchote, K. Homlamai, N. Joraleechanchai, T. Sangsanit, W. Tejangkura and M. Sawangphruk, Chem. Commun., 2023, 59, 10376–10379.
7. J. Wutthiprom, N. Phattharasupakun and M. Sawangphruk, Carbon, 2018, 139, 945–953.
8. X. Ji, K. T. Lee and L. F. Nazar, Nat. Mater., 2009, 8, 500–506.
9. W. S. Heo, W. Kwon, T. Lee, S. Chae, J. B. Park, M. Park, E. Jeong, J. H. Lee and S. G. Lee, Small Methods, 2024, 8, 2301401.
10. B. Li, M. Xie, G. Yi and C. Zhang, RSC Adv., 2020, 10, 2823–2829.
11. J. Tao, C. Zhang, X. Li, X. Chen, C. Ji, W. Wan and C. Wang, J. Mater. Chem. A, 2024, 12, 18127–18136.
12. S. Gu, R. Qian, J. Jin, Q. Wang, J. Guo, S. Zhang, S. Zhuo and Z. Wen, Phys. Chem. Chem. Phys., 2016, 18, 29293–29299.
13. Z.-X. Chen, M. Zhao, L.-P. Hou, X.-Q. Zhang, B.-Q. Li and J.-Q. Huang, Adv. Mater., 2022, 34, 2201555.
14. D. Cao, C. Tan and Y. Chen, Nat. Commun., 2022, 13, 4908.
15. A. Jozwiuk, B. B. Berkes, T. Weiß, H. Sommer, J. Janek and T. Brezesinski, Energy Environ. Sci., 2016, 9, 2603–2608.
16. X. Chen, T.-Z. Hou, B. Li, C. Yan, L. Zhu, C. Guan, X.-B. Cheng, H.-J. Peng, J.-Q. Huang and Q. Zhang, Energy Storage Mater., 2017, 8, 194–201.
17. A. Dorai, J. Kawamura and T. Omata, Electrochem. Commun., 2022, 141, 107360.
18. S. Jiao, X. Ren, R. Cao, M. H. Engelhard, Y. Liu, D. Hu, D. Mei, J. Zheng, W. Zhao, Q. Li, N. Liu, B. D. Adams, C. Ma, J. Liu, J.-G. Zhang and W. Xu, Nat. Energy, 2018, 3, 739–746.
19. Y. Wu, T. Momma, S. Ahn, T. Yokoshima, H. Nara and T. Osaka, J. Power Sources, 2017, 366, 65–71.
20. C. Zhang, Q. Lan, Y. Liu, J. Wu, H. Shao, H. Zhan and Y. Yang, Electrochim. Acta, 2019, 306, 407–419.
21. S. D. S. Fitch, G. E. Moehl, N. Meddings, S. Fop, S. Soulé, T.-L. Lee, M. Kazemian, N. Garcia-Araez and A. L. Hector, ACS Appl. Mater. Interfaces, 2023, 15, 39198–39210.
22. X. Liang, C. Hart, Q. Pang, A. Garsuch, T. Weiss and L. F. Nazar, Nat. Commun., 2015, 6, 5682.
23. Z. Tao, J. Xiao, Z. Yang and H. Wang, J. Nanomater., 2019, 2019, 6480236.
24. W. Xiao, S. Oh, T. V. M. Sreekanth, J. Kim and K. S. Yoo, ACS Appl. Mater. Interfaces, 2022, 14, 34802–34813.

---

Footnote

† Electronic supplementary information (ESI) available: Additional figures, tables, NMR/XPS spectra, electrode specifications, and gas evolution videos supporting synthesis, characterization, and mechanistic analysis of Li–S cells. See DOI: https://doi.org/10.1039/d5cc02606g

---