Enhancing the catalytic conversion of polysulfides utilizing a covalent organic framework–carbon nanotube interlayer

Abstract

Lithium–sulfur (Li–S) batteries, characterized by their exceptionally high theoretical energy density of 2600 Wh kg⁻¹, encounter significant challenges related to polysulfide shuttling and slow redox kinetics. Covalent organic frameworks (COFs) have demonstrated potential in addressing these challenges; however, traditional synthesis methods are often hindered by inefficiencies and limitations in scalability. In this study, we introduce a triazine-based COF–carbon nanotube (CTF–CNT) composite separator, synthesized via a scalable vacuum-assisted strong-acid polymerization technique. The AA-stacked CTF structure, enriched with nitrogen-active sites, establishes an electrostatic catalytic field that effectively confines polysulfides and enhances their conversion kinetics. Coupled with the improved conductivity provided by CNTs, the composite separator exhibits dual functionality: (1) superior lithium-ion transport (t_Li⁺ = 0.60, σ_Li⁺ = 5.16 × 10⁻⁴ S cm⁻¹) and (2) efficient polysulfide adsorption through chemical-electrocatalytic coupling. Under practical conditions, with a sulfur loading of 5 mg cm⁻² and an electrolyte volume of 10 μL mg⁻¹, CTF–CNT cells achieve a capacity of 599 mA h g⁻¹ after 100 cycles at 0.5C, with minimal polarization (ΔE = 281 mV). In situ Raman spectroscopy indicates full reversibility of sulfur redox reactions, whereas symmetric cell experiments exhibit stable lithium plating and stripping over a duration of 1400 hours. This study introduces a scalable materials design framework for high-energy batteries, effectively addressing shuttle suppression, kinetic enhancement, and the inhibition of lithium dendrite formation.

---

Ruizheng Zhao

Dr Ruizheng Zhao is an Associate Professor at Zhengzhou University. She received her PhD degree in 2020 from Shandong University under the supervision of Prof. Longwei Yin, and subsequently worked as a postdoctoral fellow in the group of Prof. Dongyuan Zhao at Fudan University from 2021 to 2022. She has published 50+ articles with more than 4500 citations (H-index of 28). She was awarded the Talent Development Funding Project of Shanghai in 2021, and recognized as an Excellent Young Scientist by Henan Province in 2024, etc. Her research mainly focuses on the development of advanced energy storage and conversion devices, including aqueous batteries, secondary batteries, supercapacitors, and so on.

1. Introduction

Lithium–sulfur (Li–S) batteries have emerged as formidable candidates for next-generation high-energy-density battery systems due to their exceptional theoretical specific energy of 2600 Wh kg⁻¹, which significantly exceeds the energy density metrics of traditional lithium-ion battery technologies.^1,2 The primary advantages of Li–S batteries, such as the abundance of sulfur resources, low cost, environmental friendliness, and high theoretical specific energy, render them promising options for electric vehicles and large-scale energy storage applications. Nevertheless, the practical implementation of Li–S batteries is hindered by significant challenges, most notably the polysulfide shuttle effect, which leads to lithium anode corrosion, inefficient sulfur utilization, and rapid capacity degradation.^3–5 Additionally, the poor electrical conductivity and substantial volumetric expansion of sulfur cathodes negatively affect cyclic longevity and electrochemical stability.^6,7

In recent years, significant research efforts have been directed towards addressing these challenges through multifaceted strategies aimed at enhancing the performance of Li–S batteries. In the realm of cathode engineering, the integration of conductive carbon matrices,⁸ as well as metal oxides^9,10 and sulfides,^11,12 has been shown to improve sulfur conductivity while mitigating volumetric expansion. However, these advancements often fall short of fully resolving the persistent issue of the polysulfide shuttle phenomenon. In terms of separator modifications, researchers have investigated the application of functional coatings composed of oxides,^13,14 sulfides,^15,16 phosphides,^17,18 metal–organic frameworks (MOFs),^19,20 and Mxene^21,22 to inhibit polysulfide migration. Despite demonstrating partial efficacy, these modifications continue to face challenges related to electrochemical stability, cost-effectiveness of manufacturing, and scalability of processes. Moreover, the development of composite materials that integrate chemical adsorption with physical barrier mechanisms has attracted considerable attention. Achieving optimal architectures that concurrently enhance ion transport kinetics and effectively mitigate polysulfide diffusion remains a significant scientific challenge. Covalent Organic Frameworks (COFs), a class of highly ordered porous materials synthesized through the covalent bonding of organic molecular building blocks, have exhibited remarkable potential in the domains of energy storage and conversion.^23–25 Their unique advantages include: (1) hierarchically ordered nanochannels that enhance lithium-ion transport kinetics, (2) chemical stability afforded by robust covalent linkages, ensuring sustained cycling performance, (3) exceptionally high surface areas that provide abundant electroactive sites, and (4) precisely tunable structure–function relationships facilitated by molecular-level design. These superior attributes have positioned COFs at the forefront of research in Li–S battery applications, particularly in terms of suppressing polysulfide shuttle effects and improving lithium-ion diffusion efficiency. Nonetheless, conventional COF synthesis methodologies often necessitate extended reaction times, especially under elevated temperatures and pressures.^26,27 Prolonged processing durations not only impede production efficiency but also increase operational costs, with these constraints becoming particularly significant in contexts requiring large-scale manufacturing.^28,29 Therefore, the development of optimized synthetic protocols is a crucial research priority.^30,31

Building upon the aforementioned background, this study introduces an innovative composite separator (CTF–CNT) derived from a CTF. This framework is synthesized through strong acid polymerization under vacuum conditions, facilitating its potential for large-scale production. The separator is specifically engineered for application in Li–S batteries, aiming to mitigate the issues associated with polysulfide shuttle and slow lithium-ion transport. Diverging from traditional solvothermal synthesis methods, the triazine COF, characterized by an AA-stacked layered architecture, was developed via a two-step thermal treatment under strong acid conditions, as depicted in Fig. 1a. The triazine units, which are rich in nitrogen atoms, were strategically incorporated to establish an intrinsic electrostatic catalytic field, as illustrated schematically in Fig. 1b.

Fig. 1 (a) Preparation methods of CTF and CTF–CNT. (b) Electrostatic potential distribution of the triazine COF. (c) Schematic diagram of the expected facilitation of polysulfide catalytic conversion and suppression of dendrite growth by the triazine COF/CNT composite material. (d) XRD patterns of CTF–CNT-5, CTF–CNT-10, and CTF–CNT-20. (e) FT-IR spectra of CTF–CNT-5, CTF–CNT-10, and CTF–CNT-20. (f) Nitrogen adsorption–desorption isotherms of CTF–CNT-10 at 77 K.

Owing to its highly ordered porous architecture, remarkable chemical stability, and substantial specific surface area, the CTF–CNT composite separator proficiently sequesters polysulfides through physical confinement while concurrently facilitating swift lithium-ion transport, as illustrated in Fig. 1c. This dual functionality primarily arises from the COF's systematically arranged nanochannels and robust chemical affinity for polysulfides, which enable their retention in proximity to the cathode during charge–discharge cycles, thereby mitigating their diffusion toward the anode.

In comparison to the traditional PP separator, the triazine COF-coated composite separator demonstrates a markedly improved lithium-ion transference number (t_Li⁺ = 0.60) and lithium-ion conductivity (5.16 × 10⁻⁴ S cm⁻¹). These advancements are chiefly ascribed to the effective dissociation of ion pairs. The improved lithium transport mitigates concentration polarization at the lithium metal anode, thereby enhancing cycling stability.

2. Results and discussion

2.1 Material characterization

To examine the influence of CNT content, three composite structures were engineered by integrating 5 wt%, 10 wt%, and 20 wt% of CNTs into the CTF, relative to the mass of the monomer precursors. The successful synthesis of the resulting four COF-based materials was verified using powder X-ray diffraction (PXRD) and Fourier transform infrared spectroscopy (FT-IR).

The PXRD patterns indicated that the CTF frameworks exhibited high crystallinity and were devoid of detectable impurities. As illustrated in Fig. 1d, the PXRD profile of thermally treated CTF–CNT-10 maintained a pronounced diffraction peak at 7.29°, corresponding to the (100) plane, along with two additional weaker peaks at 12.61° and 14.38°, which can be attributed to the (110) and (200) planes, respectively. A broad peak observed at 26.46°, indexed to the (002) plane, suggests an interlayer distance of approximately 3.4 Å. These diffraction characteristics are consistent with an eclipsed AA stacking model, referred to as CTF-AA.

The structural transformation indicates that thermal treatment and the subsequent removal of residual CF₃SO₃H from the initial CTF-AB phase facilitated a successful transition in the stacking mode from a staggered AB arrangement to an ordered AA configuration.³² Additionally, scanning electron microscopy (SEM) analysis corroborated the successful development of a layered morphology in the synthesized materials, as illustrated in Fig. S1.†

As illustrated in the FT-IR spectra presented in Fig. 1e, the characteristic peak observed at 1501 cm⁻¹ is ascribed to the C=N stretching vibration, which signifies the presence of imine linkages. It is noteworthy that the intensity of this peak diminishes progressively with an increase in CNT content, implying a partial coverage or interaction of the nitrogen-rich functional groups by the CNTs. Consequently, to enhance the exposure of redox-active sites, it is advisable to select materials endowed with nitrogen-rich functional groups and an optimized CNT content to achieve superior performance.

X-ray photoelectron spectroscopy (XPS) was utilized to conduct a detailed analysis of the elemental composition and chemical states present on the material's surface, as illustrated in Fig. S2.† Calibration of the spectra was performed using the C 1s peak at 284.8 eV as a reference standard. The survey spectra verified the presence of C and N elements within the CTF-based materials. The high-resolution C 1s spectrum of CTF was deconvoluted into two primary components: a peak at 284.8 eV, which corresponds to C=C bonds in benzene rings, and a peak at 285.5 eV, which is attributed to C=N bonds within the triazine units. Similarly, the N 1s spectrum displays a peak at 398.8 eV, which is also associated with the C=N bonding environment.

In the context of CTF–CNT composites, the high-resolution C 1s spectrum reveals a peak at 284.6 eV, indicative of sp²-hybridized C–C bonds within the CNT backbone, alongside a peak at 285.7 eV corresponding to sp²-hybridized C=C bonds in aromatic rings. The presence of a distinct peak at 287.8 eV, attributed to C=N bonds, further substantiates the successful incorporation of the triazine framework. Concurrently, the N 1s spectrum exhibits a peak at 400.1 eV, which is ascribed to the cyano (–CN) groups derived from the 1,4-dicyanobenzene monomer utilized during synthesis. Collectively, these findings confirm the homogeneous integration of CTF and CNT components within the composite materials.

To examine the pore structure of the pristine COF and its composites with CNTs, nitrogen adsorption–desorption measurements were conducted at 77 K. As illustrated in Fig. S3a–d,† the COF exhibits a typical type I isotherm, indicative of a microporous structure. Similarly, both CTF–CNT-5 and CTF–CNT-10 retain microporous characteristics, as depicted in Fig. 1f. In contrast, CTF–CNT-20 demonstrates a markedly different behavior: its pore size extends up to 16.3 nm, and the isotherm transitions to type IV, indicative of a mesoporous structure. This transformation suggests that excessive incorporation of CNTs modifies the intrinsic pore structure of the material, likely due to the inherent mesoporosity of the CNTs. Furthermore, a progressive decrease in specific surface area was observed with increasing CNT content, which can be attributed to the partial occupation or obstruction of the COF's micropores by the CNTs.

In the context of Li–S batteries, the incorporation of sulfur into the CNT-based cathode framework results in a transformation of the nitrogen adsorption isotherm to a type V profile, as illustrated in Fig. S3e.† This modification is accompanied by a reduction in the average pore size to approximately 7 nm and a significant decrease in the specific surface area to 0.9 m² g⁻¹. These findings suggest that sulfur molecules effectively penetrate the internal pore structure and CNT channels during the melt-diffusion process.

During the transformation of sulfur into polysulfides, particularly with increasing sulfur content, the resultant polysulfide molecules exhibit increased size and solubility within the electrolyte. High-order polysulfides, such as Li₂S₈, are especially susceptible to dissolution and diffusion towards the lithium anode, which precipitates the well-documented shuttle effect and substantial performance degradation. Consequently, the judicious selection of CNT loading in the CTF–CNT composite is essential for maintaining structural integrity and optimizing pore confinement. This approach effectively suppresses the formation of large polysulfide species, thereby enhancing the electrochemical stability of Li–S batteries.³³

The electrochemical performance of Li–S batteries was comprehensively examined to assess the practical applicability of triazine-based CTF composite separators incorporating varying loadings of CNTs. The objective was to determine the optimal composition that would enhance battery performance.

2.2 Electrochemical performance of the Li–S batteries

Rate capability assessments were initially performed on batteries assembled with separators incorporating 0%, 5%, 10%, and 20% CNTs, in addition to the pristine polypropylene (PP) separator, as illustrated in Fig. 2a. Among these configurations, the separator containing 10% CNTs demonstrated superior rate performance. The corresponding batteries achieved the highest discharge capacities of 1123.8, 918.1, 797.3, 715.1, and 670.4 mA h g⁻¹ at current densities of 0.2, 0.5, 1.0, 2.0, and 3.0C, respectively. Furthermore, upon reverting the current rate to 0.5C following high-rate cycling, the capacity was restored to 827.0 mA h g⁻¹, indicating excellent rate reversibility and structural stability.

Fig. 2 (a) Rate performance of CTF, CTF–CNT-10, CTF–CNT-5, CTF–CNT-20, and unmodified PP separators. (b) Cycling experiments of CTF–CNT, CTF, and PP separators at 0.5C. (c) Voltage-capacity curves after 50 cycles at 0.5C for CTF–CNT, CTF, and PP separators. (d) Long-cycle performance of CTF–CNT-10, CTF, and PP separators at 1C. (e) CV curves of CTF–CNT-10, CTF, and PP separators. (f) Cycling test of CTF–CNT battery at high sulfur loading (5 mg cm⁻²) and electrolyte volume (10 μL mg⁻¹). (g) Comparison of electrochemical performance of CTF–CNT with other composite materials.^34–38

The exceptional rate performance of the CTF–CNT-10 separator can be attributed to its superior bidirectional catalytic activity in facilitating both the reduction of lithium polysulfides (LiPSs) and the oxidation of Li₂S. Furthermore, the CTF separator alone demonstrated superior performance compared to the CTF–CNT-20 variant, suggesting that excessive incorporation of CNTs may interfere with the intrinsic functionality of the C=N groups within the CTF structure. These functional groups, which are formed through strong acid-mediated trimerization, effectively suppress polysulfide shuttling and mitigate the loss of active sulfur species. This, in turn, contributes to improved capacity retention and enhanced charge/discharge performance.

In comparison to the voltage hysteresis values observed in batteries utilizing separators with 5% CNTs (158 mV), 20% CNTs (168 mV), pristine CTF without CNTs (161 mV), and bare PP (144 mV), the batteries incorporating the CTF–CNT separator with a 10% CNT content demonstrated the lowest voltage hysteresis at 130 mV. This suggests diminished polarization and enhanced reaction kinetics. These findings are further supported by the galvanostatic charge–discharge (GCD) profiles presented in Fig. S4.†

Through the integration of FT-IR, specific surface area measurements, and pore size distribution analyses, it has been determined that the triazine-based covalent organic framework with a 10% CNT loading exhibits a greater number of redox-active sites (C=N). This characteristic contributes to its superior rate performance relative to frameworks with 5% or 20% CNT content. Enhanced rate performance indicates that the batteries can sustain high energy output at elevated discharge currents without experiencing significant internal resistance or substantial capacity degradation. Consequently, the CTF–CNT/PP separator with 10% CNT incorporation was identified as the optimal composition for further investigation and was directly compared with pristine CTF and unmodified PP separators. This comparative analysis highlights the superior performance of the CTF–CNT/PP separator, emphasizing its potential as an innovative nitrogen-rich coating material for advanced Li–S battery separators.

As illustrated in Fig. 2b, the battery equipped with the CTF–CNT-10 modified separator demonstrates a substantial discharge capacity of 886 mA h g⁻¹ after 50 cycles at 0.5C, markedly surpassing the performance of batteries utilizing CTF–CNT-5, CTF–CNT-20, and pristine CTF separators. Furthermore, it displays a significantly reduced overpotential, as depicted in Fig. 2c, which suggests diminished polarization and enhanced redox kinetics.

The observed enhancement can be attributed to the plentiful redox-active C=N sites and the well-developed porous architecture of the CTF, which facilitate rapid diffusion of lithium ions and efficient electron transport. The integration of 10 wt% CNTs into the CTF matrix markedly improves electrical conductivity, as the interconnected CNT network provides effective pathways for charge transport, thereby reducing resistance and enhancing rate capability.^39,40 These synergistic effects enable the CTF–CNT-10 composite to achieve optimal electrochemical performance, underscoring the significance of selecting an appropriate CNT loading for Li–S batteries.

A comprehensive comparative analysis was subsequently conducted to evaluate the long-cycle performance of batteries utilizing CTF–CNT-10-coated separators in contrast to those with CTF-coated separators. As illustrated in Fig. 2d, the battery modified with CTF–CNT-10 retains a specific capacity of 583 mA h g⁻¹ after 300 cycles at 1C, thereby exhibiting superior performance relative to batteries with pristine CTF separators (534 mA h g⁻¹) and conventional PP separators (422 mA h g⁻¹). The enhanced cycling stability of the CTF–CNT-10 is evidenced by its ability to maintain higher capacity retention and reduce capacity degradation over prolonged charge–discharge cycles. Fig. 2e presents the cyclic voltammetry (CV) profile of CTF–CNT-10, which displays characteristic redox features, including one distinct oxidation peak and two well-defined reduction peaks. The reduction peak at 2.32 V (vs. Li/Li⁺) corresponds to the initial reduction step in which elemental sulfur is converted into soluble long-chain lithium polysulfides (Li₂S_x, 4 ≤ x ≤ 8). Subsequently, the second reduction peak observed at 2.05 V (vs. Li/Li⁺) corresponds to the further reduction of polysulfide intermediates into insoluble short-chain lithium sulfides (Li₂S₂/Li₂S). The oxidation peak at 2.33 V (vs. Li/Li⁺) is associated with the reverse conversion process, wherein Li₂S₂/Li₂S undergo oxidation to regenerate S₈via a multi-step electrochemical reaction pathway. In comparison to CTF-PP and PP, the CV curve of CTF–CNT-10 exhibits a higher current density and a reduced potential gap (ΔE = 281 mV), indicative of superior redox conversion kinetics. Moreover, under conditions of high sulfur loading (5 mg cm⁻²) and limited electrolyte availability (10 μL mg⁻¹), the battery maintains a specific capacity of 599 mA h g⁻¹ after 100 cycles at 0.5C, as illustrated in Fig. 2f. This finding underscores the significant enhancement in electrochemical performance achieved through the uniform hybridization of CTF and CNT in CTF–CNT-10. Relative to other composite materials documented in the literature, CTF–CNT-10 demonstrates superior electrochemical performance, as illustrated in Fig. 2g.

In conclusion, to establish the superiority of the composite materials, we initially performed an exhaustive analysis of their material characterizations. Following this, we conducted rate and cycling performance tests for screening purposes. The final results indicate that the battery assembled with the composite material containing 10 wt% CNTs in CTF demonstrates enhanced rate capability and cycling stability. This finding directly reflects the overall electrochemical performance and underscores the advantages of the composite material.⁴¹ Consequently, the subsequent investigation into lithium-ion transport mechanisms and the adsorption-catalysis performance of the separator primarily focuses on the comparison between CTF–CNT-10 and CTF. Henceforth, CTF–CNT will specifically refer to CTF–CNT-10. The objective is to further investigate the superior catalytic performance of the composite material in comparison to the non-composited CTF.

2.3 Lithium ion conduction and polysulfide adsorption catalysis in CTF–CNTs

To assess the polysulfide adsorption capacity of the materials, a visual adsorption test was performed utilizing UV-visible spectroscopy to analyze the interaction between various COFs and LiPSs. A 0.2 M Li₂S₆ electrolyte solution was employed for the adsorption experiments. Upon the introduction of CTF, the solution exhibited a slight fading in color. In contrast, the addition of CTF–CNT resulted in the Li₂S₆ solution rapidly becoming nearly colorless, suggesting that both CTF and CTF–CNT demonstrate a superior Li₂S₆ adsorption capacity compared to the unmodified PP separator. The UV-visible (UV-Vis) absorption spectra, as depicted in Fig. 3a, further corroborate the strong adsorption capability of CTF–CNT towards Li₂S₆.

Fig. 3 (a) UV absorption spectra of CTF–CNT and CTF in Li₂S₆-containing DOL/DME solution, with inset showing optical images of materials adsorbing Li₂S₆ solution after standing for 24 hours. (b) EIS of CTF–CNT, CTF, and PP separators. (c) GITT curve for CTF–CNT. (d) CV curves of CTF–CNT at scan rates of 0.1, 0.2, 0.3, and 0.4 mV s⁻¹. (e) Linear relationship between V^0.5 of peak A and current for CTF–CNT, CTF, and PP batteries. (f) Li⁺ diffusion coefficient calculated from the CV redox peaks according to the Randles–Sevcik equation. Potentiostatic nucleation curves of Li₂S with (g) CTF–CNT, (h) CTF, and (i) CNT.

Additionally, electrochemical impedance spectroscopy (EIS) was performed on Li–S batteries following cycling with various catalytic separators. The semicircle diameter in the Nyquist plot corresponds to the interfacial charge transfer resistance (R_ct) related to polysulfide conversion. As illustrated in Fig. 3b, the EIS data indicate that the battery utilizing the CTF–CNT-modified separator demonstrates the lowest impedance. This finding corroborates the stability of battery polarization throughout the charge–discharge cycles.

To enhance our understanding of the lithium-ion conduction mechanism, we conducted galvanostatic intermittent titration technique (GITT) experiments to acquire electrochemical reaction data for CTF–CNT and CTF, as depicted in Fig. 3c and S5.† By analyzing the discharge and charge profiles of both materials, the slope depth can be utilized to quantify the internal resistance during the nucleation and activation phases of Li₂S formation. This is reflected in the difference between the open-circuit voltage (QOCV) and the closed-circuit voltage (CCV), which signifies the voltage drop attributable to internal resistance, overvoltage effects, and other contributing factors. The internal resistance of the battery is denoted as ΔR_internal, where a larger voltage difference indicates a higher ΔR_internal, signifying increased internal resistance of the battery. The specific relationship is delineated as follows eqn (1):^42,43

ΔR_internal (Ω) = |ΔV_QOCV–CCV|/I_applied (1)

The CTF–CNT battery exhibits a voltage difference of 0.0226 V during the Li₂S nucleation process, which is notably lower than the 0.0432 V observed for the CTF battery, as illustrated in the accompanying figure. This observation suggests that the CTF–CNT separator possesses superior internal resistance relative to the CTF. The CTF–CNT effectively reduces the internal resistance of the electrode during the sulfur reduction reaction, thereby enhancing the overall efficiency of the electrochemical reaction. This improvement is primarily attributed to the CTF–CNT composite material's rich porous structure and larger interlayer spacing. These features facilitate rapid electron transfer and improve lithium-ion migration efficiency, thereby significantly enhancing the battery's overall performance.

To further investigate the fast reaction kinetics, the Li⁺ diffusion coefficient (D_Li⁺) in the primary reaction potential of the Li–S battery was measured. This was done by performing CV measurements at different scan rates and calculating D_Li⁺ using the Randles–Sevcik equation.⁴⁴

At the four principal reaction potentials (designated as peaks A–D), the diffusion coefficient of lithium ions (D_Li⁺) was observed to be higher when employing the CTF–CNT separator compared to the CTF separator. This suggests that the C=N sites enhance the dissociation of ion pairs, thereby facilitating rapid lithium diffusion. Notably, the CTF–CNT separator demonstrated superior D_Li⁺ values at peaks B and C relative to the CTF separator. These peaks are associated with solid–liquid phase reactions (specifically, peak B corresponds to the transformation from L₂S₄ to Li₂S, and peak C corresponds to the transformation from Li₂S to Li₂S₄), which are critical in determining the kinetics of the reaction during both discharge and charge processes.

This study investigates the performance of Li–S batteries utilizing PP, CTF, and CTF–CNT separators, which were evaluated through CV at 0.1, 0.2, 0.3, and 0.4 mV s⁻¹. The resulting oxidation-reduction peak positions and currents are depicted in Fig. 3d and S6a, b.† Notably, the CTF–CNT separator demonstrates pronounced oxidation-reduction peaks, indicative of accelerated electrochemical reaction kinetics. This behavior is primarily attributed to the separator's effective enhancement of lithium ion conductivity. Linear fitting of the current and voltage at distinct peaks was conducted, yielding fitting curves illustrated in Fig. 3e and S6c, d.† From these curves, the lithium ion diffusion coefficients for various reaction steps were derived based on the slopes.

The findings indicate that the Li–S battery incorporating the CTF–CNT separator exhibits the highest lithium-ion diffusion coefficient during both the two-step reduction reaction in discharge and the oxidation reaction during charge, measured at 1.75 × 10⁻⁷ cm² s⁻¹, as illustrated in Fig. 3f. This suggests that the separator possesses exceptional lithium-ion conductivity, thereby significantly enhancing the rapid progression of these electrochemical reactions.

To gain a more comprehensive understanding of the catalytic properties of the CTF–CNT material, kinetic measurements were conducted to assess its catalytic capabilities. The nucleation and decomposition processes of Li₂S involve phase transitions between liquid and solid states, as well as vice versa. These processes are characterized by the presence of significant energy barriers, which ultimately govern the rates of the discharge and charge reactions.^45,46 To more intuitively examine the conversion process from soluble LiPSs to solid Li₂S, constant potential discharge/charge tests were performed to investigate the nucleation process of Li₂S on various COF electrodes.⁴⁷ As illustrated in Fig. 3g–i, the Li–S battery, utilizing a 0.2 M Li₂S₈ electrolyte, underwent a discharge process at a constant current until reaching a voltage of 2.06 V. Subsequently, it discharged at a constant voltage of 2.05 V to facilitate the nucleation and growth of Li₂S. The analysis of the constant potential current–time (I–t) curve reveals that the nucleation peak current of the CTF–CNT electrode (i = 3.03 mA) surpasses that of the CTF (i = 2.07 mA) and CNT (i = 0.776 mA) electrodes. Furthermore, the specific capacity related to the sulfur content in the electrolyte for the Li₂S precipitate on the CTF–CNT electrode (317 mA h g⁻¹) markedly exceeds that of the CTF (269 mA h g⁻¹) and CNT (166 mA h g⁻¹) electrodes. Among the electrodes studied, the CTF–CNT electrode demonstrates the highest current peak and the most pronounced Li₂S nucleation peak. Collectively, these observations suggest that the CTF–CNT electrode exhibits superior catalytic activity and enhanced reaction kinetics during the Li₂S precipitation process. These findings align with the calculated Li₂S decomposition energy barrier, corroborating that the CTF–CNT electrode effectively facilitates Li₂S transformation in Li–S batteries, thereby confirming its optimal catalytic performance for Li₂S precipitation.

To facilitate a more intuitive observation of the catalytic conversion process of polysulfides by the composite material, in situ Raman spectroscopy was utilized to monitor the transformation of polysulfides under actual charging and discharging voltages.^48,49 The in situ Raman measurements were conducted using a specially designed sample battery, depicted in Fig. 4a. For the battery incorporating the CTF–CNT electrode, at the onset of the discharge process and within the initial discharge plateau (2.6 → 2.35 V), three prominent Raman peaks (approximately 155, 218, and 475 cm⁻¹) were detected, corresponding to the original S₈ molecules. As the discharge process advanced, the intensity of these S₈ peaks gradually diminished, and several new peaks appeared, as illustrated in Fig. 4b.

Fig. 4 (a) In situ Raman testing setup for reaction monitoring. (b) Charge–discharge curve of CTF–CNT battery with in situ time-resolved Raman spectroscopy and its contour plot. (c) Charge–discharge curve of CTF battery with in situ time-resolved Raman spectroscopy and its contour plot. (d) Schematic diagram of the catalytic mechanism for CTF. (e) Schematic diagram of the catalytic mechanism for CTF–CNT.

During the second discharge plateau (ranging from 2.35 to 1.9 V), five distinct peaks were observed at approximately 398 cm⁻¹ (S₆²⁻), 460 cm⁻¹ (S₆²⁻), 202 cm⁻¹ (S₄²⁻), 285 cm⁻¹ (S₄²⁻), and 455 cm⁻¹ (S₄²⁻). These observations suggest a transformation from long-chain polysulfides to short-chain polysulfides. At the conclusion of the discharge process (1.72 V), a prominent peak at approximately 452 cm⁻¹, indicative of S₂²⁻, emerged. These findings demonstrate the conversion of sulfur into Li₂S₄ on the CTF–CNT electrode surface.

During the charging process, Li₂S₂ was ultimately reconverted to S₈, as evidenced by the reemergence of peaks at approximately 155, 218, and 475 cm⁻¹. Upon completion of charging, the peaks corresponding to Li₂S₆ and Li₂S₄ were no longer detectable, indicating a complete and highly reversible redox process occurring at the interface between the CTF–CNT electrode and the electrolyte.

In contrast, the battery equipped with the CTF electrode exhibited a more gradual decline in the intensity of the S₈ peaks during discharge, with the emergence of LiPSs peaks occurring at a comparatively later stage, as illustrated in Fig. 4c. These observations suggest that the redox kinetics were slower than those observed with the CTF–CNT electrode. Furthermore, the absence of a distinct peak at approximately 452 cm⁻¹ throughout the charge–discharge process indicates incomplete reduction, which is associated with lower sulfur utilization and diminished reversibility. During the charging phase, LiPSs peaks remained prominent at the conclusion of charging, and the S₈ peaks emerged at a higher potential relative to the CTF–CNT electrode, signifying a reduced oxidation capability. This phenomenon is likely to contribute to accelerated capacity fading during cycling.

In summary, the CTF electrode demonstrated a sluggish and incomplete redox reaction process. In contrast, the incorporation of the CTF–CNT electrode markedly enhanced the catalytic conversion of polysulfides, as depicted in the schematic diagrams presented in Fig. 4d and e.

To achieve a more comprehensive understanding of the catalytic properties of CTF–CNT, kinetic measurements were performed to assess its catalytic performance. Symmetric batteries were constructed utilizing two identical electrodes and an electrolyte containing Li₂S₈ to examine the catalytic activity. As illustrated in Fig. S7,† the CTF–CNT electrode demonstrated more distinct redox peaks, reduced overpotentials, and enhanced current responses in comparison to CNT. These findings provide compelling evidence of the superior catalytic activity exhibited by the CTF–CNT material.

To conduct a more detailed analysis of the reaction kinetics associated with the rate-limiting step during the discharge process, the deposition behavior of Li₂S on various COF electrode states was examined using an electrolyte with a concentration of 0.1 M Li₂S₈. As illustrated in Fig. 5a, linear sweep voltammetry (LSV) identified the reduction peaks of Li₂S₈. The reduction peak for the CTF–CNT electrode was observed at 2.04 V, which is higher than those for CTF (2.02 V) and CNT (2.01 V). Moreover, the CTF–CNT electrode demonstrated the highest peak current density. These findings suggest that the CTF–CNT electrode facilitates a more efficient Li₂S₈ deposition process. Further analysis of the reduction peaks in the LSV curves, as depicted in Fig. 5b, revealed that the Tafel slopes corroborated the deposition kinetics. Specifically, the Tafel slope for the CTF–CNT electrode (50 mV dec⁻¹) was lower than that for CTF (87 mV dec⁻¹) and CNT (140 mV dec⁻¹), indicating improved Li₂S₈ deposition kinetics.

Fig. 5 (a) LSV curves of the Li₂S₈ reduction process. (b) Tafel plots of the Li₂S₈ reduction process. (c) Tafel plots of Li₂S oxidation process. (d) Calculated energy barriers for Li₂S decomposition. (e) Static contact angle tests of CTF–CNT, CTF, and PP separators in electrolyte. (f) Polysulfide permeation experiments of CTF–CNT, CTF, and PP separators in an H-type electrochemical cell. (g) Potentiostatic discharge curve of CTF–CNT at 10 mV; the inset shows interfacial charge-transfer resistance before and after reaching steady state. (h) Lithium-ion transference number (t_Li⁺) comparison among PP, CTF, and CTF–CNT separators. (i) EIS curves of stainless steel symmetric cells. (j) Separator thickness and ionic conductivity comparison among PP, CTF, and CTF–CNT.

Regarding the charging process, the efficient activation of Li₂S is essential for battery reversibility, as demonstrated in Fig. 5c. The CTF–CNT electrode displayed a reduced oxidation peak potential of 2.37 V, indicating favorable Li₂S activation. Additionally, its Tafel slope was measured at 157 mV dec⁻¹, which is lower than that of CTF (180 mV dec⁻¹) and CNT (202 mV dec⁻¹), signifying enhanced Li₂S oxidation kinetics, as illustrated in Fig. 5d. The study demonstrates that CTF–CNT significantly lowers the decomposition energy barrier of Li₂S and facilitates the activation process of Li₂S. Consequently, CTF–CNT acts as a catalyst in the reduction of LiPSs/Li₂S and the oxidation reaction, thereby enhancing the electrochemical performance of the battery.

Simultaneously, the CTF–CNT composite membrane effectively combines the ordered architecture of CTF with the electrical conductivity and electrolyte affinity inherent to CNTs. Within the electrolyte, CNTs are capable of interacting with solvent molecules via hydrogen bonding or other intermolecular forces, thereby enhancing the wettability of the composite membrane. This interaction promotes rapid infiltration and uniform distribution of the electrolyte across the membrane.^50,51 As demonstrated by the contact angle measurements presented in Fig. 5e, the CTF–CNT composite membrane exhibits significantly superior wettability in the electrolyte compared to the standalone PP and CTF membranes. This enhancement is primarily attributed to the nanoscale structure of the CNTs and the increased electrolyte affinity of their surfaces.

The polysulfide adsorption capability of the materials was assessed using a visual polysulfide permeation experiment. An H-type electrochemical battery was employed, with one chamber containing a polysulfide-rich electrolyte and the other a blank electrolyte, separated by a PP membrane coated with either CTF or CTF–CNT. This arrangement allowed for direct observation of polysulfide diffusion through the coatings.^52,53Fig. 5f demonstrates that the CTF–CNT coated membrane exhibited minimal Li₂S₆ shuttle behavior after 24 h, whereas the unmodified membrane and the CTF-coated membrane showed significant shuttle effects within the same timeframe. The findings suggest that the CTF–CNT coating demonstrates a more effective suppression of polysulfide migration than the CTF coating alone. This observation further implies that the incorporation of CNTs enhances the overall structural compactness of the material, thereby contributing to its improved capacity to inhibit polysulfide diffusion.

In addition to mitigating the shuttle effect of LiPSs, the strategy of modifying separators can effectively regulate ion transport behavior within the electrolyte. Specifically, for functionalized separators, the dissociation of functional groups under an electric field consistently enhances the transport of Li⁺ ions.^54–56 The lithium-ion transference number (t_Li⁺) and ionic conductivity are critical parameters that characterize ion transport behavior. To assess the lithium-ion transference number, the Bruce–Vincent–Evans equation is frequently employed, particularly in symmetric Li|Li battery systems. The equation is as follows eqn (3):

t_Li⁺ = I_S(ΔV − I₀R₀)/I₀(ΔV − I_SR_S) (3)

As illustrated in Fig. 5g and S8,† the interfacial resistance of the CTF–CNT membrane remains approximately 30 Ω both before and after chronoamperometry testing. The steady-state currents observed for the CTF–CNT, CTF, and PP membranes are 0.0293 mA, 0.07 mA, and 0.072 mA, respectively. Utilizing metallic lithium as symmetric electrodes, the measured current predominantly reflects the transport of lithium ions through the separator. Calculations depicted in Fig. 5h reveal that the lithium-ion transference numbers for the PP, CTF, and CTF–CNT separators are 0.19, 0.44, and 0.60, respectively. Notably, the CTF–CNT membrane demonstrates a substantially higher lithium-ion transference number (0.60) in comparison to the CTF membrane (0.44) and the pristine PP separator (0.19), underscoring its enhanced capability to facilitate lithium-ion migration. These findings suggest that the CTF–CNT separator effectively promotes ion-pair dissociation and accelerates lithium-ion transport. The integration of conductive CNTs into the CTF–CNT composite membrane significantly improves electrolyte wettability and establishes continuous ion transport pathways. The inclusion of CNTs results in a lithium-ion transference number for the CTF–CNT membrane that is generally superior to that of membranes composed solely of CTF or PP. The conductive network formed by the CNTs enhances the efficiency of lithium-ion migration within the electrolyte by minimizing interference from solvated ions and promoting accelerated charge carrier mobility.

EIS was employed to investigate the ionic conductivity, further elucidating the ion transport behavior of various COF-modified separators. As shown in Fig. 5i, j and S9,† the CTF–CNT separator exhibits the highest ionic conductivity (σ) of 5.16 × 10⁻⁴ S cm⁻¹, which surpasses that of the CTF separator (4.1 × 10⁻⁴ S cm⁻¹) and the pristine PP separator (3.12 × 10⁻⁴ S cm⁻¹). Among these, the lithium-ion conductivity (σ_Li⁺) serves as a more critical parameter, as it directly reflects the efficiency of Li⁺ transport within the battery system, thereby influencing the overall electrochemical performance. The CTF–CNT separator exhibits favorable Li⁺ transport characteristics by promoting the migration of lithium ions while inhibiting the movement of anions, and this enhancement is mainly due to the presence of the C=N site.^57–59 Furthermore, the AA-stacked porous channels inherent in the COF structure offer well-defined pathways that further expedite Li⁺ transport. This phenomenon is quantitatively characterized by eqn (4).

σ = L/(R_b × A) (4)

The results presented above indicate that CTF is capable of modulating lithium ion transport, which implies its potential to partially inhibit dendritic growth. The rate capability of the symmetrical battery equipped with the CTF–CNT separator was evaluated across a range of current densities (1 to 4 mA cm⁻²), as illustrated in Fig. 6a. In comparison to the conventional PP separator, the symmetrical battery with the CTF–CNT separator demonstrated a reduced overpotential during repeated plating–stripping cycles, particularly at elevated current densities exceeding 3 and 4 mA cm⁻². This improvement is attributed to the enhanced ion transport kinetics facilitated by the CTF–CNT separator, which optimizes the deposition behavior of lithium metal. At a current density of 0.5 mA cm⁻², the symmetrical battery incorporating the CTF–CNT separator exhibited a minimal voltage hysteresis of approximately 20 mV and maintained long-term cycle stability for over 1400 h, as depicted in Fig. 6b. Following cycling at elevated current densities, the symmetrical battery employing the CTF–CNT separator demonstrated stable performance over 1000 cycles upon reverting to a current density of 0.5 mA cm⁻², maintaining an overpotential of approximately 25 mV. Conversely, the symmetrical battery utilizing the original PP separator exhibited pronounced voltage fluctuations after merely 600 hours, indicative of detrimental dendrite formation, as illustrated in Fig. 6c. The lithium deposition morphology of the symmetrical battery, examined after 20 cycles at 0.5 mA cm⁻², was characterized using SEM, as depicted in Fig. 6d. The lithium deposition on the Li foil surface with the original PP separator revealed tortuous, needle-like formations, attributed to the development of blocky lithium dendrites. Conversely, the lithium metal extracted from the battery utilizing the CTF–CNT separator demonstrates a uniform, dense, and smooth morphology, devoid of any dendritic lithium formations. In conclusion, the CTF–CNT coated separator effectively mitigates polysulfide shuttling and facilitates uniform deposition of lithium dendrites on the lithium metal anode, as illustrated in Fig. 6e.

Fig. 6 (a) Rate performance of Li–Li symmetric cells using CTF–CNT coated separator and PP separator under current densities of 1, 2, 3, and 4 mA cm⁻². (b) Long-term cycling stability of Li–Li symmetric cells with CTF–CNT coated separator and PP separator at 0.5 mA cm⁻². (c) Cycling performance of Li–Li symmetric cells at 0.5 mA cm⁻² after cycling through 0.5, 1, 2, and 3 mA cm⁻² using CTF–CNT and PP separators. (d) Morphology of Li deposition after 20 cycles at 0.5 mA cm⁻² with CTF–CNT and PP separators. (e) Schematic illustration of the suppression of polysulfide shuttling and regulation of Li dendrite deposition by CTF–CNT coated separator.

3. Conclusion

The CTF–CNT composite separator, engineered through scalable vacuum-assisted strong-acid polymerization, addresses critical challenges in Li–S batteries by synergistically integrating long-cycle stability, high-rate capability, and efficient catalytic activity. The nitrogen-rich AA-stacked CTF framework provides electrostatic catalysis to suppress polysulfide shuttling, while the embedded CNT network ensures rapid electron/Li⁺ transport, yielding a high Li⁺ transference number (0.60) and ionic conductivity (5.16 × 10⁻⁴ S cm⁻¹). This design enables exceptional cycling performance (583 mA h g⁻¹ after 300 cycles at 1C) and outstanding rate retention (670 mA h g⁻¹ at 3C with 95% capacity recovery). Catalytic mechanisms, validated by in situ Raman spectroscopy and kinetic analyses, reveal accelerated Li₂S nucleation and decomposition (Tafel slope: 157 mV dec⁻¹), achieving complete sulfur redox reversibility. Under practical high-loading conditions (5 mg cm⁻²), the separator maintains 599 mA h g⁻¹ after 100 cycles, demonstrating industrial viability. This work establishes a paradigm for multifunctional separator design, resolving the trade-off between polysulfide confinement and ion transport in high-energy-density batteries.

4. Experimental section

4.1 Synthesis of CTF materials

The CTF material was synthesized via a two-step method.

Step I: Polymerization reaction: 1,4-Dicyanobenzene (DCB, 6 mmol, 1.536 g) and trifluoromethanesulfonic acid (CF₃SO₃H, 3 mmol, 0.45 g) were sequentially added into a Pyrex tube. The tube was flash-frozen in liquid nitrogen and then flame-sealed under vacuum. The sealed tube was subsequently heated in a muffle furnace at 250 °C for 12 h with a ramping rate of 5 °C min⁻¹. After the reaction, the tube was cooled to room temperature, re-immersed in liquid nitrogen for ∼10 min, and then carefully opened in a fume hood. The resulting solid product was collected and thoroughly washed with deionized water and ethanol to remove residual acid and unreacted monomers. The purified product was dried under vacuum at 60 °C for 12 h to yield CTF-1-AB as an orange powder in near-quantitative yield.

Step II: Thermal treatment: The CTF-1-AB powder obtained from Step I was transferred into a ceramic crucible and subjected to a second thermal treatment in a tubular furnace under flowing nitrogen (100 mL min⁻¹). The temperature was ramped to 350 °C at 5 °C min⁻¹ and held for 2 h. The final product, denoted as CTF-1-AA, was obtained as a green powder with a yield of 91%.³²

4.2 Synthesis of CTF–CNT composites

The CTF–CNT composites were synthesized using the same two-step protocol described above, except that multi-walled carbon nanotubes (MWCNTs) were incorporated into the reaction mixture during Step I. Specifically, MWCNTs were added in different proportions (5 wt%, 10 wt%, and 20 wt% relative to the DCB monomer mass) before the addition of CF₃SO₃H. The resulting composites were labeled accordingly based on CNT content.

Data availability

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

Conflicts of interest

All authors declare no conflict of interest.

Acknowledgements

This work was supported by the National Key Research and Development Program of China (no. 2022YFB3807501), the National Natural Science Foundation of China (no. 22441030, 42341204, 22309165), the Excellent Youth Foundation of Henan Province (no. 242300421126), the Talent Development Funding Project of Shanghai (no. 2021030), the Joint Fund of Science and Technology R&D Plan of Henan Province (no. 232301420053), and the Postdoctoral Science Foundation of China (no. 2023M743170), the Key Research Projects of Higher Education Institutions of Henan Province (no. 24A530010).

References

1. C. Zhao, Y. Liu, F. Huo, Z. Guo, Y. Lu, B. Sun, M. Li, H. Xu, M. Zhang, H. Fan, Z. Sun, A. Cabot and Y. Zhang, Angew. Chem., Int. Ed., 2025, 64, e202502177.
2. B. Sun, Z. Sun, Y. Yang, X. L. Huang, S. C. Jun, C. Zhao, J. Xue, S. Liu, H. K. Liu and S. X. Dou, ACS Nano, 2024, 18, 28–66.
3. Z. Sun, B. Sun, J. Xue, J. He, R. Zhao, Z. Chen, Z. Sun, H. K. Liu and S. X. Dou, Adv. Funct. Mater., 2025, 35, 2414671.
4. B. Sun, Z. Sun, Z. Sun, J. Xue, J. He, Y. Chen, Y. Yan, X. Wang, J. Wu, J. Wang, H. K. Liu and S. X. Dou, Adv. Funct. Mater., 2025, 2506294.
5. C. Zhao, F. Huo, Y. Yang, J. Ruan, F. Chai, H. Xu, Y. Liu, L. Zhang, A. Cabot, Z. Sun and Y. Zhang, Adv. Funct. Mater., 2024, 34, 2402175.
6. H. Shin, D. Kim, H. J. Kim, J. Kim, K. Char, C. T. Yavuz and J. W. Choi, Chem. Mater., 2019, 31, 7910–7921.
7. K. Yang, F. Zhao, J. Li, H. Yang, Y. Wang and Y. He, Adv. Funct. Mater., 2024, 34, 2410236.
8. H. H. Nersisyan, S. H. Joo, B. U. Yoo, D. Y. Kim, T. H. Lee, J.-Y. Eom, C. Kim, K. H. Lee and J.-H. Lee, Carbon, 2016, 103, 255–262.
9. Y. Yan, W. Jing, Y. Yang, C. Fan, M. Shi, B. Ren, Y. Zou, Z. Feng, C. You, Q. Deng and R. Yang, Int. J. Energy Res., 2022, 46, 8676–8686.
10. X. Wei, Y. Luo, X. Du, L. Wu, G. Liu and J. Li, Chem. Eng. J., 2023, 454, 140152.
11. X. Yan, Y. Li, X. Zhang, X. Zhao and B. Liu, Electrochim. Acta, 2023, 468, 143181.
12. Y. Zhao, Z. Quan, N. Xu, H. Zhang and Y. Chen, ACS Appl. Energy Mater., 2025, 8, 2360–2368.
13. Y.-C. Lin, Y.-H. Wu, J.-M. Ting and S.-H. Chung, Energy Fuels, 2023, 37, 15162–15169.
14. W. Li, S. Wang, Z. Fan, S. Li and N. Newman, Cellulose, 2023, 30, 481–493.
15. S. K. Kannan, H. Hareendrakrishnakumar, J. Joseph and M. G. Joseph, Energy Fuels, 2022, 36, 8460–8470.
16. X. Liu, J.-Q. Huang, Q. Zhang and L. Mai, Adv. Mater., 2017, 29, 1601759.
17. Y. Zhang, D. Wei, Y. Liu, S. Li, W. Lei, X. He and M. Qiao, ACS Appl. Nano Mater., 2024, 7, 7172–7183.
18. H. Liu, X. Wang, Q. Wang, C. Pei, H. Wang and S. Guo, J. Nanostruct. Chem., 2024, 14, 281–292.
19. S. Lin, J. Dong, R. Chen, G. Zhang, T. Huang, J. Li, H. Zhou, L.-H. Chung, X. Hu and J. He, J. Alloys Compd., 2023, 965, 171389.
20. R. Razaq, M. M. U. Din, D. R. Småbråten, V. Eyupoglu, S. Janakiram, T. O. Sunde, N. Allahgoli, D. Rettenwander and L. Deng, Adv. Energy Mater., 2024, 14, 2302897.
21. Y. Wang, R. Niu, L. Chen, Y. Yang, H. Yu and X. Qiu, Chem. Eng. J., 2025, 512, 162305.
22. L. Yin, G. Xu, P. Nie, H. Dou and X. Zhang, Chem. Eng. J., 2018, 352, 695–703.
23. S. Cao, J. Tan, L. Ma, Y. Liu, Q. He, W. Lu, Z. Liu, M. Ye and J. Shen, Energy Storage Mater., 2024, 66, 103232.
24. R. Lv, C. Luo, B. Liu, K. Hu, K. Wang, L. Zheng, Y. Guo, J. Du, L. Li, F. Wu and R. Chen, Adv. Mater., 2024, 36, 2400508.
25. A. B. Grommet, M. Feller and R. Klajn, Nat. Nanotechnol., 2020, 15, 256–271.
26. J. Qiu, P. Guan, Y. Zhao, Z. Li, H. Wang and J. Wang, Green Chem., 2020, 22, 7537–7542.
27. Q. Qing, J. Luo, S. Liu, J. Wang, Z. Wang, X.-G. Xiong, J. Chen and Y. Lu, Nat. Commun., 2025, 16, 2571.
28. Q. X. Shi, H. J. Pei, N. You, J. Wu, X. Xiang, Q. Xia, X. L. Xie, S. B. Jin and Y. S. Ye, Chem. Eng. J., 2019, 375, 121977.
29. Y. Zhao, A. U. Mu, Z. Guo, G. Cai and Z. Chen, Cell Rep. Phys. Sci., 2025, 6, 102496.
30. J. Xue, Z. Sun, B. Sun, C. Zhao, Y. Yang, F. Huo, A. Cabot, H. K. Liu and S. Dou, ACS Nano, 2024, 18, 17439–17468.
31. Y. Yang, B. Sun, Z. Sun, J. Xue, J. He, Z. Wang, K. Sun, Z. Sun, H. K. Liu and S. X. Dou, Coord. Chem. Rev., 2024, 510, 215836.
32. Z. Yang, H. Chen, S. Wang, W. Guo, T. Wang, X. Suo, D.-e. Jiang, X. Zhu, I. Popovs and S. Dai, J. Am. Chem. Soc., 2020, 142, 6856–6860.
33. T. Xia, Z. Wu, Y. Liang, W. Wang, Y. Li, X. Tian, L. Feng, Z. Sui and Q. Chen, J. Colloid Interface Sci., 2023, 645, 146–153.
34. L. Han, Y. Li, Y. Yang, S. Sun, M. Li, J. Yue, C. Y. Chuah and J. Li, J. Colloid Interface Sci., 2022, 628, 144–153.
35. J. Xu, W. Tang, C. Yang, I. Manke, N. Chen, F. Lai, T. Xu, S. An, H. Liu, Z. Zhang, Y. Cao, N. Wang, S. Zhao, D. Niu and R. Chen, ACS Energy Lett., 2021, 6, 3053–3062.
36. K. Zhang, X. Li, L. Ma, F. Chen, Z. Chen, Y. Yuan, Y. Zhao, J. Yang, J. Liu, K. Xie and K. P. Loh, ACS Nano, 2023, 17, 2901–2911.
37. R. Meng, Q. Deng, C. Peng, B. Chen, K. Liao, L. Li, Z. Yang, D. Yang, L. Zheng, C. Zhang and J. Yang, Nano Today, 2020, 35, 100991.
38. J. Wang, W. Qin, X. Zhu and Y. Teng, Energy, 2020, 199, 117372.
39. M. Li, G. Yan, P. Zou, H. Ji, H. Wang, Z. Hu, Z. Yang, Y. Feng, H. Ben and X. Zhang, ACS Sustain. Chem. Eng., 2022, 10, 13638–13649.
40. D. Han, L. Sun, Z. Li, W. Qin, L. Zhai, Y. Sun, S. Tang and Y. Fu, Energy Storage Mater., 2024, 65, 103143.
41. S. Xia, X. Xu, W. Wu, Y. Chen, L. Liu, G. Wang, L. Fu, Q. Zhang, T. Wang, J. He and Y. Wu, Mater. Sci. Eng., R, 2025, 163, 100924.
42. L. Peng, J. Han, Y. Cao, C. Geng, Z.-Z. Pan, H. Nishihara, Q.-H. Yang and W. Lv, Adv. Funct. Mater., 2024, 34, 2310508.
43. Y. Kong, X. Ao, X. Huang, J. Bai, S. Zhao, J. Zhang and B. Tian, Advanced Science, 2022, 9, 2105538.
44. Y. Zhu, J. Yang, X. Qiu, M. Li, G. He, Q. Wang, Z. Xie, X. Li and H. Yu, ACS Appl. Mater. Interfaces, 2021, 13, 60373–60383.
45. Y.-Q. Cai, Z.-T. Gong, Q. Rong, J.-M. Liu, L.-F. Yao, F.-X. Cheng, J.-J. Liu, S.-B. Xia and H. Guo, Appl. Surf. Sci., 2022, 594, 153481.
46. G. Huang, H. Yin, K. Jia, Y. Luo, J. Chen, C. Zhong, L. Zhu and F. Wu, Chem. Eng. J., 2024, 496, 154074.
47. A. Chen, J. Xue, J. He, B. Sun, Z. Sun, L. Sun and Z. Sun, J. Alloys Compd., 2024, 1005, 176146.
48. W. Chen, B. Fan, Q. He, W. Liu, Q. Wei, Z. Xu and Y. Zou, Chem. Eng. J., 2025, 508, 161099.
49. X. Zhou, Y. Wang, Y. Gu, J. Su, Y. Liu, Y. Yin, S. Yuan, J. Ma, Z. Jin and J.-L. Zuo, Matter, 2024, 7, 3069–3082.
50. Z. Li, W. Ji, T.-X. Wang, Y. Zhang, Z. Li, X. Ding, B.-H. Han and W. Feng, ACS Appl. Mater. Interfaces, 2021, 13, 22586–22596.
51. J. Zhao, G. Yan, X. Zhang, Y. Feng, N. Li, J. Shi and X. Qu, Chem. Eng. J., 2022, 442, 136352.
52. K. Sun, C. Wang, Y. Dong, P. Guo, P. Cheng, Y. Fu, D. Liu, D. He, S. Das and Y. Negishi, ACS Appl. Mater. Interfaces, 2022, 14, 4079–4090.
53. S. Gullace, M. Abruzzese, L. Cusin, G. Saleh, S. B. Thorat, A. Gamberini, S. Bellani, A. Ciesielski, F. Bonaccorso and P. Samorì, J. Mater. Chem. A, 2024, 12, 25359–25370.
54. B. Sun, B. Jia, H. Yu, C. Yao, W. Xie and Y. Xu, J. Power Sources, 2025, 632, 236325.
55. X. Hu, G. Zhang, H. Liang, J. Li, H. Zhou, L.-H. Chung and J. He, J. Mater. Chem. A, 2025, 13, 3392–3401.
56. S.-W. Ke, Y. Wang, J. Su, K. Liao, S. Lv, X. Song, T. Ma, S. Yuan, Z. Jin and J.-L. Zuo, J. Am. Chem. Soc., 2022, 144, 8267–8277.
57. T. Li, H. Hao, J. Wang, Z. Yuan, Z. Yu and H. Liu, J. Mater. Chem. A, 2025, 13, 7939–7948.
58. S. Lv, X. Ma, S. Ke, Y. Wang, T. Ma, S. Yuan, Z. Jin and J.-L. Zuo, J. Am. Chem. Soc., 2024, 146, 9385–9394.
59. Y. Sun, Y. Wu, Q. Zhang, C. Li, L. Wang and Q. Lv, J. Mater. Chem. A, 2025, 13, 12313–12320.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ta03984c

---