A Two-Dimensional Heterostructured Covalent Organic Framework/Graphene Composite for Stabilizing Lithium–Sulfur Batteries

Abstract

The implementation of a functional separator represents a highly encouraging approach to mitigating polysulfide shuttling in lithium–sulfur (Li–S). In this study, a two-dimensional (2D) 1,3,5-triformylphloroglucinol (Tp)-p-phenylenediamine (Pa) covalent organic framework/reduced graphene oxide (rGO) functional layer was introduced to enhance the performance of the commercial separator in Li–S batteries. The resulting 2D TpPa@rGO modified separators exhibit significantly improved electronic and ionic conductivity when compared to the unmodified separator, effectively mitigating lithium polysulfide shuttling and enhancing sulfur cathode utilization. It is indicated that a heterostructured composite of a nitrogen-group-containing COF and an electronic conductive addictive is an effective modification to the separator. Consequently, the modified cell demonstrated a minimal degradation rate of only 0.12% per cycle over 350 cycles at 0.5 C.

1. Introduction

Lithium–sulfur (Li–S) batteries not only offer a high theoretical specific capacity and energy density but also present advantages such as cost-effectiveness and environmental friendliness in their sulfur cathode [1,2]. Nevertheless, there remain several challenges that impede their widespread commercialization, including capacity degradation and severe self-discharge. These challenges encompass the low electronic/ionic conductivity of sulfur and its reaction product (Li₂S) and the loss of active material caused by the “shuttle effect” of the dissolved lithium polysulfide (Li₂S_x, 4 ≤ x ≤ 8) intermediates [3,4,5]. To tackle these aforementioned challenges, the incorporation of a functional interlayer between the sulfur cathode and the conventional commercial separator has demonstrated efficacy in mitigating the “shuttle effect” of polysulfide intermediates, consequently enhancing the cycling performance of Li–S batteries. A typical method is to modify the conventional separator with functional materials [6,7,8]. The coating materials should possess high electronic conductivity to facilitate electron transport. On the other hand, the surface area should be large enough to provide sites for the chemical adsorption of polysulfides. Furthermore, the materials should possess high stability so that they will not be dissolved in the electrolyte. Carbon, polymer, metal oxides, sulfide, etc., have been used [9,10]. However, each of these examples has its own advantages and considerations. For example, carbon possesses a high surface area and electronic conductivity. Nevertheless, the functional group on carbon is poor, which cannot provide a strong affinity to the polysulfides [11,12]. It is important to develop improved coating materials to address the limitations of Li–S batteries.

Covalent organic frameworks (COFs) are porous crystalline organic polymer materials known for their low density, substantial specific surface area, tunable pores, and versatile design and functionalization capabilities [13,14]. These features enable selective Li⁺ conduction while minimizing the adverse effects of polysulfide shuttling and have garnered increasing attention for modifying separators for Li–S batteries in recent years [15,16,17]. For example, by using 1,4-phenylenediamine-2-sulfonic acid (Pa-SO₃H) and 1,3,5-triformylphloroglucinol (Tp) as precursors, Cao et al. synthesized imide-based COF (TpPa-SO₃H) nanosheets through interfacial synthesis. They subsequently converted TpPa-SO₃H into TpPa-SO₃Li via ion exchange and produced TpPa-SO₃Li/Celgard separators through vacuum filtration. The TpPa-SO₃Li COF layer possesses aligned nano-channels and negatively charged sites, which effectively facilitate Li⁺ transport and significantly impede polysulfide shuttling [17]. Nevertheless, in the context of advanced separator coatings for Li–S batteries, it is crucial to consider achieving more precise control over the porosity and improved electron/ion conductivity of the COFs. The optimal coatings should possess an appropriate thickness and an extensive surface coverage while not compromising the overall energy density of the Li–S batteries [15]. Several methods have been reported for comparing few-layered COFs, such as interfacial polymerization, gas-phase polymerization, and chemical/physical exfoliation. However, developing few-layered COF materials for Li–S batteries is a challenge in terms of scale and controllable preparation.

In recent years, heterostructures have emerged as a significant development, leveraging the strengths of two-component materials to achieve a synergistic effect greater than the sum of their individual contributions [11,18]. Combining polymers with carbon materials, focusing on improving the adsorption ability and electrical conductivity, can synergistically improve the chemical confinement and polysulfide conversion for advanced performance. Based on this, constructing a heterostructure is an extremely effective method, which can thereby simultaneously realize the advantages of achieving fast redox reaction kinetics and long-term stability for Li–S batteries [19,20]. In this work, we prepared a two-dimensional (2D) heterostructured TpPa@rGO composite as the functional interlayer for Li–S batteries. The 2D heterostructured TpPa@rGO offers robust chemisorption sites for effective polysulfide capture. Additionally, rGO exhibits an outstanding electrical conductivity, enabling the efficient utilization of sulfur. As a result, Li–S cells with a TpPa@rGO interlayer exhibited high specific capacities and long-term cycling stabilities. This work may provide some new insights into the rational design of a novel heterostructured modification layer on a separator and provide new insights into developing and opportunities to develop advanced Li–S batteries.

2. Experimental Section

2.1. Material Synthesis

Synthesis of TpPa-COF: A total of 0.45 mmol p-phenylenediamine (Pa) and 500 mg 1,3,5-triformylphloroglucinol (Tp) was dispersed into 4 mL of deionized water using ultrasound. Then, 0.3 mmol Tp was added and shaken for 20 min. The mixture was transferred into an autoclave and kept in an oven at 150 °C for 48 h. The product was washed using water 3 times and subsequently with acetone using Soxhlet extraction and then dried overnight at 60 °C.

Synthesis of the TpPa@rGO composite: A total of 16.2 mg Pa and 166.6 mg PTSA was dispersed into deionized water using ultrasound. Then, 24 mL of GO solution at a concentration of 2 mg mL⁻¹ was added to the solution, and it was shaken for 5 min. Subsequently, 21 mg of Tp was added and the solution shaken for 20 min. After that, 24 mL of GO solution was added at a concentration of 2 mg mL⁻¹, and the solution was stirred for 30 min, transferred into an autoclave, and then kept in the oven at 150 °C for 48 h. The resulting hydrogel was washed sequentially using water, hot acetone, and water.

Preparation of the modified separator: A composite material consisting of carbon black (CB, 10 mg), TpPa@rGO material (80 mg), and polyvinylidene fluoride (PVDF, 10 mg) was finely milled for 30 min to achieve a uniform powder. This mixture was subsequently dispersed in N-methyl -2-pyrrolidinone (NMP, 1.0 mL) under vigorous agitation for 30 min, forming a stable slurry. The slurry was uniformly applied onto a Celgard separator using a doctor blade technique, followed by vacuum-drying at 50 °C for an extended period (overnight). The resulting modified separators were then die-cut into discs of a 19 mm diameter. Each disc had a weight of about 4.78 mg, while the blank separator in comparison had a weight of 3.02 mg. The functionalized separators demonstrated a thickness of 18 μm and featured a TpPa@rGO loading density of approximately 0.124 mg cm⁻².

Preparation of S/Super-P cathodes: Sublimed sulfur (67 wt.%), carbon black (Super-P, 25 wt.%), and PVDF (8 wt.%) were amalgamated in NMP to form a cohesive slurry. The slurry was then coated onto aluminum foil and subjected to vacuum-drying at 60 °C overnight. The sulfur content within the cathodes was meticulously adjusted to achieve a loading range of 1.3~2.0 mg cm⁻² by modulating the thickness of the applied coating.

2.2. Electrochemical Performance Measurements

Measurement of ionic conductivity: The ionic conductivity of the COF-enhanced separators was determined utilizing Electrochemical Impedance Spectroscopy (EIS). In this setup, each electrolyte-saturated separator was positioned between two stainless steel electrodes within the coin cells. The ionic conductivity was calculated using the formula:

σ = l/(R_b × A)

where σ represents the ionic conductivity (S cm⁻¹), l represents the separator’s thickness (cm), A is the electrode area (cm²), and R_b denotes bulk resistance (Ω).

Electrochemical performance measurement of the Li–S batteries: CR-2032-type batteries were assembled employing the prepared cathode, a lithium foil anode, and the COF-enhanced separator, all within an argon-filled glovebox. The electrolyte was composed of 1 M of LiTFSI dissolved in a 1:1 volumetric mixture of DOL and DME, enriched with 1 wt% LiNO₃. Each cell was allocated 80 μL of the electrolyte, catering to a sulfur loading range of 1.2~1.5 mg cm⁻². Galvanostatic charge–discharge cycles were conducted using a LAND CT2001 tester within a voltage window of 1.7–2.8 V (vs. Li⁺/Li), calculating the specific capacity based on the active sulfur content. Cells with various compartments were cycled five times at a current density of 0.1 C.

3. Results and Discussion

The TpPa@rGO composite was synthesized using a hydrothermal method employing 1,3,5-triformylphloroglucinol (Tp) and p-phenylenediamine (Pa) as the monomers, with p-toluenesulfonic acid (PTSA) serving as the catalyst. Following the mixing of the monomers with the GO solutions, hydrogen bonding facilitated the adsorption of the oligomers of the Tp and Pa monomers onto the surface of the GO nanosheets. Subsequently, during the hydrothermal process, the oligomers underwent further growth, resulting in the formation of a COF on the GO surface while simultaneously reducing the GO. After cooling to room temperature, washing, and freeze-drying, the heterostructured TpPa@rGO was obtained. Then, the TpPa@rGO was blended with polyvinylidene fluoride (PVDF) in NMP to form a cohesive slurry, which was spread onto the commercial separator (Celgard). The heterostructured TpPa@rGO possesses several structural advantages: (1) The numerous chemisorption sites of the TpPa COF effectively capture polysulfides; (2) The interlayer space efficiently reutilizes sulfur (Figure 1a). For comparison, TpPa-COF was prepared under the same procedures but without the GO suspension.

X-ray diffraction (XRD) was carried out to elucidate the structure of the TpPa@rGO composite. As depicted in Figure S1a, the XRD pattern of TpPa-COF exhibits a distinct peak at 2θ = 4.7° and a broader peak at 2θ = 27°, corresponding to the (100) reflection plane and the π-π stacking resulting from the (001) reflection plane, respectively [21]. In contrast, the XRD pattern of TpPa@rGO displays only a broad peak at 2θ = 27° (Figure S1b). This discrepancy arises from the fact that TpPa forms rather thin layers on the rGO sheets and from the low ratio of TpPa in the TpPa@rGO powder [22,23]. In addition, rGO reduces the crystallinity of TpPa@rGO materials. The Fourier transform infrared spectroscopy (FT-IR) spectra of TpPa-COF exhibited characteristic peaks at 1628, 1585, and 1260 cm⁻¹, corresponding to the C=O, C=C, and C-N groups, respectively (Figure S2a). These peaks suggest the formation of β-ketoenamine [24]. However, the presence of β-ketoenamine only indicates the combination of two monomers rather than the formation of the framework. TpPa@rGO composites show characteristic peaks in similar places as compared to TpPa. Meanwhile, the FT-IR also demonstrates the interaction between the TpPa and rGO [25]. The observations show that hydrogen bonds form between various groups on the rGO and COF during the synthesis process. The Raman spectra of the TpPa@rGO composites combine the D and G bands of rGO with the characteristic peaks of TpPa, indicating the formation of TpPa on the surface of the rGO (Figure S2b) [26]. The X-ray photoelectron spectroscopy (XPS) results confirm that TpPa@rGO primarily comprises C, N, and O elements (Figure S3). In addition, the amount of TpPa-COF loading in TpPa@rGO is calculated using elemental analysis (Table S1). On the basis of the amount of nitrogen, the TpPa-COF loading is 21.2%. The high-resolution N 1s XPS spectra (Figure 1b) are fitted to peaks at 400.05 and 401.82 eV, which correspond to the imine group (–NH–) and protonated secondary amine (–N⁺–), respectively [27,28]. The high-resolution C 1s XPS spectra are fitted to peaks at 284.8, 286.1, 287.5, and 290.9 eV, corresponding to C-C/C=C, C-N, C-O, and N-C=O (Figure 1c) [29,30]. The amine signals and N-C=O signals result from the condensation reaction between the –CHO and –NH₂ groups of the Tp and Pa monomers.

The morphological characterization of TpPa@rGO was analyzed using transmission electron microscopy (TEM) and scanning electron microscopy (SEM). The SEM images of TpPa@rGO reveal wrinkled nanosheets that form a porous structure (Figure 2a,b), showing a similar structure as compared to rGO (Figure S4). In contrast, SEM clearly depicts the fibrous structure of TpPa-COF (Figure S5). The TEM image indicates that no particles are observed on the surface of the nanosheets (Figure 2c). High-resolution TEM (HRTEM) revealed that the COF consisted of nanoscale crystalline regions (Figure 2d). The STEM image and elemental mapping demonstrate the uniform distribution of the C, N, and O elements (Figure 2e). These findings collectively affirm that the individual graphene sheets are fully covered with TpPa.

The modification process involved directly coating the Celgard polypropylene separator (Figure S6) with the slurry of TpPa@rGO. The SEM images show that TpPa@rGO uniformly covers the surface of the Celgard separator (Figure S7). In contrast, the TpPa@rGO modified separator displays a pore-free surface (Figure 3c,d), indicating that the original pores have been entirely covered by the TpPa@rGO. This suggests a robust physical barrier against polysulfides. The loading mass of the TpPa@rGO layer is 0.56 mg cm⁻² with a thickness of ~18 μm (Figure S8). The TpPa@rGO/Celgard membrane exhibits an excellent mechanical stability without any observable cracks after bending (Figure 3a). The consistency of the TpPa@rGO/the Celgard thickness were measured using SEM images (Figure 3b) and according to thickness data from thickness measurements using thickness gauges. The contact angle test showed that the TpPa@rGO coating improved the hydrophilicity (Figure S9), which is due to the abundance of polar functional groups on TpPa@rGO. The contact angle of the TpPa@rGO/Celgard membrane (9.23°) is significantly lower compared to that of the pure Celgard separator (34.16°). It is indicated that the wettability is improved by the coating of TpPa@rGO. The TpPa@rGO/Celgard has an ionic conductivity of 1.4 mS cm⁻¹, which is higher than the Celgard separator’s conductivity of 1.0 mS cm⁻¹. It is indicated that after coating with TpPa@rGO, the separator is more conducive to the migration of Li⁺ (Figure S10).

The performance of a TpPa@rGO/Celgard separator for Li–S batteries was evaluated using coin cells, pairing the Li anode and sulfur/carbon (S/C) cathode. SEM images show the surface of the sulfur/carbon (S/C) cathode (Figure S11). The performance was compared with that of a pristine Celgard separator. The EIS results for the Li–S cells with different separators before cycling measurement are compared in Figure 4a. The semicircle observed in the high-frequency area of the Nyquist plots represented the charge transfer resistance (R_ct), while the straight line in the low-frequency region corresponded to a mass transfer process [31]. Notably, the Li–S cell using a TpPa@rGO/Celgard separator demonstrated a lower R_ct value (28.9 Ω) compared to the Li–S cell using a Celgard separator (41.1 Ω). It is indicated that electron transfer is facilitated. The charge/discharge profiles at 0.1 C are shown in Figure 4b. The Li–S cells with a Celgard separator and a TpPa@rGO/Celgard separator show initial discharge capacities of 1021 and 1456 mAh g⁻¹, respectively. The capacities of the upper plateau (C_H) and lower plateau (C_L) are compared. The upper plateau in the discharge curve corresponds to the reduction reaction of S₈ into high-order polysulfides (Li₂S_x, x = 4–8) [32]. Notably, the CH of the Li–S cell using a TpPa@rGO/Celgard separator (426 mAh g⁻¹) is much higher than that of Li–S cells using a Celgard separator (317 mAh g⁻¹), indicating a higher conversion rate from S₈ into high-order polysulfides and higher utilization of the S cathode [33]. The CL corresponds to the reduction reaction of Li₂S₄ into solid Li₂S₂/Li₂S [34,35,36]. The higher CL obtained for the cell using a TpPa@rGO/Celgard separator indicates the highest conversion rate from Li₂S₄ into solid Li₂S₂/Li₂S. Furthermore, the Li–S cell using the TpPa@rGO separator shows the lowest polarization voltage (0.13 V) between the discharge and charge curves compared to the other cells, indicating the cell with a TpPa@rGO separator shows faster redox reaction kinetics and a better reversibility. The rate performances of the cells are compared in Figure 4c. The Li–S cell using a TpPa@rGO/Celgard separator provides capacities of 1018.9, 895.9, 740.6, and 615.8 mAh g⁻¹ at rates of 0.2, 0.5, 1, and 2 C, respectively. This performance surpasses that of the Li–S cell using the Celgard separator, which experiences a rapid decline with an increasing current density (Figure S12). When the current density is restored from 2 to 0.2 C, the specific capacities of the Li–S cells using TpPa@rGO/Celgard and Celgard separators are 956 and 882 mAh g⁻¹, respectively. After 150 cycles at 0.1 C (Figure 4d), the Li–S cell using the TpPa@rGO/Celgard separator maintains a specific capacity of 849.4 mAh g⁻¹. In contrast, the capacity of the Li–S cell using a Celgard separator exhibits rapid fades from 1128 to 472.4 mAh g⁻¹. The cycling performances of the cells at a current density of 0.5 C are further compared in Figure 4e. The capacity of the cell with the TpPa@rGO/Celgard separator remained at 663 mAh g⁻¹ after cycling for 350 cycles, significantly surpassing that of the cell with the Celgard separator. At 1 C, the Li–S cell with a TpPa@rGO/Celgard separator delivers an initial specific capacity of 961.3 mAh g⁻¹, and it retains 622.1 mAh g⁻¹ after 300 cycles. In contrast, the capacity of the Li–S cell with the Celgard separator rapidly diminished to 340 mAh g⁻¹ (Figure S13). In addition, compared with the pristine TpPa-COF interlayer [37] and the rGO interlayer (Figure S14), the heterostructured TpPa@rGO also demonstrated better performance. These findings underscore the positive effect of the heterostructured TpPa@rGO layer on improving the electrochemical performance of Li–S batteries, providing enhanced stability and capacity retention during cycling. This improved performance, on the one hand, is attributed to the inhibitory effect of the TpPa@rGO/Celgard separator on the shuttling of the polysulfides. Additionally, the TpPa@rGO acted as a secondary collector, improving the utilization of sulfur and significantly enhancing the capacity.

4. Conclusions

In conclusion, a 2D heterostructured TpPa@rGO composite is coated onto a commercial Celgard separator as a functional layer for Li–S batteries. The discharge capacity and cycling performance of Li–S batteries using a TpPa@rGO/Celgard separator are greatly enhanced compared to Li–S batteries using a conventional Celgard separator. In combination with structural characterization, their improved electrochemical performance can be attributed to the adsorption of polysulfide ions by TpPa@rGO, therefore inhibiting the shuttling effect for the polysulfides. Furthermore, heterostructures offer a complementary design approach for single materials with a strong electrical ability but weak adsorption performance or a good adsorption ability but weak electrical performance. rGO significantly improves the electronic conductivity of the separator layer, and the COF pore structure promotes Li⁺ transport. This work introduces a novel approach to designing high-energy-density and long-life Li–S batteries based on COF compartments.