Mo-O-C Between MoS2 and Graphene Toward Accelerated Polysulfide Catalytic Conversion for Advanced Lithium-Sulfur Batteries.

MoS2 /C composites constructed with van der Waals forces have been extensively applied in lithium-sulfur (Li-S) batteries. However, the catalytic conversion effect on polysulfides is limited because the weak electronic interactions between the composite interfaces cannot fundamentally improve the intrinsic electronic conductivity of MoS2 . Herein, density functional theory calculations reveal that the MoS2 and nitrogen-doped carbon composite with an Mo-O-C bond can promote the catalytic conversion of polysulfides with a Gibbs free energy of only 0.19 eV and a low dissociation energy barrier of 0.48 eV, owing to the strong covalent coupling effect on the heterogeneous interface. Guided by theoretical calculations, a robust MoS2 strongly coupled with a 3D carbon matrix composed of nitrogen-doped reduced graphene oxide and carbonized melamine foam is designed and constructed as a multifunctional coating layer for lithium-sulfur batteries. As a result, excellent electrochemical performance is achieved for Li-S batteries, with a capacity of 615 mAh g-1 at 5 C, an areal capacity of 6.11 mAh cm-2 , and a low self-discharge of only 8.6% by resting for five days at 0.5 C. This study opens a new avenue for designing 2D material composites toward promoted catalytic conversion of polysulfides.

Introduction

In view of the rapid development of portable electronics and electric vehicles, there is an urgent need to develop high‐energy‐density and low‐cost energy storage systems. Lithium–sulfur (Li–S) batteries are among the most promising next‐generation battery systems because of their high theoretical capacity (1675 mAh g–1), high theoretical energy density (2600 Wh kg–1),[ ] and low cost (approximately 150 $ ton–1 of sulfur).[ ] Nevertheless, the commercialization of Li‐S batteries is limited by several challenges, such as the low utilization of sulfur, large volume expansion (≈80%),[ ] and “shuttle effect” caused by soluble lithium polysulfides (LiPSs), leading to rapid capacity fading, severe self‐discharge, low Coulombic efficiency, and poor cycling stability.[ ]

Based on theoretical calculations and experimental research, tremendous efforts have been made to mitigate the issues caused by the shuttle effect, including the construction of various nano‐structured sulfur host materials,[ , ] the design of advanced electrolytes,[ ] and the use of multifunctional separators.[ , ] Among these, functional separators have been considered a significant and efficient approach to impede the diffusion of LiPSs and improve the electrochemical performance of Li–S batteries. Owing to their high conductivity, large specific surface area, and lightweight,[ , , ] various types of carbonaceous materials have been used to mitigate the shuttle effect of LiPSs. Nevertheless, nonpolar carbon materials are usually limited in alleviating the shuttle effect of LiPSs because of their weak interactions with polar LiPSs.[ ] Therefore, several polar materials, such as, transitional metal oxides[ , , ] and sulfides[ , , ] which can adsorb LiPSs owing to their strong chemical interactions, have been applied as coated functional materials on separators. In particular, MoS2 materials have been shown to promote the catalytic conversion of LiPSs.[ ] Nonetheless, it is difficult to achieve highly effective bidirectional solid–liquid–solid conversion of sulfur–LiPSs–lithium sulfides with individual components. As a result, various composites have been used to address these issues. On the one hand, the construction of MoS2, to increase the exposure of active edge sites, combined with carbon materials is designed to boost the trapping and conversion of LiPSs in Li–S batteries.[ , ] On the other hand, researcher has focused on the micro/nanostructures of MoS2/C composites via van der Waals forces and utilize their synergistic effect to enhance the electrochemical performance of Li–S batteries.[ , , ] Although these MoS2/C composites can effectively suppress the shuttle effect of LiPSs, they are mainly bound through van der Waals bonds, and it is difficult to effectively exert the synergistic effect of the two components owing to the weak electronic pathway of the heterogeneous interface. Hence, constructing the MoS2/C composites with strongly coupled valence bonds and studying the intrinsic mechanism of LiPSs inhibition are important for the optimization of Li–S batteries.

In this study, we verified that the intrinsic electronic conductivity of MoS2 and the smooth electronic pathway of the heterogeneous interface of MoS2/C can significantly promote the transformation of LiPSs and decomposition of Li2S, respectively. An MoS2/C composite was constructed using robust MoS2 with abundant active sites covalently coupled with a 3D carbon matrix composed of nitrogen‐doped reduced graphene oxide and carbonized melamine foam (MoS2@CF‐NRGO). In this scheme, the strong interfacial connection of MoS2@CF‐NRGO ensured ultrafast electronic transfer and structural stability between MoS2 and the 3D carbon matrix. Because the composite possessed strong LiPSs chemical adsorption, rapid electronic conduction ability, and abundant catalytic sites, Li–S batteries with the MoS2@CF‐NRGO coated separator delivered excellent rate capability, cycling stability, favorable anti‐self‐discharge capacity, and high/low‐temperature performance. Even for an Li–S battery with high‐areal‐capacity sulfur loading (8.0 mg cm–2), a reversible areal capacity of 5.58 mAh cm–2 was still achieved after 50 cycles at 0.2 C with 91.3% capacity retention. More importantly, based on density functional theory (DFT) calculations and in situ Raman spectrum analyses, we found that the excellent performance originated from the pivotal role of MoS2@CF‐NRGO in the rapid anchoring‐diffusion transformation of LiPSs.

Results and Discussion

Theory‐Oriented Design

Combining theoretical calculations with experimental studies, we aimed to reveal the relationship between the intrinsic physicochemical properties of MoS2 and the anchoring‐diffusion transformation of LiPSs, explore the influence of the electron transport properties of the interface of MoS2 based composites on the dissociation kinetics of Li2S, and propose a well‐developed strategy for the MoS2/C composite toward promoted Li–S chemistry.

First‐principles calculations were performed based on DFT implemented by the Vienna Ab‐initio Simulation Package presented in Figure  and Figure S1–S5, Supporting Information. As shown in Figure 1a, pure MoS2 exhibited poor catalytic activity for the transformation of LiPSs owing to the slow reaction kinetics of the Li2S2*→Li2S* process with a high Gibbs free energy of 1.21 eV. However, pure MoS2 promoted the dissociation of Li2S, with a relatively low dissociation energy barrier of 0.79 eV. After pure MoS2 was coupled to nitrogen‐doped carbon with oxygen‐containing functional groups (NCO), the composite exhibited a lower Gibbs free energy (Figure 1c) than pure MoS2 and the NCO substrate (Figure 1b). Density of state (DOS) calculations indicated the enhanced catalytic activity of MoS2@NCO for the transformation of LiPSs originating from the enhanced electronic conductivity. As shown in Figure 1d, the energy gap of MoS2@NCO was narrower than that of pure MoS2. However, the dissociation energy barrier of Li2S increased, limiting the charging process as a result of the electrons being localized at the O atom on the interface of MoS2@NCO (Figure 1e), blocking the electron transport channel between the interface layers.

Figure 1

Energy profiles for the reduction of LiPSs on a) pure MoS2, b) NCO, c) MoS2@NCO, g) MoS2‐Edge, and h) MoS2‐Edge@NCO substrates. (Insets) The energy profiles of Li2S decomposition on corresponding substrates. d) Density of state (DOS) calculations for different models. Optimized geometries of electron distribution at the heterointerface of e) MoS2@NCO and f) MoS2‐Edge@NCO; yellow areas: charge accumulation; blue areas: charge depletion. The iso‐surface is set to 0.002 eV Å–3. i) The Gibbs free energy of the reduction of LiPSs and dissociation energy of Li2S for different models.

Therefore, an ideal MoS2/C composite for Li–S batteries could be constructed from two aspects: enhancing the electronic conductivity of MoS2 and ensuring a smooth electronic pathway of the heterogeneous interface. The DOS analysis in Figure 1d indicates that the MoS2 nanoribbon exposed numerous molybdenum atoms possessing a metallic nature because the energy band gap was zero. In addition, the electrons were concentrated near the Fermi level, which is beneficial for the anchoring and transformation of LiPSs. As shown in Figure 1g, the Gibbs free energy of the rate‐determining step was only 0.12 eV, indicating an excellent catalytic activity of the MoS2 nanoribbon for the transformation of the LiPSs. However, the dissociation of Li2S on the MoS2 nanoribbon was difficult because of the high dissociation energy barrier of Li2S (1.12 eV). As shown in Figure 1f, after coupling the MoS2 nanoribbon with NCO via the Mo–O bond (MoS2‐Edge@NCO), the composite enabled electrons to conduct smoothly from the MoS2 nanoribbon to NCO, leading to an electron‐deficient MoS2 surface, accelerating the dissociation kinetics of Li2S. As shown in Figure 1h, the MoS2‐Edge@NCO could not only easily catalyze the polysulfide conversion with a Gibbs free energy of only 0.19 eV, but also promote the dissociation of Li2S with a low dissociation energy barrier of 0.48 eV compared with other calculated samples shown in Figure 1i.

Synthesis and Characterization

Based on the above computational results, we designed the MoS2@CF‐NRGO composite and applied it to Li–S batteries as a modified layer of a polypropylene separator (MoS2@CF‐NRGO/PP) (Figure  ). First, GO was dispersed in a solution containing NH4HCO3 to obtain a GO/NH4 + composite via electrostatic self‐assembly. Subsequently, melamine foam (MF) was immersed in the above solution to prepare the MF‐GO/NH4 + composite. After annealing, a 3D carbon matrix composed of CF‐NRGO was fabricated. Second, the obtained 3D carbon matrix was added to a mixed solution of thiourea and sodium molybdate, and treated in a hydrothermal reaction with further calcination to acquire 3D porous MoS2@CF‐NRGO. MoS2 covalent coupling on the 3D carbon matrix via the Mo–O–C bond enhances the electronic conductivity between MoS2 and the 3D carbon matrix, thus improving the capability of MoS2 to catalyze LiPSs and dissociate Li2S. The morphologies of the samples were characterized using scanning electron microscopy (SEM), transmission electron microscopy (TEM), and high‐resolution TEM. As depicted in Figure 2b, CF with smooth skeletons exhibited a 3D framework architecture, which could be used as an accommodation or support system. After annealing the MF‐GO/NH4 +, the CF‐NRGO was fabricated and inherited the 3D framework architecture of CF (Figure S6a—c, Supporting Information), in which the NRGO with continuous and wrinkled nanosheet structures was uniformly dispersed in CF, providing numerous active sites for the uniform growth of MoS2. Figure 2c,d shows the SEM images of MoS2@CF‐NRGO at various magnifications. The composite maintained a 3D carbon matrix structure, which provided favorable conditions for Li+ shuttling and increased the effective infiltration of LiPSs. Moreover, the surface of the NRGO exhibited uniform growth of MoS2 without agglomeration, enabling the hybrids to possess more edge sites of MoS2, thus increasing the adsorption ability and catalytic activity of LiPSs. To further observe the detailed structures of MoS2@CF‐NRGO, Figure 2e and Figure S7, Supporting Information demonstrate that the MoS2 with nanoflower‐like morphology was composed of several lamellar nanosheets and strongly coupled on the surface of the NRGO. These results indicate that MoS2@CF‐NRGO provides a large number of adsorption and catalytic sites for the LiPSs conversion reaction and that internal synergetic effects increase the anchoring‐diffusion transformation process of LiPSs. In addition, the interlayer spacing of MoS2 was approximately 0.68 nm, corresponding to the (002) plane of MoS2 (Figure 2f,g). Meanwhile, elemental mapping (Figure 2h) not only further revealed the existence of Mo and S, but also confirmed the homogeneous distribution of N on the NRGO.

Figure 2

a) Schematic illustration of the preparation stages for MoS2@CF‐NRGO. SEM images of b) carbonized MF and c,d) MoS2@CF‐NRGO. e) Transmission electron microscopy (TEM) and f,g) HR‐TEM images of MoS2@CF‐NRGO. h) EDS mapping images of MoS2@CF‐NRGO and the corresponding images of Mo, S, and N.

The crystal phase compositions of the as‐prepared samples were identified using X‐ray diffraction (XRD) (Figure  ). From the CF‐NRGO pattern, one strong diffraction peak at approximately 26° was observed, suggesting that GO had been reduced. However, there are residual oxygen‐containing functional groups on RGO, as shown in the Fourier transform infrared spectroscopy (FTIR) pattern in Figure S8, Supporting Information,[ ] which are beneficial for the nucleation and growth of MoS2. As a comparison, the characteristic diffraction peak of the graphitic structure in MoS2@CF‐NRGO shifted slightly to a higher diffraction angle, indicating a slight increase in lattice distortion. This may be derived from the strong interaction between MoS2 and CF‐NRGO, which can be further proved by the XRD peaks of CF after hydrothermal treatment (Figure S9, Supporting Information). In addition, the characteristic diffraction peak corresponding to the (002) plane of MoS2 located at 13° was detected within MoS2@CF‐NRGO, attributed to the growth of MoS2 along the (002) crystal plane. Figure 3b displays the Raman spectra of the samples, where two typical peaks are denoted as graphitic carbon (G, at 1600 cm–1) and disordered carbon (D, at 1348 cm–1).[ ] Compared to the peak intensity ratio (I D/G = 1.00) of CF‐NRGO, the ratio of MoS2@CF‐NRGO increased to 1.09, implying strong electronic coupling between MoS2 and CF‐NRGO, which is consistent with the XRD results. In addition, MoS2@CF‐NRGO exhibited two Raman bands at 380 and 406 cm–1, corresponding to the 1 E 2g (in‐plane optical vibration of the Mo–S bond in opposite directions) and A 1g (out‐of‐plane optical vibration of S atoms) active modes, respectively.[ ] Most importantly, the intensity value of A 1g to 1 E 2g in MoS2@CF‐NRGO (2.02) was significantly higher than that in MoS2 (1.44), as shown in Figure S10b, Supporting Information, demonstrating the presence of more active edge sites of MoS2 in MoS2@CF‐NRGO.[ ]

Figure 3

a) X‐ray diffraction (XRD) patterns and b) Raman spectra of MoS2@CF‐NRGO, CF‐NRGO and GO. c) XPS spectra for O 1s of the MoS2@CF‐NRGO, and the comparison of d) Mo 3d, and e) S 2p of the MoS2@CF‐NRGO and MoS2@hydrothermal. f) FTIR spectra of the MoS2@CF‐NRGO and CF‐NRGO. g) Polysulfide adsorption test for MoS2@CF‐NRGO and CF‐NRGO and h) X‐ray photoelectron spectroscopy (XPS) spectra of Mo 3d of the MoS2@CF‐NRGO before and after adsorption with Li2S6. i) Charge curves at 2.4 V of a Li2S8/tetraglyme solution on different the surfaces of MoS2@CF‐NRGO, CF‐NRGO, and MoS2‐CF‐NRGO electrodes.

X‐ray photoelectron spectroscopy (XPS) analysis of MoS2@CF‐NRGO was performed to characterize the chemical bonding states. O 1s spectrum was deconvoluted into three peaks centered at 530.78, 531.98, and 533.12 eV (Figure 3c) for MoS2@CF‐NRGO, which can be associated with C≐O, C–O—Mo, and C–OH, respectively[ ] demonstrating the formation of C–O–Mo bonds in MoS2@CF‐NRGO. Moreover, the Mo 3d spectrum of MoS2@CF‐NRGO (Figure 3d) exhibited five fitted characteristic peaks at 232.5 eV (Mo 3d3/2 of Mo4+), 229.4 eV (Mo 3d5/2 of Mo4+), 235.9 eV (Mo 3d3/2 of Mo6+), 233 eV (Mo 3d5/2 of Mo6+), and 226.6 eV (assigned to the S 2s of divalent sulfide ions),[ ] and the S 2s spectrum of MoS2@CF‐NRGO (Figure 3e) presented three fitted characteristic peaks located at 169.2, 163.4, and 162.2 eV, corresponding to SO3 2–/SO4 2–, S2‐, and S2 2–, respectively. Compared to the high‐resolution Mo 3d and S 2s spectra of pure MoS2, the binding energies of MoS2@CF‐NRGO shifted toward high‐energy direction, indicating a change in the electronic structure and electron density loss from MoS2.[ ] This is consistent with the results of the theoretical calculations. Meanwhile, the FTIR spectra showed that the ratio of C‐O in MoS2@CF‐NRGO was larger than that in CF‐NRGO (Figure 3f), further implying the formation C–O–Mo bonds in MoS2@CF‐NRGO.[ ] Subsequently, a 23.3% MoS2 content was determined in MoS2@CF‐NRGO using thermogravimetric analysis (Figure S11a, Supporting Information). The specific surface area and permanent porosity were examined using N2 adsorption/desorption (Figure S11b, Supporting Information). Although the specific surface area of MoS2@CF‐NRGO was smaller than that of CF‐NRGO owing to the introduction of MoS2, the diameter of the holes was concentrated in the range of less than 10 nm. Therefore, MoS2@CF‐NRGO had sufficient sites for the adsorption and catalytic conversion of LiPSs but did not affect Li+ transport.

Polysulfide Adsorption and Catalytic Conversion on MoS2@CF‐NRGO

A visual adsorption experiment was conducted as shown in Figure 3g. After 24 h, the initial yellowish Li2S6 solution became transparent in the presence of MoS2@CF‐NRGO, whereas the yellow color of the Li2S6 solution remained for CF‐NRGO. Moreover, the adsorbed Li2S6 solution was analyzed by UV absorption spectroscopy (Figure S12, Supporting Information), in which the characteristic peaks corresponding to S6 2− at approximately 260 nm presented a more obvious decrease in intensity in the spectra of MoS2@CF‐NRGO,[ ] suggesting strong chemical absorption of MoS2@CF‐NRGO toward LiPSs. XPS was further used to characterize MoS2@CF‐NRGO after the adsorption experiment. First, the XPS full spectrum of MoS2@CF‐rGO (Figure S13, Supporting Information) confirmed the co‐existence of Mo, S, C, N, and O. The characteristic peaks of Mo6+ decreased after the adsorption experiment (Figure 3h), indicating a strong chemical reaction between Mo6+ and Li2S6. Meanwhile, the positions of several Mo 3d peaks and the S 2p peaks exhibited a slight shift (Figure 3h and Figure S14a, Supporting Information), confirming the strong electronic interaction between MoS2 and LiPSs. However, in the C 1s spectrum with three fitted peaks (C–C (284.7 eV), C–N (286.3 eV), and O–C≐O (288.9 eV)),[ ] only the C–N peak changed, while the other peaks did not change before or after the adsorption experiment (Figure S14b, Supporting Information), indicating that the doping of nitrogen atoms could increase the adsorption ability of the carbon substrate to Li2S6.

Furthermore, cyclic voltammetry (CV) of symmetric cells was conducted in a potential window from −1.5 to 1.5 V at a scan rate of 10 mV s−1 to evaluate the LiPSs catalytic ability of modified materials (Figure S15, Supporting Information). Compared with CF‐NRGO and MoS2‐CF‐NRGO (physical mixture of pure MoS2 and CF‐NRGO), the CV curve of MoS2@CF‐NRGO displayed a larger redox current, in which peaks C and D in the cathodic scan were assigned to the reduction of elemental sulfur to Li2S6 and Li2S6 to Li2S2/Li2S, respectively, and peaks A and B in the subsequent anodic scan represented opposite oxidation processes.[ ] In addition, Li2S precipitation and oxidation measurements were conducted to further reveal the preponderance of MoS2@CF‐NRGO in polysulfide conversion, as shown in Figure 3i and Figure S16, Supporting Information. Compared with the results of CF‐NRGO and MoS2‐CF‐NRGO, the responsivity of Li2S uncleation in MoS2@CF‐NRGO was more rapid, and the capacity of Li2S precipitation in MoS2@CF‐NRGO was higher (Figure S16, Supporting Information). More importantly, the oxidation of solid Li2S in MoS2@CF‐NRGO exhibited a more obvious enhancement than those in CF‐NRGO and MoS2‐CF‐NRGO (Figure 3i), suggesting a significantly reduced oxidation overpotential for Li2S conversion,[ ] which is consistent with the results of the theoretical calculations. The above results reveal that the synergistic effects of CF‐NRGO and MoS2 can significantly accelerate the transformation kinetics of LiPSs.

Electrochemical Performance of Li–S Batteries Constructed Using MoS2@CF‐NRGO

Before being applied to Li–S batteries, MoS2@CF‐NRGO/PP was investigated and analyzed. As shown in Figure S17a, Supporting Information, there was no appreciable delamination or deformation of MoS2@CF‐NRGO/PP after folding and bending, suggesting excellent flexibility and firm adhesion. A cross‐sectional SEM image shows an approximately 10 µm thick coating layer with an areal density of 0.3 mg cm−2 (Figures S17b and S18, Supporting Information). The SEM images of the samples (Figure S17c,d,e, Supporting Information) show that PP Celgard possessed a macroporous structure (>200 nm) (Figure S17c, Supporting Information), which is larger than the size of soluble LiPSs (<2 nm), leading to a serious shuttle effect on the LiPSs. The coating layer of CF‐NRGO and MoS2@CF‐NRGO was relatively fluffy, which provided favorable conditions for Li+ shuttling and physically blocked the shuttling of LiPSs. The above separators were then used to perform a permeation experiment a H‐type device (Figure S19, Supporting Information). Compared to PP and CF‐NRGO/PP, MoS2@CF‐rGO/PP exhibited the most effective ability to alleviate the shuttling of LiPSs. Furthermore, Li–Li symmetrical cells with different separators were used as shown in Figure S20, Supporting Information. The Li–Li symmetrical cell with MoS2@CF‐NRGO/PP showed a lower and more stable overpotential after 300 h than those of Li–Li symmetrical cells with PP and CF‐NRGO/PP, also proving favorable conditions for Li+ transfer and inhibition of lithium dendrite growth in MoS2@CF‐NRGO/PP.

Subsequently, CV curves with different Li–S battery separators were first constructed within the potential range 1.7–2.8 V at a scan rate of 0.1 mV s–1 (Figure  ). Two clear cathodic peaks located at approximately 2.3 and 2.0 V, corresponding to the reduction from cyclic S8 to soluble lithium polysulfide (Li2Sn, 4 ≤ n ≤ 8) and further conversion to Li2S2/Li2S, and two anode peaks at near 2.3 and 2.4 V can be ascribed to the delithiation of Li2S/Li2S2 and Li2Sn. Moreover, compared to the CV curves of the contrastive samples, CV profile of MoS2@CF‐NRGO configuration delivered a higher peak current and smaller polarization voltage (Figure S22, Supporting Information), indicating the enhanced transformation of S8 to soluble LiPSs and then further to insoluble products (Li2S) and facilitated the oxidation of Li2S to sulfur. In addition, the onset potential values of all redox peaks, which were determined at a current density of 10 µA cm–2 beyond the baseline voltage as shown in Figure S23, Supporting Information,[ ] were measured based on different CV profiles (Figure 4b and Figure S24, Supporting Information). The MoS2@CF‐NRGO configuration showed the largest reduction peak and the smallest oxidation peak, indicating the preferable catalytic activity of MoS2@CF‐NRGO.[ ] The galvanostatic charge/discharge curves of all the samples presented one charge plateau and two discharge plateaus, which is consistent with the CV analysis (Figure 4c). The MoS2@CF‐NRGO configuration exhibited a the longer and more stable discharge profile with a significantly smaller polarization potential, ΔE1, further demonstrating the electrocatalytic effect of the heterostructure.[ ] Meanwhile, the MoS2@CF‐NRGO configuration exhibited the lowest overpotential ΔE 2 and ΔE 3 at the second discharge plateau and the first charge plateau, respectively, indicating the rapid conversion of soluble Li2S4 to insoluble Li2S2/Li2S and the transformation of soluble Li2S to sulfur, respectively, which is consistent with the theoretical calculations analysis. As expected, the MoS2@CF‐NRGO configuration delivered outstanding rate capacities of 1274.9, 1016.9, 901.4, 801.6, 731.7, and 615.1 mAh g–1, at 0.2, 0.5, 1, 2, 3, and 5 C, respectively, which were better than those of the CF‐NRGO configuration (1012.9, 782.2, 684.2, 599, 538.1, and 481.7 mAh g–1) and PP configuration (937.2, 721.7, 624.9, 526.2, 462.5, and 379.6 mAh g–1) (Figure 4d). The cycling performance of the MoS2@CF‐NRGO configuration was also measured at 0.2 C as shown in Figure 4e. The initial reversible discharge capacity was 1120 mAh g–1, and present an excellent capacity retention of 742.9 mAh g–1 was presented after 200 cycles with 99.6% average Coulombic efficiency. In comparison, the discharge capacities of the CF‐NRGO configuration and PP configuration rapidly faded from 1103.5 to 485.6 mAh g−1 and from 890.3 to 381 mAh g−1, respectively, after 200 cycles. Meanwhile, the galvanostatic charge/discharge profiles of Li‐S batteries with different separators for the 1st, 50th, 100th, 150th, and 200th cycles at 0.2 C exhibited the typical characteristics of two plateaus (Figure S25, Supporting Information). However, the charge/discharge voltage profiles of the MoS2@CF‐NRGO configuration were flatter profile than those of the others, demonstrating the excellent electrocatalytic ability of MoS2@CF‐NRGO. The MoS2@CF‐NRGO configuration also exhibited stable performance, even at a low electrolyte/sulfur ratio of 10 µL mg–1 (Figure S26, Supporting Information). The long‐term cycling stability was also tested at 1 C, as shown in Figure S27, Supporting Information. The MoS2@CF‐NRGO configuration exhibited an average decay rate of only 0.06% per cycle after 1000 cycles, which is significantly lower than those in the CF‐NRGO configuration (0.09%) and PP configuration (0.13% after 443 cycles). The electrochemical performance of the MoS2@CF‐NRGO configuration was comparable and even superior to that of reported compounds for Li–S batteries (Table S1).

Figure 4

a) Cyclic voltammetry (CV) curves at a scan rate of 0.1 mV s−1, and b) the onset potentials for the electrochemical reaction stages of different separators. c) The first charge/discharge profiles at 0.2 C (inset: the corresponding enlarged part). d) Rate capacity and e) cycling capabilities at 0.2 C of a cell with the MoS2@CF‐NRGO/PP separator, CF‐NRGO/PP, and PP separators. f) Cycling performance at 0.5 C before and after rest of the MoS2@CF‐NRGO/PP and PP separators. g) Cycling performance at 0.2 C of different separators under 0 ℃. h) Cycling performance of the MoS2@CF‐NRGO/PP separator with high sulfur content. i) Soft‐pack cells charge a mobile phone, even in the folded state.

To further understand the improved electrochemical performance of MoS2@CF‐NRGO configuration, electrochemical impedance spectroscopy (EIS) measurements and SEM characterization were performed on the cells before and after cycling. Figure S28a, Supporting Information presents Nyquist plots before cycling of the different configurations, in which the MoS2@CF‐NRGO configuration shows the smallest semicircle in the high‐frequency region, indicating the lowest R ct and best electrolyte wettability. After 100 cycles, an additional semicircle was appeared in the Nyquist plots of all configurations (Figure S28b semicircle), which was attributed to the formed solid electrode/electrolyte interface.[ ] Compared with the contrastive samples, the MoS2@CF‐NRGO configuration showed the lowest R ct, electrode/electrolyte interface resistance, and Warburg impedance, demonstrating accelerated catalytic conversion of LiPSs and facilitated Li+ transport during cycling.[ ] Furthermore, the Li+ diffusion coefficient, D (cm2 s–1), of the MoS2@CF‐NRGO configuration before and after cycling, calculated from the EIS curves (Figure S29, Supporting Information), was higher than those of the PP and CF‐NRGO configurations (Figure S28c, Supporting Information). Li–S batteries with different separators were disassembled after 200 cycles to observe the curbing effect of multifunctional layer for LiPSs. Digital photographs and the corresponding SEM images are shown in Figure S30, Supporting Information. The clear yellow substance adhered to the surface of the PP separator due to the high concentration of insoluble Li2S2/Li2S in the electrolyte, whereas the surface of MoS2@CF‐NRGO/PP hardly exhibited a yellow substance (Figure S30a1,b1,c1, Supporting Information). Meanwhile, compared with the contrastive samples (Figure S30, Supporting Information), the surface of the lithium anode in the MoS2@CF‐NRGO configuration was the smoothest because it had the smallest deposition of Li2S2/Li2S, indicating the favorable catalytic conversion of LiPSs in MoS2@CF‐NRGO.

The suppressed shuttle effect of LiPSs and enhanced redox reaction dynamics endowed the MoS2@CF‐NRGO configuration with Mo–O–C bonds possessed excellent performance. Thus, several properties were further investigated from a practical perspective. First, the self‐discharge behavior of Li–S batteries was studied by resting for five days after 10 charge/discharge cycles at 0.5 C. As shown in Figure 4f, the discharge capacity of the MoS2@CF‐NRGO configuration after resting for five days is decreased slightly from 1015.8 to 927.9 mAh g−1, in which the average self‐discharge value was calculated to be only 8.6%. In contrast, the discharge capacity of the PP configuration is sharply decreased from 669.7 to 391.1 mAh g−1 and the self‐discharge value is reached to 46.08%. This result indicates that the MoS2@CF‐NRGO configuration can effectively relieve self‐discharge by suppressing the shuttle effect of LiPSs. From the galvanostatic charge/discharge curves shown in Figure S31, Supporting Information, the reversibility of the high‐voltage plateau was inferior for the PP configuration, in which the QH (capacity of the high plateau) retention was only 79.16% after rest. In contrast, the MoS2@CF‐NRGO configuration had a 97.25% retention rate after rest, which demonstrates its excellent LiPSs capture capability during the self‐discharge process.[ ] The self‐discharge value of the MoS2@CF‐NRGO configuration was also comparable to those reported for coating separators or host materials (Table S2, Supporting Information). In addition, the cycling performance was evaluated at high and low temperatures, as shown in Figure 4g and Figure S32, Supporting Information. Both configurations delivered an increased initial capacity at 60 ℃ and 0.2 C because of the facilitated diffusion of Li ions and electrochemical reactions at high temperatures (Figure S32, Supporting Information). Importantly, the MoS2@CF‐NRGO configuration still retained a relatively stable Coulombic efficiency because the barrier and reuse effect of the modified layer for LiPSs becomes more important at high temperatures.[ ] Moreover, the MoS2@CF‐NRGO configuration provided a higher capacity than the PP configuration with 0 ℃ at 0.2 C (Figure 4g), suggesting an enhanced catalytic effect in MoS2@CF‐NRGO. Furthermore, the cycling performance of the MoS2@CF‐NRGO configuration with various areal sulfur loading at 0.2 C was also investigated, as shown in Figure 4h. The areal capacities with 3.7, 6.4, and 8.0 mg cm–2 were 3.41, 5.30, and 6.11 mAh cm–2 and only slightly decreased after 50 cycles. Meanwhile, the cyclability of the MoS2@CF‐NRGO configuration at 4.2 mg cm–2 was maintained after 200 cycles at 1 C (Figure S33, Supporting Information). This demonstrates the excellent synergistic effect of MoS2 strongly coupled with the 3D carbon matrix to effectively alleviate the shuttle effect in high‐areal‐capacity Li–S batteries. Subsequently, soft‐pack batteries with MoS2@CF‐NRGO/PP were assembled and could charge a mobile phone even in the folding state (Figure 4i).

To better understand the shuttling effect of LiPSs, their evolution was analyzed by in situ Raman spectroscopy at the Li anode side during the discharging process (Figure  ). Figure 5a shows the construction diagram of the device, in which LiPSs can be qualitatively detected in real time using a quartz observation window. Figure 5b presents the Raman spectra of the PP separator and electrolyte used to eliminate redundant Raman peaks during the testing process. As clearly presented in Figure 5c,d, the characteristic peaks of long‐chain and mid‐chain polysulfides were observed in the Raman image and Raman spectra in the PP configuration.[ ] In particular, at the low‐voltage plateau, the peak intensity of polysulfides in the PP configuration was clearly heightened because of the penetration of massive polysulfides from the sulfur cathode to the Li anode. With continuous discharge, the polysulfide signal gradually weakened, which indicates that the polysulfides abscised from the PP separator and diffused to the lithium anode. These results demonstrate that a significant shuttle effect occurred in the PP configuration. Conversely, the in situ Raman spectra of the MoS2@CF‐NRGO/PP configuration are shown in Figure 5e,f, in which the peak intensity of the polysulfides was weak during the all‐discharge process, indicating only a small amount of polysulfide shuttling. This further confirms that MoS2@CF‐NRGO enables stronger adsorption and effective inhibition of intermediate polysulfides.

Figure 5

In situ micro‐Raman measurements of the MoS2@CF‐NRGO/PP and PP separators during discharge: a) Schematic illustration of Li‐S configurations with a sealed glass window for in situ Raman experiments. b) Raman spectra of the PP separator and Li‐S electrolyte. c,e) In situ time‐resolved Raman images of the PP and MoS2@CF‐NRGO/PP separators, respectively. d,f) Selected Raman spectroscopy of the PP and MoS2@CF‐NRGO/PP separators, respectively. The inset red curves in (c) and (e) are voltage profiles of the PP and MoS2@CF‐NRGO/PP separators, respectively.

Conclusions

DFT calculations revealed that the intrinsic electronic conductivity of MoS2 and the smooth electronic pathway of the heterogeneous interface of MoS2/C can significantly accelerate bidirectional solid–liquid–solid conversion. Then, MoS2@CF‐NRGO with covalent coupling of the Mo–O–C bond, as a multifunctional modified layer, was prepared for Li–S batteries. A series of experiments, such as anchoring/catalytic analyses and in situ Raman spectroscopy, proved that this functional separator can regulate Li2S deposition and facilitate Li2S decomposition. Li–S batteries with an MoS2@CF‐NRGO‐modified separator presented low self‐discharge (only 8.6%), excellent rate performance (615 mAh g−1 at 5 C), stable cycling performance (after 1000 cycles at 1 C with a capacity decay rate of only 0.06% per cycle), and a high areal capacity (6.11 mAh cm−2 at 0.2 C). Composites with strong covalent coupling can be extended to other materials, facilitating the development of composite electrocatalysts for high‐performance Li–S batteries.

Experimental Section

Experimental details were shown in the Supporting Information.

Conflict of Interest

The authors declare no conflict of interest.

Supporting Information

Click here for additional data file.