In‐situ growth of ultrathin sulfur microcrystal on MXene‐based 3D matrice for flexible lithium–sulfur batteries

Abstract

Lithium–sulfur batteries (Li‐S batteries) are promising next‐generation energy storage systems because of their high‐theoretical energy density. However, the commercialization of Li‐S batteries is still impeded by the aggregation of sulfur, low‐sulfur utilization, shuttling of dissolved polysulfides and sluggish reaction kinetics. Herein, we designed a hierarchically maple leaf‐like structured sulfur electrodes by in‐situ growth of ultrathin sulfur microcrystal on two‐dimensional MXene‐graphene‐cellulose nanofiber (MGN) matrice (denoted as IS‐MGN@S). The sulfur microcrystal as cathode can achieve improved kinetics than bulk sulfur due to its few layers of sulfur atoms, which is proved by the density functional theory calculations. The MXene not only confines polysulfides through strong chemisorption but also promotes the catalytic conversion of polysulfides. The introduction of graphene improves the conductivity and boosts the immobilization and conversion of polysulfides. As a result, the IS‐MGN@S cathode demonstrates remarkable electrochemical properties with a high‐initial capacity (1229 mAh g⁻¹ at 0.2C), substantial improvement in rate capability (770 mAh g⁻¹ at 2C), and stable long‐term cycling capacity. Moreover, the pouch cells with IS‐MGN@S cathode and gel electrolyte demonstrate excellent mechanical properties under mechanical damage (nail & cut tests, severe deformations), suggesting their promising applications for wearable electronic devices.

INTRODUCTION

Driven by the ever-increasing requirement for energy consumption, energy storage systems for propelling electric vehicles and storing surplus electricity generated from renewable sources have become a worldwide priority.^1–3 As one of the most promising next-generation batteries, lithium–sulfur batteries (Li-S batteries) with high-theoretical energy density (2600 Wh kg⁻¹) have attracted an enormous amount of interest thanks to the overwhelming theoretical capacity (1675 mAh g⁻¹), widespread source, environmental friendliness, low cost, and nontoxicity of sulfur.^4,5 However, Li-S batteries are still not primed for commercialization as a result of intractable issues that remain including the electrically insulating nature of sulfur,⁶ the severe shuttle effect of lithium polysulfides (LiPSs),⁷ and large volumetric contraction/expansion (~80%) between sulfur and insulating Li₂S, causing low-coulombic efficiency (CE) and fast capacity fading.⁸ To ameliorate these obstacles, much effort has been dedicated to the elaborate design and rational construction of sulfur cathodes.

One of the common strategies is to encapsulate sulfur into porous carbonaceous materials, such as micro-mesoporous carbon,⁹ hollow carbon spheres, and carbon nanofibers.^10,11 In addition, significant efforts have been focused on fabricating sulfur with functionalized hosts, which have extensive functional groups to trap the dissolved LiPSs.^12–18 However, the weak affinity between the nonpolar carbonaceous host and the polar LiPSs intermediates leads to the serious decay of capacity over a long lifespan. MXene-based materials thus are considered as the promising host candidates for Li-S batteries due to their intrinsically metallic electrical conductivity and highly sulphiphilic surface that can bind LiPSs via polar–polar interactions.^19–22

Melting-diffusion method is the most commonly used sulfur loading technique, in which solid sulfur is melted and diffused into the hosts by heat treatment. Unfortunately, high-sulfur loading via melt-diffusion often results in severe sulfur agglomeration with significantly degraded electrical conductivity. How to control the nucleation and growth of sulfur is the key to generate thin and uniformly dispersed sulfur on the host matrice, thus realizing improved kinetics and sulfur utilization.^23–26

Here, a MXene and graphene matrice was designed as two-dimensional substrate for in-situ growth of the uniform sulfur. A small amount of cellulose nanofiber (CNF) is added in the matrice to improve its mechanical strength and flexibility. By controlling the sulfur deposition, a hierarchically maple leaf-like sulfur microcrystal is successfully in-situ grown on the MXene-graphene-CNF (MGN) matrice (denoted as IS-MGN@S). We visualized the distinct sulfur crystal growth behaviors via in-situ optical microscopy and demonstrated that the sulfur microcrystal was uniform and ultrathin (~ tens to hundreds of nanometers) without aggregation. Benefiting from the unique microcrystal structure, as well as the MGN matrice with high-electrical conductivity and strong physical/chemical confinement of LiPSs, the coin-cells show a high-initial capacity (1229 mAh g⁻¹ at 0.2C), substantial improvement in rate capability (770 mAh g⁻¹ at 2C), and stable long-term cycling capacity with a low-capacity decay (0.03% per cycle within 700 cycles at 1C). More remarkably, the electrode with a high-sulfur loading (~4 mg cm⁻²) and lean electrolyte content (E/S ratio: 4.8 μL mg⁻¹) delivers an initial capacity of 1213 mAh g⁻¹ with a CE of approximately 99% at 0.2C. In addition, the structure is beneficial for developing flexible Li-S batteries by using MGN as a current collector and high-sulfur loading host. We envision that the concept to design the uniform and ultrathin sulfur microcrystal with a hierarchical maple leaf-like structure can promote the development of high-energy-density and flexible energy-storage devices for future wearable electronics.

RESULTS AND DISCUSSION Fabrication and characterization

For the comparison, we also fabricated the sulfur electrodes by ex-situ anchoring on two-dimensional MGN matrice (denoted as ES-MGN@S). The preparation of IS-MGN@S and ES-MGN@S are schematically shown in Figure S1 via in-situ and ex-situ processes. Typically, CNF was prepared by the 2,2,6,6-tetramethyl-1-piperidine oxoammonium salt-mediated oxidation pretreatment process. Detailed experimental procedures of the synthesis are shown in Figure S2 and experimental section. Ultrathin layers of MXene nanosheets can be easily obtained by sonication of the colloidal suspensions. The as-prepared GO was dispersed in stable black colloidal supernatant of delaminated MXene nanosheets to form a homogeneous MXene-GO mixed solution. Subsequently, the MXene-GO mixed solution was transferred into CNF solution by magnetic agitation to obtain MGN matrice with a two-dimensional structure. Finally, we adopted the chemical deposition method to prepare two-dimensional sulfur microcrystal anchored to the MGN matrice by the disproportionation reaction. Moreover, the capping agent Triton X-100 would limit the growth of sulfur microcrystal to form submicro-scale particles. To synthesize ES-MGN@S, 300 mg of MGN was ground with 700 mg sulfur powder and applied to a heat treatment at 155°C under nitrogen atmosphere in tube furnace.

The preparation process of Ti₃C₂T_x with few layers is schematically presented in Figure 1A, delaminated Ti₃C₂T_x was prepared by selectively etching of Al from pristine Ti₃AlC₂ MAX, followed by subsequent ultrasonic exfoliation. The scanning electron microscopy (SEM) images of Ti₃C₂T_x exhibit an accordion-like layered structure with clean surface (Figure 1B). The layers of the Ti₃C₂T_x are split from each other which demonstrate the successful etching process. The transmission electron microscopy (TEM) images further show that the exfoliated MXene consists of only a few layers of MXene nanosheets (Figure 1D). From the SEM (Figure 1C) and TEM images (Figure 1E–G), it is clear that the IS-MGN@S has a hierarchically maple leaf-like structure. The uniform and ultrathin sulfur microcrystal with a thickness of tens to hundreds of nanometers is homogenously dispersed without aggregation. According to the mechanism, CNF, which performs as an adhesive, interacts with graphene and MXene nanosheets to form MGN. The MGN has a strong affinity toward sulfur microcrystal, serves as an anchor site for the sulfur microcrystal during the subsequent steps. No sulfur is observed from the HR-TEM images, demonstrating sulfur microcrystal is covered uniformly on the inner surface of MGN (Figure 1H–J). As for the SEM and TEM images of the ES-MGN@S, it forms irregular stacks of interlaced structure which leads to a rough surface, densely covered with sulfur particles (Figure S3A-D and Figure S4A-D). Agglomeration is severe on the surface of the ES-MGN@S sample. The EDS mapping images present the inhomogeneous distribution of Ti, C, O, and S (Figure S5), suggesting the irregular stacks of interlaced structure of ES-MGN@S. The HRTEM and the corresponding selected area electron diffraction (SAED) patterns imply that the MXene nanosheets retain good crystallinity (inset of Figure 1I and Figure 1J). Moreover, from the inset of Figure 1J, the hexagonal symmetry of the hexagonal lattice structure with lattice parameter of approximately 0.23 nm matches well with the (103) lattice planes of MXene nanosheets.²⁷

FIGURE 1. (A) The schematic illustration for the synthetic procedure of the few layers Ti3C2Tx. (B) SEM, (D) TEM images of Ti3C2Tx. (C) SEM, (E–G) TEM, and (H–J) HRTEM images of the IS-MGN@S. The inset of (D) shows the Tyndall effect. Inset in (I) is the enlarged HR-TEM image of IS-MGN@S. The inset of (J) shows corresponding SAED pattern. SEM, field-emission scanning electron microscopy; TEM, transmission electron microscopy; SAED, selected area electron diffraction

We visualize the distinct sulfur crystal growth behaviors via in-situ optical microscopy and demonstrate that a growing sulfur microcrystal can form a hierarchically maple leaf-like structure in Figure 2A and Figure S6A, B. The HAADF-STEM image and EDS mappings of the IS-MGN@S further verified the highly homogenous distribution of Ti, C, O, and S elements in Figure 2B, C. The atomic force microscopy (AFM) image demonstrates that the thickness of MXene nanosheets is ~2–3 nm (Figure 2D and Figure S7A, B). The sulfur content of IS-MGN@S was measured by thermal gravimetric analysis (TGA), with results presented in Figure S8A. The X-ray diffraction (XRD) of sulfur, pristine Ti₃AlC₂, Ti₃C₂T_x, IS-MGN@S and ES-MGN@S were measured to identify the phases and crystal structures (Figure S8B-D). The sharp diffraction peaks were observed for sulfur, confirming a crystalline state. However, sulfur characteristic peaks of IS-MGN@S and ES-MGN@S composites are significantly weaker and shifted to lower angles. The front view and the top view crystal structures of Ti₃C₂T_x are shown in Figure S9. Crystallographic structures information of the IS-MGN@S was also conducted by Raman characterization in Figure S10. The peaks at around 155 and 631 cm⁻¹ are A_1g symmetry out-of-plane vibrations of Ti and C atoms. The peaks at 282, 416 and 622 cm⁻¹ are the E_g group vibrations, including in-plane modes of Ti, C, and surface functional group atoms.²⁸ The peaks appear at 1350 and 1590 cm⁻¹ can be attributed to the D and G bands of graphitic carbon. The surface chemical environment of IS-MGN@S was further performed by X-ray photoemission spectroscopy (XPS) test. The full XPS revealed four main elements, Ti, C, O, and S (Figure S11A). In the Ti 2p XPS spectra of IS-MGN@S, which display peaks at 458.9, 460.6, and 465.6 eV, in association with TiS bonding, TiO bonding, and TiO bonding, respectively (Figure 2E).²⁹ The C 1 s spectrum and S 2p peaks of IS-MGN@S are displayed in Figure 2F and Figure S11B. From the XPS results, it is indicated that the formation of TiS bond in IS-MGN@S could provide strong adsorption for LiPSs.⁹ Moreover, the existence of the reactive functional groups in IS-MGN@S provides active sites to entrap LiPSs and simultaneously to chemically regulate the conversion of sulfur. The Brunauer–Emmett–Teller and Barrett–Joyner–Halenda results of IS-MGN@S are recorded in Figure S12A, B. To reveal the chemical bonding state, an evolution of surface groups is investigated by the Fourier transform infrared (FTIR) spectra (Figure 2G and Figure S12C). For all the samples, characteristic peaks around 3440 and 1618 are attributed to the stretching vibration of OH and C═O bonds.³⁰ The peaks of MXene, IS-MGN@S and ES-MGN@S around 1420 cm⁻¹, 1360 cm⁻¹, and 1040 cm⁻¹ are ascribed to the vibrations of the N-Ti binding.³¹ To understand the LiPSs adsorption capability of CNF, graphene and MGN, UV–vis absorption spectra was used to compare the concentration change of Li₂S₆ solution. The characteristic peaks of the polysulfides solution located at 260, 280, and 350 nm are assigned to the S₆²⁻ species while 320 nm is attributed to the S₄²⁻ in UV band,³² where the absorbance of Li₂S₆ is observed in graphene and CNF samples, and in contrast negligible absorbance is seen for MGN (Figure 2H). The static adsorption test shows that the Li₂S₆ solutions are decolored to transparent solution by MGN and graphene, whereas the solutions containing CNF turn from lighter yellow to deep yellow (inset of Figure 2H). To further illustrate the absorptivity of IS-MGN@S and ES-MGN@S samples toward LiPSs, the operando optically tests were conducted in Figure 2I. The vial cell with IS-MGN@S sample remained yellowish even over 12 h of cycling/electrochemically discharging, implying that the high affinity of the IS-MGN@S to prevent the dissolution of LiPSs. However, regarding ES-MGN@S, the electrolyte color changed from colorless to yellowish after discharge for 1 h, and completely changed into dark yellow after discharge for 12 h.

FIGURE 2. (A) Snapshot of a growing sulfur crystal from marginal main area by in-situ optical microscopy at room temperature. (B) HAADF-STEM image of IS-MGN@S and (C) EDS mappings of S (yellow), C (red), Ti (purple) and O (blue) in low magnification. (D) Representative AFM image of individual MXene nanosheets. (E) XPS survey spectra of Ti 2p and (F) C 1 s. (G) FTIR spectra of IS-MGN@S and ES-MGN@S samples. (H) UV–vis absorption spectra of the Li2S6 solution before and after the addition of MGN, GO and CNF samples. The inset image shows a photograph of the Li2S6 solutions with MGN, GO and CNF absorbents. (I) The vial cells using IS-MGN@S and ES-MGN@S as sulfur hosts were operated with Li metal anode. HAADF-STEM, high-angle annular dark-field imaging-scanning transmission electron microscopy; EDS, energy-dispersive spectroscopy; AFM, atomic force microscopy; XPS, X-ray photoemission spectroscopy; FTIR, Fourier transform infrared; GO, graphene oxide; CNF, cellulose nanofiber; UV, ultra violet

Kinetics properties of the hosts

Figure S13A and Figure S13D exhibit the cyclic voltammetry (CV) profiles of the IS-MGN@S and ES-MGN@S electrodes at a scanning rate of 0.1 mV s⁻¹. The recorded CV profiles of IS-MGN@S and ES-MGN@S cathodes consist of two cathodic peaks (C₁ and C₂), implying the reduction of S₈ to long-chain LiPSs (Li₂S₈ → Li₂S₆ → Li₂S₄) and subsequent transformation to short-chain LiPSs (Li₂S₄ → Li₂S₂ → Li₂S).³³ The sharp anodic peaks (A₁ and A₂) of the IS-MGN@S and ES-MGN@S cathodes are ascribed to the oxidation of LiPSs to sulfur (Li₂S₂/Li₂S to Li₂S₈/S). Compared with ES-MGN@S electrode, CV profiles of IS-MGN@S show no distinguishable shift and low-current density of the redox peaks, which indicate the decreased cell polarization and improved electrochemical stability. To evaluate the ion transfer kinetic properties, ion diffusion coefficients of IS-MGN@S and ES-MGN@S electrodes are investigated through CV at the scanning rates ranging from 0.1 to 1.5 mV s⁻¹ (Figure 3A and Figure S14). As shown in Figure S13B, the IS-MGN@S electrode displays sharp cathodic peaks appear at 2.32 V (C₁) and 2.01 V (C₂) for all scan rates, implying the quick transformation of sulfur to LiPSs. Corresponding anodic peaks are presented at nearly 2.37 V (A₁) and 2.40 V (A₂) resulting from the oxidation of LiPSs to sulfur (Figure S13C). The ES-MGN@S electrode shows more severe decrease in onset potential of C₁ (≈ 2.28 V) and C₂ (≈ 1.95 V) peaks with an increased scan rate, and more extreme increase of A₁ (≈ 2.49 V) and A₂ (≈ 2.52 V) peaks than the IS-MGN@S (Figure S13E, F). The integration of A₁ and A₂ peaks of ES-MGN@S at high-scan rate is ascribed to the increased polarization of relatively slower oxidation reaction than the reduction reaction. The degree of change on the initial potentials of C₁, C₂, A₁ and A₂ peaks are also much lower in the IS-MGN@S electrode, demonstrates the accelerated redox reaction of IS-MGN@S (Figure 3B and Table S1).³⁴ The CV results indicate good stability, demonstrating a high reversibility of the redox reactions for the IS-MGN@S electrode. Furthermore, it shows that MGN could prevent sulfur microcrystal from dissolving in the electrolyte due to the extensive functional groups which have strong adsorbing ability to anchor sulfur species. Further characterization of the kinetic properties of IS-MGN@S and ES-MGN@S electrodes are analyzed by the calculation of b values from the Equation (1)³⁴ [Image Omitted. See PDF]where i demonstrate the measured current, and v corresponds to the scan rate. The b value is correlated to the reaction kinetics. For instance, b = 0.5 or 1.0 imply that the reaction is dependent on the semi-infinite linear diffusion or controlled by surface-diffusion. The b values of IS-MGN@S electrode are 0.72, 0.57, 0.58 and 0.78 in peaks C₁, C₂, A₁ and A₂, respectively. In comparison with IS-MGN@S electrode, the ES-MGN@S electrode shows lower b values (0.68, 0.37, 0.52, and 0.49) in Figure 3C (the higher b value is ascribed to enhanced reaction kinetics). These results strongly support that the sluggish lithiation reaction is kinetically enhanced owing to the highly conductive structure with homogenous dispersion of sulfur microcrystal.

FIGURE 3. (A) Rate-dependent CV curves of IS-MGN@S electrode. (B) Potential difference (ΔE) of IS-MGN@S and ES-MGN@S electrodes between the discharge–charge curves with various scanning rates. (C) The b-Value of IS-MGN@S and ES-MGN@S electrodes obtained by calculating the slopes of the plots (derived from the CV data measured with various scanning rates). (D–H) Density functional theory calculations. The charge density difference of S1–S8 adsorbed on (D) Ti3C2Tx and (E) graphene. (F) Adsorption energies of S1 to S8. (G) Ratio of vdW interaction of S1 to S8. (H) Adsorption energies of Li2S–Li2S8. CV, cyclic voltammetry

Density functional theory (DFT) calculations were performed to compare the binding strengths of sulfur microcrystal and bulk sulfur with MXene/graphene. For simplicity, different numbers of sulfur atoms S_n (S₁, S₂, S₄, S₆, and S₈) are absorbed on the MXene/graphene surface, as a reasonable structure model to qualitatively study the difference between the properties of sulfur microcrystal and bulk sulfur. MXene/graphene, S_n, and the corresponding adsorption structure before/after optimization are shown in Figures S15-16. The anchoring effects of MXene/graphene are evaluated by calculating adsorption energies (E_ad) for S₁–S₈ according to the equation E_adsorption = (E_{MXene/Graphene+Sn}−E_Sn−E_{MXene/Graphene})/n, where E_{MXene/Graphene+Sn} and E_{MXene/Graphene} are the total energies of MXene or graphene with/without S_n, respectively, and E_Sn is the total energy of S_n. A negative E_ad means favorable adsorption. For MXene, the binding energies of S₁, S₂, S₄, S₆, and S₈ are −2.222, −0.434, −0.144, −0.123, and − 0.125 eV/atom, respectively (Figure 3F, Table S2). A similar trend also happened for graphene, where the binding energies are −1.732, −0.151, −0.122, −0.090, and − 0.094 eV/atom, respectively. Consequently, the adsorption strength is expected to decrease with the increase of sulfur atoms. In other words, sulfur microcrystal (IS-MGN@S) as cathode can achieve remarkable capability than bulk sulfur (ES-MGN@S) due to its few layers sulfur atoms. The nature of the interaction between MXene/graphene and S_n is investigated by analyzing the charge density difference of S_n adsorption (Figure 3D, E), based on the formula Δρ = ρ(Ti₃C₂T_x/graphene +S_n)-ρ(Ti₃C₂T_x/graphene)-ρ(S_n). It could be seen that with the decrease of sulfur atoms, the charge transfer between S_n and the substrate increases apparently, especially in S₁, where there is strong covalent TiS or CS bonding between S_n and the substrate. However, the charge transfer is very weak for S₈, and the interaction is mainly Van der Waals force. This phenomenon is further confirmed by calculating the ratio of chemical and physical interaction (Van der Waals force) according to the formula $\text{RvdW}=\left(E_{\mathit{ads}}^{\mathit{vdW}}-E_{\mathit{ads}}^{\mathit{no}\mathit{vdW}}\right)/E_{\mathit{ads}}^{\mathit{vdW}}\times 100\%$ (Figure 3G).³⁵ For MXene/graphene, the contribution of Van der Waals force increases sharply with sulfur atoms, approaching 100% for S₈, which is opposite to the trend of adsorption energy.

Based on the above simulation results, the strong adsorption energy of low atom S_n comes from the substantial increase of chemical bonding between it and the substrate, whose essence may be the improvement of unsaturation and reactivity. Thus, a rational understanding is that sulfur microcrystal as a cathode can achieve remarkably improved performance than bulk sulfur, due to its few layers of sulfur atoms, increased surface reactivity, and effectively enhanced chemical adsorption. Moreover, we also calculate the adsorption energy between MXene/graphene and LiPSs (Figure 3H and Figure S17A, B). It can be seen that the adsorption energy of graphene surface is between 0 to −1 eV, while that of MXene is between −1.5 and − 4.0 eV. This phenomenon indicates that MXene/graphene can further limit the shuttle effect of LiPSs through strong adsorption, which is consistent with adsorption experiment results. To summarize, the reasonable design of both adsorbates and substrates, that is the uniform and ultrathin sulfur microcrystal with MGN matrice, can enhance adsorption strength and suppress the shuttling effect, thus ensuring the remarkable capability of Li − S batteries.

Coin-cells performances of S cathode

The discharge capacity of IS-MGN@S is 1227 mAh g⁻¹ at a current density of 0.2C (1C = 1675 mA g⁻¹) in the first cycle, which indicates the high utilization of sulfur (Figure 4A). It can be found that IS-MGN@S electrode achieves a stable cycling ability of 1114 mAh g⁻¹ after 50 cycles and remains around 1046 mAh g⁻¹ with a CE of ~99% even after 100 cycles. The IS-MGN@S electrode shows high-specific capacity and good cycling stability because of the synergy effects of each component. The MGN not only improves the conductivity and boosts the immobilization and conversion of polysulfides. In contrary, rapid decay is observed for the ES-MGN@S electrode with 1213 mAh g⁻¹ and only remains 522 mAh g⁻¹, because of the low conductivity and serious shuttle effects. The low-initial capacity loss of IS-MGN@S electrode compared to the ES-MGN@S electrode suggests that the MGN matrice not only confines polysulfides through strong chemisorption but also promotes the catalytic conversion of polysulfides. After deep cycling, the structure of IS-MGN@S electrode is well maintained without sulfur aggregation, manifesting high robustness and redox efficiency against repeated cycling.

FIGURE 4. (A) Cycling capacitance and CE of IS-MGN@S and ES-MGN@S electrodes at 0.2C rate. (B) Rate performance of IS-MGN@S and ES-MGN@S electrodes with various current densities. (C) Charge–discharge curves of IS-MGN@S electrode with various current densities. (D) Comparison of the potential difference between the charge–discharge plateaus at different current densities. (E) Cycling performance and (F) rate capacity of the IS-MGN@S with sulfur mass loading of ~4 mg cm−2 and E/S ratio of 4.8 μL mg−1. (G) Long-term cycling performance and CE of IS-MGN@S and ES-MGN@S electrodes over 700 cycles at 1C rate. (H) Radar graph of the reported MXene-based electrode materials for Li-S batteries. CE, Coulombic efficiency

In addition to the superior cyclability, the rate capabilities of IS-MGN@S and ES-MGN@S electrodes are also depicted at different current densities from 0.2 to 2C rate in Figure 4B. With further cycling at 0.2, 0.5 and 1C, IS-MGN@S electrode shows reversible discharge capacities of 1229, 1098, and 928 mAh g⁻¹ for each 10 cycles, achieving a high-capacity retention. Even at a high-current density of 2C, the IS-MGN@S still achieves the high capacity of ~800 mAh g⁻¹, indicating good rate capability compared with ES-MGN@S electrode (412 mAh g⁻¹). Increasing rates lower the discharge voltage and raise the charging voltage, demonstrating the typical increase of the Ohmic overpotential at high-current operation (Figure 4C and Figure S18). Moreover, the IS-MGN@S electrode still exhibits a discharge capacity of ~1008 mAh g⁻¹ in another 10 cycles when the current density is switched abruptly from 2 to 0.5C, suggesting its better recovery capability than that of ES-MGN@S electrode. Nevertheless, with increasing current density applied to the ES-MGN@S electrode, it shows serious capacity degradation. When the current density increases to 2C, the discharge capacity of the ES-MGN@S electrode irreversibly suffers from considerable capacity loss (597 mAh g⁻¹). The results demonstrate that the IS-MGN@S electrode has good electrochemical reversibility depending on the hierarchically maple leaf-like structure.

The polarizations of the IS-MGN@S and ES-MGN@S electrodes were further investigated by calculating the potential difference (ΔE) between the discharge–charge curves with various current densities from 0.2 to 2C. The polarization versus rate plots of IS-MGN@S and ES-MGN@S electrodes are displayed in Figure 4D. The ΔE of IS-MGN@S electrode configuration are 225, 239, 244 and 258 mV, much smaller than those of the ES-MGN@S electrode (243, 269, 301, and 337 mV) at all current rates. These results indicate a kinetically enhanced reaction process for the IS-MGN@S electrode with a small barrier.³⁶

Furthermore, the sulfur mass loading (>4 mg cm⁻²) and low-electrolyte content (E/S ratio < 5 μL mg⁻¹) are the critical factors for the practical applications of Li–S batteries.³⁷ In light of this, the cycling and rate performances of IS-MGN@S cathodes with higher sulfur loadings (~4 mg cm⁻²) and low-electrolyte content (E/S ratio: 4.8 μL mg⁻¹) have been further investigated. The cycling performance of IS-MGN@S cathode with higher sulfur loadings delivers an initial capacity of 1213 mAh g⁻¹ at 0.2C (Figure 4E). It can be found that IS-MGN@S electrode remain a capacity of 913 mAh g⁻¹ with a CE of ~99%. The rate performance of IS-MGN@S cathode with higher sulfur loadings showed in Figure 4F, in which high-reversible capacities of 1125, 876, 684, and 543 mAh g⁻¹ are delivered under 0.2, 0.5, 1, and 2C, respectively. Switching back to 0.5C rate results in restoring the discharge capacity of ~670 mAh g⁻¹ for each 20 cycles, achieving a high-capacity retention.

In addition to the rate capabilities, another important parameter in practice is the long-term cyclability of the batteries. Figure 4G compares the capacities based on IS-MGN@S and ES-MGN@S electrodes with ~2 mg cm⁻² sulfur loadings, it shows an initial capacity of 984 mAhg⁻¹ (0.032% capacity loss per cycle, CE = 99%) and a high-capacity retention even at 1C rate after 700 cycles. In striking contrast, the cell with the ES-MGN@S electrode undergoes severe capacity fade after 700 cycles, ending up with a low capacity of 348 mAhg⁻¹ and a high-capacity-fading rate (0.091% capacity loss per cycle, capacity retention of 36%). Comparing the recent reports (Figure 4H and Table S3), the IS-MGN@S electrode shows high-initial capacity, good rate capability as well as among the best long-term cycling stability for Li–S batteries. Based on the aforementioned analysis, the considerable enhancement of the IS-MGN@S electrode can be attributed to the integrated configuration, which facilitates redox-conversion capability, buffers the volume change of sulfur, and effectively promotes sulfur utilization and LiPSs trapping. As shown in Figure S19A-C, we have done the parallel experiments with high-sulfur content of ~77% to prove that the sulfur can provide excellent cycling performance at 1C rate with uniform morphology. Furthermore, FESEM can observe that the entire morphology and microstructure of the IS-MGN@S electrode are well preserved after the cycling test (Figure S20A). For a comparison, some large sulfur agglomerates are observed on the ES-MGN@S (Figure S20B), demonstrating the dissolution loss of active sulfur into the electrolyte due to the lack of confinement and weak binding between MGN and thick sulfur. Long-term cycling stability further ascertains that applying the IS-MGN@S electrode shows advantages at suppressing shuttle phenomenon and improving the sulfur utilization, owing to the anchoring effects of MGN toward LiPSs.

The electrochemical impedance spectroscopy (EIS) measurements were conducted at open circuit voltage (Figure S21). For the IS-MGN@S and ES-MGN@S electrodes, we can observe two large depressed semicircles. The high-frequency region of the semicircle could reflect the charge transfer process which dominates the reduction reaction.³⁸ The middle frequency region of semicircle could be ascribed to the mass transport.³⁹ For ES-MGN@S, it has a much higher charge transfer resistance than that of IS-MGN@S electrode. The post-mortem analysis indicates that the ES-MGN@S cathode suffers from severe agglomeration of bulky sulfur on the surface.

Electrochemical performance of Li-S pouch cells

To measure the flexibility and the cycling stability of the Li-S pouch cells, the cells with different deformations (45°, 90°, 180°) at 0.2C rate are shown in Figure 5A. Figure 5B exhibits the charge/discharge curves of Li-S pouch cell with IS-MGN@S electrode in the 1st, 11st, 21st, and 31st cycles at 0.2C rate with various folding angles. As excepted, the pouch cell shows a discharge capacity of 1168 mAh g⁻¹ with unfolded state. With the cycling performance at bending state for 200 times, the cell still exhibits a reversible capacity in another 5 cycles when the state is switched abruptly from bending to flatting, suggesting its better recovery capability (Figure 5C, D). The Li-S pouch cell results indicate that the IS-MGN@S cathode holds great potential for practical flexible batteries.

FIGURE 5. (A) Cycling performance and (B) charge–discharge curves of Li-S pouch cell based on IS-MGN@S cathode at different folding angles. (C) Photographs of Li-S pouch cell without/with bending states. (D) Cycling performance of Li-S pouch cell based on IS-MGN@S cathode at flatting and bending state

High safety of gel electrolyte Li–S pouch cells

Safety and flexibility are the main concerns of wearable electronics. Ensuring high safety and flexibility, however, remains a tough task for Li-S batteries that use flammable liquid electrolyte and extremely fragile Li metal.⁴⁰ As a result, integrating an MGN host for large sulfur loading through vacuum filtration, flexible lithium anode (Figure 6A) and gel electrolyte (Figure 6B) offers great feasibility for addressing above problems. The flexible lithium anode was realized by a facile route through calendaring two pieces of Li metal foils and one piece of nickel foam to strengthen the tension and flexibility of Li anode (Figure 6 A1, A2). Interestingly, the self-supporting IS-MGN@S cathode can also be deformed into ~160° without damages in structure (Figure S22A). The SEM cross-sectional image exhibits that the thickness of the flexible IS-MGN@S film is ~20 μm (Figure S22B). As illustrated in simple physical experiments, as-fabricated gel electrolyte is found to be in semitransparent and gelatinous state with highly flexible (Figure 6C). It can be easily curled by exerting a small external force, and then its shape can well recover after the force was released. The light-emitting diodes are lightened up by pouch cells with gel electrolyte in various folding angles for more than 40 times without obvious change in brightness (Figure S23A-D). More importantly, the extreme conditions tests including mechanical damage and severe deformations have been applied to investigate the safety behaviors of the flexible Li-S pouch cell in practical applications. The LEDs are powered by the pouch cell integrating IS-MGN@S cathode, gel electrolyte and robust Li anode in a stable and safe manner without the threat of fuming when subjected to nail penetration test (Figure 6D–G). Surprisingly, the broken cell can maintain stable energy out-put for over 2 h in the air without rapid failure by internal short-circuit when the cell was cutting into small pieces (Figure 6H, I). Furthermore, the flexibility & cut tests have been applied in the broken cell to investigate the stability and safety under severe external stimuli (Figure 6J–L), which shows the potential applications in wearable electronics.

FIGURE 6. Schematic illustration of (A) soft-packaged flexible Li-S batteries, (B) gel electrolyte, and A1, A2 two pieces of Li metal foils and one piece of nickel foam were calendered together to form a Li-nickel foam-Li flexible sandwich structure. (C) Photographs of the gel electrolyte at bending states. Photographs of light-emitting diodes lit by pouch cells based on IS-MGN@S cathode at harsh external stimuli (D–G) nail test, (H, I) cut test, and (J–L) bend & cut test

CONCLUSION

To summarize, we have investigated the growth of ultrathin sulfur microcrystal on MGN substrate via in-situ optical microscopy and DFT calculations. The homogeneous dispersion of sulfur microcrystal on the surface of MGN matrice promotes the electron and ion conductivity of sulfur. Moreover, the MGN matrice can effectively prevent LiPSs within the electrode from shuttling and dramatically improves cycling stability. Considering the ultrathin sulfur microcrystal grown on MGN matrice without aggregation, thus achieving improved kinetics than bulk sulfur. Correlating the sulfur microcrystal evolution at micro-scale during liquid-phase synthesis, as well as controlling the growth of sulfur microcrystal on MGN matrice, we developed an advanced IS-MGN@S electrode. The cathode based on the IS-MGN@S electrode demonstrates remarkable electrochemical properties with a high-initial capacity, substantial improvement in rate capability, and stable long-term cycling capacity with a low-capacity decay (0.03% per cycle within 700 cycles at 1C). More remarkably, the cathode with a high-sulfur loading (~4 mg cm⁻²) and lean electrolyte content (E/S ratio: 4.8 μL mg⁻¹) delivers a high capacity of 1213 mAh g⁻¹ at 0.2C; even at 1C rate, the cathode still shows superior rate capacity of 684 mAh g⁻¹. Furthermore, the soft-packaged flexible Li-S batteries with gel electrolyte are further fabricated and demonstrate excellent mechanical and electrochemical properties under mechanical damage (nail & cut tests, severe deformations). This design brings up new opportunities to explore the flexible Li-S batteries. We envisage that our work not only provides insights into optimized sulfur cathode for high-energy-density Li-S batteries with high-sulfur loading and lean electrolyte, but also would hold great application potential in wearable power sources and portable electronics.

ACKNOWLEDGMENTS

This work was supported by National Key Research and Development Program of China (2019YFA0705700), National Natural Science Foundation of China (51774017 and 51904016) and Key Program of Equipment Pre-Research Foundation of China (6140721020103). Also, the authors want to thank from Beijing Zhongkebaice Technology Service Co., Ltd. for the support of experimental instruments. The materials characterizations analysis studies were supported by Shiyanjia Lab (www.shiyanjia.com). Jun Xia, Weixin Chen and Yang Yang contributed equally to this work.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

AUTHOR CONTRIBUTIONS

Jun Xia, Weixin Chen and Yang Yang contributed equally. Jun Xia and Shichao Zhang conceived the project and designed the experiments. Jun Xia, Xianggang Guan, Tian Yang and Mingjun Xiao performed the electrochemical studies and characterizations. Weixin Chen and Yang Yang contributed the DFT calculations. Jun Xia, Yalan Xing and Guangmin Zhou analyzed the results in which Shichao Zhang contributed. Jun Xia, Weixin Chen and Yang Yang co-wrote the manuscript in which Guangmin Zhou and Shichao Zhang made the revision. All of the authors discussed the results and commented on the manuscript.