English

• Lithium-sulfur batteries (LSBs) are considered to be promising next-generation rechargeable systems owing to their ultrahigh theoretical specific capacity (1675 mAh/g), cost-effective and environmental compatibility [1-3]. However, the sluggish kinetics of sulfur cathodes and the shuttle effect severely hamper the rate capability and lifespan of LSBs in practical implementation [4,5]. Catalytic sulfur reduction reaction (SRR) is critical to solve the above problems [6-10]. Single-atom catalysts (SACs) with nearly 100% atom utilization and impressive activities have attracted wide attention in energy-storage devices [11]. Typically, a stable planar porphyrin-like M-N₄ structure can be easily formed by atomically dispersed Fe/Co/Ni on N-doped carbons (M-NC) to serve as catalytic active sites [12-14]. However, the electron distribution around M-N₄ configurations and high electronegativity of N atoms disfavor the adsorption of sulfur species on active sites, necessitating higher energy for the conversion of lithium polysulfides (LiPSs) during SRR [15]. In this regard, the electronic properties of SACs play the crucial role in determining electrocatalytic activity and stability, and their rational regulation is highly desired to construct efficient catalytic systems for SRR.

  The construction of asymmetric coordination environment has been considered to be an effective strategy to regulate the electronic structure of central metal atoms, thus optimizing the adsorption and conversion for LiPSs [16]. Accordingly, the related regulations mainly focus on adjusting neighboring heteroatoms, changing the coordination number of metal sites, and introducing substrate-mediated strain on the supported M-N₄ site [17-19]. Despite changes of the first coordination shell directly induce near-range electronic regulation (NRER) of SACs, it is either restricted by the precise atomic-level doping, or by the surface morphology and geometry of substrates, leading to diverse coordination structures under unpredictability and uncontrollability of the strain. Therefore, the accurate tuning of electronic structures around SACs remains a huge challenge in the design of high active and stable catalysts.

  Long-range electronic regulation (LRER) of central metal sites has been reported by introducing alien atoms around M-N₄ sites, and proposed as promising method to provide rich electronic structures [20-22]. However, the reported LRER usually is not beyond the range of three atoms around single-atom sites, and also faces the problem of precise doping. Actually, the interface engineering of two-dimensional nanosheets would change the coordination environment of SACs via charge transfer and strain, and cause more widespread LRER owing to the interlayer free-electrons [16,23]. Inspired by this, the construction of two-dimensional heterostructures of M-N₄ and supports is expected to achieve precise electronic regulation of SACs through predictable interlayer charge transfer. However, the M-N₄ structures usually exist in carbon systems in the reported experimental strategies, and scarcely are investigated in non-carbon supports.

  In this work, we proposed Ti₃C₂T_x MXene-mediating NRER and LRER around Co SACs to promote the catalytic activity and stability in SRR. Benefiting from the special interactions between 1‑butyl‑3-methimidazolium cobalt tetrachloride ([Bmim]₂[CoCl₄], Co-IL) and Ti₃C₂T_x MXene, N-doped carbons successfully support Co single atoms on MXene surfaces, and assemble 3D Co-NC/MXene network. Therein, Ti₃C₂T_x MXene substrate induces a stable heterostructure with enhanced charge transfer, allowing for both NRER via adjacent N atoms and LRER via surrounding C atoms to generate very different electronic effects, in turn, the ideal interactions with polysulfide intermediates to boost adsorption-catalysis in SRR process. Therefore, combined with the advantages of densely-distributed active sites and continuous conductive pathway by the network structure, the 3D Co-NC/MXene based sulfur cathodes show reversible capacities of 1346 mAh/g at 0.2 C, long cycling stability with capacity decay of merely 0.015% per cycle over 600 cycles at a high current density of 4 C, and superior areal capacities of 5.0 mAh/cm², outperforming state-of-the-art sulfur cathodes.

  DFT simulation was performed to evaluate the NRER and LRER around Co-N₄ on different substrates. Fig. 1a shows the optimized geometric models: Co-NC/MXene and Co-NC/G are simulated by a Co-N₄-embedded monolayer graphene respectively supported on a Ti₃C₂T_x MXene layer and a monolayer graphene. It is shown that the charge density distribution (CDD) mainly concentrated on Co atom and the first coordination shell of N atoms in Co-NC/G, and suggests the near-range electronic effect. In Co-NC/MXene, besides NRER, the electron depletion of surrounding C atoms in Co-NC layer is induced by the electron accumulation of MXene, and expands to remote coordination microenvironment across C atoms more than 5, indicating the long-range electronic effect. Specifically, the LRER arises from the transfer of ~0.70 |e| from Co-NC layer to MXene support in Co-NC/MXene, about 6 times larger than the transferred charge of ~0.06 |e| from Co-NC layer to graphene support in Co-NC/G. Bader charge analysis (Fig. S1 in Supporting information) reveals a more electron-deficient Co center as well as the coordinated N atoms in Co-NC/MXene as compared with those in Co-NC/G. The total density of states (TDOS) shows a significant increase at Fermi level of Co-NC layer in Co-NC/MXene than that in Co-NC/G (Fig. 1b), and indicates that introducing Ti₃C₂T_x MXene support induces more active electrons to promote the electrochemical reaction. Hence, the electrochemical activity of Co-NC/MXene heterostructure would be enhanced by the widespread LRER. Moreover, Co-NC/MXene has more negative of interlayer binding energies (E_b, Table S1 in Supporting information) than Co-NC/G, and demonstrates more stable 2D heterostructure to stand up against the volume changes during charge/discharge.

  Figure 1

  

  Figure 1.  Theoretical analyses of the electronic structure of the catalysts and their interactions with polysulfides. (a) Optimized geometric structure of Co-NC/MXene and Co-NC/G, and the corresponding CDD as the front view, where cyan and yellow colors represent the depletion and accumulation of electron densities, respectively (isosurfaces: 0.001 e/bohr³). (b) TDOS of Co-NC layer in Co-NC/G and Co-NC/MXene. (c) PDOS of Co-d orbitals for different structure, the red arrow and black solid line represent the positions of d-band center and Fermi level (EF), respectively. (d) PDOS of Co-d and S-p states after Li₂S interacting with Co-NC/G and Co-NC/MXene.

  DownLoad: Full-Size Img PowerPoint

  The NRER and LRER bring dual active sites with promoted adsorptions of Co SACs to LiPSs in catalytic SRR. The d-band center of Co can serve as an effective descriptor of catalytic activity towards LiPSs [24]. As shown in Fig. 1c, the projected density of states (PDOS) analysis exhibits an upward shift of the d-band center in Co-NC/MXene (−0.55 eV) compared to Co-NC/G (−0.91 eV). The shift results in a decrease in antibonding orbital electron occupation to enhance the interactions of Co-NC/MXene with LiPSs [25]. As the adsorption of Li₂S, the p orbitals of S atoms mainly hybridize with d_z² and d_xz/_yz orbitals of Co atoms in Co-NC/G (Fig. 1d). By contrast, the hybridization of S-p orbital in Co-NC/MXene proceeds not only with d_z² and d_xz/_yz orbitals of Co atoms, but also with the horizontal d_x²-y² and d_xy orbitals. The increased orbital hybridization necessitates lower energy for the formation and evolution of polysulfide intermediates.

  The computational results indicate that Co-NC/MXene heterostructures are imparted with the thiophilic and lithiophilic sites for catalytic SRR. Importantly, MXene-mediating NRER and LRER drives the Co-NC/MXene to be more stable and active than the counterpart of Co-NC/G. The merits motivate the construction of Co-NC/MXene heterostructures to boost SRR kinetics feasibly.

  We developed a reliable method to support Co-NC on MXene substrates that are further assembled into 3D hierarchical porous network. Fig. 2a illustrates the synthetic procedure of 3D Co-NC/MXene. Ti₃C₂T_x MXene was first prepared by etching Ti₃AlC₂ (MAX) with HCl and LiF, and a subsequent exfoliation process. And Co-IL was synthesized by coordination reaction between 1‑butyl‑3-methimidazolium chloride ([Bmim]Cl) and cobalt chloride (CoCl₂) (see equation in Fig. S2 in Supporting information), in which the anions and cations are combined by hydrogen bonding and electrostatic interactions. The Co-IL was utilized as precursor to disperse and anchor the isolated Co atoms on MXene by ligand separations, therein, the optimal alkyl chain length is required to fulfill densely-distributed Co single atoms. Then, a hybrid hydrogel is generated by simply mixing MXene and Co-IL, as presented by optical photos in Fig. S3 (Supporting information). The control samples of CoCl₂/MXene and [Bmim]Cl/MXene were separately prepared by blending CoCl₂ and [Bmim]Cl with MXene suspension in the same molar ratio as Co-IL/MXene. It is found that weak gelation can be observed for [Bmim]Cl/MXene, and weaker for CoCl₂/MXene, indicating the crucial role of Co-IL precursor during gelation process. To study the gelation mechanism, Fourier transform infrared (FTIR) spectroscopy was performed on Ti₃C₂T_x MXene, Co-IL and Co-IL/MXene (Fig. S4 in Supporting information). The transmission bands around 3200–2800, 1650–1440, and 1200–1100 cm^-1 are assigned to C—H stretching vibrations in [Bmim]⁺, skeleton vibrations of the imidazolium, and the C—C-N vibration, respectively [26]. After assembling with MXene, the peak intensity indicative for the C—H stretching and imidazolium vibration decreases obviously compared with C—C-N vibration. That is because the π–π interaction between [Bmim]⁺ and MXene, which has been examined in our previous work [27]. Together with the interactions between [Bmim]⁺ and [CoCl₄]^2- in Co-IL, MXene nanosheets were linked into 3D Co-IL/MXene network. Finally, thermal treatment turns Co-IL into N-doped carbons incorporated with Co single atoms, which closely contacts with MXenes and produces abundant interfaces, thereby obtaining 3D Co-NC/MXene.

  Figure 2

  

  Figure 2.  The synthesis and representative electron microscopy images of 3D Co-NC/MXene. (a) Schematic of synthesis, (b) SEM image, (c) TEM image, (d) high-resolution TEM image, and (e) the aberration-corrected HAADF-STEM of Co-NC/MXene. (f) HAADF-STEM image and color mapping show the distribution of Ti, O, C, N, Co element, respectively.

  DownLoad: Full-Size Img PowerPoint

  The morphology evolution of Co-NC/MXene was examined by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). As presented in Fig. S5 (Supporting information) and Fig. 2b, 2D nanosheets interweave together to build a 3D Co-IL/MXene network which is well-maintained during thermal reduction to obtain 3D Co-NC/MXene. The TEM images in Fig. 2c and Fig. S6 (Supporting information) show that Co-IL/MXene and Co-NC/MXene display more cross-section morphologies of 2D nanosheets which possesses ultrathin property and intersects each other, but nanosheets in controlled Ti₃C₂T_x MXene are mostly lying flat. It further attests the gelation process between Co-IL and MXene to generate interwoven networks. High-resolution TEM image (Fig. 2d) shows the (110) plane of Ti₃C₂T_x while no crystal structure of Co species can be detected. To validate atomic dispersion of Co in 3D Co-NC/MXene, aberration-corrected HAADF-STEM was performed. As displayed in Fig. 2e, isolated bright dots exist on and outside (110) plane of Ti₃C₂T_x, and demonstrate single atom Co sites. High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and corresponding energy-dispersive spectroscopy (EDS) elemental mapping further confirm the existence of Co element, and suggest a homogeneous distribution of Ti, O, C, N, and Co across the nanosheets (Fig. 2f). Therefore, Co single atoms embedded in N-doped carbons are successfully supported on MXene surfaces, and assemble 3D Co-IL/MXene network.

  The N₂ absorption-desorption isotherm analysis was then performed to determine the network structure of 3D Co-NC/MXene. As presented in Fig. S7 (Supporting information), Co-NC/MXene possesses a specific surface area (SSA) of ~75.5 m²/g and pore volume (PV) of 0.2151 cm³/g, much higher than CoCl₂/MXene derived Co/MXene (SSA: ~15.1 m²/g, PV: ~0.0589 cm³/g), [Bmim]Cl/MXene derived NC/MXene (SSA: ~35.2 m²/g, PV: ~0.1433 cm³/g) and original Ti₃C₂T_x (SSA: ~14.0 m²/g, PV: ~0.0546 cm³/g). Importantly, Co-IL derived carbons contribute to plentiful mesopores, as illustrated by pore size distribution plots of Co-NC/MXene. This meso/macroporous structure and increased SSA superiority of Co-NC/MXene is conducive to the formation of ample active heterointerfaces, effectively carrying sulfur, physically confining LiPSs and creating a continuous pathway for the transport of electrons. X-ray diffraction (XRD, Fig. S8 in Supporting information) pattern of Co-NC/MXene shows a negative shift of (002) diffraction peak compared to pure MXene, indicating an expansion of interlayer spaces after Co-NC introduction. No characteristic peaks associated with metallic Co or Co-based composites can be identified from XRD pattern. However, without the introduction of [Bmim]Cl, the XRD pattern of Co/MXene (Fig. S9 in Supporting information) show diffraction peaks assigned to Co (JCPDS No. 05–0727) and anatase TiO₂ (JCPDS No. 73–1763). That is, Co-IL precursor forms a protective layer on MXene supports to block outside oxygen and water from causing decomposition damage to MXene, and produces the well-distributed Co SACs during the thermal treatment.

  X-ray photoelectron spectroscopy (XPS) demonstrates the composition and chemical state of elements in Co-NC/MXene, and verifies the special MXene-mediating NRER and LRER. Full survey spectrum (Fig. S10 in Supporting information) confirms the presence of Ti, Co, O, N, C, and F elements, and their contents are listed in Table S2 (Supporting information). The content of N is about 4 times that of Co, consistent with their atomic ratio in the Co-IL precursor, which can be attributed to the strong interaction between N and Co, and the atom economic thermal-nitridation. The Co-NC/G was prepared as control by assembling Co-IL with graphene and a subsequent calcination process according to our previous work [28]. The XPS survey spectrum of Co-NC/G reveals the existence of Co element, but no crystal structures of Co species were detected by TEM images or XRD pattern (Figs. S11 and S12 in Supporting information), indicating atomically dispersed Co in Co-NC/G. High-resolution Co 2p spectra (Fig. 3a) verify that the Co component is present in oxidation states rather than metallic states in both Co-NC/MXene and Co-NC/G [14]. Relative to Co-NC/G, Co-NC/MXene shows the positive shifts of ~0.7 eV for the binding energies of Co 2p_3/2 and Co 2p_1/2, and of ~0.8 eV for N 1s spectra (Fig. 3b) [29]. This testifies the more electron-deficient Co and N atoms in Co-NC/MXene induced by the concerted NRER and LRER, and is expected to alter the adsorption and conversion of LiPSs, agreeing well with the calculation results.

  Figure 3

  

  Figure 3.  Structural characterizations of 3D Co-NC/MXene. XPS spectra of (a) Co 2p, (b) N 1s and (c) C 1s for Co-NC/MXene and Co-NC/G. (d) Co K-edge XANES spectra and (e) corresponding k³-weighted FT of EXAFS spectra. (f) Fitting curves of the EXAFS in R-space and optimized coordination configuration of Co-NC/MXene and Co-NC/G. (g) WT for k³-weighted EXAFS signals at Co K-edge.

  DownLoad: Full-Size Img PowerPoint

  Further, LRER of MXene is clarified by the high-resolution C 1s spectra (Fig. 3c). Except a ~0.2 eV positive shift of C—C and C—O species, the binding energy of C—N species in Co-NC/MXene, which originate from Co-IL derivatives, is ~0.5 eV higher than that in Co-NC/G, verifying the LRER occurrence to deplete the electrons of surrounding C atoms in Co-NC/MXene. In addition, Fig. S13 (Supporting information) presents high-resolution XPS in Ti 2p region, which is quite similar to typical Ti₃C₂T_x MXene, confirming the well-maintained structure of Ti₃C₂T_x protected by Co-IL precursor [30].

  The electronic state and coordination environment of Co in both Co-NC/MXene and Co-NC/G were further probed by X-ray absorption spectroscopy (XAS). As shown in Co K-edge X-ray absorption near-edge structure (XANES) spectra (Fig. 3d), both of Co-NC/MXene and Co-NC/G exhibit an edge energy between that of Co foil and CoO, revealing the valance state of Co between 0 and +2. However, the edge energy of Co-NC/MXene shows a positive shift in comparison with the case of the Co-NC/G, implying its oxidation state is increased by LRER, which is consistent with the XPS and calculation results. Fig. 3e displays the k³-weighted Fourier transform (FT) of extended X-ray absorption fine structure (EXAFS) spectra. Only one prominent peak centered at 1.4 Å appears for the Co-NC/MXene network, which is similar to Co-NC/G, ascribing to the scattering of Co−N coordination. No Co−Co (2.2 Å) or Co−O (1.7 Å) peak can be observed, verifying that the single-atom Co sites are coordinated by N species in carbons. EXAFS fittings were also carried out and the result (Fig. 3f, Fig. S14 and Table S3 in Supporting information) indicates that the coordination number and bond length of Co−N are 4.05 and 1.91 Å in Co-NC/MXene, 3.79 and 1.89 Å in Co-NC/G, respectively. It means that the Co sites adopts the same Co−N₄ configuration in both Co-NC/MXene and Co-NC/G, in accordance with the ratio in XPS results. As compared to that on carbon support, meanwhile, LRER from the surrounding C atoms causes Co−N bond on MXene support to be stretched by 1.1%. Wavelet transform (WT) EXAFS spectra were used to determine Co local coordination environment both in radial distance and k-space. As shown in Fig. 3g, both Co-NC/MXene and Co-NC/G show a contour plot similar to that of CoPc with a maximum intensity at ~6 Å^-1, and no Co−Co (8.1 Å^-1) or Co−O (7.5 Å^-1) coordination can be detected when compared to the WT contour plots of Co foil and CoO. The results confirm that the Co species are atomically dispersed in Co−N₄−C structure for both Co-NC/MXene and Co-NC/G. Inductively coupled plasma mass spectroscopy (ICP-MS) analysis reveals the Co loading is 2.15 wt% in as-prepared Co-NC/MXene and 2.65 wt% in Co-NC/G (Table S4 in Supporting information). Considering the larger SSA of 350.8 m²/g for Co-NC/G, the density of Co single atom in Co-NC/G is only 1/3 of that in Co-NC/MXene (Fig. S15 in Supporting information).

  Thus, introducing MXene substrates not only initiates NRER and LRER on Co SACs, but also enables to obtain the densely-populated Co single atoms, which can serve as high-efficient sulfur redox mediator.

  To analyze the catalytic mechanism of the concerted NRER and LRER from Co-NC/MXene, adsorption towards LiPSs was first evaluated by soaking catalysts in Li₂S₆/dioxolane (DOL)/dimethoxyethane (DME) solution for 8 h (Fig. 4a). More obvious decolorization of Li₂S₆ solution can be identified for Co-NC/MXene than for Co-NC/G, indicating stronger LiPSs affinity of Co-NC/MXene. Then, the Li₂S₆-absorbed Co-NC/MXene precipitate was collected for XPS analyses. The XPS survey (Fig. S16 in Supporting information) clearly shows the peaks for Li 1s and S 2p, indicating a successful adsorption of Co-NC/MXene towards Li₂S₆. Li atoms primarily form Li-N bonds, as validated by the XPS peaks at 55.6 eV (Fig. 4b). Note that Co-NC/G-Li₂S₆ also shows Li-N bond similar to Co-NC/MXene-Li₂S₆, but with a higher binding energy, suggesting the LRER in Co-NC/MXene weakens the interaction of N atoms with LiPSs. In the high-resolution S 2p spectra (Fig. 4c), the peaks at binding energies of 161.4, 163.1, 166.7 and 168.7 eV are ascribed to bridging sulfur (S_B⁰), terminal sulfur (S_T^-1), thiosulfate and polythionate, respectively [31]. Remarkably, Co-NC/MXene can spontaneously split Li₂S₆ into thiosulfate to a large extent. The resultant thiosulfate can further capture newly formed long-chain LiPSs to generate polythionates and initiate their conversion to short-chain LiPSs during reduction process [32], but polythionates is almost absent in Co-NC/G-Li₂S₆. Moreover, compared with Li₂S₆, Co-NC/G causes positive shifts in binding energies of S_B⁰ and S_T^-1 and implies strong adsorption to Li₂S₆. However, the shift of S_B⁰ or S_T^-1 in Co-NC/MXene is less pronounced than that in Co-NC/G, which is associated with stronger interactions of Co sites with existing thiosulfate/polythionate. Therefore, LRER effectively mediates LiPSs absorption and chemical conversion in Co-NC/MXene, and leads to the high catalytic efficiency of Co SACs in SRR.

  Figure 4

  

  Figure 4.  Interaction of 3D Co-NC/MXene catalyst with LiPSs and electrocatalytic SRR activity. (a) Photograph of the samples soaked in a Li₂S₆/DOL/DME solution after 8 h, XPS spectra of (b) Li 1s and (c) S 1s for Li₂S₆-absorbed Co-NC/MXene. (d) The calculated adsorption energy (E_a) of various Li₂S_n (n = 1, 2, 4, 6 and 8) and S₈ species on Co-NC/G and Co-NC/MXene. (e) Free energy diagram for the conversion of LiPSs on Co-NC/G and Co-NC/MXene. (f) Li₂S deposition profile. (g) CV curves of Li₂S₆ symmetric cells using different active materials. (h) EIS spectra of LSBs based on Co-NC/G/S and Co-NC/MXene/S cathodes. (i) TDOS of Co-NC/G and Co-NC/MXene models.

  DownLoad: Full-Size Img PowerPoint

  The optimized configurations of Li₂S₆ anchored on Co-NC/MXene confirm that the absorption results from thiophilic Co and lithiophilic N sites (Fig. S17 in Supporting information), agreeing well with the XPS results. Additionally, the optimized configurations of various LiPSs and sulfur adsorbed on Co-NC/G and Co-NC/MXene were simulated (Figs. S17 and S18 in Supporting information), and gave the corresponding adsorption energies (E_a, Fig. 4d). The Co-NC/MXene demonstrates more negative E_a values than Co-NC/G, suggesting LRER-mediating superior capability to anchor polysulfides in Co-NC/MXene, and to prevent the dissolution and diffusion of LiPSs, consistent with the results in visual adsorption test.

  It is followed that Co SACs Co-NC/MXene are brough with the favorable kinetics in catalytic SRR through MXene-mediating LRER. DFT calculation demonstrates that the discharging process is thermodynamically more favorable on Co-NC/MXene than on Co-NC/G (Fig. 4e). The largest increase of Gibbs free energy is obtained for the reduction from Li₂S₂ to Li₂S, which is the rate-determining step (RDS) for the whole discharge process. Accordingly, Co-NC/MXene shows a relatively lower barrier in the RDS than Co-NC/G, indicating the conversation of LiPSs to Li₂S is kinetically more beneficial. Experimentally, the Li₂S nucleation behaviors of different catalysts was studied by a potentiostatic method at 2.05 V after the first galvanostatic discharging at 0.112 mA. As shown in Fig. 4f, dark-color part in the profiles refers to the reduction of Li₂S₈ and Li₂S₆, and light-color area corresponds to the Li₂S nucleation [33]. The Co-NC/MXene exhibits higher capacity (78 mAh/g) of Li₂S precipitation than that of Co-NC/G (51 mAh/g) according to Faraday's law. Moreover, the Co-NC/MXene electrode displays earlier current peaks and more intense current response. This proves that the nucleation behavior of Li₂S is greatly promoted by Co-NC/MXene, corresponding with the results of theoretical calculations. Cyclic voltammetry (CV) tests of symmetrical batteries containing Li₂S₆ were then performed to study the catalysis of Co-NC/MXene on SRR kinetics (Fig. 4g). Two dominant peaks appear during cathodic and anodic scan, which can be attributed to the reduction of Li₂S₆ to Li₂S₂/Li₂S and reversible oxidation of Li₂S₂/Li₂S to Li₂S₆ at the electrodes [34]. Compared to Co-NC/G, Co-NC/MXene illustrates higher current response, indicating higher electrocatalytic activity towards polysulfide conversion.

  To directly evaluate the concerted NRER and LRER impact of electrocatalysts on device performances, coin cells were assembled from different cathodes with a sulfur loading of ~68 wt% (Fig. S19 in Supporting information). The CV tests of Co-NC/MXene and Co-NC/G based sulfur cathodes were performed at a scan rate of 0.05 mV/s between 1.7–2.8 V (Fig. S20 in Supporting information). Two reduction peaks appear at 2.2–2.3 and 1.9–2.1 V during the discharge process, while one oxidation peak appears at ~2.4 V during charge process, corresponding to the formation of soluble LiPSs, insoluble Li₂S and decomposition of Li₂S, respectively [35]. It is found that Co-NC/MXene/S electrode shows larger current response and smaller voltage polarization as compared to Co-NC/G/S electrode, indicative of the superior catalytic conversions to alleviate the polarization. The electrochemical impedance spectroscopy (EIS) curves of the electrodes and the corresponding equivalent circuit accessories are shown in Fig. 4h and Fig. S21 (Supporting information). The semicircle diameter at a medium and high frequency corresponds to the charge transfer resistance (R_ct). According to the equivalent circuit fitting results, Co-NC/MXene/S cathodes show much smaller R_ct value (127 Ω) than that of Co-NC/G/S cathodes (202 Ω). It reveals that Co-NC/MXene-mediated sulfur cathodes possess shortened ion/electron diffusion path for fast mass/charge transfer process, which is related with the well-designed heterointerface and 3D porous structure, leading to improved conductivity and electrocatalytic activity than Co-NC/G/S cathodes. Besides, the superior activity is also demonstrated by the increase in DOS (Fig. 4i) near Fermi level of Co-NC/MXene compared to Co-NC/G, confirming that the Co-NC layer supported on MXene can accelerate charge transfer. The diffusion coefficient of Li⁺ (D_Li) can be calculated based on equation of $D_{\mathrm{Li}}=\frac{R^2 T^2}{2 A^2 n^4 F^4 C^2 \sigma^2}$ where R, T, A, n, F, C, and σ are the gas constant, room temperature, surface area, number of transfer electrons per molecule, Faraday constant, concentration of Li⁺ in electrode, and slope of Z'~ω^−0.5, respectively. As shown in Fig. S22 (Supporting information), the Co-NC/MXene/S cell shows smaller σ than Co-NC/G/S cell, and the rapid mass transfer benefits from the MXene interlayers owing to LRER. Therefore, LRER drives Co-NC/MXene catalysts more favorable for interfacial kinetics than Co-NC/G.

  The battery performances of Co-NC/MXene/S cathodes were further evaluated. Fig. 5a presents the charge/discharge curves of Co-NC/MXene and Co-NC/G based sulfur cathodes at the initial cycle in a voltage window of 1.7–2.8 V. Both cathodes show two plateaus during discharge, which is attributed to the stepwise reduction from S₈ to long-chain Li₂S₈ at ~2.3 V, then to short-chain Li₂S₄, further to Li₂S₂/Li₂S at ~2.1 V, and finally to Li₂S, as well as an oxidation reaction from Li₂S to S₈ during reversible charging, agreeing well with the CV results [36]. Under an area sulfur loading of ~1.5 mg/cm² and electrolyte/sulfur (E/S) ratio of 10 mL/g, the Co-NC/MXene/S cathodes display a much higher discharge capacity of 1402.5 mAh/g at 0.2 C compared with 995.9 mAh/g for Co-NC/G/S cathodes. Moreover, lower potential difference was obtained for Co-NC/MXene/S (141 mV) between anodic and cathodic scan at 0.2 C compared with Co-NC/G/S (201 mV). Even though increased current density (from 0.2 C to 4 C) causes more severe polarization, Co-NC/MXene/S still shows narrower voltage gaps of the two cathodes due to its superior catalytic activity for accelerated redox kinetics (Fig. 5b). A lower Li₂S nucleation overpotential of 18 mV for Co-NC/MXene/S (Fig. 5c) than that of 39 mV for Co-NC/G/S further verifies the fast kinetics, which is in accordance with results from the Li₂S nucleation tests and the DFT calculations [37].

  Figure 5

  

  Figure 5.  Electrochemical performances of the Co-NC/MXene/S cell. (a) Charge/discharge curves at 0.2 C of Co-NC/MXene and Co-NC/G based sulfur cathodes. (b) Potential difference between the anodic and cathodic scan at different C rates. (c) Magnified discharge curves showing the overpotential of Li₂S nucleation. (d) Rate capability and (e) cycling stability at 0.5 C of Co-NC/MXene, Co-NC/G, NC/MXene and Co/MXene based sulfur cathodes. (f) Long-term cycling stability of Co-NC/MXene based sulfur cathodes at 4 C. (g) Cycling performance of Co-NC/MXene based sulfur cathodes with an increased sulfur loading of 3.8 mg/cm². (h) Comparison of the stability of Co-NC/MXene with reported sulfur cathodes based on metal single atoms (1 C = 1675 mA/g).

  DownLoad: Full-Size Img PowerPoint

  Rate performance was evaluated under different current densities from 0.2 C through 4.0 C and back to 0.2 C. As displayed in Fig. 5d, the Co-NC/MXene/S cell shows outstanding rate capability with average capacities of 1346, 1079, 881, 783 and 637 mAh/g at the current density of 0.2, 0.5, 1, 2 and 4 C, respectively. When the current density switches back to 0.2 C, a reversible capacity of 1276 mAh/g can be restored, signifying high stability of Co-NC/MXene. By contrast, the Co-NC/G exhibit inferior capacities, especially at high rates. Furthermore, NC/MXene/S and Co/MXene/S cell were fabricated, whose rate performances are also inferior to that of Co-NC/MXene/S, highlighting the great contribution of SACs.

  High-efficiency SRR process significantly mitigates the shuttling effect of LiPSs, thus leading to enhanced cycling performance. As displayed in Fig. 5e, the Co-NC/MXene/S cathode shows excellent cycling stability with much lower capacity decay of 0.026% per cycle at 0.5 C for 100 cycles, compared with that of 0.168% for Co-NC/G/S, 0.116% for NC/MXene/S and 0.118% for Co/MXene/S, respectively. The long-term cycling performance of Co-NC/MXene was evaluated at a high current density of 4 C (Fig. 5f). A capacity retention of 90.5% with a low-capacity decay rate of 0.015% per cycle can still be obtained after cycling over 600 cycles. The SEM image shows well-preservation of the 2D lamellar structure and 3D networks after cycling, indicating a robust architecture of Co-NC/MXene (Fig. S23 in Supporting information). High sulfur loading and low E/S ratio are the key to achieve high energy density in the practical application of LSBs [38]. When S loading was further increased to ~3.8 mg/cm², the resultant battery using Co-NC/MXene-supported S cathode and low E/S ratio of 7 mL/g displays the highest capacities at different rates compared with other cathodes (Fig. S24 in Supporting information). It shows reversible areal capacities of 5.0 mAh/cm² at 0.2 C, which exceeds commercial lithium-ion batteries (4.0 mAh/cm²) and achieves stable SRR electrochemistry under 100 cycles (Fig. 5g and Fig. S25 in Supporting information). Such excellent cycling stability outperforms most advanced sulfur cathodes (Fig. 5h) [39-43].

  Combining the experimental and calculation results, the remarkable rate and cycling performances of Co-NC/MXene/S cathode mainly come from three levels: (1) The Co-IL was rationally selected as precursor to prepare the well-stabilized and densely-distributed active Co-N₄ sites, protects MXene supports from decomposition, and initiates the assembling of 3D network through its special interactions with MXene support; (2) The abundant and stable 2D heterointerfaces of Co-NC layer with MXene support induce a widespread NRER and LRER, which achieves more favorable electronic structure to tune the adsorption and conversion of LiPSs, and accordingly results in higher catalytic activity and stability than carbon substrate loaded Co-NC; (3) The as-obtained 3D Co-NC/MXene network is equipped with highly active surfaces, simultaneous lithiophilicity and sulfiphilicity, continuous conductive channel, and hierarchical porosity, which increases sulfur utilization and promotes mass transfer to obtain LSBs with high capacities and superior stability. Therefore, impressive electrochemical performances have been achieved, and demonstrates the superiority of well-designed structure of highly active and stable Co-NC/MXene catalyst.

  In summary, we have developed a facile Co-IL assisted strategy to construct novel Co-NC/MXene catalysts, in which N-doped carbons can anchor the densely-distributed active sites of Co-N₄ on MXene substrates and further assemble them into 3D porous network. Compared with Co-NC/G, the far-reaching LRER of surrounding C atoms is imparted by interlayer charge transfer between Co-NC layer and MXene support. It is demonstrated that the concerted NRER and LRER around Co SACs contribute significantly to catalyze SRR from LiPSs. In mechanism, the MXene-mediating LRER weakens the adsorption of N sites to Li species, and readily initiate the dissociation of polysulfides and favor the conversion kinetics. Besides the enhanced catalytic activity, LRER induces robust coupling of heterostructures to give stable 3D Co-NC/MXene networks, hence, Co-NC/MXene offers continuous conductive pathways and rapid mass transfer channels to boost adsorption-catalysis in SRR process. Consequently, the Co-NC/MXene-based sulfur cathodes demonstrate high reversible capacities of 1346 mAh/g at 0.2 C and excellent rate performance to deliver a capacity of 637 mAh/g even at high current density of 4 C. Furthermore, the obtained LSBs exhibit an outstanding cycling stability with capacity decay of merely 0.015% per cycle at 4 C for 600 cycles, and superior areal capacities of 5.0 mAh/cm² under a high sulfur loading and a low E/S ratio. The electrochemical performances of Co-NC/MXene/S are superior to Co-NC/G/S, NC/MXene/S and Co/MXene/S, and also outperform most advanced sulfur cathodes. This work reveals the importance of charge transfer-induced LRER for improved catalysis for LSBs and provides guidance for future structural design of highly efficient SACs in various catalytic process.

  Declaration of competing interest

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

  CRediT authorship contribution statement

  Haiyan Wang: Writing – original draft, Methodology, Investigation, Funding acquisition, Formal analysis, Conceptualization. Hucheng Zhang: Writing – review & editing, Resources, Methodology, Investigation, Funding acquisition. Lijing Wang: Software, Investigation, Formal analysis, Data curation. Yonghui Li: Software, Investigation, Formal analysis, Data curation. Tianhao Zhang: Software, Formal analysis, Data curation. Zhansheng Lu: Resources, Methodology, Investigation, Funding acquisition. Hao Jiang: Writing – review & editing. Chunzhong Li: Writing – review & editing. Jianji Wang: Writing – review & editing, Supervision.

  Acknowledgments

  This work was supported by the National Natural Science Foundation of China (Nos. 21573059, 12274118 and 22208088), Henan Center for Outstanding Overseas Scientists (No. GZS2023007), Special Project for Fundamental Research in University of Henan Province (No. 22ZX013). The simulations were performed on resources provided by the High-Performance Computing Center of Henan Normal University. The authors would like to thank Shiyanjia Lab (www.shiyanjia.com) for the XAS analysis.

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110373.

  
  
• 1. [1]

     G. Zhou, H. Chen, Y. Cui, Nat. Energy 7 (2022) 312–319. doi: 10.1038/s41560-022-01001-0

  2. [2]

     M. Zhao, H.J. Peng, B.Q. Li, J.Q. Huang, Acc. Chem. Res. 57 (2024) 545–557. doi: 10.3390/systems12120545

  3. [3]

     Y. Song, L. Song, M. Wang, W. Cai, Sci. Bull. 69 (2024) 2013–2016. doi: 10.1016/j.scib.2024.04.024

  4. [4]

     Y. Liu, M. Zhao, L.P. Hou, et al., Angew. Chem. Int. Ed. 62 (2023) e202303363. doi: 10.1002/anie.202303363

  5. [5]

     Y.W. Song, L. Shen, N. Yao, et al., Angew. Chem. Int. Ed. 63 (2024) e202400343. doi: 10.1002/anie.202400343

  6. [6]

     Z. Shen, X. Jin, J. Tian, et al., Nat. Catal. 5 (2022) 555–563. doi: 10.1038/s41929-022-00804-4

  7. [7]

     X.Y. Li, S. Feng, M. Zhao, et al., Angew. Chem. Int. Ed. 61 (2022) e202114671. doi: 10.1002/anie.202114671

  8. [8]

     Q. Yang, J. Cai, G. Li, et al., Nat. Commun. 15 (2024) 3231. doi: 10.1038/s41467-024-47565-1

  9. [9]

     Y. Song, X. Wei, L. Song, et al., Energy Storage Mater. 70 (2024) 103491. doi: 10.1016/j.ensm.2024.103491

  10. [10]

      L. Chen, Y. Sun, X. Wei, et al., Adv. Mater. 35 (2023) 2300771. doi: 10.1002/adma.202300771

  11. [11]

      Z. Liang, J. Shen, X. Xu, et al., Adv. Mater. 34 (2022) e2200102. doi: 10.1002/adma.202200102

  12. [12]

      Y. Zhang, C. Kang, W. Zhao, et al., J. Am. Chem. Soc. 145 (2023) 1728–1739. doi: 10.1021/jacs.2c10345

  13. [13]

      F. Zhang, Z. Tang, L. Zheng, et al., Appl. Catal. B: Environ. 334 (2023) 122876. doi: 10.1016/j.apcatb.2023.122876

  14. [14]

      H. Tian, H. Tian, S. Wang, et al., Nat. Commun. 11 (2020) 5025. doi: 10.1038/s41467-020-18820-y

  15. [15]

      Y. Qiu, L. Fan, M. Wang, et al., ACS Nano 14 (2020) 16105. doi: 10.1021/acsnano.0c08056

  16. [16]

      W. Li, C. Liu, C. Gu, et al., J. Am. Chem. Soc. 145 (2023) 4774–4783. doi: 10.1021/jacs.2c13596

  17. [17]

      S. Zhang, X. Ao, J. Huang, et al., Nano Lett. 21 (2021) 9691–9698. doi: 10.1021/acs.nanolett.1c03499

  18. [18]

      Y. Zhang, L. Jiao, W. Yang, C. Xie, H.L. Jiang, Angew. Chem. Int. Ed. 60 (2021) 7607–7611. doi: 10.1002/anie.202016219

  19. [19]

      J. Yang, Z. Wang, C.X. Huang, et al., Angew. Chem. Int. Ed. 60 (2021) 22722–22728. doi: 10.1002/anie.202109058

  20. [20]

      L. Jiao, J. Zhu, Y. Zhang, et al., J. Am. Chem. Soc. 143 (2021) 19417–19424. doi: 10.1021/jacs.1c08050

  21. [21]

      C. Hu, Y. Zhang, A. Hu, et al., Adv. Mater. 35 (2023) 2209298. doi: 10.1002/adma.202209298

  22. [22]

      Y. Ding, T. Yan, J. Wu, et al., Appl. Catal. B: Environ. 343 (2024) 123553. doi: 10.1016/j.apcatb.2023.123553

  23. [23]

      S. Gbadamasi, M. Mohiuddin, V. Krishnamurthi, et al., Chem. Soc. Rev. 50 (2021) 4684–4729. doi: 10.1039/d0cs01070g

  24. [24]

      H. Wang, S. Xu, C. Tsai, et al., Science 354 (2016) 1031–1036. doi: 10.1126/science.aaf7680

  25. [25]

      Z. Han, S. Zhao, J. Xiao, et al., Adv. Mater. 33 (2021) e2105947. doi: 10.1002/adma.202105947

  26. [26]

      D. Wei, F. Ma, R. Wang, et al., Adv. Mater. 31 (2018) 1707583.

  27. [27]

      H. Wang, L. Wang, H. Zhang, et al., Chem. Eng. J. 431 (2021) 134179.

  28. [28]

      H. Wang, H. Du, H. Zhang, et al., J. Energy Chem. 76 (2023) 67–74. doi: 10.3390/engproc2023055067

  29. [29]

      R. Chu, T.T. Nguyen, Y. Bai, N.H. Kim, I.H. Lee, Adv. Energy Mater. 12 (2022) 2102805. doi: 10.1002/aenm.202102805

  30. [30]

      D.K. Lee, Y. Chae, H. Yun, C.W. Ahn, J.W. Lee, ACS Nano 14 (2020) 9744–9754. doi: 10.1021/acsnano.0c01452

  31. [31]

      Y. Wang, R. Zhang, J. Chen, et al., Adv. Energy Mater. 9 (2019) 1900953. doi: 10.1002/aenm.201900953

  32. [32]

      Z. Xiao, Z. Yang, Z. Li, P. Li, R. Wang, ACS Nano 13 (2019) 3404–3412. doi: 10.1021/acsnano.8b09296

  33. [33]

      H. Gu, W. Yue, J. Hu, et al., Adv. Energy Mater. 13 (2023) 2204014. doi: 10.1002/aenm.202204014

  34. [34]

      Y. Li, X. Lei, Y. Yuan, et al., ACS Cent. Sci. 6 (2020) 1827–1834. doi: 10.1021/acscentsci.0c00899

  35. [35]

      S. Gu, H. Jiang, X. Li, et al., Energy Storage Mater. 53 (2022) 32–41. doi: 10.1016/j.ensm.2022.08.048

  36. [36]

      X. Zhou, R. Meng, N. Zhong, et al., Small Methods 5 (2021) e2100571. doi: 10.1002/smtd.202100571

  37. [37]

      T. Feng, T. Zhao, N. Zhang, et al., Adv. Funct. Mater. 32 (2022) 2202766. doi: 10.1002/adfm.202202766

  38. [38]

      M. Zhao, Y.Q. Peng, B.Q. Li, et al., J Energy Chem. 56 (2021) 203–208. doi: 10.1016/j.jechem.2020.07.054

  39. [39]

      Y. Ding, Q. Cheng, J. Wu, et al., Adv. Mater. 34 (2022) e2202256. doi: 10.1002/adma.202202256

  40. [40]

      R. Yan, Z. Zhao, M. Cheng, et al., Angew. Chem. Int. Ed. 62 (2022) e202215414.

  41. [41]

      D. Zhang, S. Wang, R. Hu, et al., Adv. Funct. Mater. 30 (2020) 2002471. doi: 10.1002/adfm.202002471

  42. [42]

      R. Wang, J. Yang, X. Chen, et al., Adv. Energy Mater. 10 (2020) 1903550. doi: 10.1002/aenm.201903550

  43. [43]

      G. Li, W. Qiu, W. Gao, et al., Adv. Funct. Mater. 32 (2022) 2202853. doi: 10.1002/adfm.202202853

• 

  Figure 1  Theoretical analyses of the electronic structure of the catalysts and their interactions with polysulfides. (a) Optimized geometric structure of Co-NC/MXene and Co-NC/G, and the corresponding CDD as the front view, where cyan and yellow colors represent the depletion and accumulation of electron densities, respectively (isosurfaces: 0.001 e/bohr³). (b) TDOS of Co-NC layer in Co-NC/G and Co-NC/MXene. (c) PDOS of Co-d orbitals for different structure, the red arrow and black solid line represent the positions of d-band center and Fermi level (EF), respectively. (d) PDOS of Co-d and S-p states after Li₂S interacting with Co-NC/G and Co-NC/MXene.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  The synthesis and representative electron microscopy images of 3D Co-NC/MXene. (a) Schematic of synthesis, (b) SEM image, (c) TEM image, (d) high-resolution TEM image, and (e) the aberration-corrected HAADF-STEM of Co-NC/MXene. (f) HAADF-STEM image and color mapping show the distribution of Ti, O, C, N, Co element, respectively.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Structural characterizations of 3D Co-NC/MXene. XPS spectra of (a) Co 2p, (b) N 1s and (c) C 1s for Co-NC/MXene and Co-NC/G. (d) Co K-edge XANES spectra and (e) corresponding k³-weighted FT of EXAFS spectra. (f) Fitting curves of the EXAFS in R-space and optimized coordination configuration of Co-NC/MXene and Co-NC/G. (g) WT for k³-weighted EXAFS signals at Co K-edge.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Interaction of 3D Co-NC/MXene catalyst with LiPSs and electrocatalytic SRR activity. (a) Photograph of the samples soaked in a Li₂S₆/DOL/DME solution after 8 h, XPS spectra of (b) Li 1s and (c) S 1s for Li₂S₆-absorbed Co-NC/MXene. (d) The calculated adsorption energy (E_a) of various Li₂S_n (n = 1, 2, 4, 6 and 8) and S₈ species on Co-NC/G and Co-NC/MXene. (e) Free energy diagram for the conversion of LiPSs on Co-NC/G and Co-NC/MXene. (f) Li₂S deposition profile. (g) CV curves of Li₂S₆ symmetric cells using different active materials. (h) EIS spectra of LSBs based on Co-NC/G/S and Co-NC/MXene/S cathodes. (i) TDOS of Co-NC/G and Co-NC/MXene models.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  Electrochemical performances of the Co-NC/MXene/S cell. (a) Charge/discharge curves at 0.2 C of Co-NC/MXene and Co-NC/G based sulfur cathodes. (b) Potential difference between the anodic and cathodic scan at different C rates. (c) Magnified discharge curves showing the overpotential of Li₂S nucleation. (d) Rate capability and (e) cycling stability at 0.5 C of Co-NC/MXene, Co-NC/G, NC/MXene and Co/MXene based sulfur cathodes. (f) Long-term cycling stability of Co-NC/MXene based sulfur cathodes at 4 C. (g) Cycling performance of Co-NC/MXene based sulfur cathodes with an increased sulfur loading of 3.8 mg/cm². (h) Comparison of the stability of Co-NC/MXene with reported sulfur cathodes based on metal single atoms (1 C = 1675 mA/g).

  下载: 全尺寸图片 幻灯片
•