Combined High Catalytic Activity and Efficient Polar Tubular Nanostructure in Urchin‐Like Metallic NiCo2Se4 for High‐Performance Lithium–Sulfur Batteries

Urchin‐shaped NiCo2Se4 (u‐NCSe) nanostructures as efficient sulfur hosts are synthesized to overcome the limitations of lithium–sulfur batteries (LSBs). u‐NCSe provides a beneficial hollow structure to relieve volumetric expansion, a superior electrical conductivity to improve electron transfer, a high polarity to promote absorption of lithium polysulfides (LiPS), and outstanding electrocatalytic activity to accelerate LiPS conversion kinetics. Owing to these excellent qualities as cathode for LSBs, S@u‐NCSe delivers outstanding initial capacities up to 1403 mAh g−1 at 0.1 C and retains 626 mAh g−1 at 5 C with exceptional rate performance. More significantly, a very low capacity decay rate of only 0.016% per cycle is obtained after 2000 cycles at 3 C. Even at high sulfur loading (3.2 mg cm−2), a reversible capacity of 557 mAh g−1 is delivered after 600 cycles at 1 C. Density functional theory calculations further confirm the strong interaction between NCSe and LiPS, and cytotoxicity measurements prove the biocompatibility of NCSe. This work not only demonstrates that transition metal selenides can be promising candidates as sulfur host materials, but also provides a strategy for the rational design and the development of LSBs with long‐life and high‐rate electrochemical performance.


Introduction
The low energy density and relatively high price of traditional lithium-ion batteries (LIBs) are dramatically limiting effective approach is to host sulfur at the cathode in carbonbased materials with high conductivity, such as porous structures of graphene, [8] carbon spheres, [9] carbon nanotubes, [10] and nanofibers. [11] These carbon-based materials can accelerate electron transfer, but are not able to suppress LiPS shuttling due to a weak chemical interaction between nonpolar carbons and polar LiPS. Therefore, LSBs based on carbon suffer from serious capacity fading. [12] On the other hand, polar materials, such as TiO 2 and MnO 2 , strongly bind LiPS and efficiently confine LiPS to the cathode, achieving notable improvements in cycling stability. [13,14] However, such semiconducting oxides are characterized by insufficient electrical conductivities, what results in inferior rate capabilities. In terms of structure, hollow nanomaterials, such as nanotubes, nanospheres or nanocubes, have been demonstrated advantageous in LSBs because of their large pore volumes and surface-to-volume ratios, which mitigate the detrimental effect of the volume expansion and provide an effective physical confinement for LiPS. [5,15] Besides, the use of electrocatalysts have been demonstrated effective to accelerate the conversion of soluble long-chain LiPS into solid phases of sulfur and Li 2 S 2 /Li 2 S. [16][17][18] Overall, high-performance LSB cathodes require materials with excellent electrical conductivity, significant polarity to ensure a strong polysulfide affinity, high catalytic activity toward sulfide redox reactions and with hollow nanostructures to relieve volumetric expansion during charge/discharge (as shown in TOC).
Transition metal sulfides/selenides (TMS/TMSe) have attracted much attention for energy storage in recent years. TMS (e.g., CoS 2 and VS 2 ) have been proved as efficient catalysts in several energy conversion fields such as photovoltaics, solar-light to fuel photoconversion and electrochemical hydrogen evolution. [19,20] Their high catalytic activity has been related to the abundance of defects on the surface of TMS due to the moderate electronegativity differences between transition metals and sulfur, the variable oxidation state of sulfur, and the potential formation of sulfur-sulfur and also metalmetal bonds. [21,22] TMS are also highly stable catalysts in reactions involving sulfur. [23] Besides, TMS have shown a strong bonding ability for LiPS owing to their polar character. [17,24] TMSe display similar crystallographic structures, high defect densities and polar character to TMS owing to the relatively similar electronegativity and ionic radius of S and Se . However, the electrical conductivity of TMSe is much higher than the corresponding TMS. Se is characterized by electrical conductivities (1 × 10 −3 S m −1 ) many orders of magnitude higher than S (5 × 10 −28 S m −1 ). [25] Thus, it is reasonable to speculate that TMSe would be promising hosts for LSBs because of their polarity, potential high catalytic activity, and high electrical conductivity. To our knowledge, this is the first work in which bimetallic selenides are reported as S host for LSBs. NiCo 2 Se 4 (NCSe) was specifically selected as the host material owing to its metallic nature and synergistic effect between Ni/Co atoms. [26,27] The compound was prepared in the form of urchin-like structures through a two-step hydrothermal process. We thoroughly studied the performance of LSBs based on urchin-like NCSe (u-NCSe) both experimentally and though theoretical calculations. Results presented in this manuscript show the benefits of a highly conductive and polar bimetallic selenide with a tubular structure for rapid electron transfer, enhanced confinement of LiPS, mitigation of volume expansion effects, and a catalytic enhancement of the electrochemical reaction kinetics.

Results and Discussion
The synthesis strategy to produce S@u-NCSe is schematically shown in Figure 1 (details can be obtained in the Experimental Section). u-NCSe was produced using two hydrothermal reaction steps. [7,28] In the first step, Ni 0.33 Co 0.67 (CO 3 ) 0.5 OH urchinlike particles having an average diameter of 8-10 µm and containing solid nanoneedles of 200 nm diameter were produced (Figure 2a; Figure S1, Supporting Information). [28] In a second step, such precursor nanostructures were selenized to u-NCSe (Figure 2b,c), which crystallized in the NiCo 2 Se 4 phase, as indicated by XRD (JCPDS No. 81-4821) and HRTEM characterization (Figure 2e,h). [27] u-NCSe displayed hollow tubular structures as observed from SEM and TEM micrographs (Figure 2c,d). The hollow structure was originated from the differential diffusivity of the metals and selenium through the growing NiCo 2 Se 4 shell, via the nanoscale Kirkendall effect. [29,30] The surface of the u-NCSe nanotubes is very rough, which translates into high effective surface areas and provides additional sites for electrochemical reactions as compared to the bulk counterpart (b-NCSe; Figure S2, Supporting Information). Within the experimental error, energy-dispersive spectroscopy (EDS) and electron energy loss spectroscopy (EELS) elemental maps showed the relative atomic content of Ni, Co, and Se to match well with stoichiometric NiCo 2 Se 4 , with the three elements homogeneously distributed within u-NCSe (Figure 2f,g).
X-ray photoelectron spectroscopy (XPS) spectra of the samples exposed to air are shown in Figure S3 (Supporting Information). Ni 2p and Co 2p spectra display two pairs of spinorbit doublets, 2p 3/2 and 2p 1/2 , and two shake-up satellite peaks (marked "Sat."). [31] In the Ni 2p spectra ( Figure S3a, Supporting Information), the peaks located at 853.6 eV (Ni 2p 3/2 ) and www.afm-journal.de www.advancedsciencenews.com 871 eV (Ni 2p 1/2 ) are assigned to Ni 2+ , and the peaks at 856.3 eV (Ni 2p 3/2 ) and 874.2 eV (Ni 2p 1/2 ) to Ni 3+ . [28] Similarly, in the Co 2p spectra ( Figure S3b, Supporting Information), the peaks located at 778.9 eV (Co 2p 3/2 ) and 793.9 eV (Co 2p 1/2 ) are related to Co 3+ and those at 781.2 eV (Co 2p 3/2 ) and 797.5 eV (Co 2p 1/2 ) to Co 2+ . [28] Se 3d peaks are located at 59.3 (Se 3d 3/2 ) and 54.8 eV (Se 3d 5/2 ) in agreement with Se 2− in a metal selenide environment ( Figure S3c, Supporting Information). [32] The XPS spectra show the presence of occupied states at the Fermi level as it corresponds to a metal or a highly degenerated semiconductor ( Figure S3d, Supporting Information). Additionally, the calculated band structure and density of states of NCSe showed no gap of states at the Fermi level, demonstrating its metallic character (Figure 2i). [27] Sulfur was introduced within u-NCSe by a melt-diffusion process (see the Experimental Section for details). The product or S@u-NCSe morphology resembles the original urchin-like structure of u-NCSe (Figure 3a,b), but with the hollow structure partially filled with sulfur. Attempts to completely fill the tubes with sulfur were not considered since we believe that remaining internal voids in the porous structure are advantageous to accommodate the volumetric change during the charge/ discharge process and trap polysulfides, favoring the cycling stability. [33] XRD analysis demonstrates the presence of crystalline cubic sulfur (JCPDS No. 08-0247) within the S@u-NCSe nanocomposite ( Figure 3c) and the retention of the NiCo 2 Se 4 crystal structure. [7] S@u-NCSe contains ≈70 wt% of sulfur as measured by thermogravimetric analysis (TGA; Figure 3d). In addition, with the incorporation of sulfur, the value of Brunauer-Emmett-Teller (BET) specific surface area reduced from 22.4 m 2 g −1 (u-NCSe) to 1.7 m 2 g −1 (S@u-NCSe), and the overall pore volume decreased from 0.20 to 0.017 cm 3   indicating the successful filling of the u-NCSe porous structure by S ( Figure S4, Supporting Information). Four-point probe method was applied to obtain electrical conductivities of the host materials before and after sulfur fusion ( Figure S5, Supporting Information). u-NCSe and b-NCSe exhibited relatively high electrical conductivities, 287.7 and 295.1 S cm −1 , respectively, well above that of Super P (9.5 S cm −1 ). [34] After fusion with sulfur, S@u-NCSe showed electrical conductivities up to of 24.4 S cm −1 , well above that of S@b-NCSe (16.9 S cm −1 ) and nearly sixfold above that of S@Super P (3.9 S cm −1 ). The higher electrical conductivity of S@u-NCSe compared to S@b-NCSe can be explained by the hollow tubular nanostructure, which allows storing a large amount of sulfur but partially conserving a network of avenues for charge transport.
The material adsorption ability plays a vital role in the confinement of LiPS. We tested this adsorption ability by immersing 20 mg of u-NCSe into a LiPS (≈Li 2 S 4 , 10 × 10 −3 m) solution. For comparison the same test was carried out with b-NCSe and also with Super P, a carbon material typically used as an electrode additive. Upon immersion, clear differences in color were observed in as-prepared solutions (Figure 4a). This color change was quantitatively followed by UV-vis spectroscopy, monitoring the absorbance intensity in the 400-500 cm −1 region associated to Li 2 S 4 ( Figure 4b). [35][36][37] The color of Li 2 S 4 solution after the addition of u-NCSe and b-NCSe was much lighter than that of the solution containing Super P, inferring a stronger chemical interaction of LiPS with NCSe. [38] The color of the solution containing u-NCSe was clearer than that of b-NCSe, most probably due to the much higher surface area of the former. The colors of the solutions with or without addition of Super P were nearly the same, indicating the weak Li 2 S 4 adsorption ability of Super P.
XPS analysis confirmed the strong interaction of LiPS with NCSe. Figure 4c,d exhibits high-resolution Ni 2p 3/2 and Co 2p 3/2 XPS spectra of u-NCSe before and after adsorption test. The last denoted as u-NCSe/Li 2 S 4 . Compared with the original Ni 2p 3/2 and Co 2p 3/2 spectra, electron binding energies in u-NCSe/Li 2 S 4 shifted to higher values, indicating the interaction of S with surface Ni and Co. [39] We further verified the strong interaction between NCSe and intermediate LiPS species by density functional theory (DFT). Figure S6 (Supporting Information) exhibits the binding energies and atomic structures between LiPS (Li 2 S 2 , Li 2 S 4 , and Li 2 S 6 ) and the (110) and (001) surfaces of NCSe. Figure 4e displays the relaxed adsorption structure of Li 2 S 4 on the two selected NCSe facets. Li preferentially binds to Se sites and S to Ni and Co ions. Compared with the previous reports on graphitic carbon, [40] the lower sulfur binding energies on the surface of NCSe ( Figure 4f) indicates a stronger adsorption of soluble LiPS, which favors an enhanced electrochemical performance. Interestingly, (110) surface shows lower binding energies than (001) surface, demonstrating a higher anchor strength to soluble LiPS of the former.
To better understand the role of Ni and Co within u-NCSe, we produced and characterized the structural and functional properties of the selenides of the constituent elements. XRD patterns of Ni and Co selenides matched well with NiSe and Co 3 Se 4 crystal phases ( Figure S7e,f, Supporting Information). Figure S7 (Supporting Information) shows the dandelion-liked NiSe (and its precursor) and nanoneedle-shaped Co 3 Se 4 (and  its precursor) produced from the same process used to obtain u-NCSe. [28] Notice the geometry of the elemental selenides significantly differed from that of u-NCSe, which can be considered a first main effect of combining both elements into a selenide. u-NCSe was characterized by higher electrical conductivities than NiSe and Co 3 Se 4 ( Figure S7g, Supporting Information), which is explained by a synergistic effect between the two transition metals, Ni and Co, as reported previously. [27,41,42] Besides, u-NCSe presented much higher LiPS adsorbabilities as displayed in Figure S7h (Supporting Information). This higher adsorbability can be explained by a higher concentration of defects in the bimetallic selenide, which could act as adsorption/catalytic sites. [43,44] Overall, the combination of Ni and Co within a single selenide structure influence the morphology of the obtained materials and increased electrical conductivity and LiPS adsorbability.
Cyclic voltammetry (CV) tests in symmetric cells using an electrolyte containing 0.5 mol L −1 Li 2 S 6 and 1 mol L −1 LiTFSI dissolved in DOL/DME (v/v = 1/1) were carried out to study the electrocatalytic activity of u-NCSe, b-NCSe, and Super P (see details in the Experimental Section). As illustrated in Figure 5a, u-NCSe and b-NCSe electrodes displayed two pairs of reversible redox peaks, named I, II, III, and IV, and associated to the following forward and reverse chemical reactions, respectively [16]   peak current densities, indicating higher redox activity and accelerated reaction kinetics during liquid-to-solid (Li 2 S ↔ S 6 2− ↔ S 8 ) conversion. [38,39] This higher activity should have associated a reduction of soluble LiPS in the electrolyte, having a positive influence in the cycling stability of u-NCSe-based cells, as shown below. Besides, the CV curve of u-NCSe without Li 2 S 6 addition exhibited a nearly rectangular shape ( Figure S8, Supporting Information) that indicated a pure capacitive contribution, thus pointing at Li 2 S 6 as the unique electrochemically active specie. Electrochemical impedance spectroscopy (EIS) analysis of symmetric cells showed NCSe samples to be characterized by much lower charge-transfer resistance (R ct ) than Super P, i.e., a much faster charge transfer at the NCSe-polysulfide interface than at Super P-polysulfide interface (Figure 5b). [16,45,46] CV curves of Li-S coin cells based on S@Super P, S@b-NCSe and S@u-NCSe containing similar amounts of S (Figures S9 Figure 5. Polysulfide redox activity of u-NCSe a) CV profiles and b) EIS spectra of symmetrical cells with different host materials using an electrolyte containing 0.5 mol L −1 Li 2 S 6 and 1 mol L −1 LiTFSI dissolved in DOL/DME (v/v = 1/1) . c) CV profiles of Li-S cells with different electrodes. d) Corresponding peak voltages and onset potentials of asymmetrical Li-S cells obtained from the CV curves. e) CV curves of S@u-NCSe electrode at various scan rates. f) Plot of CV peak current for peaks I, II, and III versus the square root of the scan rates. g) Potentiostatic discharge profile at 2.05 V on different electrodes with Li 2 S 8 catholyte. h) Potentiostatic charge profile at 2.40 V for evaluating dissolution kinetics of Li 2 S. and S10, Supporting Information) were shown in Figure 5c. Two cathodic peaks (peak I and peak II) were identified during reduction of S 8 into long-chain LiPS (Li 2 S x , 4 < x < 8) and their subsequent conversion to insoluble products (Li 2 S 2 and Li 2 S), respectively. The anodic peak (peak III) accounts for the multistep oxidation conversion of short-chain Li 2 S 2 /Li 2 S to LiPS and eventually to sulfur. [46] Reduction peaks measured from cells based on S@u-NCSe systematically exhibited the highest potentials (peak I at 2.32 and peak II at 2.07 V) and current densities among the different materials tested (S@b-NCSe at 2.26 and 2.02 V, S@Super P at 2.2 and 1.92 V), as shown in Figure 5d. However, the peak voltage and onset potential of oxidation peaks displayed inverse results, indicating that u-NCSe can effectively increase the polysulfides redox reaction kinetics. [17,47] Besides, the enhanced catalytic activity of u-NCSe was also confirmed by changes in onset potentials, taken at a current density of 10 µA cm −2 beyond the baseline current ( Figure S11, Supporting Information). As illustrated in Figure 5d, among the three kinds of electrode tested, S@u-NCSe exhibited the highest onset potentials of reduction peaks and the lowest onset potentials of oxidation peaks, evidencing the capacity of u-NCSe to electrocatalytically accelerate the reaction kinetics. [17,48] CV curves of S@u-NCSe ( Figure S12a, Supporting Information) almost overlapped in the first cycle, showing no obvious peak shifts or current changes, which indicated good stability and high reversibility. [49] The lithium ion diffusion coefficient was evaluated qualitatively from CV tests under different scanning rates, in the range from 0.1 to 0.4 mV s −1 (Figure 5e,f). A linear relationship was obtained between the reduction and oxidation peak currents and the square root of scanning rates, demonstrating the reaction to be diffusion-limited. Thus, the lithium ion diffusivity can be calculated using the classical Randles-Sevcik equation [38,47] where I p is the peak current, n is the number of charge transfer, A is the geometric electrode area, D + Li is the lithium ion diffusion coefficient,C + Li is the concentration of lithium ions in the cathode, and ν is the scan rate. S@u-NCSe electrodes showed the sharpest slopes ( Figure S13, Supporting Information), thus the highest lithium ion diffusivity. We hypothesize this higher lithium ion diffusivity to be related to the relief of the shuttle effect and the improved catalytic activity of the u-NCSe host towards LiPS conversion demonstrated above, avoiding the high viscosity electrolyte caused by LiPS dissolution and the deposition of a thick insulating layer on the electrode. [38] During charge/discharge processes, the overpotential of LSBs was mainly caused by the sluggish kinetics of the oxidation/reduction of insulated solid Li 2 S. [38,39,50] To further demonstrate the catalytic effect of u-NCSe, Li 2 S nucleation and dissolution experiments were conducted with a Li 2 S 8 / DOL-DME solution (details can be found in the Experimental Section). [51] Figure 5g shows potentiostatic discharge profiles that demonstrate that CP/u-NCSe electrodes displayed faster responsivity toward Li 2 S nucleation than CP/Super P. Based on the Faraday's law, CP/u-NCSe electrodes also exhibited larger capacities of Li 2 S precipitation (151.1 mAh g −1 ) and shorter nucleation and growth times than CP/Super P electrodes (74.6 mAh g −1 ). These results demonstrate that u-NCSe hosts can significantly reduce overpotential for the initial Li 2 S nucleation and promote kinetics for subsequent Li 2 S precipitation. [52][53][54] A similar strategy was used to study the kinetics of Li 2 S dissolution (Figure 5h). Potentiostatic charge curves of CP/u-NCSe exhibited higher current densities than CP/Super P, indicating a lower oxidation overpotential for Li 2 S dissolution. Moreover, the calculated dissolution capacity of CP/u-NCSe (743 mAh g −1 ) was much higher than for CP/Super P electrodes (389 mAh g −1 ). Overall, these results verified the superior electrocatalytic effect of u-NCSe hosts in reducing polarization and promoting redox kinetics of LiPS conversion reaction. [39] Electrochemical performance was further analyzed through galvanostatic charge-discharge tests (Figure 6). Chargedischarge curves of S@Super P, S@b-NCSe and S@u-NCSe at 0.1 C showed one charge plateau and two discharge plateaus, consistently with CV. S@u-NCSe showed lower polarization potential (ΔE = 152 mV) than S@b-NCSe (ΔE = 205 mV) and S@Super P electrodes (ΔE = 222 mV). [17,46] The voltage gap ΔE between the oxidation and the second reduction plateaus introduced a hysteresis in the redox reaction.
Discharge curves showed two plateaus, corresponding to the reduction of sulfur to soluble LiPS (S 8 → S 6 2− → S 4 2− ) and the subsequent conversion to insoluble products (S 4 2− → Li 2 S 2 → Li 2 S). The associated capacity of the two discharge plateaus was defined as Q1 and Q2, respectively (Figure 6a). The ratio between Q2 and Q1 (Q2/Q1) can be interpreted in terms of the catalytic ability for LiPS conversion reaction: sluggish kinetics during the solid → liquid → solid process and shuttle effect caused by diffusion of soluble LiPS give rise to capacity fading during Q2 stage. Thus, the higher Q2/Q1, the better catalytic ability. [25,55] As shown in Figure 6b, the Q2/Q1 of S@u-NCSe was 2.8, much higher than that of S@b-NCSe (2.32) and S@ Super P (1.88). This high ratio also proved the superior catalytic activity towards polysulfides redox reaction of u-NCSe.
Associated with the ability of u-NCSe to accelerate the charge transfer and promote conversion of polysulfides, S@u-NCSe showed the largest capacity among the different electrodes tested. All discharge curves at different current rates exhibited two evident discharge plateaus (Figure 6c). The electrochemical capacity of the cell with S@u-NCSe at various current densities from 0.1 C to 5 C is shown in Figure 6d. The initial discharge capacity was 1403 mAh g −1 , and stabilized to an average capacity of 1330 mAh g −1 at 0.1 C. Even at high current rates of 5 C, the capacity still remained stable at 626 mAh g −1 , which is significantly higher than the one obtained for S@Super P electrodes (5 mAh g −1 ; Figure S14, Supporting Information) under the same conditions. Moreover, when the current rate was turned back to 0.2 C, the average capacity of the cell with S@u-NCSe returned to the same approximate value of 1060 mAh g −1 , implying a remarkable electrochemical stability. [7,56] Energy efficiency, the ratio of energy output/input (E = ∫UI dt) upon voltage polarization cycles, is a pivotal parameter in largescale electrochemical energy storage systems. [17] S@u-NCSe electrodes were characterized with much higher and stable energy efficiencies than S@Super P, especially at high current rates ( Figure 6e). As an example, S@u-NCSe retained 85.6% efficiency at 5 C, much higher than the 71.3% for S@Super P. The significant improvement in energy efficiency arised from the lower polarization potential, associated with the exceptional catalytic properties of u-NCSe, as discussed above.
EIS analyses were carried out to gain understanding of the enhanced electrochemical performance of S@u-NCSe electrodes. Figure 6f shows the Nyquist plot obtained from a fresh S@u-NCSe coin cell and the same cell after 100 cycles at 1 C.  In the high frequency region, the fresh electrode showed a semicircle corresponding to the charge-transfer resistance, and a linear dependence in the low frequency region that reflected the diffusion of lithium ions into the electrode. After 100 cycles, the impedance plot changed to two poorly-resolved semicircles at high and middle frequencies and a lineal dependence at low frequencies. [57,58] Apparently, the charge-transfer resistance decreased after cycling, which should be associated with the activation process. Moreover, comparing with the other two types of electrode tested, S@b-NCSe and S@Super P ( Figure S15, Supporting Information), S@u-NCSe electrodes showed the lowest charge-transfer resistance (R ct ).
The long-term cycling stability of the NCSe-based batteries was evaluated at a high current density of 3 C (Figure 6g). After 2000 cycles, S@u-NCSe electrodes delivered a capacity of 480 mAh g −1 , involving a 0.016% average capacity decay per cycle. Meanwhile, a high and steady Coulombic efficiency above 99.7% was obtained. It is worth mentioning that a negligible capacity was obtained from pure u-NCSe under the same measuring conditions, as shown in Figure S16 (Supporting Information). In contrast, S@Super P electrodes delivered a considerably low capacity after 400 cycles (294 mAh g −1 ), suffering from a rapid capacity fading (0.11% average capacity decay per cycle), as well as a low Coulombic efficiency (average about 97.1%) at 1 C ( Figure S14c, Supporting Information).
For practical applications, high energy density Li-S batteries require increasing the sulfur loading. Therefore, we studied the performance of S@u-NCSe electrodes at a higher sulfur loading, 3.2 mg cm −2 . Figure S17a (Supporting Information) displays galvanostatic charge-discharge curves of a S@u-NCSe electrode at different current rates. One charge plateau and two discharge plateaus were clearly observed at all current rates, up to 3 C, demonstrating the low polarization between charge and discharge processes. At this high sulfur loading, we measured average reversible capacities of S@u-NCSe electrodes from 1169 mAh g −1 at 0.1 C to 522.8 mAh g −1 at 3 C, which corresponded to areal capacities of 3.65 and 1.63 mAh cm −2 , respectively. This high rate performances even at high sulfur loadings was consistent with the high electrical conductivity and superior catalytic properties of this material. Long-term cycling tests at 1 C showed S@u-NCSe electrodes loaded with 3.2 mg cm −2 of sulfur to maintain 557 mAh g −1 after 600 cycles, i.e., a 74.3% capacity retention, involving a 0.043% average capacity decay per cycle. Additionally, a high and steady Coulombic efficiency above 98.8% was consistently obtained (Figure 6h), indicating an excellent cycling stability.
Electrochemical results of S@u-NCSe cathodes for LSBs are compared to other state-of-the-art TM-based materials in Table S1 (Supporting Information). To illustrate the favorable electrochemical performance of S@u-NCSe cathodes and the promising practical application of related LSBs, one S@u-NCSe coin cell was used to light up a "LSB"-shaped LED panel containing 47 LEDs (voltage: 2-2.2 V), as shown in Figure 6i.
Finally, to further demonstrate that u-NCSe effectively confines LiPS and minimizes the shuttle effect in LSBs, coin cells were disassembled after 200 cycles at 1 C to inspect their membrane, cathodic integrity and anodic corrosion. Separators from S@u-NCSe coin cells exhibited much lighter color compared to those from S@Super P (Figure 6k). This observation probed that u-NCSe better confined LiPS, avoiding its diffusion during charge/discharge processes. [7,59] Consistently with the lighter color of the separator, Li metal foils from S@u-NCSe coin cells showed less corrosion and fewer Li 2 S species deposited at their surface than S@Super P coin cells, as shown by SEM and EDS analyses in Figure 6k,l. [48] Thus, the use of u-NCSe as host cathode material greatly relieved the LiPS shuttle effect and minimizes the irreversible losses of active sulfur in LSBs, leading to a superior stability during long-term cycling (Figure 6g). Besides, the crystal structure and morphology of S@u-NCSe after cycling was analyzed. HAADF-STEM and SEM micrographs showed the original tubular nanostructure to be conserved after the cycling (Figure 6j,l). Additionally, HRTEM and XRD analysis probed the NCSe crystal structure to be conserved (Figure 6j; Figure S18, Supporting Information), indicating an excellent stability towards lithiation/delithiation cycles.
Biological security is an important parameter for application of energy storage materials. Thus, we analyzed the biocompatibility of S@u-NCSe by measuring through MTT assays the cytotoxicity of this material against the human hepatocellular carcinoma cell line HepG2 (please refer to the Supporting Information for details). [60] Figure S19 (Supporting Information) showed the viability of the cultured cells in the presence of S@u-NCSe at concentrations ranging from 0.001 to 1000 µg mL −1 . Even though a gradual decreasing trend was observed with increasing concentrations, cell viabilities above 85% even at S@u-NCSe concentration of 1000 µg mL −1 were obtained, indicating that S@u-NCSe composites have a negligible cytotoxicity.

Conclusions
In summary, we developed urchin-like NiCo 2 Se 4 nanostructures serving as polar host with catalytic effect for cathode of LSBs. Comprehensive kinetic investigations revealed that u-NCSe promoted redox kinetics of LiPS conversion reaction, and effectively decreased polarization during charging and discharging processes. A strong LiPS adsorbability was confirmed simultaneously by experimental results and DFT calculations. u-NCSe was characterized by a beneficial hollow structure to relieve volumetric expansion and a superior electrical conductivity to improve electron transfer. Owing to these excellent qualities, S@u-NCSe delivered impressive rate performance with 1330 and 626 mAh g −1 at 0.1 C and 5 C, respectively. More significantly, a reversible capacity of 480 mAh g −1 was retained after 2000 cycles at 3 C and, even at high sulfur loading (3.2 mg cm −2 ), 557 mAh g −1 capacity was delivered after 600 cycles at 1 C. Additional cytotoxicity measurements demonstrated the u-NCSe biocompatibility. This work provides a strategy for the rational design and development of LSBs with long-life and high-rate performance in addition to insights into transition metal selenides as sulfur host material.

Experimental Section
Synthesis of u-NCSe: u-NCSe was synthesized by a two-step synthesis process, from selenization of Ni 0.33 Co 0.67 (CO 3 ) 0.5 OH precursor obtained by a simple hydrothermal process. First, 5 × 10 −3 m NiCl 2 ·6H 2 O (98%, Alfa Aesar) and CoCl 2 ·6H 2 O (98%, Alfa Aesar) with molar ratio of 1:2 were dissolved into 30 mL of deionized (DI) water, and then 300 mg of urea (99%, Acros Organics) added, using an ultrasounds bath for 3 min to form a homogeneous solution. This solution was then poured into a Teflon lined stainless steel autoclave of 50 mL volume and heated at 130 °C for 8 h. After naturally cooling to ambient temperature, the Ni 0.33 Co 0.67 (CO 3 ) 0.5 OH precipitate was centrifuged, washed, dried and recovered. Subsequently, 50 mg of as-obtained Ni 0.33 Co 0.67 (CO 3 ) 0.5 OH were dispersed in 25 mL of deionized water using an ultrasonic bath and then 150 mg of Na 2 SeO 3 (99%, Alfa Aesar) and 4 mL of N 2 H 4 ·H 2 O (98%, Sigma Aldrich) were incorporated under vigorous stirring. The mixture was finally poured into a Teflon-lined stainless steel autoclave of 50 mL volume and heated at 180 °C for 8 h. After cooling naturally to ambient temperature, the precipitate was centrifuged, washed, dried and recovered. NiSe and Co 3 Se 4 nanostructures were synthesized following the same synthesis protocol.
Synthesis of b-NCSe: b-NCSe was obtained in just one synthesis step. 37 mg of NiCl 2 ·6H 2 O, 75 mg of CoCl 2 ·6H 2 O, and 150 mg of Na 2 SeO 3 were dissolved into 25 mL of deionized water and then 4 mL of N 2 H 4 ·H 2 O were dropped into the solution under vigorous stirring conditions. The resulting solution was finally poured into a Teflonlined stainless steel autoclave of 50 mL volume and heated at 180 °C for 8 h.
Synthesis of S@u-NCSe and S@b-NCSe: u-NCSe and sulfur powder (99.98%, Sigma Aldrich) (1:3, weigh ratio) were mixed and heated at 155 °C for 12 h in a glass bottle under Ar atmosphere. In order to remove the redundant sulfur not incorporated into u-NCSe, the powder was immersed in a 10 mL CS 2 and ethanol solution (1:4, volume ratio) for 10 min twice. S@b-NCSe was obtained using the same process.
Synthesis of S@Super P: Super P (99%, Alfa Aesar) and sulfur powder (3:7, weigh ratio) were well mixed and heated at 155 °C for 12 h.
Materials Characterization: X-ray diffraction (XRD) patterns were recorded at room temperature using a Bruker AXS D8 Advance X-ray diffractometer with Cu K radiation (λ = 1.5106 Å) operating at 40 kV and 40 mA. The morphology and microstructure were examined by TEM (ZEISS LIBRA 120) and FESEM (ZEISS Auriga) equipped with an energydispersive X-ray spectroscopy detector operated at 20 kV. High-resolution TEM (HRTEM) and scanning TEM (STEM) studies were carried out using a field emission gun FEI Tecnai F20 microscope at 200 kV with a point-to-point resolution of 0.19 nm. High angle annular dark-field (HAADF) STEM was combined with electron energy loss spectroscopy in the Tecnai microscope by using a GATAN QUANTUM filter. X-ray photoelectron spectroscopy measurements were carried out in normal emission using an Al anode XR50 source operating at 150 mW and a Phoibos 150 MCD-9 detector. TGA (PerkinElmer Diamond TG/DTA instrument.) experiments were performed to estimate the content of S in prepared composites. The specific surface area and analysis of the pore size distribution were performed by Brunauer-Emmett-Teller method (Tristar II 3020 Micromeritics system). UV-vis absorption spectra were recorded on a PerkinElmer LAMBDA 950 UV-vis spectrophotometer. Electrical conductivities were measured using a four-point probe station (Keithley 2400, Tektronix).
Li-S Cell Assembly and Measurements: S@host composites (S@u-NCSe; S@b-NCSe; S@Super P), Super P, and PVDF binder (weight ratio = 8:1:1) were dispersed in N-methyl pyrrolidone (NMP, 99.5%, Acros Organics) to form a slurry which was coated on aluminum foils and dried at 60 °C overnight. The coated aluminum foil was then punched into small disks with a diameter of 12.0 mm. Sulfur loading was about 1.0-1.1 mg cm −2 . High-loading tests were applied using 3.2 mg cm −2 of sulfur. Electrochemical measurements were conducted in standard 2032 coin-type cells. In LSBs assemblies, lithium foils were used as counter electrode and Celgard 2400 membranes as separators. The electrolyte used was 1.0 m lithium bis(trifluoromethanesulfonyl) imide (LiTFSI) (99%, Acros Organics) dissolved in a mixture of 1,3-dioxolane (DOL, 99.5%, Alfa Aesar) and 1,2-dimethoxy ethane (DME, 99%, Honeywell) (v/v = 1/1) and containing 0.2 m of LiNO 3 (99.98%, Alfa Aesar). For each coin cell, 20 µL of electrolyte was used, high-loaded coin cells added 45 µL. The cells were galvanostatically cycled within a voltage range of 1.7-2.8 V using a Neware BTS4008 battery tester at different C rates. Cyclic voltammetry measurements were performed on a battery tester BCS-810 from Bio Logic at a scan rate of 0.1-0.4 mV s −1 and electrochemical impedance spectroscopy tests were performed using a sinusoidal voltage with amplitude of 10 mV in the frequency range 100 kHz to 10 mHz.
Preparation of Li 2 S x (Like Li 2 S 4 , Li 2 S 6 , and Li 2 S 8 , x = 4, 6, or 8) Solutions for Adsorption Test and Kinetic Study: Sulfur and Li 2 S (99.9%, Alfa Aesar) in the molar ratio x − 1:1 were added to appropriate amounts of DME and DOL (volume ratio of 1:1) under vigorous magnetic stirring overnight until a dark brown solution was formed. 20 mg of Super P, b-NCSe, or u-NCSe were poured into 3.0 mL 10 × 10 −3 m Li 2 S 4 solution, respectively, and the mixture was stirred for homogenization overnight.
Symmetric Cell Assembly and Measurements: Electrodes for symmetric cells were fabricated in the same way as electrodes for LSBs. Two pieces of the same electrode (average loading about 0.5 mg cm −2 ) were used as identical working and counter electrodes with 40 µL of electrolyte containing 0.5 mol L −1 Li 2 S 6 and 1 mol L −1 LiTFSI dissolved in DOL/DME (v/v = 1/1). For comparison, symmetric cells with electrolyte 1 mol L −1 LiTFSI dissolved in DOL/DME (v/v = 1/1) were also assembled and tested. In all cases, CV measurements were performed at scan rate of 40 mV s −1 .

Measurement of Nucleation and Dissolution of Li 2 S:
The nucleation and dissolution of Li 2 S were tested in 2032 coin cells, where 1 mg of u-NCSe or Super P loaded on the carbon papers was applied as work electrode, Li foil worked as the counter electrode, 20 µL of 0.25 m Li 2 S 8 dissolved in DOL/DME (v/v = 1:1) solution with 1.0 m LiTFSI was used as catholyte, and 20 µL of 1.0 m LiTFSI in DOL/DME (v/v = 1:1) solution as anolyte. The cells were held at 2.19 V for 2 h to reduce higher order LiPS to Li 2 S 4 . And then held them at potential of 2.05 V until current decreased to 10 −2 mA for Li 2 S nucleation and growth. [51] In order to analyze the Li 2 S dissolution, fresh cells were first discharged at a current of 0.10 mA to 1.80 V, and subsequently discharged at 0.01 mA to 1.80 V for full transformation of S species into solid Li 2 S. After this discharge, cells were potentiostatically charged at 2.40 V for the dissolution of Li 2 S into LiPS until charge current was below 10 −5 A. [39]

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.