CoNiO2 /Co4 N Heterostructure Nanowires Assisted Polysulfide Reaction Kinetics for Improved Lithium-Sulfur Batteries.

The "shuttle effect" of soluble polysulfides and slow reaction kinetics hinder the practical application of Li-S batteries. Transition metal oxides are promising mediators to alleviate these problems, but the poor electrical conductivity limits their further development. Herein, the homogeneous CoNiO2 /Co4 N nanowires have been fabricated and employed as additive of graphene based sulfur cathode. Through optimizing the nitriding degree, the continuous heterostructure interface can be obtained, accompanied by effective adjustment of energy band structure. By combining the strong adsorptive and catalytic properties of CoNiO2 and electrical conductivity of Co4 N, the in situ formed CoNiO2 /Co4 N heterostructure reveals a synergistic enhancement effect. Theoretical calculation and experimental design show that it can not only significantly inhibit "shuttle effect" through chemisorption and catalytic conversion of polysulfides, but also improve the transport rate of ions and electrons. Thus, the graphene composite sulfur cathode supported by these CoNiO2 /Co4 N nanowires exhibits improved sulfur species reaction kinetics. The corresponding cell provides a high rate capacity of 688 mAh g-1 at 4 C with an ultralow decaying rate of ≈0.07% per cycle over 600 cycles. The design of heterostructure nanowires and graphene composite structure provides an advanced strategy for the rapid capture-diffusion-conversion process of polysulfides.

Introduction

Lithium–sulfur (Li–S) batteries have attracted much attention because of their much higher theoretical capacity (1672 mAh g−1) and energy density (2567 Wh kg−1) than those of conventional Li‐ion batteries, such as LiCoO2–graphite and LiFePO4–graphite.[ , ] However, the intermediates polysulfides (Li2S , n = 4–8) can dissolve in the electrolyte and deposit on the counter electrode, resulting in “shuttle effect”. It will consume electroactive species, reducing Coulombic efficiency and capacity. Large amounts of polysulfides react with Li anode disproportionation, destroying the stability of the electrode interface, which greatly increases the threat of Li dendrites.[ , ] In addition, the poor electron conductivity of sulfur and slow reaction kinetics seriously hinder the rapid start‐up and response of Li–S batteries.[ ]

In order to tackle the above‐mentioned issues, various efforts have been made in exploring new electrode structures and materials. Dispersing active sulfur into porous carbon matrix to form composite cathode is one of the main strategies.[ , , ] Among the carbon based materials, graphene has attracted extensive attention. Its excellent conductivity and large specific surface area are used as electronic network to reduce the sulfur interface resistance and physical barrier to inhibit the diffusion of polysulfides, respectively.[ ] Unfortunately, some inherent problems still exist in the graphene–sulfur (G–S) composite architectures. Nonpolar carbon surface has a poor affinity for polar hydrophilic polysulfides. Such weak capture ability results in limited cyclic stability and high sulfur loading.[ ] Functional groups (such as hydroxyl, carboxyl) modified graphene can enhance the adsorption of polysulfides. However, it will reduce the conductivity of the material and increase the complexity of preparation.[ , ] Moreover, the aggregation and stacking properties of graphene also need to be emphasized, which may cause unnecessary obstacles to ion diffusion.[ ]

The combination of polar transition metal compounds (TMCs, such as oxides, nitrides) with graphene has proved to be an effective strategy to inhibit “shuttle effect” and enhance conversion of Li2S .[ ] However, most TMCs have only one or two functions, which are relatively single and unable to integrate all the advantages. For instance, some metal oxides have strong polysulfide adsorption, while their poor electrical conductivity is not conducive to rapid charge transfer and will lead to severe electrode passivation.[ , , ] In contrast, metal nitrides with metal properties exhibit weak polysulfide affinity, resulting in a significant reduction in confinement effect and catalytic activity.[ , ] Recently, the design of TMC materials with different functions into a unified heterojunction structure has become a research focus to alleviate the above problems. For example, Luo et al. prepared heterophase V2O3–VN hollow structure.[ ] The multifunctional heterostructure was proved to be beneficial to chemical anchoring and catalytic conversion of polysulfides. Similarly, Xiao and co‐workers developed ZnS–SnS heterojunction with uniform cubic morphology to suppress the “shuttle effect”.[ ] Afterward, Wu et al. designed a twinborn holey Nb4N5–Nb2O5 heterostructure to combine the merits of electronic conducting (Nb4N5) with polysulfide adsorption (Nb2O5).[ ] Although these heterojunctions achieve good cathode performance through synergy effect, the large structure and size (usually above 100 nm in size) limit the overall energy density improvement.[ ] This is because these “inactive” additives do not provide additional reaction capacity, only surface layers play the role of adsorption and catalysis.[ ] Excessive usage can increase the overall weight of the battery. As a result, the design of heterojunction morphology is very important.

It can be seen that in addition to conductivity, adsorption and catalysis, another feature of an ideal G–S cathode additive is the minimization of the amount of material used. Ultrafine nanoparticles or ultrathin nanosheets are powerful structural designs, but neither of them can well alleviate the issue of graphene agglomeration. Previous studies have shown that the introduction of 1D structure can effectively alleviate the agglomeration of graphene layers.[ , ] More importantly, the high aspect ratio shape has a lower penetration threshold and is easy to form a continuous conductive network.[ , ] Therefore, it will be a promising challenge to construct heterojunctions with different characteristics for Li2S on TMC nanowires as graphene based sulfur cathode medium.

In this study, a graphene composite matrix supported by CoNiO2/Co4N heterostructure nanowires was prepared from the point of view of physical properties and structure. The in situ formed heterostructure greatly enhanced the polysulfide reaction kinetics and reduced the “shuttle effect” of Li–S batteries (Figure ). We have studied the interface and synergistic effect of the CoNiO2/Co4N heterostructure theoretically and experimentally. On one hand, it combined the strong polysulfide anchoring ability of CoNiO2 and the excellent electrical conductivity of Co4N. On the other hand, as the active site, the heterogeneous interface had stronger ability of ion diffusion and polysulfide transformation (Figure 1b). As a result, the unique CoNiO2/Co4N–G–S cathode delivered high initial discharge capacity of 1198 mAh g−1 at 0.2 C and 688 mAh g−1 at 4 C, and favorable cycling stability over 600 cycles with an ultraslow capacity decay of ≈0.07% per cycle.

Figure 1

a) A comparison of the conventional Li–S battery with “shuttle effect” and the CoNiO2/Co4N based improved cell. b) Schematic illustration of interface mechanism of heterostructure and polysulfides.

Results and Discussion

The phase and composition information of as‐prepared samples was first characterized by X‐ray diffraction (XRD). As shown in Figure S1a in the Supporting Information, the pattern of precursor was well matched with standard diffraction peaks of NiCo2O4 (JCPDS No. 20‐0781), without any contamination observed. After nitridation, a mixed phase appeared (Figure ), which meant the precursor had undergone a significant phase transition. The diffraction peaks at 36.8°, 42.8°, and 61.7° could be readily assigned to CoNiO2 phase (JCPDS No. 10‐0188), and the diffraction peaks at 44.1° and 51.7° were indexed to cubic Co4N structure (JCPDS No. 15‐0806).[ , ] This binary mixing state was further investigated by X‐ray photoelectron spectra (XPS). It was not difficult to see from Figure S2a in the Supporting Information that the Co 2p3/2 orbit could be separated into three peaks, indicating Co existed in more than one chemical state.[ ] The peaks at 779.3 and 782.4 eV were well attributed to the Co—Co and Co—N bonds, respectively.[ , , ] The existence of Co—Co metal bond was due to the similar atomic structure of Co4N and metallic Co (0), and the N element was located in the middle of the unit cell.[ , ] The peak located at 780.8 eV was assigned to the Co—O bond, which was derived from CoNiO2.[ ] In the N 1s XPS spectra (Figure S2b, Supporting Information), the peak at 398.2 eV verified the existence of Co—N.[ , ] The energy‐dispersive X‐ray spectroscope (EDS) result (Figure S3, Supporting Information) also demonstrated the incorporation of N. The oxidation states of Ni 2p3/2 orbit in Figure S2c in the Supporting Information could be deconvolution derived from divalent and trivalent Ni ions.[ ] These results indicated the formation of CoNiO2/Co4N heterostructure, which combined the characteristics of CoNiO2 and Co4N concurrently.[ ] The XRD of the CoNiO2/Co4N–G–S composite (Figure 2a) confirmed the presence of layered graphene (Figure S1b, Supporting Information) and orthorhombic sulfur (Figure S1c, Supporting Information), which could be used as polysulfide host and active materials, respectively. From thermogravimetric (TG) analysis in Figure S4 in the Supporting Information, it could be seen that the sulfur content in CoNiO2/Co4N–G–S, CoNiO2–G–S, Co4N–G–S, and G–S were 66.7%, 66.5%, 66.9%, and 66.6%, respectively. More importantly, the characteristic peaks of CoNiO2/Co4N had not changed significantly, indicating that this preparation method would maintain the stability of heterostructure.

Figure 2

Phase characterization of CoNiO2/Co4N heterostructure nanowires. a) XRD patterns. b,c) TEM image and diameter distribution. d) HRTEM result and corresponding lattice spacing. e) SAED pattern and rotational integral.

The transmission electron microscopy (TEM) and scanning electron microscopy (SEM) images observed in Figure 2b and Figure S5 in the Supporting Information show an overall morphology of the CoNiO2/Co4N sample. It reveals that the heterojunction material presented a uniform nanowire‐like structure with an average diameter of 8–12 nm (Figure 2c). These nanowires were overlapped with each other and would form a mixed 3D conductive framework with layered graphene (Figure S6, Supporting Information). When graphene was combined with CoNiO2/Co4N, CoNiO2 and Co4N samples, their Barrett–Emmett–Teller (BET) surface areas were up to 621.3, 665.6, 620.2 m2 g−1, respectively (Figure S7a–c, Supporting Information). Similar specific surface area ensured uniform loading of sulfur and comparability of corresponding Li–S batteries. Meanwhile, the partially graphitized nature of as‐obtained graphene based composite was confirmed by Raman technology in Figure S7d–f in the Supporting Information. Two prominent peaks at about 1352 and 1587 cm−1 corresponded to the characteristic D and G bands of carbon structure, respectively. Figure 2d depicts a high‐resolution TEM (HRTEM) image with clear lattice fringes for further revealing as‐prepared CoNiO2/Co4N heterostructure. The lattice spacing of 0.217 and 0.205 nm corresponded to the (200) plane of NiO and (111) plane of CoNiO2, respectively. A higher magnification and clearer HRTEM image is shown in Figure S8 in the Supporting Information. The distinct heterointerface could be easily observed. The resulting interface would be used not only as an active site for the adsorption and conversion of soluble polysulfides, but also as a channel for rapid Li‐ion diffusion, thus greatly improving the electrochemical performance of Li–S batteries.[ , ] The rotational integral from selected area electron diffraction (SAED) pattern (Figure 2e) were consistent with XRD result, which further indicated the formation of CoNiO2/Co4N. XPS characterization revealed that the ratio of CoNiO2 to Co4N in the heterojunction was ≈1.8:1.

The electronic conductivities of CoNiO2, Co4N, and CoNiO2/Co4N were calculated via partial density of states (DOS) plots. As shown in Figure , pristine CoNiO2 was a typical semiconductor, while Co4N had metallic characteristics. It was noteworthy that after engineering the in situ heterojunction, not only the band structure of Co4N remained intact, but also the band structure of the heterostructure crossed the Fermi level, indicating the excellent conductive properties of CoNiO2/Co4N. Such conductivity would also facilitate the rapid transfer of electrons and polysulfide conversion.[ ]

Figure 3

DOS of optimized geometries of CoNiO2/Co4N, CoNiO2, and Co4N.

Subsequently, a series of visual adsorption experiments were designed to investigate the anchoring behavior of as‐prepared samples to soluble polysulfides. Equal amounts of CoNiO2, Co4N, CoNiO2/Co4N were added to the brown‐yellow solution of Li2S6, respectively, and their changes with time were observed (Figure ). All the samples could decolorize Li2S6 solution in varying degrees within 0.5 h, indicating that they all had certain adsorption characteristics for polysulfides. After 3 h, the color of Li2S6 solution with adding pure CoNiO2 and heterostructure had changed significantly, especially the former, which was almost completely clarified. However, the Co4N sample did not reach the transparent effect until 8 h later, indicating that the anchoring ability of Li2S6 is in the order of CoNiO2, CoNiO2/Co4N, and Co4N, from strong to weak. This difference could also be confirmed by density functional theory (DFT) (Figure 4b–e). As can be seen from Figure 4e, the adsorption energies of Li2S6 on CoNiO2, Co4N and CoNiO2/Co4N interface were calculated to be 5.69, 4.48, and 4.48 eV respectively, which was basically consistent with the results of the above visual test. Moreover, Li2S4 showed similar results. Interestingly, for Li2S8, the heterojunction revealed a higher binding energy than monomeric CoNiO2 or Co4N, implying tighter anchoring and rapid adsorption. All these adsorption values were much higher than those on the surface of pure graphene. This might be because the interaction between the carbon and polysulfides was only weak van der Waals forces (Figure S9, Supporting Information).[ , ] In contrast, the adsorption of polysulfides on Ni—Co substrates resulted from chemical bonding (Figure 4b–d). It should be noted that these bonds did not significantly change the structure of polysulfide molecules, which meant that the irreversible decomposition of Li2S can be avoided.[ ]

Figure 4

a) Visual Li2S6 adsorption test over time. Optimized geometries of Li2S6 on different substrates: b) CoNiO2 (200) surface, c) Co4N (111) surface, and d) corresponding CoNiO2/Co4N heterojunction interface. e) Contrast diagram of calculated adsorption energies.

The XPS results in Figure S2 in the Supporting Information further analyzed the interaction between the CoNiO2/Co4N and Li2S6. After interacting with Li2S6, an additional peak appeared at 780.2 eV of the Co 2p3/2 spectrum, indicating the formation of Co—S bond.[ , ] The shift of the N 1s orbital to higher binding energy (blue‐shift) indicated the transfer of electrons from the Co—N to the terminal sulfur.[ , ] The Ni 2p3/2 peak shifted slightly toward the lower binding energy (red‐shift), which was similar to previous literatures, meaning that a small amount of electrons transferred from the polysulfide to the Ni atom.[ , , ] Such tight affinity of CoNiO2/Co4N heterostructure to polysulfides would not only effectively improve the anchoring ability, but also promote charge transfer and facilitate the rapid conversion of sulfur species.

In order to explore the effect of CoNiO2/Co4N heterostructure on polysulfide reaction kinetics, the fully assembled Li–S batteries were first tested by cyclic voltammogram (CV). In general, the strong anodic peak at about 2.3–2.4 V is related to the conversion of short‐chain Li2S2/Li2S to long‐chain polysulfides and elemental sulfur.[ , ] Two cathodic peaks (≈2.35 and ≈2.05 V) are attributed to the reduction of sulfur to long‐chain polysulfides and subsequent formation of Li2S2/Li2S.[ , ] As shown in Figure , all CV curves showed typical redox peaks for Li–S batteries and no additional peaks appeared, indicating that none of the three samples obstructed normal polysulfide reactions. The CoNiO2/Co4N exhibited the lowest anode potential and the highest cathode potential, and the voltage polarization was as low as 0.263 V. Even the CoNiO2–G–S electrode with highest polarization showed less polarization than pure graphene without any additives (Figure S10, Supporting Information). Considering the conductivity of CoNiO2 and graphene, the polarization difference fully indicated that CoNiO2 had a good catalytic function of polysulfides.[ ] Meanwhile, the corresponding Tafel plots and slopes of all peaks were also calculated to quantify the catalytic activities (Figure 5b,c and Figure S11, Supporting Information). Evidently, the CoNiO2/Co4N–G–S electrode exhibited the lowest slopes at the oxidation of Li2S and the reduction of sulfur to long‐chain sulfur species, implying the greatest improvement in the kinetics.[ , ] For the cathodic peak at ≈2.0 V, the slope of CoNiO2/Co4N–G–S was not much different from that of the lowest Co4N–G–S electrode, indicating that the charge in the heterostructure could also be rapidly transformed for this reaction.

Figure 5

Verification of catalytic properties of CoNiO2/Co4N heterostructure. a) CV curves of the CoNiO2/Co4N–G–S, CoNiO2–G–S, and Co4N–G–S cathodes in a potential window of 1.7–2.6 V at 0.1 mV s−1. Corresponding Tafel plots and fitted slope values for b) the oxidation peak and c) the reduction peak at ≈2.4 V from three CV curves. c) Local magnification of charge–discharge curves of different cathodes. d) Schematic diagram of symmetrical cell. e) CV contrast profiles of Li2S6 symmetric cells. f) Nyquist plots of different symmetric cells.

The conversion kinetics of as‐prepared samples was more obvious in symmetrical cell based on Li2S6 catholyte (Figure 5d).[ ] From the CV curves in Figure 5e, except for graphene, the other electrodes showed two distinct redox peaks and high response current, indicating easy conversion between Li2S and Li2S2/Li2S.[ ] The redox peak of CoNiO2/Co4N sample provided the highest intensity and smallest polarization, which was in good agreement with the electrochemical impedance spectroscopy (EIS) results (Figure 5f) for symmetric cells. The charge transfer resistance of symmetrical cell increased in the order of CoNiO2/Co4N

In addition to adsorption and conductivity, Li‐ion diffusion is another important factor affecting the dynamics of Li–S system.[ ] To better understand the S8/S8 2− redox properties of CoNiO2/Co4N nanowires, DFT calculated the migration of Li‐ion at the heterostructure interface and graphene surface (Figure ). The Li‐ion diffusion barrier of graphene exhibited higher barrier with peak value of 0.32 eV, while the peak value of CoNiO2/Co4N was as low as 0.321 eV. The low diffusion energy meant that the ions migrate effectively on the heterojunction substrate, which ensured the strong interfacial ion transfer kinetics.[ ] Moreover, CV analysis of CoNiO2/Co4N–G–S, Co4N–G–S and CoNiO2–G–S electrodes in the range of 0.1–0.9 mV s−1 were carried out in Figure 6e–g. It could be seen that all the reaction peak currents were linear with the square root of the scanning rate. According to Randles–Sevcik equation (see Supporting Information for details), the diffusivity of Li‐ion can be simply judged by the slope of the line.[ , , ] As shown in Figure 6h–j, all slopes of CoNiO2/Co4N displayed the highest values based on linear fitting, indicating a faster ion diffusion capability. Overall, combined with the analysis of CV and symmetric cells, it was not difficult to find that CoNiO2/Co4N heterostructure did have the most suitable catalytic activity, which would help to avoid the formation of “dead sulfur”.[ ]

Figure 6

The illustration of Li ion migration pathways and corresponding barriers profles: a,b) graphene and c,d) CoNiO2/Co4N heterostructure. e–g) CV of as‐obtained three electrodes at various scan rates. h–j) The linear fits of the peak currents from CV curves.

The above theoretical calculations and experimental designs have preliminarily evaluated the advantages of CoNiO2/Co4N heterostructure. In order to verify the performance improvement of the actual Li–S battery, a series of electrochemical tests were carried out. Figure shows the voltage‐capacity diagrams at various C‐rates from 0.2 to 4 C. All profiles showed two distinct discharge platforms and a wide charging platform, which were consistent with CV curves. It indicated that the CoNiO2/Co4N–G–S electrode could complete the reversible conversion of polysulfides well even at the high rate of 4 C.[ ] At 0.2, 0.5, 1, 2, and 4 C, capacities of 1198, 941, 860, 765, and 618 mAh g−1 were achieved, respectively. In this voltage range, CoNiO2/Co4N itself only exhibited a capacity of ≈13 mAh g−1 (Figure S12, Supporting Information), much lower than that of CoNiO2/Co4N–G–S, indicating the efficient utilization of sulfur in the composite electrode. Furthermore, with the increase of rate, the polarization (ΔE) value and the increase range of the heterostructure based cell were the lowest among all the comparison samples, indicating that CoNiO2/Co4N provided the most significant improvement in cell dynamics.

Figure 7

Electrochemical performance of CoNiO2/Co4N–G–S and other control cathodes. a) Charge–discharge profiles of CoNiO2/Co4N–G–S at various current densities. b) Comparing the potential polarizations of different electrodes. c) Shuttle current test. d) Cyclic properties of different cathodes at 0.5 C. e) The rate conversion and the subsequent constant current cycle. f) Cyclic stability of the CoNiO2/Co4N–G–S electrode at high rate. g) EIS curves at 20th cycle. (h,i) Cyclic performance of CoNiO2/Co4N–G–S with different sulfur loading at 0.5 C.

To further verify the inhibition effect of the prepared samples on the “shuttle effect”, shuttle current tests were performed. The typical Li–S battery was discharged at 0.5 C to 2.38 V, and then adjusted to constant potential mode. The value of stabilized current can be called shuttle current.[ , ] As shown in Figure 7c, the G–S electrode unsurprisingly showed the maximum shuttle current. On the contrary, the values of CoNiO2/Co4N–G–S, and CoNiO2–G–S were the lowest, and the value of Co4N–G–S was in the middle, indicating the strong polysulfide adsorption of the former two. This was basically consistent with the above theoretical calculation and visual experiment.

Figure 7d compares the cycling performance of different cathodes (1.0 mg cm−2) at 0.5 C. The CoNiO2/Co4N–G–S electrode provided a better cycling stability, maintaining 896 mAh g−1 at 100th cycle with a high Coulombic efficiency of 99.2%. The retention rate of CoNiO2/Co4N–G–S as high as 90.2% was much higher than 72.8% of Co4N–G–S, 76.5% of CoNiO2–G–S, and 56.4% of G–S. Surprisingly, CoNiO2 had a high adsorption for polysulfides, but showed unsatisfactory cycle stability, which might be related to its poor electrical conductivity.[ , ] The rate performances of all the samples were evaluated by current density conversion and cycling (Figure 7e). With the increase of the cycling rate, the decrease of CoNiO2/Co4N–G–S capacity was obviously smaller than that of the other three electrodes. After switching back to 0.2 C, its capacity could be recovered to 997 mAh g−1, indicating the superior reversibility and rapid charge–discharge of the CoNiO2/Co4N–G–S cathode. In particular, during the following 100 cycles of 2 C, the heterostructure sample showed a capacity decrease of 12.8%, which was the lowest among all samples. In contrast, the CoNiO2–G–S, Co4N–G–S, and G–S cells lost almost 23.9%, 31.4%, and 46.7%, respectively. Figure 7f shows a higher rate (4 C) and longer cycle test of CoNiO2/Co4N–G–S. Since there was no previous cycle, its initial capacity was up to 688 mAh g−1. The Coulombic efficiency rapidly increased from 95.7% in the first cycle to 99.6% in the 15 cycles. The reversible capacity retained a high value of 389 mAh g−1 at 600th with a stabilized Coulombic efficiency of nearly 99%, which corresponds to a low decaying rate of ≈0.07% per cycle. Even after 1100 cycles, it still retained considerable capacity, with a corresponding decline rate of only 0.056% per cycle (Figure S13, Supporting Information). The charge–discharge curve of CoNiO2/Co4N–G–S remained stable without any obvious change in polarization, suggesting a homogeneous deposition of insoluble Li2S2/Li2S products without over thick accumulation of insulating sulfides.[ ] Such high rate and low capacity attenuation were superior to most of the previous and recent literatures (Figure S14, Supporting Information).[ − ]

The EIS spectra of all electrodes after cyclic testing are displayed in Figure 7g. Due to excellent electrical conductivity, CoNiO2/Co4N–G–S and Co4N–G–S showed approximately small charge transfer resistance (R ct). The slightly larger R ct of CoNiO2–G–S further proofed its catalytic ability of polysulfide conversion. In comparison with G–S, all the other electrodes showed much smaller R ct at high‐frequency region. Moreover, the heterostructure demonstrated a steeper straight line at low‐frequency, suggesting faster ion migration that was basically agree with the foregoing analysis.[ ] This boosted reaction kinetics not only benefited from the modification of the conductivity of the material, but also from the catalytic effect on polysulfides in the liquid phase.

The construction of high loading electrode is another challenge facing the practical application of high energy density Li batteries.[ ] Figure 7h shows the cyclic performance of CoNiO2/Co4N–G–S cathode with different sulfur loading at 0.5 C. Areal capacities of 1.37, 1.97, 2.50, and 2.20 mAh cm−2 were obtained for the cathodes with 1.6, 2.4, 3.1, and 3.7 mg cm−2 loadings, respectively, and the corresponding capacity was 858, 821, 807, and 601 mAh g−1 (Figure 7i). Due to the acceleration of polysulfide redox kinetics, all high loading cells showed small voltage polarization (Figure S15, Supporting Information). After 100 cycles, the impressive capacity of 602, 472, 483, and 303 mAh g−1 were still maintained.

Post cycling analysis could further prove that heterojunction alleviated the “shuttle effect”. Figure . shows the morphology and elemental analysis of the Li anode after cycling. From Figure. 8a,b, it was not difficult to see that the surface of the Li anode in G–S cell presented a porous structure and was covered with Li dendrites and “dead Li”, which is similar to previous literatures.[ , ] The rough surface might result from the disproportionation reaction caused by the diffusion of a large number of soluble polysulfide compounds toward the anode. The deposited sulfur species modulating the growth morphology of metallic Li.[ ] In contrast, for CoNiO2/Co4N–G–S cell, the Li anode surface was bright without obvious structural damage. A small amount of polysulfide was involved in strengthening the construction of solid–electrolyte interphase film, which made the metal interface more stable and smooth during cycling.[ ] The corresponding EDS spectra (Figure. 8c,d) show the same results. The content of sulfur on the Li surface of CoNiO2/Co4N–G–S battery was much less than that of G–S battery, which meant that the CoNiO2/Co4N could effectively inhibit the diffusion of polysulfides and slow down the “shuttle effect”. In addition, the XRD analysis (Figure S16, Supporting Information) after cycling showed that the CoNiO2/Co4N heterojunction still maintained its original phase, indicating that its overall structure did not change significantly under Li–S electrochemical reactions. This excellent stability would greatly promote the long‐term operation of Li–S batteries. Such good electrochemical properties and post cycling analysis further corroborated the effectiveness of CoNiO2/Co4N heterostructure nanowires design.

Figure 8

SEM of cycled Li–S anodes: a) G–S and b) CoNiO2/Co4N–G–S. Corresponding EDS results of c) G–S and d) CoNiO2/Co4N–G–S.

Conclusions

In summary, to alleviate the polysulfide incompatibility and kinetic delay of graphene host of Li–S batteries, CoNiO2/Co4N heterostructure nanowires medium was successfully prepared. In this design, the nanowire structure would effectively avoid graphene stacking and agglomeration. The band structure of CoNiO2 was accurately tuned by introducing Co4N, which effectively improved the electrical conductivity. Benefiting from the interface effect, as‐obtained heterostructure also retained the higher polysulfide adsorption and electrocatalytic activity of CoNiO2, which greatly reduced the “shuttle effect”. Meanwhile, CoNiO2/Co4N–G–S cathode showed a stronger Li‐ion diffusion behavior. Therefore, the overall performance of cell was enhanced due to the alleviated polysulfide diffusion and improved electrochemistry reaction kinetics. As a result, the CoNiO2/Co4N based Li–S cell achieved high reversibility capacity of 1198 mAh g−1 at 0.2 C, high rate performance (688 mAh g−1 at 4 C), and excellent cyclic stability of 600 cycles with low capacity attenuation of ≈0.07% each cycle on average. The coupling mechanism of graphene supported by heterostructure nanowires will provide new insights and possibilities for the practical application of Li–S batteries in the future.

Experimental Section

Preparation of CoNiO2/Co4N Heterogeneous Nanowires

The NiCo2O4 nanowire precursor was synthesized by simple hydrothermal‐calcination method: 3 mmol Ni(CH3COO)2·4H2O and 6 mmol Co(CH3COO)2·4H2O were added with 1.1 mmol cetyltrimethylammonium bromide (CTAB) and 18 mmol urea into 20 mL of deionized water with stirring to a homogeneous solution. The solution was transferred into 50 mL Teflon‐lined autoclave and maintained at 110 °C for 6 h. After cooling to room temperature, the precipitate was centrifuged and washed. The dried product was annealed at 350 °C for 3 h to obtain the NiCo2O4 nanowires. Subsequently, as‐prepared oxide precursor was nitrided at 350 °C for 1 h in a mixed atmosphere composed of NH3 (150 sccm) and Ar (100 sccm) to form CoNiO2/Co4N heterogeneous nanowires.

Preparation of CoNiO2/Co4N–G–S Composite Cathode

Graphene was prepared by classic Hummers method and reduction at high temperature (1000 °C). After that, the graphene and CoNiO2/Co4N were mixed and dispersed in deionized water at a weight ratio of 2:1. After ultrasonic and stirring, the mixture was freeze‐dried and mechanically ground. The process of loading sulfur was similar to the CS2 dissolution method in the literature.[ ] As‐obtained CoNiO2/Co4N–G composite was wetted with sulfur solution (CS2/N‐methy‐pyrrolidinone (NMP), 1:1) and dried at 50 °C for 12 h. Note that the content of sulfur in this compound was controlled at 67%. For comparison experiments, other composite cathodes (such as CoNiO2–G–S, Co4N–G–S) were prepared by the same process.

Materials Characterization

The morphology and size of the as‐prepared samples were conducted by SEM (Hitachi S‐8100) and TEM (Tecnai G2 F30). The SAED was performed with TEM system at high acceleration voltage of 200 kV. The elemental analyses were collected by XPS (Thermo Scientific K‐Alpha) and EDS (Thermo Fischer, F200x). XRD patterns were characterized for the crystal structure and crystallinity using a Shimadzu X‐ray‐6000 diffractometer with Cu K radiation. The Raman spectra were characterized by Labram HR‐800) with optical micrographs (Nikon A1). The BET surface area was obtained by N2 adsorption measurement on a Kubo X1000 equipment at 77 K. TG analysis was performed using NETZSCH ASAP2020 thermal analyzer.

Visual Polysulfde Adsorption Test

Sulfur and Li2S (5:1 in molar ratio) were mixed and stirred in 1,2‐dimethoxyethane (DME) solvent to obtain a brown‐yellow Li2S6 solution (8.0 mmol L−1). Subsequently, the equivalent amount of CoNiO2/Co4N heterostructure, CoNiO2, and Co4N were added into the above Li2S6 solution (5.0 mL), respectively.

Electrochemical Measurements

Typically, the CoNiO2/Co4N sample, Super P, and polyvinylidene fluoride (PVDF) binder were mixed in NMP solvent with a mass ratio of 8:1:1 to obtain cathode slurry. Then, this slurry was coated onto Al foil current collector using a film applicator. The loading of active sulfur was controlled to 1.0−3.7 mg cm−2 by adjusting the thickness of the coating. After vacuum drying, as‐prepared cathode, Celgard 2400 separator, and Li anode were assembled in coin cell and tested at room temperature. The electrolyte was LiTFSI (1 mol L−1) in 1,3‐dioxolane (DOL) and DME (1:1, volume ratio) with 2 wt% LiNO3. The ratios of electrolyte versus sulfur (E/S) were 15 µL mg−1 (for typical sulfur loading of 1.0 mg cm−2) and 10 µL mg−1 (for high sulfur loadings of 1.6–3.7 mg cm−2), respectively. CV and EIS were measured VSP potentiostat (Bio‐Logic Corp.) and CHI 760D electrochemical analyzer (ChenHua Corp.). The galvanostatic charge–discharge cycling was performed using a Land Battery Tester. The same assembly process was used for the comparison samples.

For symmetrical cells, two identical electrodes were fabricated by a method similar to previous literatures.[ , ] The weight ratio of each additive material (CoNiO2/Co4N, CoNiO2, Co4N, and graphene), conductive agent and PVDF was 7:2:1. Using the above conventional LiTFSI based electrolyte as the solvent, a 0.5 mol L−1 Li2S6 solution was prepared as the new electrolyte. The CV scan rate was 50 mV s−1.

Theoretical Calculations

The optimize geometries and electronic properties of all the investigated structures in this study were calculated at the DFT. The heterojunction was constructed by combining the CoNiO2 (111) surface with Co4N (200) surface based on the experimental observations. Vienna ab initio simulation package (VASP) were employed in the simulations using the projector augmented wave (PAW) potentials with a plane wave cutoff of 500 eV. The Perdew Burke Ernzerhof (PBE) form of the exchange correlations functional was employed in the simulation. Spin‐polarized calculations were performed for all the structures except for the graphene. The k‐point meshes in the simulations were generated using the VASPKIT tool with the grid separation of 0.4 Å−1 for the geometry optimizations and 0.1 Å−1 for the DOS calculations. All the structures were optimized by using the conjugate gradient method, in which the convergence for total energy and interaction force was set to be 10−4 eV and 10−3 eV Å−1, respectively.

The binding strength E b of Li2S4, Li2S6, and Li2S8 on the five investigated substrates were calculated as follows: E b = (E sub + E ps) − E sub+ps, where E sub+ps, E ps, and E sub denote the calculated energies of the total adsorption system, adsorbates, and substrates, respectively. For evaluating the diffusion barrier of Li on the surfaces of graphene and heterojunction, the transitional state was located using the Nudged Elastic Band method.

Conflict of Interest

The authors declare no conflict of interest.

Supporting Information

Click here for additional data file.