1 Introduction
The developments of portable electronic devices and sustainable energy storage solutions have obtained significant progress and received growing scientific interest [
1]. Notablely, traditional lithium-ion batteries suffer from some limitations, include of low energy density (200−250 Wh·kg
−1), environmental toxicity and high production costs [
2]. Due to its high theoretical energy density (2600 Wh·kg
−1), resource saving and environmentally friendly, the lithium-sulfur (Li−S) batteries could be one of efficient clean energy and energy storage solutions in future [
3]. However, some obstacles of Li−S systems should be solved: Firstly, the volume expansion of the S material in lithiation/delithiation process could result in the exfoliation of sulfur materials from the current collector and infrastructural damage of cathode electrodes [
4]. Secondly, low Li
+/e
− conductivity of solid S and lithium polysulfides (LiPSs) could bring about tardy reaction kinetics [
5]. Lastly, the high solubility of long-chain LiPSs can lead to poor cycling performance and irreversible loss of active LiPSs [
6].
For the last few years, some strategies have been employed to address these issues, including introducing hollow structure designs, conductive composites, modified separator, and electrolyte engineering. Typically, the introduction of carbon-based materials as sulfur host materials or MWCNT (multi-walled carbon nanotube) as interlayer can enhance the Li
+/e
− transmission performance in S cathode and effectively restrict the migration of free LiPSs [
7]. Furthermore, the graphene or few-layer redox graphene oxide (rGO) with outstanding electrical conductivity and abundant oxygen-functional-groups [
8], can provide effectively transport network for rapid transport of Li
+/e
− [
9]. Xu
et al. [
10] proposed the Co
9S
8@Graphene nanocomposites to restrain the disclosed LiPSs, which could provide the deposition sites for LiPSs and suppress the formation of “dead sulfur”. Moreover, transition metal oxides, nitrides, and carbides, especially sulfides were considered as catalytical active materials to promote conversion kinetics of LiPSs, through the robust chemical bonds and “sulfurophilic” sites with LiPSs [
11], such as MoS
2 [
12], ZnS [
13], CoS [
14], Co
9S
8 [
15], FeS
2 [
16], and NiS
2 [
17]. In particular, cobalt-based sulfides with metallic or semimetallic characteristics demonstrate higher conductivity, superior polarizability, and sulfurophilic properties and can achieve good rate performance and excellent cycling performance [
18], i.e., heterostructure CoMoS
3/CoS [
19], Co
9S
8/C-nanotubes [
9]. Furthermore, the CoS can promote the nucleation of Li
2S and further decrease the decomposition energy barrier of LiPSs. However, the inevitable volume expansion of S cathodes will lead to the leakage of LiPSs [
20] and introducing modified separators/interlayers with abundant catalytic reaction sites as physical barrier layers can enhance the immobilization efficacy of LiPSs and the electron/Li
+ ion transport properties [
21,
22], i.e., Co
9S
8/CoO-graphene [
23] and Co
9S
8 [
24]. Consequently, transition metal composites with N, P, or O heteroatoms not only can provide the polar adsorption and catalytically active sites for LiPSs transformation, but also can enhance the interface properties [
25]. Furthermore, the metal−sulfur (M−S) bonds between metal and sulfur atoms can further serve as a catalytic center to facilitate Li
+ ion transfer, redox kinetics, and polysulfide conversion [
26]. However, the composite materials with the same anions and cations could synergistically improve LiPSs conversion catalytic performance, due to the interfaces of different components [
27]. Thus, exploring synthetic strategies for homologs is particularly important for efficient catalysis, combining polar materials@carbon, which would be beneficial for increasing active sites and specific surface area, as well as accelerating charge transfer [
28].
In this work, we systematically designed and synthesized Co4S3/Co9S8@rGO composites through simple hydrothermal and high or low-temperature annealing processes, which has been used as sulfur host materials and interlayer. The designed Co4S3/Co9S8@rGO well agglomerate the rGO with the high conductivity and abundant catalytic reaction sites of Co4S3/Co9S8 with nano-size structure. Consequently, the Co4S3/Co9S8@rGO electrode with Co4S3/Co9S8@rGO interlayer exhibits lower charge transfer resistance, more abundant reaction sites and strong LiPSs adsorption effect, benefiting for catalytic conversion of LiPSs and Li2S. Ascribing to these benefits, the Li−S batteries of S@Co4S3/Co9S8@rGO cathodes with Co4S3/Co9S8@rGO-modified separators delivered a high discharge capacity of 1346.13 mAh·g−1 (0.1C) and long-term cycling stability (over 1000 cycles at 2C) with a decay rate of 0.038% per cycle. The electrode with sulfur loading of 2.99 mg cm−2, displayed an initial capacity 1168.16 mAh·g−1 and even under lean electrolyte (E/S: 4 μL·mg−1), the electrodes still demonstrate a high discharge capacity of 843.31 mAh·g−1 at 0.2C and cycling stability.
2 Experimental section
Synthesis of Co4S3/Co9S8@rGO composite. The synthesis of Co
4S
3/Co
9S
8@rGO was prepared by hydrothermal method and thermal treatment. Initially, reduced graphene oxide (0.2 g, rGO; Chengdu Jiacai Technology Co., Ltd.) was placed in a Teflon-lined stainless steel reactor containing 20 mL of deionized water (DI water) and stirred more than 15 minutes [
29]. Secondly, polyvinyl pyrrolidone (0.2 g, PVP; Aladdin), Na
2S
2O
3·5H
2O (1.24 g, Aladdin: purity >99.5%), and Co(NO
3)
3·6H
2O (2.91 g) were successively mixed in DI water in a beaker and magnetically stirred for more than 15 minutes [
29]. Subsequently, the mentioned mixture was heated at 180 °C (3 h), then filtered and washed by DI water or ethanol. After vacuum drying (80 °C, 10 h), the obtained mixture was further heated at 750 °C (1 h) with an Ar atmosphere, and the Co
4S
3/Co
9S
8@rGO was prepared [
29].
Synthesis of S@Co4S3/Co9S8@rGO composite. The S@Co
4S
3/Co
9S
8@rGO was prepared through a thermal method. Firstly, Co
4S
3/Co
9S
8@rGO was uniformly mixed with S powder (sulfur content: 70 wt%; Co
4S
3/Co
9S
8@rGO content: 30 wt%) [
29]. Then, the precursor was kept in Teflon-lined stainless steel autoclave and treated under 155 °C for 12 h to prepare the S@Co
4S
3/Co
9S
8@rGO composite (sulfur content: 70 wt%).
Preparation of Co4S3/Co9S8@rGO modified Celgard. A mixture of Co
4S
3/Co
9S
8@rGO (0.09 g), polyvinylidene difluoride (PVDF, 0.01 g), and N-methyl pyrrolidone (NMP) was mixed by ball milling (2 h, rotational speed: 200 r min
−1). Subsequently, the homogeneous blend was meticulously coated onto separator (Celgard 2400) [
29]. Last, the Co
4S
3/Co
9S
8@rGO@separators after drying (70 °C, 24 h) were successfully fabricated and the loading of Co
4S
3/Co
9S
8@rGO was 0.87 mg·cm
−2.
Computational methods. The Quantum Espresso (QE) program was adopted in the first-principles calculations, and the Perdew−Burke−Ernzerhof (PBE) generalized gradient approximation was utilized to calculate the exchange-correlation functional [
30−
32]. The interactions between electrons and ions were modeled via the projector-augmented-wave (PAW) method, while van der Waals interactions were accounted for using the vdw-DF functional [
33]. A kinetic energy cutoff of 25 Ry was adopted. For the four-layer
p (2 × 2) Co
4S
3(100) and
p (2 × 2) Co
9S
8(111) slabs, a Monkhorst-Pack k-point mesh of 2 × 2 × 1 was used with a vacuum region exceeding 12 Å. The definition of adsorption energy was as follows [
34]:
Ead =
E(ad/surf) −
E(ad) −
E(surf). Therein:
E(ad/surf) is the total binding energies between adsorbate and surface,
E(ad) and
E(surf) are binding energies of free adsorbate in vacuum and clean surface [
34], respectively.
Electrochemical characterization. To reveal its electrochemical performance, the S@Co4S3/Co9S8@rGO (80 wt%), carbon fiber (10 wt%), and PVDF (10 wt%) were mixed in N-methyl pyrrolidinone (NMP) with the addition of 1−2 drops of isopropanol. And then the mixture was uniformly blended through ball milling (2 h, 200 r·min−1) to obtain a slurry and coated onto aluminum foil and finally dried (75 °C, 15 h) to prepare the cathode electrode. In lithium-sulfur batteries, a lithium foil (thickness: 1 mm) was used as the anode and the electrolyte was 0.5 mol/L lithium trifluoromethanesulfonate and 0.5 mol/L LiNO3 in DME/DOL (vol%:vol% = 1:1). The Neware BTS4000 (Shenzhen, China) is used to carried out the galvanostatic tests at 25 °C and the range of cutoff voltage is 1.8−2.6 V. The sulfur content was 70% and sulfur loading was 1−1.5 mg·cm−2.
3 Results and discussion
The Co4S3/Co9S8@rGO was synthesized by hydrothermal and thermal treatment, as shown in Fig. 1(a). Firstly, PVP, Na2S2O3, Co(NO3)2·6H2O and rGO were stirred in DI water to achieve a homogeneous mixture, which was loaded in a hydrothermal reactor and heated at 180 °C for 3 h to prepare the precursor of Co4S3/Co9S8@rGO. After being filtered and washed, the as-prepared precursor of Co4S3/Co9S8@rGO was vacuum dried and heated (750 °C, 1 h) in a tubular furnace to obtain the final production of Co4S3/Co9S8@rGO. The scanning electron microscope (SEM) images and transmission electron microscope (TEM) images are adopted to reveal the morphology and structure of Co4S3/Co9S8@rGO. As shown in Figs. 1(b) and (c), the rGO with thin layer structure is well maintained and cross-linked to form a network, and the Co4S3/Co9S8 was grown on the surface of rGO, which was uniformly distributed across the rGO (Figs. S1a and b), enabling it to maintain a higher active surface area. As shown in Figs. 1(d) and (e), the results of TEM display that the Co4S3/Co9S8 nanoparticles have a sheet structure and the distribution of nanoparticles size has a wide range (an average particle size: 50−340 nm). The high-resolution TEM image reveals that the Co4S3/Co9S8 have well crystallinity and obviously crystal lattice [Fig. 1(f)]. The interplanar spacings of 1.95 Å and 1.7 Å could due to the (102) and (110) crystal planes of Co4S3, respectively, and the interplanar spacings of 2.36 Å and 1.64 Å belong to the (311) and (600) crystal face of Co9S8, confirming that Co4S3 and Co9S8 coexist in rGO [Figs. S1(c) and (d)]. The results of the selected area electron diffraction pattern [Fig. 1(g)] demonstrate that the presence of the (102) plane could belong to Co4S3 and the (311) and (733) planes could belong to Co9S8, further confirming the existence of the two components. The elemental mapping images of the Co4S3/Co9S8 show the equidistribution of Co and S atoms on rGO in the as-prepared composite material (Fig. S2).
The X-ray diffraction [XRD, Fig. 2(a)] is proposed to investigate the crystal structure and components of the synthesized samples, which reveals that the main diffraction peaks could correspond to Co
4S
3 (PDF#02-1458) and Co
9S
8 (PDF#02-1459), and the amorphous peak detected between 17° and 28° could due to the rGO. The result of S@Co
4S
3/Co
9S
8@rGO (S content: 70%) XRD is shown in Fig. S3 and the main diffraction peaks well consistent with S (PDF#08-0247). The X-ray photoelectron spectroscopy (XPS) was employed to analyse valence states and chemical bonding. The survey XPS spectra of Co
4S
3/Co
9S
8@rGO is shown in Fig. S4 and the results demonstrate that the as-prepared material was composed of Co, S, and C. In Fig. 2(b) of Co 2p spectra, the peaks located at 786.25 and 803.86 eV are the satellite peaks, which could be ascribed to the high-spin Co
2+/Co
3+ compounds and configuration interaction between Co
9S
8 and Co
4S
3 [
33]. The double peaks at 779.28 (Co 2p
3/2) and 794.57 eV (Co 2p
1/2) may due to Co
3+ in Co
9S
8, and the double peaks of 782.05 (Co 2p
3/2) and 798.37 eV (Co 2p
1/2) may due to Co
2+ in Co
4S
3 species [
35,
36]. As shown in Fig. 2(c) of the S 2p, the double peaks at 163.1/164.3 eV are attributed to S 2p
3/2/2p
1/2 formed metal-sulfur bonding in sulfide species, whereas those at 168.3 and 169.5 eV should correspond to satellite peaks, due to the combined effects of Co-S bonding and high S content in Co
4S
3/Co
9S
8@rGO [
37]. In Fig. 2(d) of the C 1s, these peaks at 284.2, 285.1 and 288.3 eV can be belonged to the C=C, C−O, and O−C=O bonds in rGO or partially derived from air contamination during sample preparation, respectively [
38].
To further understand the electrochemical property of Co
4S
3/Co
9S
8@rGO, the Co
4S
3/Co
9S
8@rGO was employed as sulfur host materials to prepare the S@Co
4S
3/Co
9S
8@rGO cathode, and work as the materials to modify the separators. In the electrochemical impedance spectroscopy (EIS) plots of Fig. 3(a), the intercept with the Z’ axis is the
Rs impedance, representing the electrolyte resistance. Meanwhile, the mid-frequency semicircle corresponds to the charge transfer resistance (
Rct), including of interface resistance of Li
+ transferring from the electrode/electrolyte interface, electrolyte/interlayer interface and inserting into the electrodes [
39]. Although the values of
Rs in fresh cell with/without Co
4S
3/Co
9S
8@rGO (in Table S1) are 2.28 and 2.33 Ω, respectively, the value of
Rct in fresh cell with Co
4S
3/Co
9S
8@rGO interlayer is obviously reduced to 22.56 Ω, compared with 34.9 Ω of electrodes without Co
4S
3/Co
9S
8@rGO interlayer. After 30 cycles (Fig. S5), the
Rct with/without Co
4S
3/Co
9S
8@rGO electrodes is further decreased to 7.39 and 9.47 Ω. These results demonstrate that the interfacial resistance of the electrode with Co
4S
3/Co
9S
8@rGO interlayer can be reduced, thereby enhancing electrical conductivity, and accelerating electrochemical reaction kinetics [
40]. Compared with the pure S electrode at 0.1C, the charge-discharge curves of S@Co
4S
3/Co
9S
8@rGO electrode with/without interlayer [Fig. 3(b)] displays two typical discharge platforms. The first discharge platform (2.4−2.16 V) represents the reduction of S
8 into long-chain polysulfides, while the second discharge platform (2.16−1.8 V) corresponds to the conversion of long-chain polysulfides into Li
2S
2/Li
2S. The charging curves indicate the reverse process, where Li
2S
2/Li
2S is oxidized back to S
8 [
41]. The initial capacities at 0.1C were 1346 and 1277 mAh·g
−1 for the electrodes with and without the Co
4S
3/Co
9S
8@rGO interlayer, respectively, while the initial capacity of pure S cathode was 850 mAh·g
−1 (Fig. S6), which is remarkable lower than that of S@Co
4S
3/Co
9S
8@rGO electrodes. Meanwhile, the S@Co
4S
3/Co
9S
8@rGO electrode with an interlayer exhibits lower polarization voltage (Δ
E = 0.13 V) than that of without interlayer (Δ
E = 0.15 V) [
29], which can be attributed to the faster kinetics of sulfur redox reactions [
42]. The CV of the S@Co
4S
3/Co
9S
8@rGO electrode are illustrated in Fig. 3(c). The two reduction peaks indicated the processes where sulfur was first reduced to highly soluble polysulfides and then further to insoluble Li
2S
2/Li
2S species, while the two oxidation peaks represented the reverse process of Li
2S
2/Li
2S being re-oxidized back to sulfur [
43]. Compared with the electrode without interlayer, S@Co
4S
3/Co
9S
8@rGO cathode with the interlayer exhibits sharper and higher peak current densities, indicating that the incorporation of the Co
4S
3/Co
9S
8@rGO interlayer could effectively enhance the conversion rate of lithium polysulfides [
44]. As shown in Fig. 3(d), after activation during the first cycle, the subsequent cycles demonstrate high repeatability, indicating its excellent cycle stability [
45]. Furthermore, lithium-ion diffusion performance was proposed to further investigate the influence of Co
4S
3/Co
9S
8@rGO modified through the CV test at various scan rates [Figs. S7(a) and (b)]. The results of the cyclic CV curves with/without the Co
4S
3/Co
9S
8@rGO interlayer display that the oxidation peaks currents and reduction peaks currents become larger at higher scan rate and the diffusion coefficient of Li
+ could be estimated using the Randles−Sevcik equation [
46]:
where
I,
A,
n,
CLi+, and
represent the peak current, electrode area, number of electrons transferred of the reaction, lithium ions concentration, and voltage scan rate [
46], respectively. From Figs. S7(c)−(f), the lithium ions diffusion coefficients for the four peaks were estimated and presented in Table S2. The peaks with separator modification exhibited a higher diffusion coefficient than those without, and the D
Li+ was directly proportional to the slope of the
I vs.
plot [
47]. These results confirmed that the Co
4S
3/Co
9S
8@rGO interlayer would facilitate the lithium ions diffusion rate, and the smaller Tafel plots of the electrodes with interlayer [Figs. S8(a)−(d)] further confirm the faster electrochemical reaction process, occurring at the electrode/electrolyte interfaces [
48]. To investigate the adsorption performance towards LiPSs, 30 mg of rGO or Co
4S
3/Co
9S
8@rGO were added to vials containing 5 mL of a 0.0125 M Li
2S
8 solution (Fig. S9), respectively. After 84 h, the results reveal that the solution containing Co
4S
3/Co
9S
8@rGO had turned almost colorless, while the other two remained reddish-brown with no changes, indicating that the Co
4S
3/Co
9S
8@rGO has strong chemical interaction and adsorption towards polysulfides.
The Co
4S
3/Co
9S
8@rGO not only has strong adsorption towards polysulfides, but also sever as catalyst to accelerate the transformation of LiPSs and inhibit its leakage [
29,
49]. To further investigate its catalytic performance for the redox reactions of LiPSs, Li
2S
6 symmetric battery tests and Li
2S nucleation tests were conducted on the as-prepared materials [
49]. The structure of symmetric batteries is shown in Fig. S10a, where both the working electrode or counter electrode were loaded with Co
4S
3/Co
9S
8@rGO (or rGO) onto the surfaces of carbon sheets (current collectors) [
50]. Additionally, the other symmetric battery was constructed featuring the two sides of separator coated by the Co
4S
3/Co
9S
8@rGO material, as depicted in Fig. S10(b), which would exhibit the importance for double catalytic strategy (Co
4S
3/Co
9S
8@rGO act as the sulfur host and interlayer materials). As shown in Figs. 4(a) and (b), during the forward scan, it was observed that Co
4S
3/Co
9S
8@rGO exhibited a maximum oxidation peak at 0.236 V, which was faster and had a sharper peak shape than rGO (at 0.296 V). When examining the CV curves with the modified interlayer [Fig. 4(c) and Fig. S10(b)] was its structure of symmetric batteries], it was found that the area of the curve increased, and the potential of the maximum oxidation peak (0.165 V) was lower than those without the modified interlayer. Additionally, a second oxidation peak appeared at 1.5 V, which could due to the Co
4S
3/Co
9S
8@rGO interlayer. These results suggested that Co
4S
3/Co
9S
8@rGO had faster redox reaction kinetics [
51]. The Li
2S nucleation experiments were executed to give insights on the liquid-solid transformation kinetics of LiPSs [
52]. The assembled Li−S battery was firstly discharged at 0.6 mA down to 2.11 V and then discharged at the potential of 2.05 V for Li
2S nucleation after stewing 2 minutes [
29], and cut off until the discharge current reached 1.0 × 10
−3 mA [
52]. In Figs. 4(d)−(f), the results confirmed that the deposition rate (
t = 150 s) and current response (
Ipeak = 1.458 mA) of the Co
4S
3/Co
9S
8@rGO electrode with an interlayer are significantly faster and higher than those without an interlayer (
t = 300 s,
Ipeak = 1.054 mA) and the rGO electrode (
t = 630 s,
Ipeak = 0.833 mA). Notably, the Li
2S deposition capacity of the Co
4S
3/Co
9S
8@rGO electrode with an interlayer [644.56 mAh·g
s−1, Fig. 4(d)] was lower than those without an interlayer [950.3 mAh·g
s−1, Fig. 4(e)] and the rGO electrode [707.06 mAh·g
s−1, Fig. 4(f)], which could due to the adsorption between the Co
4S
3/Co
9S
8@rGO interlayer and LiPSs, suppressed migration and lower concentration of LiPSs in electrolyte, in consistent with the fastest deposition rate (
t = 150 s) and the highest current response (
Ipeak = 1.458 mA). Consequently, the double catalytic strategy is effective to enhance conversion reaction kinetics of LiPSs, and facilitating rapid Li
2S nucleation [
53]. Ex situ XPS measurements were conducted at different potentials during the first cycling process at 0.1C to investigate the reaction mechanism of Co
4S
3/Co
9S
8@rGO, which simultaneously serves as both cathode and modified separator in lithium-sulfur batteries. Figure S11 displays the corresponding XPS spectra of the S@Co
4S
3/Co
9S
8@rGO cathode under various voltage states during the discharge/charge process. Compared with the point a (fresh cathode), the peaks at 163.9 and 162.3 eV in S 2p can correspond to the S and Li
2S
2 after discharging to the point b (2.20 V), respectively, where a shift of binding energy (+0.8 eV) of S is detected, indicating lithium polysulfide generation and few formed oxidized sulfide during the reaction process [
54]. At discharge points c (2.00 V), the peak emerged at 160.6 eV corresponds to Li
2S, demonstrating the progressive conversion of Li
2S
2 to Li
2S toward the end of discharge. The S was also detected at discharge termination point d (1.80 V), implying incomplete transformation to Li
2S
2/ Li
2S. Upon charging to points e (2.40 V) and f (2.60 V), the peaks of Li
2S disappeared and the intensity of S peaks increased, confirming the oxidation of Li
2S/Li
2S
2 to S.
In order to further investigate the Co
4S
3/Co
9S
8@rGO in promoting electrochemical properties, the cycling properties of S@Co
4S
3/Co
9S
8@rGO were comprehensively studied in Fig. 5. The S@Co
4S
3/Co
9S
8@rGO electrode with the interlayer delivered a capacity of 1118.39 mAh·g
−1 at 0.2C (at 0.1C, the initial capacity is 1346.13 mAh·g
−1) [Fig. 5(a)] and a capacity of 913.28 mAh·g
−1 was maintained after 100 cycles , which was higher than those of the electrodes without the interlayer (838.16 mAh·g
−1 after 100 cycles) and pure S electrode (Fig. S12). At 0.5C [Fig. 5(b)], the S@Co
4S
3/Co
9S
8@rGO with interlayer displayed a high specific capacity of 1014.61 mAh·g
−1 and a capacity of 755.17 mAh·g
−1 was maintained after 500 cycles with a cycle decay rate of 0.05%, which is better than that of the electrode without interlayer (504.4 mAh·g
−1 after 300 cycles). The separators of the pure S and the S@Co
4S
3/Co
9S
8@rGO electrode was compared after 200 cycles at a rate of 0.2C, respectively (Fig. S13), which can provide the direct evidence on leakage of LiPSs. It should be noted that the separator of the pure S electrode displayed significantly yellowed, while the separator of the S@Co
4S
3/Co
9S
8@rGO electrode demonstrated dark brown, indicating that Co
4S
3 and Co
9S
8 could effectively inhibit the dissolution of LiPSs [
55]. Figure 5(c) illustrates the rate performance of the S@Co
4S
3/Co
9S
8@rGO electrode and the electrode with the interlayer exhibited high discharge capacities of 1311.06, 1078.71, 850.39, 720.09, and 640.26 mAh·g
−1 at 0.1C, 0.2C, 0.5C, 1C and 2C, which are better than those of the electrodes without interlayer, respectively. As the discharge rate was recovered to 0.2C, a reversible capacity of 860.55 mAh·g
−1 was obtained, significantly outperforming the electrode without interlayer. In Fig. 5(d), the discharge/charge curves of the electrode with Co
4S
3/Co
9S
8@rGO interlayer at various current densities are unveiled, which revealed a high coulombic efficiency of 99.91% and well rate capability. Even at 2C [Fig. 5(e)], a reversible discharge capacity of 387.21 mAh·g
−1 after 1000 cycles was remained with an extremely low cycle decay rate of 0.038%. For better understanding the interaction between Co
4S
3/Co
9S
8@rGO and S species, the SEM images of the electrodes with/without interlayer were shown in Fig. S14 after 500 cycles at 2C. Compared with the morphological structure of carbon nanofibers in the electrode, it can been observed that the carbon nanofibers in the electrode with interlayer were interlaced and wrapped by more polysulfides [Fig. S14(a), the red regions] and the carbon nanofibers in the electrode without interlayer demonstrated clear boundaries and isolated linear shapes [Fig. S14(b), the red regions]. Additionally, for the as-prepared Co
4S
3/Co
9S
8@rGO (in the blue regions), larger plate-like regions were formed by Co
4S
3/Co
9S
8@rGO and polysulfides in the electrode with interlayer [Fig. S14(a), the blue regions]. Meanwhile, Co
4S
3/Co
9S
8@rGO displayed scattered and stacked conformation in the electrode without interlayer [Fig. S14(b), the blue regions]. The results of elemental mappings [Figs. S14(c1)−(c3), (d1)−(d3)] further demonstrated uniform distributions of Co, S and C elements in the electrodes with/without interlayer. Furthermore, the comparative analysis revealed that the electrochemical performance of Co
4S
3/Co
9S
8@rGO demonstrated competitiveness against recently reported metal sulfide-based lithium-sulfur batteries (As shown in Table S3). These results confirmed that the Co
4S
3/Co
9S
8@rGO electrodes with interlayer worked as a catcher of LiPSs and finally enhances cycling stability of Li−S batteries [
56].
First principles calculations were conducted to provide deeper insights into the chemical interactions between Co4S3/Co9S8 and LiPSs. As displayed in Fig. 6, the optimal structures of Co4S3 and Co9S8 as well as the binding of Li2Sn (n = 2, 4, 6 or 8) on the flat Co4S3 (100) and Co9S8 (111) surface are viewed. In Fig. 6(a), the adsorption binding energies of Li2Sn (n = 8, 6, 4, 2) on Co4S3 (100) surface are −1.66, −1.51, −2.47 and −2.04 eV, respectively, and the binding energies of Li2S4 and Li2S2 are more negative than those of Li2S8 and Li2S6, indicating the stronger chemical interactions between Co4S3 and the short-chain LiPSs. The calculation results of Co9S8 (111) surface are −1.68 (Li2S8), −1.52 (Li2S6), −2.07 (Li2S4) and −2.55 eV (Li2S2), respectively, thus facilitating the deposition of short-chain polysulfides. Consequently, these results indicated that the Co4S3/Co9S8 can afford stronger adsorption of Li2S4 and Li2S2 and improve catalytic kinetic, which are consistent with the results of Li2S6 symmetric battery and Li2S nucleation analyses. To delve into the application of Co4S3/Co9S8@rGO under high sulfur loading, the electrodes with a sulfur loading of 1.52, 2.3, and 2.99 mg·cm−2 were conducted at 0.2C and the initial discharge capacities at a rate of 0.1C were 1329.36, 1343.27 and 1168.16 mAh·g−1 [Fig. S15(a)], respectively. After 200 cycles, the reversible capacities were 821.37, 738.5 and 681.02 mAh·g−1, respectively. In Fig. S15(b), rate performance under high sulfur loading were conducted to reveal its potential adaptability for high power output, fast charging/discharging property, and adaptive load. The results reveals that the electrodes with the sulfur loading of 2.0 mg·cm−2 delivered an excellent electrochemical performance at 2C. Although the capacity was relatively low at a sulfur loading of 3.28 mg·cm−2 under discharge rate of 0.5C, 1C, and 2C, when the current density reverted to 0.2C, the capacity of the electrode recovered to 1026.62 mAh·g−1. Additionally, the ratio of E/S (electrolyte/sulfur) has directly impact on cost, high energy density and shuttle effect of LiPSs, thus the electrochemical properties of the Li−S batteries under lean electrolyte would provide important application prospect analysis and should be further operated. As depicted in Fig. S16(a), the electrodes with a sulfur loading of 2.83 mg·cm−2 delivered a high discharge capacity of 843.31 mAh·g−1 at 0.2C (with an initial discharge capacity of 1000.19 mAh·g−1 at 0.1C) under an E/S ratio of 4 μL·mg−1 and the capacity of 729.65 mAh·g−1 was maintained after 60 cycles, corresponding to a low capacity decay rate of 0.022% per cycle. As revealed in Figure S16b, the reversible capacities of 619.04 mAh·g−1 (sulfur loading: 2.02 mg·cm−2) and 681.24 mAh·g−1 (sulfur loading: 1.79 mg·cm−2) at 0.2C were remained after 200 cycles, demonstrating excellently electrochemical properties under lean electrolyte condition. These results demonstrate that Co4S3/Co9S8@rGO exhibits excellent cycle stability and reversibility under high sulfur loading and lean electrolyte usage, highlighting its significant potential for development in Li−S batteries. However, due to the high synthesis temperature (750 °C) and high costs, we developed an alternative preparation process of cobalt sulfides@rGO with 200 °C for 2 hours. The results of XRD [Fig. S17(a)] revealed that the as-prepared composites are Co3S4, Co4S3 and Co9S8. When cobalt sulfides@rGO was adopted as S host and interlayer, the modified electrode delivered an initial discharge capacity of 1017.78 mAh·g−1 at 0.2C and maintained a reversible capacity of 808.44 mAh·g−1 after 200 cycles [Fig. S17(b)]. This electrochemical performance demonstrated comparability with the high-temperature synthesized Co4S3/Co9S8@rGO, confirming its practical viability for application and advantages of low-temperature synthesis.
4 Conclusions
In summary, we proposed Co4S3/Co9S8 nanoparticles as the low-cost catalyst for lithium polysulfides (LiPSs) conversion, which were deposited on the surface of redox graphene (rGO) by a simple hydrothermal method and thermal treatment and worked as double catalytic strategy (including of sulfur host materials and interlayer). In the CV and Tafel plots tests suggest that the Co4S3/Co9S8@rGO can optimize the interfaces between electrode and electrolyte, and accelerate the Li+/e− electrical transport properties. Moreover, Li2S6 symmetric battery and Li2S nucleation tests further confirm that the double catalytic strategy can effectively accelerate the reaction kinetics of LiPSs and nucleation of Li2S, inhibiting its migration and shuttle effect. The first principles calculations also indicate that the active components of Co4S3 and Co9S8 have stronger chemisorption of Li2S4 and Li2S2 than those Li2S8 and Li2S6, which can finally restrain the migration of short-chain LiPSs. Facilitating from these advantages, the Li−S batteries with Co4S3/Co9S8@rGO interlayer delivered a capacity of 1346.13 mAh·g−1 at 0.1C and a reversible capacity of 913.28 mAh·g−1 after 100 cycles, and even at 0.5 and 2C, a capacity of 755.17 mAh·g−1 was maintained after 500 cycles and a reversible discharge capacity of 387.21 mAh·g−1 after 1000 cycles was remained with a cycle decay rate of 0.038%, respectively. Under the sulfur loading of 2.99 mg·cm−2, the electrode still delivered a discharge capacity 1168.16 mAh·g−1 at 0.2C and a reversible capacity of 681.02 mAh·g−1 after 200 cycles. Even under lean electrolyte (E/S: 4 μL·mg−1), the electrodes still demonstrate a high discharge capacity of 843.31 mAh·g−1 at 0.2C and a reversible capacity of 681.24 mAh·g−1 at 0.2C were remained after 200 cycles (sulfur loading: 1.79 mg·cm−2), respectively. At last, in order to reduce the cost of preparation, a cobalt sulfides@rGO was also obtained through a lower-temperature treat (200 °C for 2 hours), which displayed electrochemical comparability with the high-temperature synthesized Co4S3/Co9S8@rGO. Consequently, our work provides insights on the catalytic effect of LiPSs conversion, highlighting the efficacy of a double catalytic strategy and the application of cost-effective catalysts in Li−S system.