Electrocatalysis of polysulfide conversion by sulfur-deficient MoS₂ nanoflakes for lithium–sulfur batteries

Abstract

Lithium–sulfur batteries are promising next-generation energy storage devices due to their high energy density and low material cost. Efficient conversion of lithium polysulfides to lithium sulfide (during discharge) and to sulfur (during recharge) is a performance-determining factor for lithium–sulfur batteries. Here we show that MoS_2−x/reduced graphene oxide (MoS_2−x/rGO) can be used to catalyze the polysulfide reactions to improve the battery performance. It was confirmed, through microstructural characterization of the materials, that sulfur deficiencies on the surface participated in the polysulfide reactions and significantly enhanced the polysulfide conversion kinetics. The fast conversion of soluble polysulfides decreased their accumulation in the sulfur cathode and their loss from the cathode by diffusion. Hence in the presence of a small amount of MoS_2−x/rGO (4 wt% of the cathode mass), high rate (8C) performance of the sulfur cathode was improved from a capacity of 161.1 mA h g⁻¹ to 826.5 mA h g⁻¹. In addition, MoS_2−x/rGO also enhanced the cycle stability of the sulfur cathode from a capacity fade rate of 0.373% per cycle (over 150 cycles) to 0.083% per cycle (over 600 cycles) at a typical 0.5C rate. These results provide direct experimental evidence for the catalytic role of MoS_2−x/rGO in promoting the polysulfide conversion kinetics in the sulfur cathode.

---

Broader context

Among the alternatives proposed to succeed lithium-ion batteries, lithium–sulfur batteries have drawn the most interest because of the very high theoretical capacity of the sulfur cathode (about 1672 mA h g⁻¹). In addition, sulfur also has the benefits of being low cost, naturally abundant and environmentally benign. The development of lithium–sulfur batteries is however met with several technical challenges. The insulating properties of sulfur and its discharge products (Li₂S₂ and Li₂S) resulted in a slow discharge/charge process and a low practical capacity. The intermediate products formed during battery discharge and charge, i.e. lithium polysulfides (Li₂S_n, where 3 ≤ n ≤ 8), are electrolyte soluble. The loss of sulfur electrochemically as dissolved lithium polysulfides is the cause of rapid capacity fading during cycling. This article reports the development of an electrocatalyst, MoS_2−x/reduced graphene oxide (MoS_2−x/rGO), which can accelerate the kinetics of polysulfide conversion reactions to insoluble products. Sulfur deficiencies in the MoS₂ nanoflakes were found to be the catalytic centers. The fast conversion of soluble polysulfides can lower their accumulation in the cathode, and hence their effusion from the electrode. Consequently lithium–sulfur batteries using this catalyst in the sulfur cathode could increase the battery rate performance and cycle stability.

Introduction

Among the next-generation rechargeable batteries proposed to succeed lithium-ion batteries, lithium–sulfur batteries have drawn the most interest because of the high theoretical capacity of the sulfur cathode (1672 mA h g⁻¹, about 10 times of that of a typical lithium-ion battery cathode).^1–4 Sulfur also has the advantages of low cost, natural abundance and environmental benignity. The development of lithium–sulfur batteries is however met with several technical challenges. The insulating properties of sulfur and its discharge products (Li₂S₂ and Li₂S) result in a slow discharge/charge process and a low practical capacity. The intermediate products formed during battery discharge and charge, i.e. lithium polysulfides (Li₂S_n, where 3 ≤ n ≤ 8), are electrolyte soluble and as such can migrate to the lithium metal anode and deposit there.^5,6 The loss of electrochemically active lithium polysulfides leads to a rapid capacity fading during cycling.

The strategies developed to date to address these challenges consist mostly of the following: (i) new cathode designs to increase the electrode conductivity and polysulfide retention,^7–10 (ii) new electrolyte formulations,^11,12 separator structure^13–15 and binder chemistry^16,17 to minimize polysulfide migration, and (iii) surface engineering of the lithium metal anode to protect against passivation by the lithium polysulfides migrating from the cathode.^18–20 Although substantial progress has been made, these strategies are still far from realizing the full potential of the lithium–sulfur batteries.

Some recent research has considered the alternative of improving the kinetics of polysulfide conversion in the sulfur cathode.^21–23 During battery discharge and charge, the conversion between sulfur and its end products (Li₂S₂ and Li₂S) has to occur via lithium polysulfides as the intermediate products which are soluble in most lithium–sulfur battery electrolytes used today.²⁴ Since sulfur, Li₂S₂ and Li₂S are insoluble, accelerating the rates of conversion of soluble lithium polysulfides (to S, Li₂S₂ or Li₂S) can reduce the presence of polysulfides in the electrolyte, and hence their impact on the battery performance. This could improve both the sulfur utilization and the battery cycle stability. Although the use of polar compounds such as Magnéli phase Ti₄O₇ (2 × 10³ S cm⁻¹),²³ metal-like TiC (10⁴ S cm⁻¹)²⁵ and CoS₂ (6.7 × 10³ S cm⁻¹)²² as conductive sulfur hosts with good polarity for polysulfide adsorption has been known for some time, the use of platinum, nickel²¹ and cobalt²⁶ as the “catalysts” for polysulfide conversion is a relatively recent development. As such the catalysis of polysulfide conversion is still in an early phase of research.

In the search for catalysts which can provide good performance at low cost, we discovered MoS₂ to be a strong candidate. MoS₂ has been shown to be highly effective for the catalysis of several industrially important reactions such as the hydrogen evolution reaction (HER), the oxygen reduction reaction (ORR) and the oxygen evolution reaction (OER).^27–30 MoS₂ with sulfur deficiencies, in particular, has drawn the most research interest because of the high electrochemical activity associated with the presence of sulfur deficiencies.^31,32 Indeed, our previous work on using MoS₂ as the lithium-ion battery anode has revealed some behavior of the lithium–sulfur batteries, but without the issues of low sulfur conductivity and polysulfide shuttle in discharge and charge.³³ Deficiencies such as MoS₂ edge sites and terrace surfaces have shown good electrochemical activity for Li₂S deposition.³⁴ Herein, rGO decorated with few-layer MoS₂ nanoflakes with a controlled amount of sulfur deficiency (MoS_2−x/rGO) was used to catalyze the polysulfide conversion in a sulfur cathode. The MoS₂ nanoflakes were prepared by the sonication assisted liquid phase exfoliation of commercial MoS₂ powder in N-methyl-2-pyrrolidone (NMP). The amount of sulfur deficiencies could be varied by changing the time and temperature in a heat treatment in hydrogen. The experimental results confirmed the involvement of surface sulfur deficiencies in the polysulfide conversion reactions and their catalytic effect on the kinetics of polysulfide conversion. In the presence of a small amount (4 wt% of the cathode mass) of MoS_2−x/rGO in the sulfur cathode, the sulfur cathode exhibited both high-rate capability (capacity of 826.5 mA h g⁻¹ at an 8C rate) and good cycle stability (capacity fade rate of 0.083% per cycle for 600 cycles at a 0.5C rate). These performances place MoS_2−x/rGO as one of the best (if not the best) polysulfide conversion catalysts reported to date.

Experimental section

Chemicals

N-Methyl-2-pyrrolidone (NMP, 99.5 wt%), polyvinylidene fluoride (PVDF, 99.5 wt%), sulfur (99.5 wt%), lithium sulfide (Li₂S, 99.98 wt%), 1,3-dioxolane (DOL, 99.8 wt%), 1,2-dimethoxyethane (DME, 99.5 wt%), molybdenum(VI) oxide (MoO₃, 99.5%), lithium bis(trifluoromethanesulfonyl) imide (LiTFSI, 99.95 wt%) and lithium nitrate (LiNO₃, 99.99 wt%) from Sigma Aldrich; molybdenum(IV) sulfide (MoS₂, 99 wt%) from Alfa Aesar; and Super-P carbon (99.5 wt%) from Timcal were used as received.

Preparation of MoS₂ nanoflakes and MoS₂/GO composite

Few-layer MoS₂ nanoflakes were prepared by the sonication-assisted exfoliation of commercial MoS₂ powder in NMP.³⁵ In brief, 100 mg MoS₂ powder was dispersed in 20 mL NMP, and sonicated for 5 hours under ambient conditions. After centrifugation at 10 000 rpm for 5 minutes, the supernatant containing the MoS₂ nanoflakes was diluted with 30 mL water to form the MoS₂ stock solution. A graphene oxide (GO) sample prepared using a modified Hummer's method³⁶ was added to this solution and sonically homogenized for 10 minutes. The composite formed as such (MoS₂/GO) was recovered by vacuum filtration.

Preparation of MoS_2−x/rGO and MoS₂/rGO composites

MoS₂ nanoflakes with sulfur deficiencies (sulfur-deficient MoS₂ nanoflakes) were formed by heating the MoS₂/GO composite prepared above in a 10% H₂/Ar mixture. Different combinations of reaction temperature and time were used for the preparation. A MoS₂/rGO composite without the sulfur deficiencies was also prepared for performance comparison. Here the rGO was separately prepared by heating a GO sample in a 10% H₂/Ar atmosphere at 600 °C for 6 hours. The rGO was dispersed into the MoS₂ stock solution to a rGO/MoS₂ mass ratio of 8 : 2 (the same ratio as that of rGO to MoS_2−x in MoS_2−x/rGO), and sonically homogenized for 5 hours. The MoS₂/rGO composite was then recovered by vacuum filtration.

Preparation of rGO/S, MoS₂/rGO/S and MoS_2−x/rGO/S composites

rGO/S, MoS₂/rGO/S and MoS_2−x/rGO/S composites (the actual cathode materials for the lithium–sulfur test batteries) were prepared by the conventional melt-diffusion method. In brief sulfur powder and rGO (or MoS_2−x/rGO or MoS₂/rGO/S) in a 75 : 25 mass ratio were homogenized by grinding; and then sealed in a vial with Ar. The mixture was then heated at 155 °C for 5 hours to distribute the sulfur uniformly in rGO (or in MoS₂/rGO/S or MoS_2−x/rGO).

Materials characterization

The morphology of the composites in this study was examined by field emission scanning electron microscopy (FESEM) on a JEOL JSM-6700F SEM, by transmission electron microscopy (TEM) on a JEOL 2100F microscope, and by high-resolution TEM (HRTEM) on a JEOL 2100F system. The composite crystal structures were determined by X-ray diffraction (XRD) on a BRUKER D8 ADVANCE (Germany) instrument using Cu K_α radiation. X-ray photoelectron spectroscopy (XPS) analysis of the samples was performed on a Kratos AXIS Ultra DLD surface analyzer using a monochromatic Al K_α radiation source at 15 kV (1486.71 eV). The XPS peak locations were corrected by referencing the C 1s peak of adventitious carbon to 284.5 eV. Spectral deconvolution was carried out using the XPS Peak 4.1 software. The rGO and sulfur contents of the composites were analyzed by thermogravimetry (TGA) on a Shimadzu DTG-60H analyzer in air (for the measurement of the rGO content) or in N₂ (for the measurement of the sulfur content) at a temperature ramp rate of 10 °C min⁻¹.

Adsorption properties of lithium polysulfides

Li₂S and sulfur in amounts corresponding to the nominal stoichiometry of Li₂S₆ were added to a 1 : 1 (v/v) DOL/DME mixture and stirred overnight at 60 °C. The concentration of the Li₂S₆ solution prepared as such was 3 mmol L⁻¹, and was used as the stock solution for adsorption measurements. 10 mg rGO, MoS₂/rGO or MoS_2−x/rGO was added to 2 mL each of the lithium polysulfide stock solution. The mixtures were vigorously stirred to facilitate adsorption.

Cell assembly and electrochemical measurements

Symmetric electrochemical cells were assembled by the following procedure: 80 wt% active material (MoS_2−x/rGO, MoS₂/rGO, or rGO) and 20 wt% PVDF binder were homogenized in NMP to form a consistent slurry, which was then uniformly applied to an Al foil. The foil was cut into 1 cm × 1 cm sheets. The active material loadings on the sheets were about 2–4 mg. CR2025 coin cells were assembled in an Ar-filled M Braun glove box by using two coated Al sheets as the cathode and anode, a Celgard 2400 separator, and 50 μL electrolyte of 1 M LiTFSI and 0.2 M Li₂S₆ in a 1 : 1 (v/v) DOL/DME mixture. The counter electrode after the test was disassembled from the cell, rinsed with DOL thrice to remove the lithium salt on the surface; and then evacuated overnight at room temperature for ex situ analysis on the next day. Lithium–sulfur test batteries were assembled by a slightly different procedure: an NMP slurry of 80 wt% active material (MoS_2−x/rGO/S, MoS₂/rGO/S or rGO/S), 10 wt% Super P and 10 wt% PVDF was applied onto an Al foil to a loading of ∼1.5 mg cm⁻². CR2025 coin cells were assembled using the coated Al foil as the cathode, a lithium metal foil anode, a Celgard 2400 separator, and 50 μL 1 M LiTFSI and 2 wt% LiNO₃ solution in DOL/DME (1 : 1 v/v) as the electrolyte. A Neware battery tester was used to regulate the cell discharge and charge. The cathode specific capacities were normalized only by the mass of sulfur, as per the common practice. Cyclic voltammetry (CV) and electrochemical impedance measurements were carried out on an Autolab type III electrochemical workstation.

Results and discussion

Fig. 1A shows the major steps in the preparation of MoS_2−x/rGO. The MoS₂ nanoflakes and GO (Fig. 1B) were co-dispersed in water; and then the mixed solid phase recovered by filtration was heated in a reducing hydrogen atmosphere at high temperature. The nanocomposite formed as such consisted of MoS_2−x nanoflakes on a thin film of rGO. rGO, a common substrate for electrochemical devices,^37–39 was used here as a flexible and conductive catalyst support. Fig. 1C is the TEM image of the MoS₂ nanoflakes formed by the liquid phase exfoliation of bulk commercial MoS₂ particles shown in Fig. S1 (ESI†). MoS₂ could be exfoliated into nanoflakes very easily by this procedure and formed a uniform dispersion in the solvent (Fig. S2, ESI†). The HRTEM image in Fig. 1D shows the layer-like structure of the MoS₂ nanoflakes, which were about 3–5 nm in thickness and consisted of 6–8 layers. The lattice spacing of 0.62 nm matches well with the (002) diffraction of hexagonal MoS₂.⁴⁰ The small MoS₂ nanoflakes were well dispersed on the rGO sheets. The high temperature treatment in hydrogen removed some sulfur atoms in the MoS₂ nanosheets to result in the formation of sulfur deficiencies. The geometric compatibility between the two 2D nanomaterials (rGO and MoS₂ nanosheets) should improve the quality of the interfacial contact, and the 2D-on-2D construction also allowed a good exposure of the sulfur deficiencies on the catalyst surface for the conversion of adsorbed polysulfides.

Fig. 1 (A) Schematic of the synthesis of the MoS_2−x/rGO composite and the conversion of Li₂S_x on the MoS_2−x/rGO surface. TEM images of (B) a thin GO film and (C) MoS₂ nanoflakes. (D) HRTEM image of MoS₂ nanoflakes.

The effects of heat treatment temperature and time on the structure of the MoS_2−x/rGO composite were analyzed by XRD and XPS. The XRD patterns of samples prepared under different conditions are quite similar (Fig. 2A). The broad diffraction at around 2θ = 20–30° can be attributed to the disorderly stacked rGO sheets. The diffraction peaks of the MoS₂ nanoflakes are in good agreement with the 2H phase of MoS₂ (PDF#37-1492),⁴¹ and hence the phase purity of MoS₂ was good. The most intense MoS₂ peak was the (002) peak at 2θ ∼ 15°, suggesting [001] as the crystal growth direction. Fig. 2B shows the expanded view of the MoS₂ (002) peak. There was a slight shift of this peak to lower 2θ values with the increase in treatment severity (higher temperature or longer heat treatment time). The shift indicates an increase in the lattice parameter⁴² caused most likely by the removal of sulfur by hydrogen. The resultant reduction of Mo to a lower oxidation state with a larger atomic radius led to the increase in the lattice parameter.

Fig. 2 (A and B) XRD patterns and (C) Mo 3d XPS spectra of MoS₂ nanoflakes and MoS_2−x/rGO composites formed by different combinations of reaction temperature and time in a hydrogen atmosphere.

The surface compositions of MoS_2−x/rGO composites prepared under different heat treatment conditions were characterized by XPS. The total molybdenum (Mo) and sulfur (S) contents of the samples as analyzed by XPS are summarized in Table S1 (ESI†). The Mo : S ratio of the as-synthesized MoS₂ nanoflakes was 33.2 : 65.4, close to the 1 : 2 ratio in stoichiometric MoS₂. The Mo : S ratio increased with the increase in reaction temperature and reaction time; indicating the progressive removal of the sulfur element. The Mo 3d spectra were deconvoluted to determine the stoichiometric MoS₂ (red), the sulfur-deficient MoS₂ (blue) and the MoO₃ (green) contents of various samples (Fig. 2C).^43,44 Specifically the Mo 3d_5/2 and 3d_3/2 doublets at ∼229.5 eV and 232.5 eV were deconvoluted into two sets of peaks. The first set of peaks with binding energies of 232.6 eV and 229.5 eV could be attributed to stoichiometric MoS₂, while the second set at lower binding energies (232.2 eV and 229.1 eV) could be assigned to sulfur-deficient MoS₂. The peak distinctively upstream of the MoS₂ peaks may be attributed to MoO₃ (green).⁴⁴ The appearance of MoO₃ could be attributed to the oxidation of some low oxidation state Mo atoms in MoS₂, as per the previous report.³⁵ The results showed that the increase in temperature and reaction time increased the amount of sulfur deficiency. The presence of MoO₃ in the 700 °C sample could be due to the oxidation of Mo metal clusters (which is highly susceptible to atmospheric oxidation). Thermal annealing of MoS₂ in a hydrogen environment could lead to the removal of sulfur atoms as H₂S gas and hence the formation of sulfur deficiencies. The excess Mo could also form Mo metal clusters.⁵⁹ The Mo metal clusters were oxidized to MoO₃ after the sample was removed from the heating chamber. MoO₃ could also be formed directly by the substitution of sulfur in MoS₂ with oxygen from GO during the heat treatment. Although the 700 °C sample contained a high sulfur deficiency content, the deep reduction of MoS₂ led to an unstable nanoflake structure (Fig. S3, ESI†) caused by the excessive expansion of the lattice parameter. The MoS_2−x/rGO composite with the highest sulfur deficiency content which could still preserve the nanoflake structure was prepared at 600 °C for 6 hours (x = 0.42).

The morphology of the stable high sulfur-deficiency MoS_2−x/rGO composite (prepared at 600 °C for 6 hours) was examined by both FESEM and TEM. The FESEM image in Fig. 3A shows that the composite mirrored the laminate structure of rGO synthesized under the same conditions (Fig. S4A, ESI†). TEM images (Fig. 3B and C) confirm the presence of MoS_2−x nanoflakes on the rGO sheets. The 0.62 nm lattice spacing in the HRTEM image of a MoS_2−x nanoflake sample (Fig. 3D) is the same as that of the (002) planes of hexagonal MoS₂ in Fig. 1D, indicating that the sulfur deficiencies did not alter the native MoS₂ structure. The rGO content in the MoS_2−x/rGO composite as calculated by TGA (Fig. S5, ESI†) was about 78 wt%. For comparison, a composite containing MoS₂ and rGO (reduced from GO by heating in hydrogen at 600 °C for 6 hours), MoS₂/rGO, was also analyzed. The rGO content in the latter was similar, 77 wt%. TGA nonetheless detected higher thermal stability for rGO in MoS_2−x/rGO to suggest the stronger interaction between rGO and MoS₂ when the latter was sulfur deficient. Since sulfur deficiencies can render the MoS_2−x surface more electron rich,⁴⁵ and rGO formed below 600 °C has general p-type characteristics,⁴⁶ electron transfer from rGO to MoS_2−x may occur to develop a stronger bond between the two at their interface.⁴⁷ Such electron coupling is expected to contribute positively to the charge transfer at the MoS_2−x/rGO interface.

Fig. 3 (A) FESEM image, (B and C) TEM images and (D) HRTEM image of the MoS_2−x/rGO composite prepared by heating MoS₂/GO in a hydrogen atmosphere at 600 °C for 6 hours.

The catalytic effect of MoS_2−x on the polysulfide redox reactions was first revealed by CV in symmetric cells with identical working and counter electrodes in a 0.2 M Li₂S₆ electrolyte (Fig. 4A). MoS₂/rGO and rGO prepared under the same conditions were used as the experimental controls (Fig. 4B and C). The CV of a Li₂S₆-free electrolyte was also measured to correct for capacitive contributions. The voltammogram of the MoS_2−x/rGO electrode in the Li₂S₆ electrolyte exhibited high reversibility with four distinct peaks at −0.047 V, −0.39 V, 0.047 V and 0.39 V respectively (Fig. 4A). The MoS₂/rGO electrode exhibited remnants of these peaks as broad redox features at −0.31 V, −0.61 V, 0.31 V and 0.61 V (Fig. 4B). For the rGO electrode, only a very drawn-out reduction peak at −1.22 V and a very drawn-out oxidation peak at 1.22 V were detected (Fig. 4C).

Fig. 4 Cyclic voltammograms of symmetric cells with identical electrodes of (A) MoS_2−x/rGO, (B) MoS₂/rGO and (C) rGO in electrolytes with and without 0.2 M Li₂S₆ at 3 mV s⁻¹. (D) Multi-cycle voltammograms of the MoS_2−x/rGO symmetric cell at 3 mV s⁻¹. (E) Electrochemical impedance spectra of the symmetric cells. (F) Voltammograms of the MoS_2−x/rGO symmetric cell at different scan rates.

Fig. 4D shows the first five cycles of the MoS_2−x/rGO electrode in CV. The nearly perfect superimposition of the peaks suggests good stability of the sulfur-deficient MoS_2−x/rGO electrode. In the first cathodic scan from zero potential between the electrodes, only the peak at −0.39 V (peak a) appeared. The cathodic peak at −0.047 V (peak d) emerged only from the second scan onwards. Since Li₂S₆ was the only electrochemically active species in the electrolyte, it is reasonable to assume that Li₂S₆ was reduced to Li₂S (or Li₂S₂) on the working electrode, and oxidized to sulfur on the counter electrode in the cathodic scan. Hence the reduction of Li₂S₆ on the working electrode which manifested in peak a was complemented by the oxidation of Li₂S₆ on the counter electrode. Peak b in the following anodic scan was due to the reconstitution of Li₂S₆ by the oxidation of Li₂S (or Li₂S₂) on the working electrode. Similarly, peaks c and d identical in shape to peaks a and b were due to the oxidation of Li₂S₆ to sulfur, and the reduction of sulfur to Li₂S₆ on the working electrode respectively. Therefore, the peaks at −0.39 V/0.047 V and −0.047 V/0.39 V‡ were paired redox features of the symmetric cell. The reactions are summarized in Fig. S6 (ESI†). The sharpness of the peaks and the narrow peak separation in each redox pair indicate good electrochemical reversibility and facile polysulfide conversion. It should be mentioned that the two paired redox peaks related to the Li₂S₆ conversion reaction were absent in a previous study on CoS₂ using the symmetric cell.²² The high scan rate (50 mV s⁻¹) and the narrow voltage range (from −0.7 V to 0.7 V) used in that study could have suppressed the detectability of these redox features. When the electrodes were MoS₂/rGO without the surface sulfur deficiencies, the broadened peaks and the increased peak separation are indications of reduced electrochemical reversibility and slower reactions (Fig. 4B).§ Electrochemical reversibility and conversion kinetics were the lowest with the rGO electrodes, resulting in the merging of peaks (Fig. 4C).

The polarity-induced adsorption between the polysulfides and a polar sulfide surface may have contributed to the more facile kinetics of polysulfide conversion on MoS₂/rGO and MoS_2−x/rGO (the apolar rGO surface is antagonistic to polysulfide adsorption).⁴⁸ This was demonstrated by a simple visual adsorption test (Fig. S8, ESI†) where the adsorption of Li₂S₆ on MoS₂/rGO and MoS_2−x/rGO completely decolorized the polysulfide solution. Electrochemical impedance spectroscopy (Fig. 4E) also registered the smallest charge transfer resistance (the size of the high frequency semicircle in the Nyquist plot) for the MoS_2−x/rGO symmetric cell. In the CVs measured at different scan rates (Fig. 4F), there were some slight shifts of the redox peaks with the increase in the scan rate. However, the peak separation in the MoS_2−x/rGO cell at a high scan rate of 9 mV s⁻¹ was still significantly narrower than the peak separations in the MoS₂/rGO or rGO cells at 3 mV s⁻¹. All the above are evidence for the greatly enhanced kinetics of polysulfide conversion on the MoS_2−x/rGO surface.

For additional insights into the reactions of polysulfides on the MoS_2−x/rGO surface, the counter electrodes of symmetric cells with MoS_2−x/rGO or rGO electrodes after scanning from 0 V to −1.4 V were examined by FESEM.¶Fig. 5 shows the FESEM images of the rGO (A–C) and the MoS_2−x/rGO (D–F) counter electrodes before and after scanning to −1.4 V. The larger number of sulfur particles and their more uniform distribution on the MoS_2−x/rGO surface could only come from the presence of more electrochemically active sites for polysulfide conversion. XPS was also used to analyze the surfaces of the counter electrodes of symmetric cells after scanning from 0 to −0.14 V, and from −0.14 V to 0 V. Fig. 5G and H show, respectively, the S 2p XPS spectra of the rGO and MoS_2−x/rGO counter electrodes. The 163.5 eV and 164.7 eV peaks could be attributed to the sulfur deposited on the counter electrodes, while the very prominent peak at ∼169 eV to the S–O bond in oxidized sulfur species such as –SO_x.²¹ Since sulfur deposition on the counter electrode was mostly completed when the symmetric cells were scanned to −0.39 V (Fig. 4D), the sulfur deposit would be extensively oxidized when the cells were scanned to −1.4 V. The stronger S–O peak from the MoS_2−x/rGO cell can then be used as an indirect evidence for more sulfur formation in this cell (Fig. 5G and H). When this symmetric cell was returned to 0 V from −1.4 V, the decrease in the S–O peak intensity was caused by the electrochemical reduction of the oxidized sulfur species on the counter electrode. In contrast, the S–O peak from the rGO symmetric cell underwent very minor intensity changes from −1.4 V to 0 V, an indication of the limited sulfur presence on the rGO surface. The more extensive sulfur formation and reduction reactions in the MoS_2−x/rGO cell could only be caused by the existence of catalytically more active sites on the MoS_2−x/rGO surface. There was also evidence in the Mo 3d XPS spectra for the interaction between MoS_2−x and polysulfides during the polysulfide conversion reactions (Fig. 5I). When the symmetric cells were scanned to −1.4 V, the sulfur-deficient MoS₂ component (blue curve) in MoS_2−x/rGO was significantly diminished in intensity. It is believed that the deficiencies in MoS_2−x rendered the surface of MoS_2−x/rGO electron-rich. The XPS results of Fig. 5I indicate a weaker XPS signal from the sulfur deficiencies after sulfur deposition, suggesting the electron transfer from the former to the latter.^43,49,50 It has been reported in the oxygen reduction reaction (ORR) research that oxygen adsorption on an oxygen-deficient surface would elongate the O–O bond for an easier reduction.⁵⁶ The sulfur deficiencies on the MoS_2−x/rGO surface may likewise facilitate the reduction of sulfur to polysulfides, probably through the involvement of some metastable S_x˙⁻ species.^57,58 When the symmetric cell was scanned back to 0 V, XPS showed that the sulfur deficiencies were restored, and hence the reversibility of the overall process. These changes establish the correspondence between sulfur deficiency and the extent and reversibility of polysulfide conversion, and provide indirect proof for sulfur deficiencies as the origin of enhanced catalytic activity in polysulfide electrochemical reactions.

Fig. 5 FESEM images of (A) the pristine rGO electrode and (B and C) the rGO counter electrode removed from the symmetric cell after scanning to −1.4 V; (D) the pristine MoS_2−x/rGO electrode and (E and F) the MoS_2−x/rGO counter electrode after scanning to −1.4 V. XPS spectra of (G) the rGO and (H and I) MoS_2−x/rGO counter electrodes of symmetric cells after scanning to −1.4 V, or after scanning to −1.4 V and returning to 0 V.

The actual performance of MoS_2−x/rGO as a catalyst in lithium–sulfur batteries was evaluated in coin cells using a MoS_2−x/rGO/S composite cathode and a lithium metal anode. Coin cells with a MoS₂/rGO/S or rGO/S cathode were also assembled for comparison. The sulfur contents in the composites as assayed by TGA were about 75 wt% (Fig. S9, ESI†). Fig. 6A shows the typical voltammograms and the galvanostatic discharge–charge voltage profile of the MoS_2−x/rGO/S cathodes between 1.8 and 2.6 V. Since lithiation of MoS₂ occurs below 1.5 V vs. Li/Li⁺, the MoS₂ nanoflakes would not have contributed to any capacity in the 1.8–2.6 V voltage range.⁵¹ The two cathodic peaks at about 2.3 V and 2.0 V could be associated with the reduction of sulfur to soluble long-chain lithium polysulfides (Li₂S_x, 4 ≤ x ≤ 8), and the subsequent conversion of the latter to insoluble short-chain polysulfides (Li₂S₂/Li₂S) respectively. The two peaks in the reverse anodic scan at about 2.3 V and 2.4 V represent the reverse reactions of the conversion of short-chain polysulfides to sulfur.^52,53 The two distinct discharge voltage plateaus (at ∼2.34 V and 2.12 V) at the 0.5C rate could be attributed to the conversion of sulfur to long-chain lithium polysulfides, and the formation of Li₂S₂/Li₂S from the latter. The reverse of these reactions occurred during charge to form two corresponding voltage plateaus at ∼2.23 V and 2.35 V respectively. The galvanostatic discharge and charge curves are therefore in agreement with the voltammograms.^54,55 The generally intense peaks on MoS_2−x/rGO/S indicate the great extent of the polysulfide conversion due to the fast electrode kinetics.

Fig. 6 (A) Cyclic voltammograms at 0.1 mV s⁻¹ and the representative galvanostatic discharge–charge voltage profile at 0.5C, (B) comparison of rate performance at different C-rates, (C) galvanostatic discharge–charge curves and (D and E) cycle stability of rGO/S, MoS₂/rGO/S and MoS_2−x/rGO/S cells in the 1.8–2.6 V voltage range at 0.5C (1C = 1600 mA g⁻¹ based on the mass of sulfur).

Fig. 6B compares the electrochemical performance of the MoS_2−x/rGO/S, MoS₂/rGO/S and rGO/S cathodes at different C-rates (from 0.2C to 8C; 1C = 1600 mA g⁻¹). The first cycle discharge capacities of rGO/S, MoS₂/rGO/S and MoS_2−x/rGO/S at the 0.2C rate were 1210.5 mA h g⁻¹, 1243.2 mA h g⁻¹ and 1310.5 mA h g⁻¹ respectively. The capacity difference deviated more at higher rates, and was 826.5 mA h g⁻¹ for MoS_2−x/rGO/S (63.1% of its 0.2C capacity), 473.3 mA h g⁻¹ for MoS₂/rGO/S (38.1% of its 0.2C capacity) and 161.1 mA h g⁻¹ for rGO/S (13.3% of its 0.2C capacity) at the 8C rate, which was the test limit of this study. The higher affinity of MoS_2−x/rGO for polysulfide adsorption and the catalytic effect of sulfur deficiencies in MoS₂ for polysulfide conversion are expected to be the contributive factors although their respective contributions are difficult to resolve at this time. Fig. 6C shows the galvanostatic discharge and charge curves of the cells at different C-rates. An increase in the C-rate caused the charge voltage plateaus to shift positively and the discharge voltage plateaus to shift negatively. The voltage plateaus at the 8C rate were clearly visible in the MoS_2−x/rGO/S cell, due to the more facile electrode kinetics on MoS_2−x/rGO.

The cyclabilities of the MoS_2−x/rGO/S, MoS₂/rGO/S and rGO/S cathodes at the typical 0.5C rate are compared in Fig. 6D. Not only were the rGO/S and MoS₂/rGO/S cathodes lower in initial capacity (1013.3 mA h g⁻¹ and 1033 mA h g⁻¹), they also exhibited more severe capacity fading endings with 445.3 mA h g⁻¹ and 576.4 mA h g⁻¹ after 150 cycles. In contrast, the MoS_2−x/rGO/S cathode exhibited both higher discharge capacity and greater cycle stability (initial discharge capacity of 1159.9 mA h g⁻¹ and capacity of 819.9 mA h g⁻¹ after 150 cycles). The long-term cycling performance of MoS_2−x/rGO/S at the 0.5C rate was also evaluated (Fig. 6E). After 600 cycles of continuous cycling, a discharge capacity of 628.2 mA h g⁻¹ remained (a capacity fade rate of 0.083% per cycle). The Coulombic efficiency was as high as 99.6%. Cycle stability was thereof another benefit of the catalysis of polysulfide conversion in the sulfur electrode. A higher conversion rate of soluble polysulfides to insoluble sulfur products could decrease their accumulation in the cathode and consequently, their loss from the cathode by diffusion. Greater cycle stability was therefore realized by suppressing a major capacity loss mechanism. Compared to other catalysts in use today for the lithium–sulfur batteries (Table S2, ESI†), MoS_2−x/rGO is clearly a well-rounded choice with a strong performance in almost all functional categories.

Conclusions

In this study, we demonstrated the effectiveness of MoS_2−x/rGO as a catalyst for polysulfide conversion in a sulfur cathode. It was confirmed that the surface sulfur deficiencies participated in the polysulfide conversion and catalyzed the kinetics of polysulfide redox reactions. When a small amount of MoS_2−x/rGO (4 wt%) was added to the sulfur cathode, high-rate performance and good cycle stability of the batteries were obtained. The high rate performance could be attributed to the acceleration of the polysulfide conversion kinetics on the surface sulfur deficiencies. The fast conversion of soluble polysulfides decreased their accumulation in the sulfur cathode and inhibited their subsequent loss from the cathode by diffusion. The suppression of this loss mechanism led to a more sustained cyclability. The study here not only presented a catalyst candidate which is among the best reported to date, but it also provided experimental evidence for and some new insights into the origin of the catalytic effects.

Acknowledgements

H. L. acknowledges the National University of Singapore for his research scholarship.

References

1. W. Zhou, B. Guo, H. Gao and J. B. Goodenough, Adv. Energy Mater., 2016, 6, 1502059.
2. J. Li, L. Yang, S. Yang and J. Y. Lee, Adv. Energy Mater., 2015, 5, 1501808.
3. J. Zhang, H. Hu, Z. Li and X. W. Lou, Angew. Chem., Int. Ed., 2016, 55, 3982–3986.
4. M. Wild, L. O'Neill, T. Zhang, R. Purkayastha, G. Minton, M. Marinescu and G. J. Offer, Energy Environ. Sci., 2015, 8, 3477–3494.
5. Y. X. Yin, S. Xin, Y. G. Guo and L. J. Wan, Angew. Chem., Int. Ed., 2013, 52, 13186–13200.
6. Q. Pang, X. Liang, C. Y. Kwok and L. F. Nazar, Nat. Energy, 2016, 1, 16132.
7. X. Ji, K. T. Lee and L. F. Nazar, Nat. Mater., 2009, 8, 500–506.
8. H.-J. Peng, J.-Q. Huang, M.-Q. Zhao, Q. Zhang, X.-B. Cheng, X.-Y. Liu, W.-Z. Qian and F. Wei, Adv. Funct. Mater., 2014, 24, 2772–2781.
9. G. Zheng, Q. Zhang, J. J. Cha, Y. Yang, W. Li, Z. W. Seh and Y. Cui, Nano Lett., 2013, 13, 1265–1270.
10. X. Wang, T. Gao, X. Fan, F. Han, Y. Wu, Z. Zhang, J. Li and C. Wang, Adv. Funct. Mater., 2016, 26, 7164–7169.
11. S. Zhang, K. Ueno, K. Dokko and M. Watanabe, Adv. Energy Mater., 2015, 5, 1500117.
12. J. Chen, K. S. Han, W. A. Henderson, K. C. Lau, M. Vijayakumar, T. Dzwiniel, H. Pan, L. A. Curtiss, J. Xiao, K. T. Mueller, Y. Shao and J. Liu, Adv. Energy Mater., 2016, 6, 1600160.
13. J. Balach, T. Jaumann, M. Klose, S. Oswald, J. Eckert and L. Giebeler, Adv. Funct. Mater., 2015, 25, 5285–5291.
14. S. H. Chung and A. Manthiram, Adv. Mater., 2014, 26, 1360–1365.
15. J.-Q. Huang, Q. Zhang, H.-J. Peng, X.-Y. Liu, W.-Z. Qian and F. Wei, Energy Environ. Sci., 2014, 7, 347–353.
16. P. Bhattacharya, M. I. Nandasiri, D. Lv, A. M. Schwarz, J. T. Darsell, W. A. Henderson, D. A. Tomalia, J. Liu, J.-G. Zhang and J. Xiao, Nano Energy, 2016, 19, 176–186.
17. T. Nakazawa, A. Ikoma, R. Kido, K. Ueno, K. Dokko and M. Watanabe, J. Power Sources, 2016, 307, 746–752.
18. G. Zheng, S. W. Lee, Z. Liang, H. W. Lee, K. Yan, H. Yao, H. Wang, W. Li, S. Chu and Y. Cui, Nat. Nanotechnol., 2014, 9, 618–623.
19. W. Xu, J. Wang, F. Ding, X. Chen, E. Nasybulin, Y. Zhang and J.-G. Zhang, Energy Environ. Sci., 2014, 7, 513–537.
20. Y. Yang, G. Zheng and Y. Cui, Energy Environ. Sci., 2013, 6, 1552.
21. H. Al Salem, G. Babu, C. V. Rao and L. M. Arava, J. Am. Chem. Soc., 2015, 137, 11542–11545.
22. Z. Yuan, H. J. Peng, T. Z. Hou, J. Q. Huang, C. M. Chen, D. W. Wang, X. B. Cheng, F. Wei and Q. Zhang, Nano Lett., 2016, 16, 519–527.
23. Q. Pang, D. Kundu, M. Cuisinier and L. F. Nazar, Nat. Commun., 2014, 5, 4759.
24. Y. Gorlin, M. U. M. Patel, A. Freiberg, Q. He, M. Piana, M. Tromp and H. A. Gasteiger, J. Electrochem. Soc., 2016, 163, A930–A939.
25. H. J. Peng, G. Zhang, X. Chen, Z. W. Zhang, W. T. Xu, J. Q. Huang and Q. Zhang, Angew. Chem., Int. Ed., 2016, 55, 12990–12995.
26. Y.-J. Li, J.-M. Fan, M.-S. Zheng and Q.-F. Dong, Energy Environ. Sci., 2016, 9, 1998–2004.
27. K. Chang, Z. Mei, T. Wang, Q. Kang, S. Ouyang and J. Ye, ACS Nano, 2014, 8, 7078–7087.
28. J. Kibsgaard, Z. Chen, B. N. Reinecke and T. F. Jaramillo, Nat. Mater., 2012, 11, 963–969.
29. D. Kiriya, P. Lobaccaro, H. Y. Nyein, P. Taheri, M. Hettick, H. Shiraki, C. M. Sutter-Fella, P. Zhao, W. Gao, R. Maboudian, J. W. Ager and A. Javey, Nano Lett., 2016, 16, 4047–4053.
30. M. Asadi, B. Kumar, C. Liu, P. Phillips, P. Yasaei, A. Behranginia, P. Zapol, R. F. Klie, L. A. Curtiss and A. Salehi-Khojin, ACS Nano, 2016, 10, 2167–2175.
31. Y. Yin, J. Han, Y. Zhang, X. Zhang, P. Xu, Q. Yuan, L. Samad, X. Wang, Y. Wang, Z. Zhang, P. Zhang, X. Cao, B. Song and S. Jin, J. Am. Chem. Soc., 2016, 138, 7965–7972.
32. H. Li, C. Tsai, A. L. Koh, L. Cai, A. W. Contryman, A. H. Fragapane, J. Zhao, H. S. Han, H. C. Manoharan, F. Abild-Pedersen, J. K. Norskov and X. Zheng, Nat. Mater., 2016, 15, 48–53.
33. G. Ji, Y. Yu, Q. Yao, B. Qu, D. Chen, W. Chen, J. Xie and J. Y. Lee, NPG Asia Mater., 2016, 8, e247.
34. H. Wang, Q. Zhang, H. Yao, Z. Liang, H.-W. Lee, P.-C. Hsu, G. Zheng and Y. Cui, Nano Lett., 2014, 14, 7138–7144.
35. A. Jawaid, D. Nepal, K. Park, M. Jespersen, A. Qualley, P. Mirau, L. F. Drummy and R. A. Vaia, Chem. Mater., 2016, 28, 337–348.
36. W. Tang, B.-M. Goh, M. Y. Hu, C. Wan, B. Tian, X. Deng, C. Peng, M. Lin, J. Z. Hu and K. P. Loh, J. Phys. Chem. C, 2016, 120, 2600–2608.
37. B. Qu, C. Ma, G. Ji, C. Xu, J. Xu, Y. S. Meng, T. Wang and J. Y. Lee, Adv. Mater., 2014, 26, 3854–3859.
38. J. Wu, M. Liu, P. P. Sharma, R. M. Yadav, L. Ma, Y. Yang, X. Zou, X. D. Zhou, R. Vajtai, B. I. Yakobson, J. Lou and P. M. Ajayan, Nano Lett., 2016, 16, 466–470.
39. F. Wu, J. Li, Y. Su, J. Wang, W. Yang, N. Li, L. Chen, S. Chen, R. Chen and L. Bao, Nano Lett., 2016, 16, 5488–5494.
40. M. Thripuranthaka, R. V. Kashid, C. Sekhar Rout and D. J. Late, Appl. Phys. Lett., 2014, 104, 081911.
41. X. Fan, P. Xu, Y. C. Li, D. Zhou, Y. Sun, M. A. Nguyen, M. Terrones and T. E. Mallouk, J. Am. Chem. Soc., 2016, 138, 5143–5149.
42. J. Xiao, X. Chen, P. V. Sushko, M. L. Sushko, L. Kovarik, J. Feng, Z. Deng, J. Zheng, G. L. Graff, Z. Nie, D. Choi, J. Liu, J.-G. Zhang and M. S. Whittingham, Adv. Mater., 2012, 24, 2109–2116.
43. D. M. Sim, M. Kim, S. Yim, M.-J. Choi, J. Choi, S. Yoo and Y. S. Jung, ACS Nano, 2015, 9, 12115–12123.
44. I. S. Kim, V. K. Sangwan, D. Jariwala, J. D. Wood, S. Park, K.-S. Chen, F. Shi, F. Ruiz-Zepeda, A. Ponce, M. Jose-Yacaman, V. P. Dravid, T. J. Marks, M. C. Hersam and L. J. Lauhon, ACS Nano, 2014, 8, 10551–10558.
45. L.-p. Feng, J. Su and Z.-t. Liu, RSC Adv., 2015, 5, 20538–20544.
46. N. D. K. Tu, J. Choi, C. R. Park and H. Kim, Chem. Mater., 2015, 27, 7362–7369.
47. X. Wang, G. Li, M. H. Seo, F. M. Hassan, M. A. Hoque and Z. Chen, Adv. Energy Mater., 2015, 5, 1501106.
48. Q. Pang and L. F. Nazar, ACS Nano, 2016, 10, 4111–4118.
49. H.-J. Peng, Z.-W. Zhang, J.-Q. Huang, G. Zhang, J. Xie, W.-T. Xu, J.-L. Shi, X. Chen, X.-B. Cheng and Q. Zhang, Adv. Mater., 2016, 28, 9551–9558.
50. E. Gracia-Espino, G. Hu, A. Shchukarev and T. Wågberg, J. Am. Chem. Soc., 2014, 136, 6626–6633.
51. J. Ye, W. Chen, Q. Chen, Z. Yu and J. Y. Lee, Electrochim. Acta, 2016, 190, 538–547.
52. L. Wang, Y. Wang and Y. Xia, Energy Environ. Sci., 2015, 8, 1551–1558.
53. C. Wang, X. Wang, Y. Yang, A. Kushima, J. Chen, Y. Huang and J. Li, Nano Lett., 2015, 15, 1796–1802.
54. K. Park, J. H. Cho, J.-H. Jang, B.-C. Yu, A. T. De La Hoz, K. M. Miller, C. J. Ellison and J. B. Goodenough, Energy Environ. Sci., 2015, 8, 2389–2395.
55. J. Liu, W. Li, L. Duan, X. Li, L. Ji, Z. Geng, K. Huang, L. Lu, L. Zhou, Z. Liu, W. Chen, L. Liu, S. Feng and Y. Zhang, Nano Lett., 2015, 15, 5137–5142.
56. F. Cheng, T. Zhang, Y. Zhang, J. Du, X. Han and J. Chen, Angew. Chem., Int. Ed., 2013, 52, 2474–2477.
57. H. L. Wu, L. A. Huff and A. A. Gewirth, ACS Appl. Mater. Interfaces, 2015, 7, 1709–1719.
58. D.-H. Han, B.-S. Kim, S.-J. Choi, Y. Jung, J. Kwak and S.-M. Park, J. Electrochem. Soc., 2004, 151, E283.
59. D. Kiriya, P. Lobaccaro, H. Y. Y. Nyein, P. Taheri, M. Hettick, H. Shiraki, C. M. Sutter-Fella, P. Zhao, W. Gao, R. Maboudian, J. W. Ager and A. Javey, Nano Lett., 2016, 16, 4047–4053.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7ee01047h

‡ These voltage values are potential differences between two redox reactions on two electrodes, and as such cannot be associated with half reactions.

§ We have also prepared another composite (MoS_2−x/rGO-2) by mixing rGO and MoS_2−x in water, and used it as the electrodes of a symmetric cell (Fig. S7, ESI†). Its redox performance was slightly inferior to the MoS_2−x/rGO electrode due to the poor mixing between MoS_2−x and rGO by this preparation route, which increased the likelihood of nanoflake aggregation. Nonetheless, sharper and narrower redox peaks were still obtained relative to the MoS₂/rGO electrode. The positive effect of sulfur deficiencies on polysulfide conversion was again demonstrated.

¶ The counter electrode instead of the working electrode was analysed because it was where sulfur was deposited. The Li₂S or Li₂S₂ on the working electrode is prone to decomposition by atmospheric moisture, which could increase the difficulty in result interpretation.

---