Sodium and Lithium Storage Properties of Spray-Dried Molybdenum Disulfide-Graphene Hierarchical Microspheres

Abstract

Developing nano/micro-structures which can effectively upgrade the intriguing properties of electrode materials for energy storage devices is always a key research topic. Ultrathin nanosheets were proved to be one of the potential nanostructures due to their high specific surface area, good active contact areas and porous channels. Herein, we report a unique hierarchical micro-spherical morphology of well-stacked and completely miscible molybdenum disulfide (MoS₂) nanosheets and graphene sheets, were successfully synthesized via a simple and industrial scale spray-drying technique to take the advantages of both MoS₂ and graphene in terms of their high practical capacity values and high electronic conductivity, respectively. Computational studies were performed to understand the interfacial behaviour of MoS₂ and graphene, which proves high stability of the composite with high interfacial binding energy (−2.02 eV) among them. Further, the lithium and sodium storage properties have been tested and reveal excellent cyclic stability over 250 and 500 cycles, respectively, with the highest initial capacity values of 1300 mAh g⁻¹ and 640 mAh g⁻¹ at 0.1 A g⁻¹.

Introduction

In recent years, room temperature sodium-ion batteries (SIBs) have been the object of significant interest for their potential application in large-scale energy storage systems. This is mainly caused by concerns about insufficient lithium ores to satisfy the increasing demands for lithium-ion batteries (LIBs), as well as because sodium is a relatively cheaper option compared to lithium, which could be significant in large-scale applications such as grid storage. In addition, the electrochemical principles of SIBs are identical to those of LIBs¹. However, some anode materials which are suitable for LIBs may not be well compatible for SIBs. For instance, graphite, which is a commercial anode material in LIBs, delivers unsatisfactory electrochemical behavior in SIBs unless with specialized ether-based electrolyte solvents such as diethylene glycol dimethyl ether (DEGDME), tetraethylene glycol dimethyl ether (TEGDME) etc^2,3. Thus it could be of research interest to explore the compatibility of various classes of anode materials for LIBs and SIBs.

Amongst various anode materials, molybdenum disulfide is one of the earliest compounds studied for rechargeable LIBs due to its layered structure, which can intercalate Li⁺ between the MoS₂ layers^4,5,6,7,8,9. Although the capacity of MoS₂ has been greatly improved (theoretical capacity =670 mAh g⁻¹), large volume changes occur during charge-discharge cycling, which results in poor cycling stability. Several methods have been reported to successfully improve the cycling stability, such as exfoliation and restacking of MoS₂ layers¹⁰, introducing polymers between the MoS₂ layers,^11,12 and addition of graphene sheets to form composites^13,14. On the other hand, there have been only a few reports on the Na-ion storage of MoS₂^{15,16,17,18,19,20,21,22}, which involves the intercalation of 1 Na⁺ per MoS₂. Considering the advantages of graphene as a highly conductive and stable material, Wang et al.¹⁸ reported MoS₂—graphene nanocomposite with an improved electrochemical performance when compared to pristine MoS₂, however, there is no clear evidence of well-ordered stacking of the MoS₂ and graphene nanosheets among themselves. Three-dimensional (3D) hierarchical micro-spherical architectures with nanostructures as building blocks are considered to be electrochemically and structurally stable morphology that could lead to improved practical application of such active materials in battery systems²¹. Such microspheres exhibit high effective contact areas between active material and the electrolytes, leads to enhanced electrochemical performance (cyclic stability and rate performance) along the sides of short ionic diffusion pathways and resist volume changes due to ‘nanosheets’ as sub-units²³.

In this work, we are reporting structurally and electrochemically optimized MoS₂-graphene composites with a unique micro-spherical morphology synthesized via the spray-drying technique, which is an industrial-scale synthesis procedure for a large-scale production of composite powders with a controllable narrow particle size distribution and nano/micro-spherical morphology²⁴. We found that the well-ordered and highly miscible hierarchical stacking of graphene-like MoS₂ and graphene nanosheets could be an efficient structure to enhance the interfacial effect between graphene-like MoS₂ and graphene, thus taking full advantages from both MoS₂ and graphene components and achieving high capacity, excellent cycle life and high rate capability as electrode materials for LIBs and SIBs.

Results and Discussion

After preparation of composite material with various proportions of MoS₂ and graphene oxide, two proportions of MoS₂: graphene oxide i.e., 80:20 and 60:40 were optimized based on structural and electrochemical aspects. The synthesis of the MoS₂-graphene composites with two different graphene ratios is briefly described in Fig. S1 in the Supporting Information. Solutions containing exfoliated MoS₂ and graphene oxide were mixed in 80:20 (MoS₂-G1) and 60:40 (MoS₂-G2) ratios to form a homogenous mixture. The graphene oxide sheets and MoS₂ sheets were well dispersed among each other and no aggregation were observed after leaving the solution undisturbed overnight. The solutions for the samples were then pumped through a nozzle using a peristaltic pump into a custom-made spray-drying reactor at 350 °C. The resultant black fluffy powders were collected using a cyclone collector attached to the spray drying reactor. The products were then annealed at 800 °C in H₂/Ar gas flow for 2 h to fully reduce the graphene oxide. The samples were then characterized using X-ray diffraction (XRD) to determine the phase of the material as shown in Fig. 1(a). All of the three samples, MoS₂, MoS₂-G1 and MoS₂-G2, show peaks which can be indexed to hexagonal MoS₂ (ICDD# 37-1492) and no impurity peaks can be observed. In order to determine the amount of graphene present in the samples after annealing, thermo-gravimetric analysis (TGA) was performed on the samples. The samples were loaded into alumina crucibles and heated in flowing air at the rate of 5 °C/min up to 700 °C. Based on the assumption that all the MoS₂ is converted into MoO₃ at 700 °C, the carbon content estimated for MoS₂-G1 and MoS₂-G2 is 13 wt% and 26 wt%, respectively, as shown in Fig. 1(b). The composites were also characterized using X-ray photoelectron spectroscopy (XPS) to determine the elemental compositions. Figure 1(c) shows the survey scans of the three samples and the inset tables indicate the atomic percentages of the elements present in the sample. It should be noted that the samples were subjected to surface etching using ion beams before the XPS characterization. No significant impurity elements were detected from the scans of any of the three samples. The carbon detected on the bare MoS₂ sample can be attributed to adsorbed CO₂ on the surface of the samples. The ratio of C to O is roughly 1:2. The atomic percentages (%) of C in MoS₂-G1 and MoS₂-G2 agree well with the results from TGA, where the later has a higher amount of graphene sheets present. For all three samples, the ratio of Mo to S is roughly 1:2, which suggest there is minimal or no oxidation of MoS₂. This is further confirmed by Raman spectroscopy, as shown in Fig. 1(d). All the peaks observed in the Raman spectra below 1000 cm⁻¹ can be attributed to hexagonal phase MoS₂, which is in well agreement with the literature elsewhere²⁵. On the other hand, the D and G bands of carbon at 1331 cm⁻¹ and 1597 cm⁻¹, respectively, can be observed on the spectra of MoS₂-G1 and MoS₂-G2.

Figure 1

(a) XRD patterns, (b) TGA curves, (c) XPS survey scans, with the insets showing the atomic percentages of the elements in the samples and (d) Raman spectra for the MoS₂, MoS₂-G1 and MoS₂-G2 samples.

The morphology of the MoS₂, MoS₂-G1 and MoS₂-G2 samples was investigated using field emission scanning electron microscopy (FESEM), and the images are shown in Fig. S2. All three samples have identical morphology. At low magnification, spheroidal particles with diameters ranging from 1–3 micrometers can be observed. Upon closer inspection at higher magnification, the spheres are found to be made of crumpled sheets and the kinks can be clearly observed. The same is represented in schematic representation as shown in Fig. 2. The samples were further investigated using transmission electron microscopy (TEM) and the images for MoS₂-G2 are shown in Fig. 3. TEM analysis shows similar sphere-like morphology. Selected area electron diffraction (SAED) analysis of the sample was performed on a small area inside the sphere and a large area covering the whole sphere. The SAED patterns shown in Fig. 3(b,c) correspond to the areas marked 1 and 2, respectively in Fig. 3(a). The SAED pattern of the area marked 1 (Fig. 3(b)) shows diffuse bright dots due to the single crystalline nature of the MoS₂ nanosheets. In addition, the spherical morphology of the sample enables the sheets to naturally stack on top of each other rather than in a random stacking. When the SAED pattern was collected from a larger area covering the sides of the sphere, rings were observed instead of the bright dots. Both of the SAED patterns can be indexed to the hexagonal MoS₂ phase consistent with the XRD results. High resolution TEM (HRTEM) was also used to study the distribution of the MoS₂ and graphene nanosheets. Figure 3(e) corresponds to the area marked by a red circle in Fig. 3(d). From the image, it can be observed that the graphene nanosheets and MoS₂ nanosheets are stacked on top of each other, forming a sandwich-like structure. Such a unique spherical microstructure with inter-stacked graphene and MoS₂ nanosheets is due to a combination of several factors. The miscibility and stability of the graphene oxide and the MoS₂ nanosheets in aqueous solution is very important to provide the sandwich-like stacking, while the spray-drying process is crucial in providing the spherical morphology of the end product. The thickness of the stacks of graphene and MoS₂ nanosheets was determined to be from 3–15 layers by studying 10 random areas using HRTEM. Figure 3(f) is an enlarged image of Fig. 3(e), showing the d-spacing of the MoS₂ and graphene nanosheets, which were measured to be 0.63 nm and 0.34 nm, respectively. TEM analysis of MoS₂ and MoS₂-G1 was also carried out, yielding similar results. The images and diffraction patterns are presented in Fig. S3 and S4. To further confirm the well-ordered distribution of graphene (carbon), EDX elemental mapping was performed for MoS₂, MoS₂-G1 and MoS₂-G2 samples and represented in Figs S5-S7, respectively and corresponding elemental compositions were tabulated in Tables S1-S3.

Figure 2

Schematic illustration of MoS₂-Graphene composite microsphere with hierarchical assembly of MoS₂ and graphene nanosheets.

Figure 3

(a) TEM image of MoS₂-G2 microspheres, (b, c) selected area diffraction (SAED) patterns of corresponding regions marked 1 and 2, respectively, with the patterns indexed to the hexagonal phase, (d) TEM image of MoS₂-G2 sample, (e) HRTEM image of marked region in (d) and (f) magnified image of region from (e), revealing the lattice d-spacing values of MoS₂ (0.63 nm) and graphene (0.34 nm).

Computational studies were performed to further understand the interfacial behaviour of MoS₂ and graphene. A (5 × 5) single graphene layer containing 50 carbon atoms was used to match a (4 × 4) MoS₂ monolayer containing 16 Mo and 32 S atoms. The lattice mismatch between the graphene and the MoS₂ monolayer is only 1.3%. Plane-wave basis VASP code was used to perform all the calculations^26,27, implementing the Perdew-Burke-Ernzerhof (PBE) exchange correlation functional²⁸. A damped van der Waals correction is also incorporated, based on Grimme’s scheme²⁹, to better describe the non-bonding interaction between the graphene and the MoS2 monolayer. In an all-electron description, the projector augmented wave method is used to describe the electron-ion interaction^30,31. The cut-off energy for plane waves was chosen to be 500 eV and the vacuum space is at least 18 Å, which is large enough to avoid the interaction between periodical images. A Monkhorst pack mesh of k-points (3 × 3 × 1) and (5 × 5 × 1) is used respectively to sample the two-dimensional Brillouin zone for geometry optimization and for calculating the charge density. The convergence of the tolerance force on each atom during structure relaxation was set to 0.005 eV/Å. Figure 4(a) presents a top view of the fully relaxed graphene-MoS₂ geometry. The equilibrium distance between the graphene layer and the top of the MoS₂ monolayer is calculated to be 3.34 Ǻ. The interface adhesion energy, E_ad, was obtained according to the following equation,

Figure 4

(a) Top view of the optimized graphene-MoS₂ interface, and (b) a side view of the three-dimensional charge density difference plot for the interface between a graphene sheet and a MoS₂ monolayer. Red, yellow and green balls represent Mo, S, and C atoms, respectively. Purple and orange isosurfaces represent charge accumulation and depletion in the 3D space with an isovalue of 0.001 e/Ǻ³.

Where E_comb, E_graphene, and E_MoS2 represent the total energy of the relaxed hybrid graphene-MoS₂ complex, the pure graphene sheet and the MoS2 monolayer, respectively. The interface binding energy is as high as −2.02 eV for the whole model interface, which indicates very high stability. To characterize the electron coupling at the graphene-MoS₂ interface, three-dimensional (3D) charge density difference plots were calculated by subtracting the electronic charge of the hybrid graphene-MoS₂ nanocomposite from those of the separate graphene layer and the MoS₂ monolayer, as shown in Fig. 4(b). Clearly, there is significant charge transfer from the graphene layer to the top of MoS₂ surface in the ground electronic state.

The samples were studied for their lithium storage properties and the results are plotted in Fig. 5 and Fig. S8. All three samples were first cycled at the low current density of 0.1 A g⁻¹ over 50 cycles. In the first discharge, MoS₂-G2 shows the highest capacity at 1300 mAh g⁻¹, while MoS₂-G1 and MoS₂ show 800 mAh g⁻¹ and 630 mAh g⁻¹, respectively. Large irreversible capacity is observed for all three samples, as the first charge capacities are 945, 660, and 480 mAh g⁻¹ for MoS₂-G2, MoS₂-G1 and MoS₂, respectively as shown in Fig. S10(a). This irreversible capacity can be ascribed to the formation of a solid electrolyte interphase (SEI) layer, which is widely known to occur below 1 V. All three samples showed stable cycling behavior for 50 cycles. The capacities are 800, 630 and 470 mAh g⁻¹ for MoS₂-G2, MoS₂-G1, and MoS₂, respectively, at the end of cycling. The samples were further tested for their rate performances up to the current density of 5 A g⁻¹. The MoS₂-G2 sample showed the best rate performance, retaining 590 mAh g⁻¹ at 5 A g⁻¹. The MoS₂-G1 and MoS₂ managed to retain 435 and 387 mAh g⁻¹, respectively, at 5 A g⁻¹. When the rate was recovered to 0.5 A g⁻¹, all the samples showed capacity recovery, where MoS₂-G2, MoS₂-G1 and MoS₂, recovered 820, 680, and 560 mAh g⁻¹, respectively. It should be noted that the rate performances improved with increasing graphene content in the samples. This can be justified by the increased conductivity provided by the graphene nanosheets in the samples. In order to investigate the long-term cycling stability, the samples were tested at 50 mA g⁻¹ for the initial 5 cycles, then at 1 A g⁻¹ up to 250 cycles. The MoS₂ samples recorded a stable capacity of 500 mAh g⁻¹ up to the 75^th cycle and then the capacity gradually faded over 50 cycles. The capacity after the 125^th cycle is negligible. This could be attributed to huge volume expansions of pristine MoS₂ during charge-discharge process and gradually results in pulverization of electrodes over cycle life. The spherical shape of the MoS₂ sample yielded improved cycling stability and rate performances when compared to the bulk MoS₂, which has been reported by us previously¹⁵. Both the MoS₂-G2 and MoS₂-G1 samples exhibit stable cycling over 250 cycles, retaining 780 and 700 mAh g⁻¹, respectively. From the cycling tests in lithium half-cells, it can be noted that graphene nanosheets in the samples are crucial for improving both the capacity and the cycling performance.

Figure 5

(a) Cycling performances in lithium half-cells of all the samples for 50 cycles at 100 mA g⁻¹; (b) rate capability of all the samples from 0.05 to 5 A g⁻¹; and (c) cycling performances of all samples for 250 cycles at 1 A g⁻¹. The voltage range is 0.01–3 V.

Furthermore, the samples were studied for their sodium storage properties in room-temperature sodium half-cells using similar testing conditions to those for lithium cells (see Fig. 6 and Fig. S9). In the first discharge process (Fig. S9), three voltage plateaus are observed at around ~0.95 V, ~0.65 V and ~0.25 V, which are corresponding to the formation of intermediate Na_xMoS₂, remaining Na_1-x ions reacting with MoS₂ and reduction of Mo⁴⁺ to metallic Mo along with formation of Na₂S particulates, respectively. These observations are in consistent with the reports elsewhere^16,17,18. However, successive discharge profiles show sloping curves instead of plateaus, which represents the phenomenon of conversion reaction. The same reaction mechanism is expected to happen with Li as standard electrode potential difference (~−0.3 V) from the plateau voltages of initial charge curves (compare Fig. S8 and S9)¹⁶. Figure. 6 represents cyclic profile, at 0.1 A g⁻¹, all three samples show stable cycling behavior over 50 cycles. The initial discharge and charge capacities are 640 and 400, 620 and 420 and 430 and 280 mAh g⁻¹, for MoS₂-G2, MoS₂-G1, and MoS₂, respectively. The irreversible capacities of about 35% (see Fig. S10(b)) can be due to the formation of SEI layers. After 50 cycles, both MoS₂-G2 and MoS₂-G1 recorded 340 mAh g⁻¹, while the MoS₂ sample recorded 240 mAh g⁻¹. The sodium storage capacities of all samples are significantly lower compared to their lithium storage, because only 1 Na⁺ is reacted per MoS₂, based on the capacity of MoS₂-G2. This could be due to the sluggish kinetics of the Na⁺ reaction with MoS₂. Moreover, in contrast to the lithium cells, the increased amount of graphene sheets in MoS₂-G2 did not yield any increment in capacity when compared to MoS₂-G1. Then, the samples were also tested for their rate performances, from 0.05 A g⁻¹ to 5 A g⁻¹. Both MoS₂-G2 and MoS₂-G1 showed similar performances, retaining about 230 mAh g⁻¹ at 5 A g⁻¹, while the MoS₂ sample retained 74 mAh g⁻¹. Although the capacity for both MoS₂-G2 and MoS₂-G1 at 5 A g⁻¹ is only 230 mAh g⁻¹, this result is significantly better compared to other Na-ion battery anode materials^18,19. The long-term cycling stability of the three samples was also tested, where the samples were cycled at 50 mA g⁻¹ for the initial 5 cycles and then at 1 A g⁻¹ for up to 500 cycles. The MoS₂ sample showed a slightly consistent capacity up to the 125^th cycle, recording 240 mAh g⁻¹ and then the capacity gradually decreased to 128 mAh g⁻¹ at the 200^th cycle. The capacity further decreased to 70 mAh g⁻¹ at the 500^th cycle. The MoS₂-G1 sample showed capacity of 375 mAh g⁻¹ at the 120^th cycle, then a gradual decrease in capacity was observed up to the 500^th cycle, where 251 mAh g⁻¹ was retained. The MoS₂-G2 sample showed slightly better performance compared to MoS₂-G1, recording the capacity of 420 mAh g⁻¹ at the 130^th cycle and then gradually decreased to 300 mAh g⁻¹ at the 500^th cycle. The cycling stability is impressive considering that the capacity retention after 500 cycles is 93% of the capacity at the 6^th cycle (320 mAh g⁻¹). These excellent electrochemical properties in both LIBs and SIBs could be attributed to various reasons such as: (i) the well-ordered stacking and excellent miscibility of the MoS₂ layers and graphene sheets, which leads to improved conductivity and thereby, improved cycling performance and rate capability; (ii) suppression of volume changes by the structurally stable nanosheets in microspheres during cycling; and (iii) good penetration of electrolyte into and among the MoS₂ and graphene nanosheets. In addition, these microspheres with hierarchical nanostructures benefit from both micro-materials (high tap density) and nanomaterials (short Li/Na diffusion pathways), leading to densely packed electrodes and improved battery life in practical applications^{32,33,34,35,36}. Such hierarchical microstructures can lead to a new class of electrode materials that could be potential candidates for LIBs or SIBs with enhanced cycle life.

Figure 6

(a) Cycling performances in sodium half-cells of all the samples for 50 cycles at 100 mA g⁻¹; (b) rate capability of all the samples from 0.05 to 5 A g⁻¹; and (c) cycling performances of all samples for 500 cycles at 1 A g⁻¹. The voltage range is 0.01–3 V.

In summary, a molybdenum disulfide (MoS₂)-graphene composite with unique hierarchical microsphere morphology was prepared by the spray-drying technique. The composite microspheres consist of well-ordered stacks with MoS₂ and graphene nanosheets with a high interfacial binding energy (−2.02 eV). Under testing for their lithium and sodium storage properties, MoS₂-graphene (26 wt.%) microspheres presented excellent cycling stability and rate capability, with initial discharge capacities of 1300 mAh g⁻¹ and 640 mAh g⁻¹ at 0.1 A g⁻¹ in LIBs and SIBs, respectively. Notably, in SIBs at 1 A g⁻¹, MoS₂-G2 showed 93% capacity retention after 500 cycles. These enhanced electrochemical features are attributable to the unique hierarchical composite microspheres with uniform distribution of graphene nanosheets among MoS₂ layers.