Rapid and mass-producible synthesis of high-crystallinity MoSe2 nanosheets by ampoule-loaded chemical vapor deposition

Na Liu a, Woong Choi b, Hyeongi Kim c, Chulseung Jung a, Jeonghun Kim a, Soo Ho Choo a, Yena Kwon a, Byeong-Seon An a, Seongin Hong a, Seongjoon So c, Cheol-Woong Yang *a, Jaehyun Hur *c and Sunkook Kim *a
aSchool of Advanced Materials Science and Engineering, Sungkyunkwan University, Suwon, 16419, Republic of Korea. E-mail: cwyang@skku.edu; seonkuk@skku.edu
bSchool of Advanced Materials Engineering, Kookmin University, Seoul, 02707, Republic of Korea
cDepartment of Chemical and Biological Engineering, Gachon University, Seongnam-si, Gyeonggi, 13120, Republic of Korea. E-mail: jhhur@gachon.ac.kr

Received 9th December 2019 , Accepted 4th February 2020

First published on 21st February 2020

MoSe2 is an attractive transition-metal dichalcogenide with a two-dimensional layered structure and various attractive properties. Although MoSe2 is a promising negative electrode material for electrochemical applications, further investigation of MoSe2 has been limited, mainly by the lack of MoSe2 mass-production methods. Here, we report a rapid and ultra-high-yield synthesis method of obtaining MoSe2 nanosheets with high crystallinity and large grains by ampoule-loaded chemical vapor deposition. Application of high pressure to an ampoule-type quartz tube containing MoO3 and Se powders initiated rapid reactions that produced vertically oriented MoSe2 nanosheets with grain sizes of up to ∼100 μm and yields of ∼15 mg h−1. Spectroscopy and microscopy characterizations confirmed the high crystallinity of the obtained MoSe2 nanosheets. Transistors and lithium-ion battery cells fabricated with the synthesized MoSe2 nanosheets showed good performance, thereby further indicating their high quality. The proposed simple scalable synthesis method can pave the way for diverse electrical and electrochemical applications of MoSe2.


Transition-metal dichalcogenides (TMDs) have attracted considerable attention as their two-dimensional (2D) layered structures can provide promising properties.1 Among the various types of TMDs, MoS2 is the most extensively studied material because of the naturally occurring single crystals. Monolayer and multilayer MoS2 can be promising for applications in electronics,2,3 optoelectronics,4,5 biosensors,6 and flexible devices.7,8 In addition, bulk or nanostructured MoS2 is also attracting attention as a negative electrode material for lithium-ion batteries (LIBs) because of its high theoretical capacity (670 mA h g−1) compared with that of commercial graphite (372 mA h g−1).9,10 Apart from MoS2, a variety of other TMDs, such as MoSe2, have recently received significant attention for use as promising electrodes in LIBs because of their outstanding properties during electrochemical reactions.11 For example, the lamellar structure of MoSe2 enables facile intercalation of metal ions such as Li+, allowing volume change mitigation and fast diffusion kinetics during electrochemical reactions because it can facilitate facile strain relaxation and reduce the barrier for Li+ intercalation.12,13 Along these lines, a number of studies that use MoSe2 as a great electrode candidate in LIBs have been reported.14–16 However, the investigation of MoSe2 has been limited by the lack of naturally occurring crystals and large-area synthesis methods.17 Therefore, for the practical application of MoSe2 in electrochemical reactions, the development of a mass-production method for growing large-area MoSe2 with high quality and crystallinity is essential.

To date, various methods for synthesizing MoSe2 single crystals, thin films, and nanostructures have been used. High-quality single crystals of MoSe2 and thin films of MoSe2 with large grains having sizes of hundreds of micrometers can be grown by chemical vapor transport; but it is a slow process lasting a few days to two weeks.18 Molecular beam epitaxy,19 atomic layer deposition (ALD),20 and chemical vapor deposition (CVD)21–23 can provide MoSe2 thin films with high uniformity and crystallinity, but they suffer from a low deposition rate and small grain size. Unlike thin films, which have been mainly optimized for electronic and optoelectronic device applications, nanostructures composed of vertically oriented nanosheets can provide excellent electrochemical properties because of their exposed reactive edge sites.24,25 MoSe2 nanostructures with a flower-like morphology can be synthesized with high quality and yield using a chemical solution approach.26 However, this approach is difficult to scale-up, and the size of the synthesized MoSe2 nanostructures is usually smaller than 1 μm.

In this study, we demonstrate a rapid ultra-high-yield method of synthesizing vertically oriented MoSe2 nanosheets with high crystallinity and large grains by ampoule-loaded CVD. The high pressure in an ampoule-type tube induces rapid reactions resulting in vertically stacked MoSe2 nanosheets with large grains of sizes up to ∼100 μm and a high yield of ∼15 mg h−1. Raman spectroscopy, X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD), and transmission electron microscopy (TEM) are used to verify the high crystallinity and investigate the growth mechanism of the MoSe2 nanosheets. To further confirm the high quality of our MoSe2 nanosheets, the electrical properties of MoSe2 transistors and the electrochemical properties of LIBs with a MoSe2/graphite composite negative electrode are measured.

Experimental section

Configuration of quartz tube assembly

The quartz tube assembly consisted of two half-open (one open and one closed ends) quartz tubes having lengths of 430 mm. Quartz tube I had an indented closed end with a 2 mm hole (Fig. S1a), while quartz tube II had a flat closed end with a 2 mm hole (Fig. S1b). Quartz tube I (25 mm in diameter) can be inserted into quartz tube II (28 mm in diameter) to form the quartz tube assembly (Fig. S1c and d).

Synthesis of MoSe2 nanosheets

To synthesize MoSe2 nanosheets, MoO3 (99%, Sigma Aldrich) and Se (99.9%, Sigma Aldrich) powders were used as the Mo and Se precursors, respectively. MoO3 (0.5 g) and Se (1.0 g) powders were placed into two separate quartz boats and then placed into quartz tube I (Fig. S1a). The boats containing MoO3 and Se were placed approximately 2 and 18 cm from the open end of quartz tube I, respectively. Silicon wafers (2 × 2 cm2) with a 300 nm-thick SiO2 layer were used as substrates. After cleaning with acetone and isopropanol, a silicon substrate was vertically placed at the closed end of quartz tube II with its oxidized side facing toward the open end (Fig. S1b). Quartz tube I was horizontally inserted into quartz tube II (Fig. S1c and d). Finally, the quartz tube assembly was placed at the center of a two-zone furnace CVD system (Fig. S1e and f).

When the furnace was pumped down to a pressure below 4 × 10−3 Torr, Ar (65 sccm) and H2 (15 sccm) were introduced to the system as a carrier and a reducing gas, respectively. At the beginning of the crystal growth, the valve of the rotary pump was closed, and thus the pressure in the CVD chamber continuously increased as gas was injected. When the pressure of the CVD chamber was higher than 800 Torr, the safety unit opened automatically to vent the chamber, maintaining the pressure at approximately 800 Torr. As the gases were easily injected into the quartz tube assembly but were difficult to remove, the pressure in the assembly was relatively high (≥800 Torr). The temperature of the furnace was ramped to 850 °C in 30 min. After the crystal growth for 1 h, the furnace was cooled to room temperature naturally.


MoSe2 crystals synthesized on SiO2/Si substrates were observed using optical microscopy (BX51M, Olympus Co.) and a field-emission scanning electron microscopy (SEM) (S-4800, HITACHI) coupled with energy-dispersive X-ray spectroscopy (EDS). The microstructure of MoSe2 was further characterized using TEM (JEM ARM 200F, JEOL). TEM samples were prepared using a focused ion beam system (NX2000, HITACHI). Raman spectra (inVia Raman microscopy, Renishaw) were acquired with a laser excitation at 523 nm. XPS (ThermoVG, UK) measurements were performed with Al Kα X-ray radiation (1486.6 eV). The working pressure in the ultrahigh vacuum chamber during the measurement was maintained below 3 × 10−9 mbar. Binding energies were calibrated by C 1s at 284.5 eV. XRD (D8 ADVANCE, Bruker Corporation) patterns were obtained using Cu Kα radiation. Atomic force microscopy (AFM) (XE7, PSIA) images were obtained in noncontact mode. The electrical properties of the devices were measured using a semiconductor characterization system (Keithley 4200).

Device fabrication

MoSe2 nanosheets, mechanically exfoliated from synthesized MoSe2 crystals using Scotch tape, were transferred onto highly doped silicon substrates with a 300 nm-thick SiO2 layer. After the SiO2/Si substrates were immersed in acetone for 1 h to remove any chemical residue formed during the transfer process, they were rinsed with isopropyl alcohol and blown dried with argon. To form source and drain electrodes, Ti (20 nm)/Au (100 nm) thin films deposited by electron-beam evaporation were patterned by conventional photolithography and lift-off method. Silicon substrates were used as global back-gate electrodes. The MoSe2 devices were passivated with 40 nm Al2O3 films using ALD to prevent the effect of O2 and H2O in the ambient atmosphere. The source and drain contact regions were patterned by conventional photolithography. The exposed Al2O3 was removed with a buffered oxide etchant. All transistors were annealed in an Ar environment at 200 °C for 2 h to minimize the contact resistance of the devices.

Electrochemical measurement

The MoSe2@graphite composite was prepared using a planetary ball-milling machine (Pulverisette 5, Fritsch). The mixture of the as-grown CVD MoSe2 and graphite (100 mesh, Sigma Aldrich) in a weight ratio of 7[thin space (1/6-em)]:[thin space (1/6-em)]3 was placed in a zirconium oxide bowl (80 cm3) with zirconium oxide balls with two different diameters of 3/8′′ and 3/16′′ in a weight ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]20 and sealed under an Ar atmosphere in a glove box. The mixture was milled at a bowl rotation speed of 300 rpm for 48 h with a 30 min rest after every 1 h of milling. Separately, an as-grown CVD MoSe2 electrode was prepared for comparison. The electrochemical performance was measured using coin-type half cells (CR 2032). To prepare the working electrode, a slurry consisting of 70 wt% of active materials (ball-milled MoSe2@graphite), 15 wt% of conductive carbon (Super P), and 15 wt% of polyvinylidene fluoride dissolved in N-methyl-2-pyrrolidinone was cast onto a Cu current foil by doctor blading. This film was dried in a vacuum oven at 70 °C overnight. A circular disc-shaped electrode (a diameter of 12 mm) was punched, and the typical mass of the film was ∼1 mg. The coin cell was assembled under an Ar atmosphere in a glove box. A polyethylene separator soaked with an electrolyte consisting of 1 M of LiPF6 in ethylene carbonate/diethylene carbonate (1[thin space (1/6-em)]:[thin space (1/6-em)]1 in volume) was placed on the working electrode, followed by the placement of a lithium foil as a counter electrode. Galvanostatic charge/discharge measurements were performed using a battery cycler (WBCS3000, WonAtech) in the voltage range of 0.0–3.0 V vs. Li/Li+ at various current densities. Cyclic voltammetry was carried out to analyse the electrochemical reaction mechanism using a ZIVE MP1 potentiostat analyzer (WonATech) in the same voltage range at a scanning rate of 0.1 mV s−1. EIS measurements were performed using the ZIVE MP1 potentiostat analyzer (WonATech) with an alternating-current amplitude of 10 mV in the frequency range of 100 mHz to 100 kHz after 50 cycles.

Results and discussion

The CVD system used to synthesize MoSe2 nanosheets is illustrated in Fig. 1a. An ampoule-shaped quartz tube assembly with MoO3 and Se precursors is placed at the center of the CVD system. The temperature profile during the growth process is shown in Fig. 1b. The furnace is heated to 850 °C in 30 min, maintained at 850 °C during the process (5–60 min), and then cooled to room temperature naturally. The color of the synthesized crystals shown in Fig. 1c–f changes with growth time: the crystals are brown (5 min), gold (15 min), silver (30 min), and then black (60 min). The black color of the 60 min sample is probably due to diffuse scattering from the as-synthesized crystals. When the 60 min sample is pressed with a roller, it regains the typical silver color of MoSe2 crystals (Fig. S2). Corresponding SEM images of the synthesized crystals are shown in Fig. 1g–j. In the 5 min sample, many crystallites of approximately 10 μm in size are observed. The atomic ratio of Mo and O measured by EDS (Fig. S3), indicates that these crystallites are molybdenum oxide (MoOx, x ≤ 3). In the 15 min sample, vertically grown nanosheets that are 15–20 μm in size are observed. EDS measurements (Fig. S4) suggest that MoOx and MoSe2 coexist in these samples. When the growth time increases to 30 min, smooth and uniform nanosheets that are 40–50 μm in size are observed. EDS measurements indicate that these nanosheets comprise MoSe2 with a small quantity of MoOx (Fig. S5). In the 60 min sample, the MoSe2 nanosheets obtained are as large as 100 μm in size. More SEM images of MoSe2 nanosheets in the 30 min and 60 min samples are shown in Fig. S6 and S7. The growth temperature is set as 850 °C based on the results of preliminary experiments. SEM images in Fig. S8a–d show MoSe2 grown at 750, 800, 850, and 900 °C, respectively. As growth temperature increases from 750 to 850 °C, the density and grain size of the MoSe2 nanosheets increase. But, MoSe2 grown at 850 and 900 °C does not show significant difference. Regardless of temperature, all MoSe2 show strong A1g Raman mode at 242 cm−1 with similar intensity and shape, suggesting the high crystallinity of MoSe2 (Fig. S8e).
image file: c9nr10418f-f1.tif
Fig. 1 (a) and (b) Schematic image of the CVD system and temperature profile for bulk MoSe2 crystal growth, respectively. (c–f) Photographs and (g–j) SEM images of the CVD-grown crystals on SiO2 substrates at growth times of 5, 15, 30, and 60 min, respectively.

To investigate the effect of growth time, Raman spectra of synthesized crystals were acquired using a 523 nm excitation laser (Fig. 2a and S9). The detected Raman bands of the 5 min sample are consistent with those of Mo4O11.27 The Raman bands of the 15 min sample are consistent with those of MoO2.28 MoSe2 exhibits two characteristic Raman active modes: a strong A1g mode representing the out-of-plane vibrations of the Se atoms and a weak E12g mode corresponding to the in-plane vibration of the Mo and Se atoms.21 The A1g mode at 242 cm−1 appears in the 10 min sample, and its intensity increases with increasing growth time (Fig. S9a). After 30 min, the A1g mode becomes dominant. Moreover, the weak E12g mode at 284 cm−1 and B12g mode (belonging to bulk MoSe2) at 352 cm−1 are observed in the sample with growth times exceeding 30 min. The Raman characterization data suggest that MoO3 is reduced to Mo4O11 at the beginning of the growth process (initial 5 min) and is further reduced to MoO2 with longer growth times. This result also suggests that MoSe2 is produced by the reaction of MoO2 with Se and that MoSe2 becomes the dominant species after the growth time of 30 min.

image file: c9nr10418f-f2.tif
Fig. 2 (a) Raman spectra; (b) XRD patterns and (c) XPS spectra of the CVD-grown crystals at growth times of 5, 15, 30, and 60 min.

The phase composition of CVD-grown crystals is also investigated by XRD. Fig. 2b and S10 shows XRD patterns of crystals with growth times of 5–60 min. For the 5 min sample, diffraction peaks of Mo4O11 (Joint Committee on Powder Diffraction Standards (JCPDS) card no. 72-0448) and MoO2 (JCPDS card no. 65-1273) are detected along with very weak diffraction peaks of MoSe2 (JCPDS card no. 65-3481). As the growth time increases to 10 min, the intensity of the Mo4O11 and MoO2 diffraction peaks decreases, while that of the MoSe2 diffraction peaks increases (Fig. S10). For the samples with growth times exceeding 30 min, the diffraction peaks of Mo4O11 and MoO2 almost vanish, but those of MoSe2 become sharper and higher, suggesting that high-crystallinity MoSe2 is the dominant phase. The positions of the XRD peaks and corresponding patterns are presented in Table S1.

We also investigated the chemical compositions of the CVD-grown crystals with XPS measurements. Fig. 2c and S11 show that the Mo 3d spectra comprise four sets of peaks corresponding to Mo6+ 3d at 232.7 and 235.8 eV (MoO3, orange lines), Mo5+ 3d at 231.0 and 234.1 eV (Mo4O11, blue lines), Mo4+ 3d at 229.4 and 232.5 eV (MoO2, green lines), and Mo4+ 3d at 229.0 and 232.1 eV (MoSe2, red lines).29 In addition, two peaks correspond to Se2− 3d at 54.6 and 55.4 eV (MoSe2, pink lines).30 The atomic fraction of each crystal phase, which can be calculated using the area of the related peaks, is presented in Table S2. In the initial 20 min, with increasing growth time, the atomic fraction of MoO3 decreases, that of Mo4O11 changes slightly, and that of MoO2 increases, suggesting that MoO3 is reduced to MoO2 through Mo4O11. The atomic fraction of Mo4O11 remains relatively the same, probably because Mo4O11 is an intermediate phase during the reduction. The intensity of the XPS peaks corresponding to Se2− 3d remains very low until the growth time reaches 20 min, suggesting a very low fraction of MoSe2 (Fig. S11). Indeed, the atomic fraction of MoSe2 is as low as 8.43% in the 20 min sample. When the growth time is larger than 30 min, the atomic fraction of MoSe2 is dominant, that of MoO3 is vanished, and those of Mo4O11 and MoO2 are considerably decreased. The atomic fraction of MoSe2 is considerably increased to 70.47% in the 30 min sample and 76.75% in the 60 min sample. Additionally, for the 30 and 60 min samples, the Se2− 3d peaks are clearly observed, reflecting the significant amount of high-crystallinity MoSe2. XPS analysis suggests that MoO3 is reduced to Mo4O11 and MoO2 during the initial 20 min of growth. After 30 min, most of MoOx is transformed into MoSe2, and the transformation continues with the growth time. These results are consistent with the Raman and XRD measurements.

The crystal structure of the 60 min sample was further characterized using TEM. The low-magnification TEM image in Fig. 3a shows a multilayer MoSe2 nanosheet. The high-resolution TEM image in Fig. 3b shows the periodic atomic arrangement, demonstrating the high crystallinity of the MoSe2 nanosheet.31 The selected area electron diffraction pattern (SADP) in Fig. 3c reveals a hexagonal structure consistent with that of single-crystal 2H-MoSe2.32Fig. 3d shows a cross-sectional scanning TEM (STEM) image of a MoSe2 nanosheet transferred onto a SiO2/Si substrate. The layered structure expected of a MoSe2 nanosheet is clearly observed in the high-angle annular dark-field (HAADF) STEM image shown in Fig. 3e. Fig. 3f shows a HAADF-STEM image with a temperature-scale contrast, while a corresponding line profile is shown in Fig. 3g. As the atomic number of Mo is larger than that of Se, Mo atoms are brighter than Se atoms in the HAADF-STEM image. Fig. 3f clearly shows that the MoSe2 monolayer comprises one layer of Mo atoms sandwiched by two neighboring layers of Se atoms. The distance between Mo atomic layers of MoSe2, measured using the line profile in Fig. 3g, is approximately 0.667 nm. These results are consistent with those from previous reports, thereby confirming the high crystallinity of our CVD-grown MoSe2 nanosheets.33

image file: c9nr10418f-f3.tif
Fig. 3 (a) Low- and (b) high-magnification TEM images of a 60 min MoSe2 nanosheet. (c) SADP of (b). (d) Cross-sectional STEM image of a MoSe2 nanosheet on a SiO2 substrate. (e and f) HAADF-STEM images of MoSe2 nanosheets with gray and temperature contrast scales, respectively. (g) Line profile of (f).

Based on the above analysis of the synthesized MoSe2, we propose a possible growth mechanism of MoSe2 nanosheets in two stages. In stage 1, MoO3 is reduced to form Mo4O11 and MoO2 by hydrogen. MoSe2 was not obtained in this study unless hydrogen gas was introduced into the CVD system (Fig. S12). During the initial 5 min of growth, a small amount of the MoO3 is evaporated and then deposited onto the surface of the SiO2/Si substrate. The deposited MoO3 is partly reduced to MoO2 by Mo4O11. With increasing growth time, MoO3 is continuously deposited on the substrate and transformed into MoO2. Until the growth time reaches 20 min, the reduction reaction of MoO3 is dominant. In stage 2, MoO2 reacts with Se to form MoSe2. After 30 min of growth, MoO3 is mostly transformed into MoO2, which reacts with Se to form MoSe2. At longer growth times, the crystallinity and amount of MoSe2 increase. The two-stage reaction mechanism for the synthesis of MoSe2 may be expressed as:

Stage 1:

4MoO3 + H2 → Mo4O11 + H2O(1)
Mo4O11 + 3H2 → 4MoO2 + 3H2O(2)

Stage 2:

MoO2 + 3Se → MoSe2 + SeO2(3)

The special structure of the ampoule-shaped quartz tube assembly used in this study plays a key role in the growth of bulky MoSe2 nanosheets. Gases are easily injected into the quartz tube assembly, but cannot escape, which results in a relatively high pressure (≥800 Torr) in the assembly. The growth rate of the MoSe2 nanosheets can be expressed using the Hertz–Knudsen equation:

image file: c9nr10418f-t1.tif(4)
where α is the evaporation coefficient of the vapor molecules on the surface, P is the vapor pressure, m is the mass of the particles, k is the Boltzmann constant, and T is the temperature.34 At a definite temperature, the growth rate of the MoSe2 nanosheets increases with the pressure. Moreover, based on the nucleation theory, the rate of nucleation can be expressed:
image file: c9nr10418f-t2.tif(5)
where ns is the number of nucleation sites and ΔG* is the free energy of formation of a nucleus at the top of the nucleation barrier.35 As the pressure increases, the ns increases, leading to an enhancement in nucleation. Hence, the ampoule-shaped quartz tube assembly used in this study provides high pressure in a relatively enclosed structure, which enables the mass production of large-area MoSe2 nanosheets. The synthesized MoSe2 nanosheets are scraped off from the SiO2 substrates and then collected and weighed (Fig. S13). The yield of MoSe2 nanosheets is as high as approximately 15 mg h−1.

To investigate the electrical properties of the synthesized MoSe2 nanosheets, we fabricate transistors by transferring MoSe2 nanosheets onto a SiO2/Si substrate using a mechanical exfoliation method. Optical and AFM images of a fabricated MoSe2 transistor are shown in Fig. 4a and b, respectively. The thickness of the MoSe2 nanosheet used as a channel material is approximately 24.5 nm and the channel length and width is about 4.5 and 8.3 μm, respectively. Fig. 4c shows the transfer characteristics of a MoSe2 transistor with Al2O3 passivation at Vds = 1 V. The MoSe2 transistor exhibits n-type conduction with an on/off current ratio of approximately 3 × 103 and a field-effect mobility of approximately 34 cm2 V−1 s−1. Fig. 3d shows the output characteristics of the MoSe2 transistor at different Vgs values. The IdsVds curves show linear characteristics and robust current saturation at low and high drain voltages, respectively. The excellent electrical properties of the MoSe2 transistors are comparable with the previous reports and further confirm the high crystallinity of the synthesized MoSe2 nanosheets.23,31,36–38

image file: c9nr10418f-f4.tif
Fig. 4 (a) and (b) Optical microscopy and AFM images of the MoSe2 device fabricated on the SiO2/Si substrate, respectively. In (b), the height profile along the white dashed line indicates that the thickness of the MoSe2 nanosheet is approximately 24.5 nm. (c) Transfer characteristics at Vds = 1 V on a log scale. The on/off current ratio is over 103, while the field-effect mobility is approximately 34 cm2 V−1 s−1. (d) Output characteristics at different Vgs values.

Finally, we investigate the high-quality MoSe2 as a new anode material for LIBs. Despite the structural benefits of TMDs for use in LIBs, recent studies on most conversion-based 2D materials have revealed a limited cyclability and rate capability upon the utilization of a standalone 2D material as an active component because of its low electric conductivity, its large volume change during cycling, and the formation of thick and unstable solid-electrolyte interface (SEI) films.39–41 To address these issues, one of the most promising strategies is to incorporate an appropriate carbon-based material into a 2D material.13,14 In the design of the ideal nanocomposite between MoSe2 and carbonaceous material, the miscibility in the preparation step and stability during cycling are two essential factors to be considered for the realization of high-performance LIBs.14 Recently, several studies have reported the physical similarity between 2D materials (e.g., MoS2 and MoSe2) and layered graphite, which tends to promote their mixing behavior and cycling stability.14,42 The mixing is typically facilitated using a ball milling process, where solid lubrication of the 2D materials and layered graphite simultaneously occurs through the friction and lamination against the balls.42,43 Inspired by this concept, we synthesized a 2D hybrid nanocomposite consisting of the CVD-grown MoSe2 and graphite (c-MoSe2@G) using a ball milling process and evaluated its electrochemical performance (half-cell). The results were compared with those of a bare as-grown CVD MoSe2 (c-MoSe2).

Fig. 5a and b compare the galvanostatic voltage profiles of c-MoSe2 and c-MoSe2@G under an applied voltage of 0.0–3.0 V (vs. Li/Li+). The first discharge (lithiation) and charge (delithiation) capacities of c-MoSe2 were measured to be 289.3 and 230.5 mA h g−1, which were increased to 1134 and 814 mA h g−1 for c-MoSe2@G, respectively. Obviously, an enhancement in the specific capacity is associated with a significant increase in the number of reaction sites because of the size reduction in and further exfoliation of c-MoSe2@G during the ball milling. The capacity of c-MoSe2@G at the second and third cycles were higher than those of c-MoSe2, suggesting a synergistic effect of mixing of the two similar layered structures. The high affinity between c-MoSe2 and graphite is thought to provide additional reaction sites in MoSe2 during the milling process. Fig. 5c and d show the differential capacity profiles (DCPs) of c-MoSe2 and c-MoSe2@G for the initial three cycles, respectively. In the first discharge process, c-MoSe2 exhibits three main peaks at 0.9, 0.7, and 0.4 V. The peak at 0.9 V can be assigned to Li insertion into the layered MoSe2 lattice accompanied by the formation of LixMoSe2 (i.e., MoSe2 + xLi+ + xe → LixMoSe2, which is the phase transformation from the 2H to the 1T structure). The peaks at 0.7 and 0.4 V are associated with the decomposition of LixMoSe2 into Mo metal particles embedded in a Li2Se matrix (conversion reaction) and formation of an SEI film (i.e., LixMoSe2 + Li+ + e → Mo + Li2Se), respectively. Accordingly, the overall reaction during the discharge, can be expressed as MoSe2 + 4Li+ + 4e → Mo + 2Li2Se. In the first charge process, the peak at 1.4 V corresponds to the partial oxidation of Mo to MoSe2, whereas the distinct peak at 2.1 V indicates the oxidation of Li2Se to Se (i.e., Li2Se → 2Li+ + Se + 2e). All of these reaction mechanisms are in good agreement with previous reports, demonstrating that well-defined MoSe2 was synthesized by CVD.44–46 In the subsequent cathodic sweeps (second and third discharge cycles), all of the discharge peaks are vanished, and two new peaks emerge at 1.8 and 1.3 V, which are related to Li intercalation and the conversion of Se to Li2Se, respectively. For c-MoSe2@G, although the trends are generally identical, most of the peaks are broadened owing to the small-sized MoSe2 homogeneously distributed in the graphite matrix. However, the curves are well maintained without a noticeable shift after the second cycle, suggesting excellent reversibility of the electrochemical reactions.

image file: c9nr10418f-f5.tif
Fig. 5 (a) and (b) Initial charge/discharge profiles of c-MoSe2 and c-MoSe2@G, respectively. (c) and (d) DCPs of c-MoSe2 and c-MoSe2@G, respectively.

Fig. 6a and b compare the long-term cyclic performances at a current density of 0.1 A g−1 and rate capabilities of c-MoSe2 and c-MoSe2@G, respectively. The c-MoSe2@G nanocomposite exhibits excellent cyclic and rate performances, consistent with the DCP analysis. c-MoSe2 provides a specific capacity in the range of 300–400 mA h g−1, while the c-MoSe2@G nanocomposite provides a specific capacity of approximately 800 mA h g−1 after 50 cycles. In order to investigate the effect of crystallinity in MoSe2, we separately prepared a-MoSe2@G (commercially available MoSe2 milled with graphite from Sigma Aldrich) in which c-MoSe2 showed much higher crystallinity than that of a-MoSe2 (Fig. S14). As shown in Fig. 6a, the cyclic performance of a-MoSe2@G showed better performance than that of c-MoSe2@G for initial 5 cycles, but after 5 cycles, the capacities of a-MoSe2@G became lower than those of c-MoSe2@G. This is because defect sites that initially increased the capacity by providing increased reaction sites in a-MoSe2@G, do not contribute to the fully reversible reaction during the long-term cycling. Rather, as the cycle number increases, the capacity gradually decreased because of the irreversible reactions from many defect sites in a-MoSe2. This phenomenon is consistent with the observation of initial coulombic efficiency (ICE); ICEs of a-MoSe2@G were much lower than those of c-MoSe2@G (Table S3) which reflects the more side reactions and irreversible reactions in a-MoSe2@G. Additionally, Fig. 6b shows that the average capacity of the c-MoSe2@G nanocomposite is considerably higher than that of c-MoSe2 over all current densities of 0.1 to 3 A g−1. Remarkably, the reversible capacity of the c-MoSe2@G nanocomposite is maintained as high as ∼690 mA h g−1 (retention of 87% with respect to the specific capacity at 0.1 A g−1) even at a current density of 3 A g−1. Fig. 6c shows the electrochemical impedance spectroscopy (EIS) results for c-MoSe2 and c-MoSe2@G after 50 cycles. The kinetic differences between the two electrode are investigated by the modified Randle equivalent circuit (inset in Fig. 6c). The SEI film resistance (Rs) and charge transfer resistance (Rct) of c-MoSe2@G were 3.7 and 121 Ω, respectively, which are significantly lower than those of c-MoSe2 (8.6 and 460 Ω, respectively). The significant improvements in the cycling and rate performances are based on the synergetic effect originating from the similarity in physical structure (layered materials) between MoSe2 and graphite. Therefore, the two materials readily form a homogeneous mixture at the nanoscale via ball milling, which not only effectively suppresses volume changes and prevents electrode pulverization during the charge/discharge processes but also facilitates the conversion reaction due to the increased electrical conductivity.

image file: c9nr10418f-f6.tif
Fig. 6 (a) Cyclic performances. (b) Rate capability at different current densities (0.1–3 A g−1). (c) EIS spectra and equivalent circuit.


We investigated the large-area synthesis of high-crystallinity MoSe2 nanosheets with high yield using ampoule-loaded CVD. The high pressure in the ampoule-type quartz tube induced rapid reactions during growth, providing vertically oriented MoSe2 nanosheets with grains sizes up to ∼100 μm and yield of ∼15 mg h−1. Based on the Raman spectroscopy, XRD, XPS, and TEM analyses, we confirmed the high crystallinity of the synthesized MoSe2 nanosheets and suggested a possible growth mechanism. The transistors based on the synthesized MoSe2 nanosheets showed device performance comparable to that of transistors based on mechanically exfoliated single-crystal flakes. The LIB cells based on the composites of the synthesized MoSe2 and graphite also exhibited great electrochemical performance. These results demonstrate that our synthesis method can mass-produce large-area MoSe2 nanosheets with high crystallinity and large grains, highlighting an important step toward the applications of MoSe2 and other TMDs in electrochemical cells.

Conflicts of interest

There are no conflicts to declare.


S. Kim, J. Hur, and C.-W. Yang co-supervised this project. S. Kim and C. Jung conceived the ideas. N. Liu designed the experiments and performed the characterization. J. Kim, S. H. Choo and S. Hong carried out the growth experiments. Y. Kwon and B.-S. An performed TEM and STEM imaging experiments. N. Liu and W. Choi performed the growth mechanism analysis. J. Hur, H. Kim, and S. So performed the electrochemical experiments. J. Hur, N. Liu and W. Choi wrote and revised the manuscript. All authors read the manuscript and provided comments. This research was supported by the Korea Research Fellowship program funded by the MSIT and Future Planning through the NRF (no. 2017H1D3A1A02014116 and 2019R1F1A1057293) and by the Basic Science Research Program through the NRF funded by the Ministry of Education (no. 2017R1D1A1B03035315, 2018R1A2B2003558 and 2016R1D1A1B03931903).


  1. R. Lv, J. A. Robinson, R. E. Schaak, D. Sun, Y. F. Sun, T. E. Mallouk and M. Terrones, Acc. Chem. Res., 2015, 48, 56 CrossRef CAS PubMed.
  2. B. Radisavljevic and A. Kis, Nat. Mater., 2013, 12, 815 CrossRef CAS PubMed.
  3. S. Kim, A. Konar, W. S. Hwang, J. H. Lee, J. Lee, J. Yang, C. Jung, H. Kim, J. B. Yoo, J. Y. Choi, Y. W. Jin, S. Y. Lee, D. Jena, W. Choi and K. Kim, Nat. Commun., 2012, 3, 1011 CrossRef PubMed.
  4. O. Lopez-Sanchez, D. Lembke, M. Kayci, A. Radenovic and A. Kis, Nat. Nanotechnol., 2013, 8, 497 CrossRef CAS PubMed.
  5. W. Choi, M. Y. Cho, A. Konar, J. H. Lee, G. B. Cha, S. C. Hong, S. Kim, J. Kim, D. Jena, J. Joo and S. Kim, Adv. Mater., 2012, 24, 5832 CrossRef CAS PubMed.
  6. K. Kalantar-zadeh, J. Z. Ou, T. Daeneke, M. S. Strano, M. Pumera and S. L. Gras, Adv. Funct. Mater., 2015, 25, 5086 CrossRef CAS.
  7. W. G. Song, H. J. Kwon, J. Park, J. Yeo, M. Kim, S. Park, S. Yun, K. U. Kyung, C. P. Grigoropoulos, S. Kim and Y. K. Hong, Adv. Funct. Mater., 2016, 26, 2426 CrossRef CAS.
  8. J. Pu, Y. Yomogida, K. K. Liu, L. J. Li, Y. Iwasa and T. Takenobu, Nano Lett., 2012, 12, 4013 CrossRef CAS PubMed.
  9. Q. M. Su, S. X. Wang, M. Feng, G. H. Du and B. S. Xu, Sci. Rep., 2017, 7, 7275 CrossRef PubMed.
  10. T. Stephenson, Z. Li, B. Olsen and D. Mitlin, Energy Environ. Sci., 2014, 7, 209 RSC.
  11. A. Eftelthari, Appl. Mater. Today, 2017, 8, 1 CrossRef.
  12. J. Y. Li, Y. Hou, X. F. Gao, D. S. Guan, Y. Y. Xie, J. H. Chen and C. Yuan, Nano Energy, 2015, 16, 10 CrossRef CAS.
  13. S. H. Choi and Y. C. Kang, Nano Res., 2015, 8, 2492 CrossRef CAS.
  14. H. Kim, Q. H. Nguyen, I. T. Kim and J. Hur, Appl. Surf. Sci., 2019, 481, 1196 CrossRef CAS.
  15. L. D. Wu, P. F. Tan, Y. Liu, Y. Y. Shang, W. W. Liu, X. Xiong and J. Pan, J. Colloid Interface Sci., 2017, 491, 279 CrossRef CAS PubMed.
  16. Y. Liu, M. Q. Zhu and D. Chen, J. Mater. Chem. A, 2015, 3, 11857 RSC.
  17. R. Lieth, Preparation and crystal growth of materials with layered structures, Springer, Dordrecht, Netherlands, 1977 Search PubMed.
  18. J. S. Rhyee, J. Kwon, P. Dak, J. H. Kim, S. M. Kim, J. Park, Y. K. Hong, W. G. Song, I. Omkaram, M. A. Alam and S. Kim, Adv. Mater., 2016, 28, 2316 CrossRef CAS PubMed.
  19. A. Roy, H. C. P. Movva, B. Satpati, K. Kim, R. Dey, A. Rai, T. Pramanik, S. Guchhait, E. Tutuc and S. K. Banerjee, ACS Appl. Mater. Interfaces, 2016, 8, 7396 CrossRef CAS PubMed.
  20. M. Krbal, J. Prikryl, R. Zazpe, F. Dvorak, F. Bures and J. M. Macak, Phys. Status Solidi RRL, 2018, 12, 1800023 CrossRef.
  21. X. Lu, M. I. B. Utama, J. H. Lin, X. Gong, J. Zhang, Y. Y. Zhao, S. T. Pantelides, J. X. Wang, Z. L. Dong, Z. Liu, W. Zhou and Q. H. Xiong, Nano Lett., 2014, 14, 2419 CrossRef CAS PubMed.
  22. J. Xia, X. Huang, L. Z. Liu, M. Wang, L. Wang, B. Huang, D. D. Zhu, J. J. Li, C. Z. Gu and X. M. Meng, Nanoscale, 2014, 6, 8949 RSC.
  23. C. Jung, S. M. Kim, H. Moon, G. Han, J. Kwon, Y. K. Hong, I. Omkaram, Y. Yoon, S. Kim and J. Park, Sci. Rep., 2015, 5, 15313 CrossRef CAS PubMed.
  24. G. R. Bhimanapati, T. Hankins, Y. Lei, R. A. Vila, I. Fuller, M. Terrones and J. A. Robinson, ACS Appl. Mater. Interfaces, 2016, 8, 22190 CrossRef CAS PubMed.
  25. L. Ling, C. Wang, K. Zhang, T. Li, L. Tang, C. Li, L. Wang, Y. Xu, Q. Song and Y. Yao, RSC Adv., 2016, 6, 18483 RSC.
  26. D. Sun, S. M. Feng, M. Terrones and R. E. Schaak, Chem. Mater., 2015, 27, 3167 CrossRef CAS.
  27. M. Dieterle and G. Mestl, Phys. Chem. Chem. Phys., 2002, 4, 822 RSC.
  28. L. Kumari, Y. R. Ma, C. C. Tsai, Y. W. Lin, S. Y. Wu, K. W. Cheng and Y. Liou, Nanotechnology, 2007, 18, 115717 CrossRef.
  29. N. M. D. Brown, N. Y. Cui and A. McKinley, Appl. Surf. Sci., 1998, 134, 11 CrossRef CAS.
  30. W. A. Abdallah and A. E. Nelson, J. Mater. Sci., 2005, 40, 2679 CrossRef CAS.
  31. Y. H. Chang, W. Zhang, Y. Zhu, Y. Han, J. Pu, J. K. Chang, W. T. Hsu, J. K. Huang, C. L. Hsu, M. H. Chiu, T. Takenobu, H. Li, C. I. Wu, W. H. Chang, A. T. S. Wee and L. J. Li, ACS Nano, 2014, 8, 8582 CrossRef CAS PubMed.
  32. D. K. Hu, G. C. Xu, L. Xing, X. X. Yan, J. Y. Wang, J. Y. Zheng, Z. X. Lu, P. Wang, X. Q. Pan and L. Y. Jiao, Angew. Chem., Int. Ed., 2017, 56, 3611 CrossRef CAS PubMed.
  33. J. Xu, J. J. Zhang, W. J. Zhang and C. S. Lee, Adv. Energy Mater., 2017, 7, 1700571 CrossRef.
  34. N. M. Hwang, Non-classical crystallization of thin films and nanostructures in CVD and PVD processes, Springer, Seoul, South Korea, 2016 Search PubMed.
  35. D. Frenkel and B. Smit, Understanding molecular simulation: from algorithms to applications, Elsevier, San Diego, USA, 2001 Search PubMed.
  36. J. Baek, D. Yin, N. Liu, I. Omkaram, C. Jung, H. Im, S. Hong, S. M. Kim, Y. K. Hong, J. Hur, Y. Yoon and S. Kim, Nano Res., 2017, 10, 1861 CrossRef CAS.
  37. J.-S. Rhyee, J. Kwon, P. Dak, J. H. Kim, S. M. Kim, J. Park, Y. K. Hong, W. G. Song, I. Omkaram, M. A. Alam and S. Kim, Adv. Mater., 2016, 10, 1002 Search PubMed.
  38. H. Lee, J. Ahn, S. Im, J. Kim and W. Choi, Sci. Rep., 2018, 8, 11545 CrossRef PubMed.
  39. H. B. Shu, F. Li, C. L. Hu, P. Liang, D. Cao and X. S. Chen, Nanoscale, 2016, 8, 2918 RSC.
  40. Y. Fang, Y. Y. Lv, F. Gong, A. A. Elzatahry, G. F. Zheng and D. Y. Zhao, Adv. Mater., 2016, 28, 9385 CrossRef CAS PubMed.
  41. J. Xu, H. Tang, Y. Q. Chu and C. S. Li, RSC Adv., 2015, 5, 48492 RSC.
  42. N. Q. Hai, S. H. Kwon, H. Kim, I. T. Kim, S. G. Lee and J. Hur, Electrochim. Acta, 2018, 260, 129 CrossRef CAS.
  43. Q. H. Nguyen, J. S. Choi, Y. C. Lee, I. T. Kim and J. Hur, J. Ind. Eng. Chem., 2019, 69, 116 CrossRef CAS.
  44. J. Y. Yao, B. R. Liu, S. Ozden, J. J. Wu, S. B. Yang, M. T. F. Rodrigues, K. Kalaga, P. Dong, P. Xiao, Y. H. Zhang, R. Vajtai and P. M. Ajayan, Electrochim. Acta, 2015, 176, 103 CrossRef CAS.
  45. X. Zhao, J. H. Sui, F. Li, H. T. Fang, H. E. Wang, J. Y. Li, W. Cai and G. Z. Cao, Nanoscale, 2016, 8, 17902 RSC.
  46. X. Yang, Z. A. Zhang and X. D. Shi, J. Alloys Compd., 2016, 686, 413 CrossRef CAS.


Electronic supplementary information (ESI) available. See DOI: 10.1039/c9nr10418f
These authors contributed equally to this work.

This journal is © The Royal Society of Chemistry 2020