High-Yield Preparation and Electrochemical Properties of Few-Layer MoS2 Nanosheets by Exfoliating Natural Molybdenite Powders Directly via a Coupled Ultrasonication-Milling Process

Cost-effective and scalable preparation of two-dimensional (2D) molybdenum disulfide (MoS2) has been the bottleneck that limits their applications. This paper reports a novel coupled ultrasonication-milling (CUM) process to exfoliate natural molybdenite powders to achieve few-layer MoS2 (FL-MoS2) nanosheets in the solvent of N-methyl-2-pyrrolidone (NMP) with polyvinylpyrrolidone (PVP) molecules. The synergistic effect of ultrasonication and sand milling highly enhanced the exfoliation efficiency, and the precursor of natural molybdenite powders minimizes the synthetic cost of FL-MoS2 nanosheets. The exfoliation of natural molybdenite powders was conducted in a home-made CUM system, mainly consisting of an ultrasonic cell disruptor and a ceramic sand mill. The samples were characterized by X-ray diffraction, UV-vis spectra, Raman spectra, FT-IR, SEM, TEM, AFM, and N2 adsorption-desorption. The factors that influence the exfoliation in the CUM process, including the initial concentration of natural molybdenite powders (Cin, 15–55 g L−1), ultrasonic power (Pu, 200–350 W), rotation speed of sand mill (ωs, 1500–2250 r.p.m), exfoliation time (tex, 0.5–6 h), and the molar ratio of PVP unit to MoS2 (Rpm, 0–1), were systematically investigated. Under the optimal CUM conditions (i.e., Cin = 45 g L−1, Pu = 280 W, ωs = 2250 r.p.m and Rpm = 0.5), the yield at tex = 6 h reaches 21.6 %, and the corresponding exfoliation rate is as high as 1.42 g L−1 h−1. The exfoliation efficiency of the CUM mode is much higher than that of either the ultrasonication (U) mode or the milling (M) mode. The synergistic mechanism and influencing rules of the CUM process in exfoliating natural molybdenite powders were elaborated. The as-obtained FL-MoS2 nanosheets have a high specific surface area of 924 m2 g−1 and show highly enhanced electrocatalytic performance in hydrogen evolution reaction and good electrochemical sensing property in detecting ascorbic acid. The CUM process developed has paved a low-cost, green, and highly efficient way towards FL-MoS2 nanosheets from natural molybdenite powders.


Background
Since the discovery of graphene, two-dimensional (2D) nanomaterials have attracted highly increasing attention owing to their unique few-layer microstructures and resulting unprecedented functional properties [1][2][3][4]. Among them, the most typical 2D sample should be transition-metal dichalcogenides (TMDs), composed of covalently bonded transition-metal atoms (M) and chalcogen atoms (X = S, Se, Te) in a possible stoichiometric form of MX 2 [5], such as MoS 2 and WS 2 , of which each single-layer consists of two S atom layers and a layer of metal atoms sandwiched between the two S layers [6]. Owing to the suitable intrinsic indirect band gap (1.3 eV), MoS 2 has attracted great attention and been thought as alternative promising next-generation 2D materials. When the layered MoS 2 bulks are exfoliated into single-layered 2D nanosheets, their band gap increases from 1.3 to 1.9 eV [7,8]. The interesting nanostructures and band gaps make MoS 2 a useful active material in various fields, including energy storage and solar cells [9][10][11][12][13][14], catalysis [15][16][17][18][19], sensing [19][20][21], and electronic devices [22][23][24]. As a major member of the TMD family, MoS 2 has been known as a highly efficient catalyst for the hydro-desulfurization, reduction reaction of methyl orange by MoS 2 /montmorillonite [25] and hydrogen evolution reactions (HER) [26][27][28]. Experimental and theoretical studies have verified that the catalytic activity of MoS 2 is associated with metallic edges, whereas its semiconducting basal plane is catalytically inert [29]. Few-layer MoS 2 (FL-MoS 2 ) nanosheets can maximize the fraction of exposed edge sites and thus show highly enhanced catalytic performance [30]. FL-MoS 2 nanosheets can also be used as electrochemical biosensors for the detection of glucose [31], acetaminophen (AC) [32], double-stranded DNA [33], ascorbic acid [34], and H 2 O 2 [35] which show a high response due to the advantage of abundant exposed edges.
Micromechanical exfoliation, firstly developed to prepare graphene by exfoliating HOPG [63], can achieve isolated individual crystal planes (monolayers) with high crystal quality and macroscopic continuity, but its yield is too low to be of any practical applications. Chemical exfoliation usually necessitates intercalating various small species (i.e., organic molecules, alkali metals and transition-metal halides) into the interlayer spaces to achieve the exfoliation. In most cases, chemical exfoliation is time-consuming, sensitive to environmental conditions, and difficult to control the intercalated degree [49]. Electrochemical exfoliation, through a controllable lithiation process, has been developed to exfoliate the layered solids (i.e., MoS 2 , WS 2 , TiS 2 , TaS 2 , ZrS 2 , and graphite) as the cathode in an electrochemical setup, and the lithium-intercalation in the layered materials can be monitored and finely controlled during the discharge process [49], but high-cost and low yield limit its wide applications. Liquid-phase exfoliation (LPE) is one of the most common methods used in the exfoliation of layered solids recently [51]. The LPE process involves the following steps: (1) immersion of layered materials in a suitable solvent (i.e., NMP, DMF, THF, NVP, CAN, and DMSO), (2) exfoliation of layered materials, and (3) stabilization of exfoliated 2D materials [51]. Sufficient liquid immersion is one of the most effective and straightforward means to reduce the strength of the van der Waals attractions, where the potential energy between adjacent layers is mainly contributed by the dispersive London interactions. Although LPE can be scaled up, the solvents used are usual toxic and not environmentally friendly. Supercritical fluid exfoliation is the newly developed method to prepare few-layer nanosheets via the shearing force generated by the gas molecules in the supercritical condition [64]. To further improve the exfoliated efficiency, scientists have coupled various methods, i.e., the two-solvent grinding-assisted LPE process [65] and ionic liquid assisted grinding [35,66], to exfoliate layered solids to make few-layer nanosheets. However, scalable, efficient, cost-effective, and high-yield synthesis of FL-MoS 2 nanosheets is still challengeable.
Here, we develop a novel coupled ultrasonicationmilling (CUM) process, as shown in Fig. 1, to exfoliate natural molybdenite powders to prepare FL-MoS 2 nanosheets in a high-efficient manner via the synergistic effect of ultrasonication and sand milling. The precursor of natural molybdenite powders is directly from the molybdenite ore after a purification and superfine treatment.
Molybdenite ore is abundant in China, and using natural molybdenite as the starting material can achieve a lowcost and low-carbon approach to make FL-MoS 2 nanosheets. For the ultrasonic (U) exfoliation, the key point is the choice of suitable exfoliation solvent, and it is favorable to exfoliation when the surface tension of the solvent matches well with the surface energy of the layered bulk materials. The surface energy of MoS 2 reported is 75 mJ m −2 , and the N-methyl-2-pyrrolidone (NMP) with a surface tension of 41 mN m −1 is found to be a good solvent for the exfoliation and dispersion of MoS 2 nanomaterials [67]. NMP is therefore chosen as the exfoliation solvent in the exfoliation of natural molybdenite powders via the present CUM process. Because of the steric effect of polyvinylpyrrolidone (PVP) molecules, the exfoliating yield of FL-MoS 2 via the CUM process with the aid of PVP under the optimal conditions reaches 21.6 %, much higher than those of the reported methods. The asobtained FL-MoS 2 nanosheets by the CUM process from natural molybdenite powders show highly electrochemical Fig. 1 Schematic diagram of the coupled ultrasonication-milling (CUM) strategy for exfoliating natural molybdenite powders to make few-layer MoS 2 (FL-MoS 2 ) nanosheets: a Components of the CUM system, b basic operational principle of the CUM system, and c exfoliation of natural molybdenite towards FL-MoS 2 nanosheets by the CUM route performance for hydrogen evolution reaction and sensing applications. The newly developed CUM process combines the advantage of both liquid-phase exfoliation (LPE) and mechanical exfoliation. The synergistic effect of the shear force generated by the high-speed sand mill and the cavitation by the liquid-phase ultrasonic process can highly accelerate the exfoliation of molybdenite powders.
The key points of this work are to explore the influencing factors and possible synergistic effect during the CUM process.

Exfoliation of Natural Molybdenite Powders by CUM Process
The exfoliation of natural molybdenite powders towards FL-MoS 2 nanosheets was conducted in the home-made CUM system ( Fig. 1) using NMP as the solvent containing PVP as the exfoliation agent at~25°C. The exfoliation of natural molybdenite powders by the CUM process was conducted under the similar conditions. Typically, natural molybdenite powders (~11.25 g) and PVP (~3.9 g) were dispersed in 250 mL of NMP, forming a mixture with an initial molybdenite concentration of 45 g L −1 and a PVP concentration of 15.6 g L −1 . The above mixture was ultrasonically treated for~20 min and then was poured into the sand mill. The synergistic exfoliation under milling and cavitation was kept for 4 h using the CUM process (Fig. 1). The diameter of the horn in the ultrasonication system was 10 mm, and the ultrasonication power was 280 W. The ultrasonication acted intermittently, i.e., 3 s on and 1 s off, avoiding damage to the ultrasonic processor because of heating. The rotation speed of the sand mill was kept at 2250 r.p.m. The as-obtained FL-MoS 2 suspension was classified by centrifugation at various speeds to achieve FL-MoS 2 dispersions. Some of the FL-MoS 2 dispersions were centrifuged, washed carefully, and then freezedried to achieve FL-MoS 2 nanosheets.
The initial concentration of natural molybdenite powders (C in ,15-55 g L −1 ), ultrasonic power (P u , 200-350 W), rotation speed of sand mill (ω s , 1500-2250 r.p.m), exfoliation time (t ex , 0.5-6 h), and the molar ratio of PVP unit to MoS 2 (R pm , 0-1) were chosen as the major influencing factors that were investigated systematically during the exfoliation. The exfoliation rate (R ex , g L −1 h −1 ) of FL-MoS 2 nanosheets was measured by the FL-MoS 2 concentration (C ti ) in the upper dispersion after centrifuging the exfoliated suspension with a standing (more than~24 h) at 1500 r.p.m for 45 min: R ex = (C t1 − C t2 )/Δt = (C t1 − C t2 )/(t 1 − t 2 ), where C t1 and C t2 were the FL-MoS 2 concentrations of the upper dispersions obtained at exfoliation time t 1 and t 2 , respectively. The exfoliation yield was defined as α = C 1 /C 0 × 100 %, where C 1 was the FL-MoS 2 concentration of the upper dispersion at the ending, and C 0 was the initial concentration of natural molybdenite powders. The FL-MoS 2 concentration in the upper suspension was determined via the absorption coefficient (ε) at 670 nm according to the Beer-Lambert law (A/l = ε C) with a typically known concentration value (C' , g L −1 ). The concentration C was obtained by filtrating a certain amount (V, mL) of the target suspension, followed by washing, drying, and weighing the as-obtained solid species (m, mg): C' = m/V [44]. The typical sample was made under the following parameters: R pm = 1/2, C in = 45 g L −1 , P u = 280 W, ω s = 2250 r.p.m, and t ex = 4 h. The FL-MoS 2 dispersions and FL-MoS 2 nanosheets were also prepared by the CUM process under various parameters (C in , P u , ω s , t ex , R pm ) to investigate influencing factors.
In order to measure the exfoliation rate, a series of suspensions were sampled out for every 30 min and allowed to staying for~24 h before centrifugation at 1500 r.p.m for 45 min. The absorption coefficients of the upper dispersions (2 mL) obtained by centrifugation were measured by UV-vis-IR absorption spectra. The concentrations of FL-MoS 2 nanosheets in the upper dispersions were then calculated using the Beer-Lambert law.
For the purposes of comparison, the natural molybdenite powders were also exfoliated only using the ultrasonication or milling process. For the ultrasonication exfoliation (i.e., the U mode), 200 mL of NMP containing natural molybdenite powders (45 g L −1 ) and PVP (15.6 g L −1 ) was put in a wide-mouth bottle (250 mL) and ultrasonically treated for 4 h (280 W, 3 s on and 1 s off) using a φ10-mm horn. For the milling exfoliation using a sand mill (i.e., the M mode), the NMP mixture containing natural molybdenite powders (45 g L −1 ) and PVP (15.6 g L −1 ) was milled for 4 h using a rotation speed of 2250 r.p.m. Some of the MoS 2 dispersions obtained by centrifugation at 1500 r.p.m for 45 min after standing for about 24 h were centrifuged once again at various speeds (e.g., 8000 r.p.m, 11,000 r.p.m) to achieve FL-MoS 2 dispersions for further characterization and applications.

Characterization of FL-MoS 2
The UV-vis spectra of the FL-MoS 2 suspensions were recorded on a UV-1800PC spectrometer (Shanghai Mapada), and a quartz cell with a path length of 1.0 cm was used as the sample pool. The SEM images were obtained using a JEOL JSM-7500F scanning electron microscope, with an Au film coating (20 mA for 50 s) before SEM observation. The TEM images were measured using a field-emission transmission electron microscope (JEOL, JEM-2100) with an accelerating voltage of 200 kV. The Raman spectra were obtained using a Raman spectrometer (LabRAM HR Evolution, HOR-IBA JobinYvon, France) at 525 nm, and the powder samples were prepared on a SiO 2 /Si substrate. The X-ray diffraction (XRD) patterns were recorded on an XD-3 Xray diffractometer (Beijing Purkinje General Instrument Co., Ltd., China) with a Cu Kα irradiation (λ = 0.15406 nm). The Fourier transformed infrared (FT-IR) spectra were recorded on a Bruker-Equinox 55 spectrometer in a transmittance mode in a wavenumber range of 4000 to 400 cm −1 . The atomic force microscopy (AFM) measurement was performed on a Bruker Nano-Scope V instrument in a tapping mode. The N 2 adsorption-desorption isotherms were obtained on a Surface Area and Porosity Analyzer (ASAP 2460, Micromeritics).

Test of Electrochemical Property
The electrochemical property of the FL-MoS 2 nanosheets was characterized as electrocatalysts for hydrogen evolution reaction (HER) and as an active material for electrochemical sensors, respectively. For electrocatalytic applications, the upper FL-MoS 2 dispersion obtained by centrifugation at 8000 r.p.m for 45 min was centrifuged once again at 10000 r.p.m for 45 min, and the asobtained solid FL-MoS 2 sample was used for HER test. To prepare electrodes, 4 mg of the FL-MoS 2 nanosheets was dispersed in a mixture of Nafion (50 μL) and N, Ndimethylformamide (DMF, 950 μL) by sonication for more than 30 min to form a homogeneous ink. Before modification, the glassy carbon electrode (GCE) with a diameter of 3 mm was polished to a mirror finish using polish paper and alumina pastes of 50-70 μm and 30-50 nm, respectively, and then cleaned in an ultrasonic cleaner with acetone, alcohol, and water. Finally, 4 μL of the FL-MoS 2 ink was carefully loaded onto the surface of the above GCE electrode to prepare electrodes modified with FL-MoS 2 by allowing the liquids to evaporate under infrared light.
All electrochemical experiments were performed with a three-electrode system in which the Pt-wire electrode, saturated calomel electrode (SCE), and FL-MoS 2 -modified GCE (0.07 cm 2 ) electrode were used as the counter, reference, and working electrodes, respectively. All measurements were carried out at room temperature. The HER activity was evaluated by linear sweep voltammetry conducted with a electrochemical working station (model CS310, Wuhan Correst Instruments Co., Ltd) with a scan rate of 2 mV s −1 . The potential values with the reference of SCE electrode were calculated to be reverse hydrogen electrode (RHE):

Exfoliation of Natural Molybdenite via the CUM Process
The exfoliation of natural molybdenite powders was conducted in a home-made coupled ultrasonicationmilling (CUM) system. As Fig. 1a shows, this system consists of an ultrasonic instrument and a sand mill that are commercially available. Possible working mechanism in this CUM system can be described as Fig. 1b. Its unique feature is that the sand milling and ultrasonication can act at the same time. The major difference of between conventional wet ball milling and sand milling is their rotation speeds. The rotation speed of sand milling in this exfoliation is more than 2000 r.p.m, much higher than that of the conventional wet ball milling. Also, the special recycling in the sand milling can improve the exfoliation efficiency. With the CUM system, natural molybdenite powders can be exfoliated to fewlayer MoS 2 nanosheets via classification by centrifuging, as shown as Fig. 1c.
Natural molybdenite powders were exfoliated to FL-MoS 2 via the CUM process using the synergistic effect of ultrasonic cavitation and shearing force (Fig. 1a). Figure 2 shows the SEM images of the pristine molybdenite powders and the FL-MoS 2 nanosheets (R pm = 1/2, C in = 45 g L −1 , P u = 280 W, ω s = 2250 r.p.m, t ex = 4 h). As Fig. 2a shows, the natural molybdenite consists of micro-scale powders with an irregular morphology and their particle sizes range 1-3 μm. Figure 2b shows the SEM image of the typical FL-MoS 2 nanosheets, which were assembled to be thin films during the centrifugation by overlapping one another. The enlarged SEM image shown in Fig. 2c indicates that the FL-MoS 2 nanosheets in the thin films are flexible nanosheets with a lateral length of several hundred nanometers. Figure 3 shows the typical TEM observations of the FL-MoS 2 nanosheets derived from natural molybdenite after centrifugation at 8000-10,000 r.p.m via the CUM process (R pm = 1/2, C in = 45 g L −1 , P u = 280 W, ω s = 2250 r.p.m, t ex = 4 h). The low-magnification TEM image shown in Fig. 3a indicates that the FL-MoS 2 nanosheets are individually dispersive and have a lateral size of 161 ± 66 nm according to the statistical data based on the TEM images. The enlarged TEM images in Fig. 3b-d, g show that the MoS 2 nanosheets are very thin judged by the contrast. Figure 3e shows a selected area electron diffraction (SAED) pattern of a typical MoS 2 nanosheets of Fig. 3d. The ordered diffraction lattice indicates the MoS 2 nanosheet is single-crystal and can be readily indexed to a 2H-MoS 2 phase. Its corresponding highresolution TEM (HRTEM) image is shown in Fig. 3f, in which the highly ordered crystal lattice corroborates that the FL-MoS 2 nanosheets, obtained via the CUM process, are of a perfect single-crystal nanostructure [50,52]. The SAED pattern in Fig. 3h, corresponding to the lower  Fig. 3g, consists of two sets of ordered diffraction lattices, suggesting that two single-crystal MoS 2 nanosheets are assembled together. Figure 3i shows the HRTEM image of the particle corner in Fig. 3g and no clear defects can be found in a large area [68]. From the TEM observations, one can conclude that the MoS 2 sample obtained by exfoliating natural molybdenite powders via the CUM process takes on a thin plate-like morphology with lateral size of~160 nm, and the FL-MoS 2 nanosheets are of a perfect single-crystal structure. Figure 4 shows the typical AFM images of the FL-MoS 2 derived from natural molybdenite powders via the CUM process (R pm = 1/2, C in = 45 g L −1 , P u = 280 W, ω s = 2250 r.p.m, t ex = 4 h) with centrifugation at 10,000 r.p.m. The AFM image in Fig. 4a shows the lateral size of the MoS 2 nanosheets is about 100-250 nm, agreeing with the TEM observations (Fig. 3). Figure 4b-e show the AFM images of some typical MoS 2 nanosheets with the corresponding height profiles (Fig. 4c, e). According to Fig. 4c, e, the thickness of the MoS 2 nanosheets is about 1 nm, slightly larger than the thickness (~0.8 nm) of a single MoS 2 layer exfoliated mechanically [43,69]. This may be due to the presence of adsorbents on the nanosheets [70]. The AFM results, together with the TEM observations, corroborate that FL-MoS 2 nanosheets, even single-layer ones, are successfully achieved via the CUM process using natural molybdenite powders as the starting materials. Figure 5a shows the XRD patterns of the pristine natural molybdenite powders and the FL-MoS 2 nanosheets exfoliated by the CUM process with centrifugation at 8000-10,000 r.p.m. For the natural molybdenite powders, the diffraction peaks at around 2θ = 14.2°, 32.4°, 33.23°, 35.6°, 44.0°, 50.0°, and 58.0°are assignable to the deflections from the planes of (002), (100), (101), (102), (104), (105), and (110) of a 2H-molybdenite (β-MoS 2 or 2H-MoS 2 ) with a hexagonal structure (space group: P6 3 / mmc [194]) according to PDF # 37-1492. The cell parameters can be calculated to be a = 0.3181(3) nm, c = 1.234(1) nm, and cell vol. = 0.1082(2) nm 3 using a program UnitCell (a method of TJB Holland and SAT Redfern 1995). The FL-MoS 2 nanosheets show a similar XRD pattern, suggesting that the phase structure is not changed during the exfoliation by the CUM process, but the (002) peak broadens obviously. The calculated cell parameters of the FL-MoS 2 nanosheets are a = 0.3180(4) nm, c = 1.235(2) nm, and cell vol. = 0.1081(3) nm 3 , similar to those of the natural molybdenite powders. The obvious broadening of the (002) diffraction peak after exfoliation indicates the FL-MoS 2 nanosheets obtained by the CUM process have a small thickness [54], agreeing with the TEM, SEM, and AFM results. Figure 5b shows the UV-vis spectra of the FL-MoS 2 dispersions, i.e., (A) the supernatant just obtained by centrifugation at 8000 r.p.m for 45 min and (B) the one after standing for 42 days. One can find that the UV-vis absorption intensities at 605 and 667 nm of the FL-MoS 2 dispersions before and after standing for 42 days are almost the same. Also, the color of the dark-green FL-MoS 2 dispersion was not changed during the standing, and no sediment can be found at the bottom of the FL-MoS 2 dispersion. The peaks at 667 and 605 nm arise from direct excitonic transitions at the K point of the Brillouin zone, and those at~450 and~400 nm can be assigned to the direct excitonic transition of the M point [65,71]. This result confirms that the FL-MoS 2 dispersion obtained by the CUM process is highly stable. Figure 5c shows the typical FT-IR spectra of the FL-MoS 2 nanosheets and their precursor (i.e., the pristine molybdenite powder). The band at 475 cm −1 correspond to the characteristic vibration of the Mo-S bond due to the coupling vibration of Mo-S group, and the band at 3419 cm −1 corresponds to the stretching vibration of water molecules adsorbed on the surface, similar to literature [66]. The FL-MoS 2 nanosheets and the pristine Raman spectra are also used to characterize the thickness of FL-MoS 2 nanosheets. Figure 5d shows the Raman spectra of the pristine molybdenite powders and the solid FL-MoS 2 nanosheets obtained after centrifugation at different centrifuging speeds (i.e., one FL-MoS 2 sample obtained by centrifuging the supernatant of 1500 at 8000 r.p.m and the other one by centrifuging the supernatant of 8000 at 10,000 r.p.m). For the pristine molybdenite powders, a typical peak at 374.4 cm −1 , known as E 1 2g , originates from the in-plane vibration of the Mo-S bond, and an A 1g peak at 400.4 cm −1 from out-of-plane vibrations is also observed [53]. After exfoliation by the CUM process, the E 1 2g and A 1g peaks of the FL-MoS 2 nanosheets shift towards higher frequency (i.e., blueshift), and the peak difference between E 1 2g and A 1g is about 25 cm −1 . The blueshift of peak A 1g can be used to estimate the layer number of MoS 2 nanosheets [69]. The higher centrifugation speed achieves smaller FL-MoS 2 nanosheets. The blueshift of Raman shift indicates that the MoS 2 nanosheets exfoliated by the CUM process consist of a large amount of thin MoS 2 nanosheets with few layers [9]. Figure 5e shows the N 2 adsorption-desorption isotherms of the FL-MoS 2 nanosheets obtained by centrifuging the supernatant of 8000 at 10,000 r.p.m. The FL-MoS 2 nanosheets do not show a porous microstructure. The BET specific surface area of the FL-MoS 2 nanosheets is as high as 924 m 2 g −1 , whereas their precursor of the pristine molybdenite is only 29 m 2 g −1 .
X-ray photoelectron spectroscopy (XPS) was further used to characterize the chemical compositions of the FL-MoS 2 nanosheets derived by the CUM process. Figure 6 shows the XPS spectra of the FL-MoS 2 nanosheets obtained by centrifuging the supernatant of 1500 at 8000 r.p.m for 45 min. Figure 6a shows the survey. The major XPS peaks can be assignable to the elements of Mo, S, O, and C, and some weak peaks may correspond to Zr and Tc elements. Mo and S come from the FL-MoS 2 nanosheets. The elements of O and C are due to the substances (e.g., PVP, CO 2 ) adsorbed on the samples. The possible elements of Zr and Tc may come from the ultrasonication horn in the CUM system. Figure 6b shows the core XPS spectrum of Mo 3d. The strong peaks at around 229.0 and 232.1 eV correspond to Mo 4+ 3d 5/2 and Mo 4+ 3d 3/2 , respectively, and the weak peak at 226.1 eV corresponds to S 2s [72]. Figure 6c shows the core XPS spectrum of S 2p, and the peaks at 163.0 and 161.8 eV correspond to the S 2p 1/2 and S 2p 3/2 orbitals of divalent sulfide ions, respectively. According to the XPS spectra, Mo and S in the FL-MoS 2 are in good agreement with the binding energies of Mo 4+ and S 2− ions in the 2H-MoS 2 phase.

Factors Influencing the Exfoliation in the CUM Process
To further understand the exfoliation of natural molybdenite powders using the CUM process, we carefully designed a series of control experiments to check the factors that influence the exfoliation. The concentrations of the FL-MoS 2 nanosheets in the NMP suspensions were measured using the UV-vis-IR spectra mentioned in the experimental section. The influencing factors we checked included the exfoliation times, the initial concentrations of natural molybdenite powders, ultrasonic power, rotation speed of sand mill, and PVP concentration. Figures 7 and 8 show the typical results. Figure 7a shows the UV-vis-IR absorption spectra of the FL-MoS 2 suspensions in the CUM process (R pm = 1/2, C in = 45 g L −1 , P u = 280 W, ω s = 2250 r.p.m) with various exfoliation times (t ex = 0.5-6 h). All the UV-vis-IR spectra take on a similar profile and their absorbance values increases with the increase in the exfoliated time. Figure 7b shows the plot of the concentrations of the FL-MoS 2 nanosheets versus the exfoliation times, obtained according to the Beer-Lambert law. One can see that the concentration of FL-MoS 2 nanosheets in the NMP suspension increases linearly with the increase in exfoliation time [73]. The slope of the linear-fitting line can be seen as the exfoliation rate, and the exfoliation rate in the present condition reaches 1.42 g L −1 h −1 . Figure 7c shows the yield of FL-MoS 2 nanosheets at various exfoliation times. The yield increases linearly with the exfoliation time and reaches 21.6 % when the exfoliation time is 6 h.
The typical results on the factors that influence the exfoliation behavior in the CUM process are summarized in Fig. 8. Figure 8a-c show the effect of the initial concentrations (C in , 15-55 g L −1 ) of natural molybdenite powders on the exfoliation rate and yield of natural molybdenite powders (P u = 280 W, S r = 2250 r.p.m, R pm = 1/2). Figure 8a shows the change in the concentration of the FL-MoS 2 (i.e., supernatants obtained by centrifuging at 1500 r.p.m for 45 min after settling for more than 24 h) as the exfoliation time increases with various initial concentrations of natural molybdenite powders. The exfoliation rate and yield with various initial concentrations of natural molybdenite powders are shown in Fig. 8b, c, respectively. The increase in the initial concentration of the molybdenite powders is favorable to enhance the exfoliation rate, but the yield is not in a similar case. Low initial concentration (15 g L −1 ) achieves a higher yield (~23 %) of FL-MoS 2 nanosheets and makes better dispersion of FL-MoS 2 nanosheets than the higher ones. Considering the exfoliation amount, we chose an initial molybdenite concentration of 45 g L −1 for the following exfoliation experiments. Figure 8d-f show the effect of ultrasonic powers (P u , 200-350 W) on the exfoliation behavior of natural molybdenite powders (C in = 45 g L −1 , ω s = 2250 r.p.m, R pm = 1/2). One can see that the higher ultrasonic powers are helpful to improve the exfoliation rate and yield, but this effect is not so obvious. When the ultrasound power is 280 W, the exfoliation yield at 4 h reached highest value of~20 %. As we know, the ultrasound at a low input power corresponds to the stable cavitation, and the knocking on the right sides of plates has little ability to destroy the covalently bonded S-Mo-S, while the knocking on their profiles can easily overcome van der Waals forces between  MoS 2 nanosheets. When the input power increases, the ultrasound corresponds to inertial cavitation, and the increase in the number of cavitation bubbles leads to a small amount of useful bubbles, weakening the knocking intensity [67]. Therefore, too high or too low sonication power is not beneficial to enhance the exfoliation yield. The ultrasonication power is determined at 280 W in the following experiments. Figure 8g-i show the effect of the rotation speed (ω s , 1500-2250 r.p.m) of the sand mill on the exfoliation rate and yield at similar conditions (C in = 45 g L −1 , P u = 280 W, R pm = 1/2). The yield has the highest value when the rotation speed is 2250 r.p.m. A higher rotation speed is more favorable to overcome the van der Waals forces between the MoS 2 layers and then leads to more defects which accelerate the exfoliation with the synergistic effect of ultrasound. Considering the safety and noise of the sand mill, a suitable rotation speed of 2250 r.p.m is chosen. Figure 8j-l show the effect of PVP concentrations (R pm , 0-1) on the exfoliation behavior (C in = 45 g L −1 , P u = 280 W, ω s = 2250 r.p.m). The exfoliation rate and yield increase with the increase of the PVP concentration. PVP molecules in the CUM process act both as the dispersing agent and separating agent to prevent aggregation of the exfoliated MoS 2 nanosheets. But, if the PVP concentration is too high, the viscosity of the suspension increase rapidly, and the high viscosity leads to a terrible fluidity, resulting in a bad effect on the exfoliation [74]. The PVP content used in the CUM process is kept at R pm = 1/2 in the present experiments. Fig. 8 The effects of exfoliating parameters on the supernatant concentration (C ex ), exfoliation rate and yield of FL-MoS 2 nanosheets from natural molybdenite powders: a-c the initial concentration of natural molybdenite (C in = 15-55 g L −1 when P u = 280 W, ω s = 2250 r.p.m, R pm = 1/2); (d-f) ultrasonic powers (P u = 200-350 W when C in = 45 g L −1 , ω s = 2250 r.p.m, R pm = 1/2); g-i rotation speeds (ω s = 1500-2250 r.p.m when C in = 45 g L −1 , P u = 280 W, R pm = 1/2); j-l molar ratios of PVP to MoS 2 (R pm = 0-1 when C in = 45 g L −1 , P u = 280 W, ω s = 2250 r.p.m) To demonstrate the synergistic effect of milling and ultrasonication, we comparatively conducted exfoliation experiments under the similar conditions but using different exfoliation modes: coupled ultrasonication-milling (the CUM mode), milling (the M mode) and ultrasonication (the U mode). Figure 9 shows the exfoliation behavior using the three exfoliation modes (P u = 280 W, C in = 45 g L −1 , ω s = 2250 r.p.m, R pm = 1/2). One can see that the CUM process is more efficient in the exfoliation of natural molybdenite powders to form FL-MoS 2 . The exfoliation yield of the CUM process is~19.9 %, higher than that (18.4 %) of the M process and that (15.4 %) of the U process. The CUM process combines the advantages of both ultrasonic cavitation and shear action in the exfoliation process. During the exfoliation process, the high-speed rotating balls in the M mode knock the edges of the MoS 2 layers and lead to a large amount of defects and cracks, which are in favor of the insertion of solvent molecules into the interlayer spaces and then overcome the van der Waals forces. Also, the CUM system can realize a continuous process: the milling and ultrasonication can work at the same time. The defects and cracks caused by the milling process are helpful to improve the shearing action of ultrasonication greatly. The shear action and the knocking of cavitation cut down and exfoliate layered molybdenite powders to form smaller and thinner MoS 2 nanosheets, which are protected by the PVP molecules instantly.
Considering the above results (Figs. 7, 8, and 9), one can safely conclude that the PVP content (R pm ) has the most important effect on the exfoliation of natural molybdenite powders to FL-MoS 2 nanosheets in the CUM process when compared with the other experimental parameters, including the initial concentration of natural molybdenite powders (C in ), ultrasonic power (P u ), and rotation speed of sand mill (ω s ). The optimal experimental parameters in the CUM process can be determined as R pm = 1/2, C in = 45 g L −1 , P u = 280 W, and ω s = 2250 r.p.m judging from the overall aspects. The exfoliation yield of FL-MoS 2 nanosheets under the optimal condition for 6 h can reach 21.6 %.

Electrochemical Performance of FL-MoS 2 Nanosheets Derived by the CUM Process
Electrochemical application is one of the most important aspects for the FL-MoS 2 nanosheets. Figure 10 shows the electrocatalytic performance of the pristine molybdenite powders (A) and the FL-MoS 2 nanosheets obtained by centrifugation at 1500-8000 r.p.m (B) and at 8000-10,000 r.p.m (C), respectively. Figure 10a shows the typical polarization curves, and their corresponding Tafel plots are shown in Fig. 10b. It is clear that the higher the centrifugation speed, the thinner the FL-MoS 2 nanosheets are in the same exfoliation using the CUM process. As Fig. 10 shows, the FL-MoS 2 nanosheets obtained at 8000-10,000 r.p.m (C) have the lowest onset overpotential and the largest exchange current density at a given overpotential among the three samples (A, B, and C). By fitting the Tafel plots linearly according to the equation of η = a + b⋅log|j| (where η is overpotential, j is current density, and b is Tafel slope) [75], one can find that the FL-MoS 2 nanosheets obtained at 8000-10,000 r.p.m have a Tafel slope of 156 mV dec −1 , lower than that (165 mV dec −1 ) of the pristine molybdenite powders and that (160 mV dec −1 ) of the FL-MoS 2 nanosheets obtained at 1500-8000 r.p.m. We note that the Tafel slope of the FL-MoS 2 nanosheets obtained by the CUM process is a little larger than that of some MoS 2 nanocrystals reported [30,76]. For examples, a Tafel slope of~120 mV dec −1 was reported for the MoS 2 nanomaterials synthesized at a temperature of~500°C [76] and 140-145 mV dec −1 at a temperature of~850°C [30]. The better HER performance of MoS 2 -based composites has also been reported, resulting in a Tafel slope of 87 mV dec −1 for MoS 2 and 51 mV dec −1 for O-MoS 2 / G [26]. The above samples were synthesized through a Fig. 9 Comparison of the exfoliation yield using different exfoliation mode (U ultrasonication, M milling, and CUM coupled ultrasonication-milling) under the condition of R pm = 1/2, C in = 45 g L −1 , P u = 280 W and ω s = 2250 r.p.m conventional hydrothermal reaction using fine chemical reagents as the starting materials. Also, hybridization with other species is another reason to improve their HER performance. Therefore, fine control in the size and morphology (i.e., more edge sites and smaller size) of the MoS 2 nanosheets and hybridization with other functional species (i.e., graphene) are the possible directions to improve the electrochemical performance of the FL-MoS 2 nanosheets obtained via the present CUM process from natural molybdenite. The larger Tafel slope obtained in the present work may be due to the difference of the natural molybdenite from other synthetic MoS 2 in phase compositions. The low electrical conductivity, the possible re-stacking of FL-MoS 2 nanosheets, the residue of PVP molecules, and the relatively high-temperature treatment in electrode preparation are also possible reasons that lead to a larger Tafel slope. To improve the electrical conductivity, carbon can be used to modify FL-MoS 2 nanosheets to design highly efficient electrocatalysts for hydrogen evolution reaction and to obtain a lower Tafel slope (47 mV dec −1 ) [77]. Functional modification of the FL-MoS 2 nanosheets is under way.
The electrochemical sensing performance of the FL-MoS 2 nanosheets was also investigated.  Figure 11b shows the plots of the intensities of the anodic peak current versus the ascorbic acid concentration. The intensity of the CV peak current increases obviously with the increase in the concentration of ascorbic acid. The linear regression equation of the anodic peak obtained according to the CV curves in Fig. 11a is I (uA) = 2.08 (Vs −1 ) [AA] (mM) -4.5605, and R 2 = 0.98. The high correlation coefficient suggests a nice linearity correlation between the anodic peak current and ascorbic acid (AA) concentrations in a range of 0.5-2 mM. Unfortunately, the biosensor is of a poor performance, which is possibly caused by the low electrical conductivity of the MoS 2 nanosheets. To anchor a secondphase particles on the MoS 2 nanosheets is under our consideration to enhance their electrochemical sensing properties. Actually, MoS 2 nanosheets have been used as the supporting material to stabilize metal nanoparticles (e.g., Au) to form hierarchical nanocomposites for the detection of chemical and biological molecules [78]. The FL-MoS 2 nanosheets exfoliated by the CUM process can be a promising active material or support to construct electrochemical electrodes for the electrochemical detection of chemical and biological molecules.

Conclusions
In summary, a novel coupled ultrasonication-milling (CUM) process has been developed for the first time to exfoliate the natural molybdenite powders to synthesize FL-MoS 2 nanosheets in the NMP solution using PVP as the exfoliation agent. The factors that influence the exfoliation behavior during the CUM process have been systematically investigated. The FL-MoS 2 nanosheets obtained at the optimal conditions (i.e., initial molybdenite concentration = 45 g L −1 , ultrasonic power = 280 W, rotation speed of sand mill = 2250 r.p.m, and the mole ratio of PVP to MoS 2 = 0.5), taken on a thin single-crystal plate-like morphology, several nanometers in thickness and hundred nanometers in width. The as-obtained FL-MoS 2 nanosheets have a high specific surface area of 924 m 2 g −1 , much larger than that (29 m 2 g −1 ) of their precursor (i.e., molybdenite powders). The yield of FL-MoS 2 nanosheets obtained by the CUM process for 6 h reaches to 21.6 %, much larger than that of the literature reported. Also, the CUM process has a higher yield than the ultrasonication (U) or milling (M) process, confirming that the synergistic effect of the ultrasonication and milling process is obvious in the exfoliation of natural molybdenite powders. The as-obtained FL-MoS 2 nanosheets show highly enhanced electrocatalytic performance in hydrogen evolution reaction and good electrochemical sensing property in detecting ascorbic acid when compared with their precursor. The exfoliation of natural molybdenite via the coupled ultrasonicationmilling process developed here provides a novel, green, efficient, scalable, and cost-effective approach to prepare FL-MoS 2 nanosheets that are of promising applications as electrocatalytic and sensing-active materials.