**Direct Exfoliation of Natural SiO2-Containing Molybdenite in Isopropanol: A Cost Efficient Solution for Large-Scale Production of MoS2 Nanosheetes**

#### **Wenyan Zhao 1, Tao Jiang 1, Yujie Shan <sup>1</sup> , Hongrui Ding 2, Junxian Shi 3,\*, Haibin Chu <sup>1</sup> and Anhuai Lu 2,\***


Received: 20 September 2018; Accepted: 12 October 2018; Published: 17 October 2018

**Abstract:** The cost-effective exfoliation of layered materials such as transition metal dichalcogenides into mono- or few- layers is of significant interest for various applications. This paper reports the preparation of few-layered MoS2 from natural SiO2-containing molybdenite by exfoliation in isopropanol (IPA) under mild ultrasonic conditions. One- to six-layer MoS2 nanosheets with dimensions in the range of 50-200 nm are obtained. By contrast, MoS2 quantum dots along with nanosheets are produced using N-methyl-pyrrolidone (NMP) and an aqueous solution of poly (ethylene glycol)-block-poly (propylene glycol)-block-poly (ethylene glycol) (P123) as exfoliation solutions. Compared with molybdenite, commercial bulk MoS2 cannot be exfoliated to nanosheets under the same experimental conditions. In the exfoliation process of the mineral, SiO2 associated in molybdenite plays the role of similar superfine ball milling, which significantly enhances the exfoliation efficiency. This work demonstrates that isopropanol can be used to exfoliate natural molybdenite under mild conditions to produce nanosheets, which facilitates the preparation of highly concentrated MoS2 dispersions or MoS2 in powder form due to the volatility of the solvent. Such exfoliated MoS2 nanosheets exhibit excellent photoconductivity under visible light. Hence, the direct mild exfoliation method of unrefined natural molybdenite provides a solution for low-cost and convenient production of few-layered MoS2 which is appealing for industrial applications.

**Keywords:** natural molybdenite; MoS2 nanosheet; SiO2; liquid exfoliation; photoelectric properties

#### **1. Introduction**

Since graphene's initial discovery [1], two-dimensional (2D) transition metal dichalcogenide semiconductors (TMD) have attracted great research interest for their unique electronic and optical characteristics, which are distinctively different from those of their bulk materials [2–6]. MoS2 is the most popular member of the series, which compensates for the disadvantage caused by the absence of a band gap of graphene, and shows considerable anisotropy due to its large intrinsic band gap, resulting in novel electronic, optical, mechanical, and structural characteristics [7–11]. In many applications, such as batteries, composites, sensors, and catalytic activities, MoS2 needs to be produced on a large-scale, and preferably at a lower cost.

Different methods to prepare monolayer and few-layered MoS2 have been developed, including mechanical exfoliation [12], liquid-phase exfoliation [13–15], chemical exfoliation [16,17], and chemical vapor deposition [18,19]. Generally, studies have found that liquid-phase exfoliation has great potential for scalable production of 2D materials. Coleman et al., have indicated that layered materials can be exfoliated into ultrathin-layered 2D nanomaterials in organic solvents by sonication [14]. But for large-scale applications, it has been difficult to achieve high enough concentrations. In many cases, the exfoliating media used with high boiling points are difficult to remove. A method of mixed-solvent containing volatile solvents has been used to exfoliate TMD [20–22]. For example, the mixture of water and ethanol was demonstrated to be an effective solvent for the exfoliation of MoS2 nanosheets [23]. Nguyen et al. have reported a two-solvent grinding-assisted liquid phase exfoliation of layered MoS2, avoiding the solvent residue [24]. Ultrasonic treatment is widely used in liquid exfoliation. Ultrasonic force generates acoustic cavitation. The shear forces coming from acoustic cavitation can break the Van der Waals interactions of bulk materials, leading to the exfoliation of 2D materials [14,24], and affecting the structural characteristics of nanoparticles [25]. In many cases, ultrasound probes are utilized to exfoliate 2D materials. Their use can generate violent mechanical driving forces by concentrating energy in a small amount of dispersion liquid. In contrast, an ultrasonic water bath is milder and more cost-effective alternative, which is appealing for samples batch processing. However, it is rarely used in current liquid exfoliation methods.

The intercalation of various intercalates is another useful liquid phase exfoliation method to weaken neighboring layers arising from the interlayer expansion. Peng et al. recently demonstrated a lithium-intercalated single-crystals exfoliation method for 2D TMD nanomaterials in water by only simple manual shaking [26]. However, this chemical exfoliation method partly leads to structural deformation, and is very sensitive to the environmental conditions. To overcome the aforementioned drawback, an alternative approach used grinding/sonication-assisted Li+ intercalation in simulated sun irradiation conditions, avoiding the use of hazardous liquids such as butyllithium; however, the use of N-methyl-pyrrolidone (NMP) leads to persistent residues on the exfoliated flakes [27]. Great effort has been made to develop new exfoliation technology. Recently, a microcentrifugation surface acoustic wave device was developed to exfoliate MoS2, which applied an electric field and mechanical shear force for fast and efficient exfoliation. Similarly, the surfactant residues and expensive equipment defeat the purpose [28]. Consequently, it remains challenging to develop liquid exfoliation techniques in more convenient and cost-effective ways to obtain highly-concentrated 2D MoS2 dispersions that allow separation as precipitates for further application.

In many cases, the preparation of 2D MoS2 in the lab is based on the liquid exfoliation of MoS2 powder and single crystals. Savjani et al., [29] and Dong et al., [30] have produced MoS2 nanosheets by directly exfoliating molybdenite minerals in NMP, and showed that unrefined molybdenite could be an exfoliation source of 2D materials. However, in many cases, the molybdenite powders are purified before exfoliation, and the effects of the impurity constituents on the exfoliation have not been considered. Because natural molybdenite contains quartz, it would thus be of interest to learn whether the quartz in natural molybdenite may play the role of ball mill during the exfoliation of molybdenite, before being separated from the suspension of the MoS2 nanosheets after exfoliation. Furthermore, we wonder if a solvent whose surface tension does not harshly match with the layered bulk materials can effectively exfoliate natural molybdenite under mild conditions. Relevant studies should be performed to solve these questions.

Here, we prepared 2D MoS2 nanomaterials by liquid exfoliation of SiO2-containing molybdenite ores in the volatile solvent, isopropanol (IPA), as well as in NMP, a surfactant aqueous solution under mild water bath ultrasound conditions. The nanomaterials obtained from molybdenite have been compared with the products obtained from commercial MoS2 powder in the same experimental conditions in terms of exfoliation efficiency, composition, structural features, and photoelectric properties.

#### **2. Materials and Methods**

#### *2.1. Materials*

Natural molybdenite (NM) was collected and enriched in Dasuji diggings, Zhuozi County, Inner Mongolia. Compositional analysis by X-ray fluorescence (XRF) spectrometry shows that natural molybdenite mainly consisted of Mo, S, SiO2, Al2O3 (Table S1). The exact content of Mo is 49.71wt%, S is 27.65wt%, and SiO2 is 14.15wt%, which was determined using an inductively-coupled plasma-mass (ICP-MS) spectrometer. The NM was ball-milled (XGB planetary mill, 100 stainless steel balls of 6 mm diameter, rotation speed 500 rpm) for 1 h and sifted to obtain mineral powder with particle sizes of <45 μm. The scanning electron microscope (SEM) images are shown in Figure S1a,b. The X-ray diffraction (XRD) patterns prove that the NM was mainly formed from 2H MoS2 and quartz-phase SiO2 (Figure S2). Energy dispersive spectrometry (EDS) shows that the SiO2 and MoS2 are mixed uniformly on the micro-scale (Figure S3). Commercially-available MoS2 (CM, 99%, average 2 μm) was bought from Sigma-Aldrich (Figure S1c,d). NMP, P123, and IPA were purchased from J&K Chemicals (Beijing, China).

Preparation of control samples of NM: SiO2 associated in NM was removed by HF method; 2 g of NM was added to 20 mL mixed acid solution (16 mL 40%HF and 4 mL H2SO4) in 50 mL Teflon crucible, and maintained under magnetic stirring for 3 h at 80 ◦C; after reaction, the product was filtrated, washed with Milli-Q water several times to neutral, then dried at 60 ◦C in oven.

#### *2.2. Exfoliation Process*

0.5 g of NM or CM was added to 50 mL of exfoliation solution (IPA, NMP or P123 aqueous solutions) in a 100 mL glass vial. The mixture was batch sonicated for 16 h in an ultrasonic water bath with a frequency of 40 kHz (volume 15 L, power 400 W), taking care to have the water level higher than the level of the suspension. The ultrasonic temperature was controlled below 50 ◦C. The resulting dispersion was centrifuged for 15 min at 10,000 rpm to remove the very thick nanosheets and excess impurity. The supernatants were collected by pipette. The products derived from NM or CM were labeled as NM-X or CM-X respectively, where X is the exfoliation solvent.

#### *2.3. Characterization*

Compositional analysis of NM was achieved using a Rigaku ZSX Primus II XRF spectrometer and a Thermo X Series II ICP-MS spectrometer. The UV-vis (Ultraviolet-visible) spectra were performed by a Hitachi UV-3900 spectrometer using 1.0 cm quartz cuvette. Accurate dilutions of the dispersions were produced to obtain suitable UV-vis spectra. The XRD pattern measurements were carried out using a PANalytical X'Pert Pro diffractometer with Cu Kα radiation at 0.15406 nm (equipped with monochromator). Raman spectra were collected by a MicroRaman spectrometer (Renishaw in Via Reflex, Gloucestershire, UK), equipped with a 532 nm DPSS laser and a 2400 lines/mm grating. For the specimen preparation, the suspension was drop-casted onto a silicon substrate (wafer) and heated to 50 ◦C to form an opaque film. SEM imaging was performed using Hitachi S4800. Energy dispersive X-ray analyses were performed using Bruker QUANTAX 200 energy dispersive spectrometer. The transmission electron microscopy (TEM) images were measured using a FEI Tecnai G2F20S-TWIN. Samples for TEM imaging were prepared by drop-casting the 2D MoS2 dispersion onto 200 mesh lacey carbon-coated copper grid and dried before analysis. Nanosheets thickness measurements were carried out on a Bruker atomic force microscopy (AFM) (Dimension Icon, Bruker, Billerica, MA, USA) operating in ScanAsyst mode. A silicon nitride probe (type SCANASYST-AIR, Bruker, Billerica, MA, USA) with a curvature radius of 2 nm was used for AFM measurements. Its specifications include a force constant of approximately 0.4 N/m, and a resonant frequency of 70 kHz. Samples for AFM were prepared by diluting the samples in ethanol to a concentration in the range of 100 μg/mL, drop-casting the diluted suspension onto a clean silicon wafer piece, and heating to evaporate the solvent.

#### *2.4. Photoelectrochemical Measurement*

The working electrodes were prepared on glassy carbon (GC) electrodes, which were dip-coated with nano MoS2 dispersions (loading 18 μg/cm2). The electrochemical measurements were tested in 0.5 M Na2SO4 with a conventional three-electrode system. Pt plate and saturated Ag/AgCl were used as counter electrode and reference electrode, respectively. The photocurrents were measured on a CHI760E Chenhua electrochemical workstation. The visible light irradiation source was obtained by a 300 W Xe lamp (Beijing Trustech, Beijing, China, PLS-SXE300c) with a 420 nm cut-off filter.

#### **3. Results and Discussion**

Nano MoS2 dispersions were prepared from NM or CM using a simple bath sonication technique in IPA, NMP, or aqueous solutions of surfactant P123. After bath sonication for 16 h, suspension of different colours were obtained and labeled as NM-NMP, NM-P123, NM-IPA, CM-NMP, CM-P123, and CM-IPA respectively (Figure 1 inset). Products from NM are dark green, CM-NMP is golden yellow, CM-P123 is yellow-green, and CM-IPA is colourless and transparent. Except for CM-IPA, the Tyndall effect was observed, which indicates that the suspension was a colloidal dispersion. UV-vis spectra of the exfoliated MoS2 are shown in Figure 1. In the samples derived from NM, the characteristic absorption bands of MoS2 nanosheets at 390nm (A), 450nm (B), 610nm (C), and 670 nm (D) are clearly observed [31–33]. The excitonic peaks at 610 nm and 670 nm arise from the K point of the Brillouin zone. The peaks at ~390 nm and ~450 nm can be attributed to the direct transition from the deep valence band to the conduction band [34,35]. But in the samples derived from CM, absorption bands in the visible light region are very weak. These results from UV-vis absorption only confirm that NM can be exfoliated to produce MoS2 nanosheets in three solvents, but CM cannot be exfoliated to obtain nanosheets under mild ultrasonic conditions.

**Figure 1.** Ultraviolet-visible absorption spectra of products exfoliated in various solvents from natural molybdenite (NM) and commercial MoS2 (CM). The Photographs of different products with a red laser crossed through is shown in the inset.

Further characterization was studied using the TEM images of exfoliated MoS2 for NM-NMP, CM-NMP, and NM-IPA. TEM images of NM-NMP are shown in Figure 2a,b,c. The transparency of the nanosheets confirms that the MoS2 sheets are very thin. Clear lattice fringes with a separation of 0.2725 nm are observed for (100) planes of 2H-MoS2 in Figure 2a. The selected area electron diffraction (SAED) pattern of MoS2 nanosheets shown in the left inset of Figure 2a reveals the highly crystalline nature of the sheets with a hexagonal diffraction pattern. The diffraction spots shown by the SAED correspond to the (100), (110), (200) crystal faces of the 2H-MoS2. Figure 2b shows several erected nanosheets. The nanosheets have only a few molecular layers, with interlayer spacing of 0.623 nm corresponding to the (002) crystal plane; the atomic arrangements are shown in Figure 2b inset. In addition, MoS2 quantum dots around the nanosheets are clearly observed in Figure 2c. Figure 2d

shows the image of CM-NMP, in which quantum dots of ~5nm are observed. The left inset of Figure 2d shows SAED pattern of quantum dots which correspond to the (100), (105) crystal faces of hexagonal system, and the inset at the top right corner shows the hexagonal arrangement of 2H-MoS2. The above results indicate that MoS2 quantum dots, along with nanosheets, were obtained by exfoliating NM; meanwhile, only quantum dots were produced from CM powder in NMP. TEM images of NM-IPA are shown in Figure 2e,f. MoS2 nanosheets with lateral size ~100 nm are observed (Figure 2e), and the layer structure of the edge warpage can be clearly shown in Figure 2f (arrow indication). Lattice fringes with a separation of 0.2745 nm are responsible for (100) planes of 2H-MoS2, and the inset at the top left corner shows the hexagonal arrangement of 2H-MoS2. MoS2 nanosheets are the only products observed in NM-IPA. The majority of the exfoliated MoS2 nanosheets are close to 1–6 layers thickness. More images for the other products obtained, such as NM-P123 and CM-P123, can be found in Figure S4. Quantum dots spreading nanosheets of MoS2 were observed in NM-P123 (Figure S4a,b), and amounts of quantum dots and a few nanosheets were observed in CM-P123 (Figure S4c,d). To further ascertain the morphology and thickness of exfoliated MoS2, AFM was carried out. Figure 3 shows scan AFM images in ScanAsyst mode of MoS2 nanosheets in NM-NMP and NM-IPA samples. Figure 3a,b reveals that the thickness of nanosheets was less than 6 nm both in NM-NMP and NM-IPA. Given that the thickness of a MoS2 monolay is ~1 nm [14,23,36], this suggests that the obtained MoS2 nanosheets contain ~1 to 6 layers, which agrees with the TEM results.

**Figure 2.** The high-resolution transmission electron microscopy (HRTEM) images of products exfoliated in N-methyl-pyrrolidone (NMP) from NM (**a**, **b**, **c**), CM (**d**), and products exfoliated in isopropanol (IPA) from NM (**e**, **f**). Insets in (**a**): a selected area electron diffraction (SAED) pattern, scale bar = 5 nm-1, a low resolution image, scale bar = 20 nm; insets in (**b**, **f**): magnification of the selected area and schematic diagram of atomic arrangement; inset in (**c**): a quantum dot image, scale bar =2 nm; insets in (**d**): a SAED pattern, scale bar = 10 nm-1, a low resolution image, scale bar = 50 nm, magnification of the selected area and schematic diagram of atomic arrangement; inset in (**e**): a low resolution image, scale bar = 1 μm.

**Figure 3.** Atomic force microscopy (AFM) images of products exfoliated from NM in NMP (**a**) and in IPA (**b**). The insets show the height profiles along with the arrows. AFM images and height profile confirming the thickness of <6 nm.

To further confirm the phase, XRD and Raman spectroscopy were performed (Figure 4). Due to the presence of polymeric surfactant P123, we could not use AFM, XRD, or Raman to analyze product NM-P123 unless the samples were treated by washing or calcining, which could change the samples (Figure S5 and S6). The appearance of (002) reflection in the XRD patterns of NM-NMP and NM-IPA indicates the presence of MoS2 nanosheets derived from NM with good crystallinity (Figure 4a). Meanwhile, XRD pattern of CM-NMP exhibiting no apparent reflection also indicates that CM-NMP mainly consists of quantum dots. Raman spectra are shown in Figure 4b. The characteristic shifts near 380.5 cm-1 and 406.5 cm-1 for bulk NM, NM-NMP, and NM-IPA respectively correspond to E1 2g and A1g vibrations modes of Mo-S bands in 2H MoS2 [37]. After exfoliation, the E1 2g and A1g peak shift, and the shift difference between E1 2g and A1g, is generally only changed with the layer number of molecules. So, the reduction of the shift difference between the two peaks can be considered to be an indicator of the layer number of MoS2 nanosheets [38]. Compared to 26.0 cm-1 shift difference of the two characteristic peaks for bulk NM, the shift difference reduces to 24.7 cm-1 for NM-NMP and 24.1 cm-1 for NM-IPA respectively. On the basis of information provided by Li et al., [37], this result indicates that the thickness of MoS2 nanosheets exfoliated from NM was mostly less than 6 layers, which is consistent with the measurements of TEM and AFM.

**Figure 4.** (**a**) X-ray diffraction (XRD) patterns of NM-NMP, NM-IPA, CM-NMP and bulk NM. (**b**) Raman spectra of bulk NM, NM-NMP and NM-IPA.

MoS2 concentration in suspension was measured by atomic absorption spectroscopy. The concentrations of dispersed nano MoS2 exfoliated from NM are 596, 485, and 252 mg/L for P123, NMP, and IPA solution respectively. Compared to the concentrations of products exfoliated from CM, which are listed in Table 1, the concentration of MoS2 derived from NM is relatively high. Furthermore, nanosheets are more easily obtained from NM. The surprise result raises a question: why can NM with lower purity and larger particle sizes compared with CM be exfoliated to produce more nanosheets? In many exfoliation methods, MoS2 nanosheets were obtained using an ultrasound probe as a violent mechanical driving force because it concentrates more energy [14,36,39]. In our case, due to the dispersive energy of the ultrasonic water bath, MoS2 quantum dots separated from the crystal defects were the only product from CM [40]. However, MoS2 nanosheets were obtained under the same mild ultrasound conditions from NM. NM contains small amounts of impurities compared with CM, mainly SiO2 (Table S1). Thus, it seems reasonable to hypothesize that the quartz phase SiO2 in the NM may cause collisions with MoS2 during bath sonication, leading to cleavage of bulk layers into ultrathin sheets, which can play the role of superfine ball milling. Furthermore, compared with the added ball milling medium (e.g., ZrO2) [39], associated quartz in the NM bonds with MoS2 more closely and uniformly, and the particle size of that is much smaller, thereby causing more effective stripping. After exfoliation, most of the SiO2 was removed from dispersion by centrifugation, as observed in the EDS. Quantitative analysis shows that Si content in exfoliated nanosheets is 0.32wt%, which is much lower than that in natural molybdenite (5.27wt%). EDS analysis shows Mo : S to be 1 : 2 in all case, and reveals that the nano MoS2 prepared from NM does not contain more non-negligible impurity contaminations than that obtained from CM.

**Table 1.** The concentrations and morphologies of MoS2 exfoliated.


<sup>1</sup> No detection.

To test the aforementioned hypotheses, we prepared control samples of NM which were purified to remove SiO2 by the HF method, and compared them with the initial NM as the source powders of exfoliation in IPA. After HF treatment, the SiO2 content in molybdenite reduces from 11% to 0.65%, and the absorbance intensity of exfoliation products at ca. 670 nm plummets, as shown in Figure S7, indicating that the yield of the nanosheets from NM with SiO2 removed is considerably lower than that from initial NM. Therefore, this result supports our argument about the associated SiO2 dependent exfoliation mechanism of the mineral, as shown in Scheme 1.

**Scheme 1.** Exfoliation mechanism of natural SiO2-associated molybdenite in IPA, NMP, and P123 aqueous solution, under mild ultrasonic condition.

In addition, the exfoliation efficiency varies greatly with the solvents, resulting in a different morphology and yield of exfoliated nano-MoS2. During the wet grinding process, the choice of grinding solvent has a significant influence on the final product [24]. According to the principle of matching surface energy of solvents [41], NMP and 10 wt% P123 with a surface tension close to 40 mN/m are good solvents for exfoliating bulk MoS2; therefore, MoS2 quantum dots are easy to obtain by water bath sonication in NMP or P123. As shown in the height profile of AFM (Figure 3), NMP produces thinner nanosheets than IPA. Studies have suggested that a close match of the surface energy between the solvent and the layered material leads to the exfoliation of flakes with increased aspect ratios [42]. Thinner flakes are more liable to breakdown and produce quantum dots by ultrasonic vibration resulting from the ultrasonic water bath. Because IPA surface tension is 20.80 mN/m, far away from 40 mN/m, nano MoS2 is not successfully prepared in IPA from CM by mild water bath sonication; however, it can be prepared from NM. Due to the surface tension of IPA not being matched, the quantum dots cannot be exfoliated off from the crystal defects, but MoS2 nanosheets have been achieved in IPA via a mineral-associated quartz milling process. Though the concentrations of dispersed MoS2 nanosheets in IPA are relatively low, the highly concentrated MoS2 nanosheets dispersion or its powder form can be obtained due to the volatility of IPA, and the solvent can be recovered for recycling exfoliation to improve yields. Therefore, aggregation of MoS2 nanosheets caused by extracting exfoliated MoS2 from NMP via heating evaporation at high temperature can be avoided. Thereafter, the obtained powder from IPA is convenient for characterization and preparation of samples in applications avoiding the effect of a solvent, which could be widely applied.

Two-dimensional MoS2 has shown prior promise for various energy-related applications because of its intrinsic band-gap structure. To confirm the photoelectric properties of our exfoliated MoS2, simple light-to-electric conversion experiments were conducted. The time-dependent photocurrent of exfoliated nano MoS2 was measured at 0.6 V bias voltage under alternating darkness and visible light conditions. As shown in Figure 5, all the samples exhibited repeatable and stable responses to illumination. In contrast, the photocurrents of NM-IPA and NM-NMP are higher than those of bulk NM and CM, and the highest ΔI is generated by NM-IPA, which is around 9-fold higher than that of bare GC. The remarkable increase in photocurrent density results from the sensitivity of exfoliated MoS2 nanosheets to visible light and the efficient separation of photo-generated electron-hole pair. Unexpectedly, a lower photocurrent response of CM-NMP was observed, which is close to that of bare GC. A higher photocurrent response of MoS2 nanosheets (NM-IPA) in comparison to their bulk structure gives rise to the change in band structure from indirect to direct bandgap [43,44]. The remarkable decrease in the photocurrent density of CM-NMP (MoS2 quantum dots) was observed because of its insensitivity to visible light and the quantum confinement effect.

**Figure 5.** The transient on/off photocurrent responses vs. time plotted for NM-IPA, NM-NMP, CM-NMP, NM, CM, and bare GC in 0.01 M Na2SO4 solution under visible-light.

#### **4. Conclusions**

Few-layered MoS2 was prepared from natural SiO2-containing molybdenite by exfoliation in IPA under mild ultrasonic conditions. The exfoliation products from NM and CM in different exfoliation solvents including IPA, NMP, and aqueous solutions of P123 were compared. The results show that the yield of nano MoS2 exfoliated from the molybdenite is higher than that from CM powder. The choice of exfoliation solvent plays a crucial role. IPA can successfully exfoliate molybdenite to produce nanosheets, but cannot exfoliate commercial MoS2 powder. NMP and aqueous solutions of P123 can exfoliate NM to MoS2 nanosheets and quantum dots, and exfoliate CM to quantum dots. It was found that the MoS2 nanosheets produced from NM in IPA exhibit greater photocurrent density under visible light. The quartz phase associated in molybdenite is a major contributor to the effective exfoliation, and plays the role of superfine ball milling.

This work highlights that unrefined natural SiO2-containing NM can be used as an exfoliation source of 2D MoS2; the ability to use volatile IPA to exfoliate it will enable recirculated exfoliation to produce MoS2 nanosheets with higher yields, for cost–effective and large-scale industrial applications. However, in the exfoliation process of minerals, the solvent has an influence on the morphology of the product, which is poorly understood at present. Further research may lead to proper evaluations of the effects of different solvents on exfoliation, which help us to fully elucidate the underlying exfoliation mechanisms.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2079-4991/8/10/843/s1, Figure S1: SEM images of natural molybdenite (a), sifted natural molybdenite after ball-milled (b) and commercial MoS2 (c, d); Figure S2: XRD patterns of natural molybdenite and commercial MoS2; Figure S3: Elemental analysis of natural molybdenite using EDS shows the SiO2 and MoS2 are mixed uniformly in the micro-scale; Figure S4: TEM images of products exfoliated in P123 aqueous solution from natural molybdenite (a, b), commercial MoS2 (c, d); Figure S5: (a) XRD patterns of products exfoliated in P123 from natural molybdenite and commercial MoS2, (b) TEM images of NM-P123 powder-sample obtained by washing with deioned water several times to remove P123 after centrifuging at 1500rpm for 45min; Figure S6: (a) Raman spectrum of NM-P123 powder-sample obtained by calcining at 450◦C for 2h, (b) TGA curves of NM, CM, NM-IPA-Powder and P123; Figure S7: UV-Vis absorption spectra stack plot of MoS2 dispersions obtained from the initial NM and NM removed SiO2, Table S1: Content of natural molybdenite by component analysis using XRF.

**Author Contributions:** Conceptualization, W.Z. and J.S.; methodology, T.J., W.Z; formal analysis, T.J., Y.S. and H.D.; resources, J.S. and W.Z.; writing—original draft preparation, W.Z., T.J. and Y.S.; writing—review and editing, H.C. and J.S.; funding acquisition, A.L., J.S. and W.Z.; supervision, A.L.; project administration, A.L.

**Funding:** This research was funded by the National Basic Research Program of China (2014CB846001), National Natural Science Foundation of China (21167008) and Inner Mongolia Natural Science Foundation (2015MS0203).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Operation Mechanism of a MoS2/BP Heterojunction FET**

**Sung Kwan Lim 1,2 , Soo Cheol Kang 1,3, Tae Jin Yoo 1,3, Sang Kyung Lee 1,3 , Hyeon Jun Hwang 1,3 and Byoung Hun Lee 1,3,\***


Received: 20 September 2018; Accepted: 4 October 2018; Published: 7 October 2018

**Abstract:** The electrical characteristics and operation mechanism of a molybdenum disulfide/black phosphorus (MoS2/BP) heterojunction device are investigated herein. Even though this device showed a high on-off ratio of over 1 × <sup>10</sup>7, with a lower subthreshold swing of ~54 mV/dec and a 1fA level off current, its operating mechanism is closer to a junction field-effect transistor (FET) than a tunneling FET. The off-current of this device is governed by the depletion region in the BP layer, and the band-to-band tunneling current does not contribute to the rapid turn-on and extremely low off-current.

**Keywords:** MoS2; black phosphorus; 2D/2D heterojunction; junction FET; tunneling diode; tunneling FET; band-to-band tunneling (BTBT)

#### **1. Introduction**

Tunneling field-effect transistors (tFETs) have been studied as an alternative device for silicon MOSFET enabling very sharp turn-on which is required to reduce the operation voltage and the system power consumption. tFETs utilize band-to-band tunneling (BTBT) from a source to a channel, and an off-current is maintained using a P-N-N or N-P-P-type channel-doping profile [1–5]. When BTBT occurs in this channel-doping profile, the carriers from the source are injected directly into the channel and transported to the drain. When BTBT is not possible, the carrier cannot be injected into the drain because of the barrier formed in the channel region. In this device, the tunneling distance should be minimized to allow the tunneling current to rapidly increase. Thus, a very sharp P-N junction should be formed. The performances of experimental tunnel FETs reported in the literature have not reached their theoretical limit, primarily due to graded doping profiles and interface traps [3,6]. For an ideal BTBT current flow, an atomically sharp interface with minimal interface states is necessary. Fortunately, these requirements can be easily satisfied using transition metal dichalcogenide (TMD) materials because the various choices of band gaps and band alignment combinations make the stack of two-dimensional (2D) materials an ideal candidate for tunneling FETs [7–11]. Thus, a variety of stacks, including molybdenum disulfide (MoS2)/tungsten diselenide (WSe2), tin diselenide (SnSe2)/WSe2, MoS2/black phosphorus (BP), and SnSe2/BP have been investigated [12–19]. Most of these studies explain that the turn-on mechanism is due to the BTBT, and the turn off mechanism is due to the band misalignment.

In this work, we fabricated a heterojunction FET, using a multilayer MoS2 and a thick black-phosphorus stack with a back gate structure, and investigated the operation mechanism. This system was chosen because a MoS2/BP stack is suitable for broken bandgap device fabrication. Our analysis revealed that the operation mechanism of this heterojunction FET is quite different from what has been reported in the literature. The off-current is dominated by the depletion in the BP layer, and the subthreshold swing is related to the reduction of the depletion region. The BTBT current only contributes to the hump in the drain current.

#### **2. Materials and Methods**

The fabrication processes of the MoS2/BP heterojunction FET are shown in Figure 1. Figure 1a shows the structure of a stamp used to transfer 2D flakes. Polypropylene carbonate (PPC) (Sigma-Aldrich, CAS 25511-85-7, Sigma-Aldrich, CAS 25511-85-7, St. Louis, MO, USA) was used to pick up and transfer the flakes of MoS2 and BP at a low temperature [18,20,21]. Since the flakes are easily damaged during the detachment process, and some of 2D materials, e.g., SnSe2, hafnium diselenide (HfSe2), and BP, can be oxidized during the transfer or device fabrication [22–24], we modified the fabrication process to directly transfer the flakes to the PPC film to minimize the damage and to reduce the air exposure time. Both sides of a handmade polydimethylsiloxane (PDMS) sheet were treated with ozone plasma for 10 min to improve the adhesion of the double-sided tape to the PDMS sheet. The PPC film (15% solution in Anisole) was coated onto the stack of tape/PDMS/tape and cured on a hot plate at 100 ◦C for 10 min. Then, the PPC/tape/PDMS/tape sheet was placed on a glass slide patterned with align keys. Figure 1b–d show the rest of the device fabrication process.

**Figure 1.** (**a**) Schematic of stamp (polypropylene carbonate (PPC)/double-sided tape/ polydimethylsiloxane (PDMS)/double-sided tape/glass slide), with the 2D flake transferred directly onto the PPC film. (**b**) MoS2 transferred onto the drain electrode (5-nm/45-nm Ti/Au) and gate oxide (30-nm Al2O3). (**c**) Black phosphorus (BP) flake transferred quickly to the substrate using the same method. (**d**) Device passivated using polymethylmethacrylate (PMMA) film. (**e**) Optical image of MoS2/BP (4.2 nm/50 nm) heterojunction. (**f**) The thickness of flakes was measured using Raman spectra (using a 514-nm laser) of the molybdenum disulfide (MoS2)/BP stack. The lower panel shows the Raman spectra of the MoS2 flake (the E1 2g peak at 382.29 cm−<sup>1</sup> and the A1g peak at 406.25 cm<sup>−</sup>1).

The source and drain electrodes (5-nm Ti/45-nm Au) were formed on a 30-nm aluminum oxide (Al2O3)/highly doped P-type silicon substrate using e-beam evaporation and photolithography. In this experiment, MoS2 was used as the channel material with BP as the source material. Exfoliated MoS2 flakes were transferred to the PPC film from a bulk crystal using commercial adhesive tape, and then transferred onto the drain electrodes using a dry transfer system at 80 ◦C. The selected BP flake was also transferred onto the source electrode using the same process, while carefully overlapping the BP flake onto the MoS2 flake that was already connected to the drain electrode. Since the BP flake could be easily oxidized in air [24], a polymethylmethacrylate (PMMA, 950 K 4 A, Microchem, Westborough, MA, USA) coating was applied, followed by thermal annealing at 180 ◦C for 5 min to eliminate the solvent. Figure 1e shows an optical microscope image of the device. The thickness of the MoS2, measured with atomic force microscopy, was 4.2 nm and the BP thickness was ~50 nm. Figure 1f shows the Raman spectrum of the BP/MoS2, measured from the overlapped region. The characteristic BP peaks were observed at 360.65 (A1 g), 437.3 (B2g), and 464.4 (A<sup>2</sup> <sup>g</sup>) cm<sup>−</sup>1; however, the Raman peak of the MoS2 was not observed in this spectrum because the BP layer was very thick. The Raman spectrum of the MoS2 shown in the lower part of Figure 1f was measured from the region not overlapping the BP layer.

#### **3. Results and Discussion**

First, the electrical characteristics of the MoS2/BP diode were measured using a parameter analyzer (Keithely 4200, Santa Rosa, CA, USA). The TMD materials show different electrical characteristics depending on the thickness of the layer. When the BP is very thick, no significant gate modulation is observed, as shown in Figure 2a. In fact, this characteristic is beneficial for device operation because the high current level with a small gate modulation means that the BP layer can be used as a good contact material with a bandgap.

**Figure 2.** (**a**) Transfer characteristic of the thick-layer BP field-effect transistor (FET). (**b**) Electrical characteristics of a MoS2/BP diode following the gate voltage. Band structure of the MoS2/BP, (**c**) before contact, (**d**) at the equilibrium state, and (**e**) with a forward rectifying condition with negative bias applied to the MoS2 electrode. The holes from BP cannot overcome the high barrier at the forward bias and the electrons from MoS2 diffuses into BP, generating a depletion region. (**f**) Reverse bias condition with positive bias applied to the MoS2 electrode. The current is primarily due to the drift of minority carriers, as well as the tunneling carriers from the BP side.

Figure 2b shows the diode characteristics of the MoS2/BP heterojunction at different gate biases, from −2 to 2 V with a gate bias step of 1 V. While the potential of the BP layer is almost fixed to the source–drain bias, the Fermi level of the MoS2 layer shows a reasonable gate modulation for both single layers and multilayers [25]. As the gate bias increased from −2 V, the rectification characteristics at the MoS2/BP junction seemed to improve because the barrier height at the MoS2/BP interface increased. These characteristics can be explained more intuitively with a band diagram. The ideal band structure of a MoS2/BP stack before stacking is shown in Figure 2c. The work function of the 2D materials is measured differently depending on the measurement environment due to its high surface energy. We assumed that the Fermi levels of the MoS2 and BP are 4.53 and 4.5 eV, respectively [26,27]. After the stacking, the MoS2/BP heterojunction forms a staggered (type II) band alignment at an equilibrium state, with a very small barrier on the conduction band side, as shown in Figure 2d. Theoretically, the

effective band gap, which is the difference between the conduction band of MoS2 and the valence band of BP, formed at the MoS2/BP junction is 0.29 eV. The effective band gap is modulated by the drain bias during the diode type operation.

Even though the doping profile of a MoS2/BP junction is similar to a P-N junction, the carrier conduction mechanisms are quite different. When a negative drain bias is applied, the Fermi level of the MoS2 shifts upward (forward bias for a P-N junction) and the majority carriers from MoS2 flow into the BP layer; however, the holes in the BP layer cannot be transferred to the MoS2 layer, due to the high barrier height. As a result, the recombination of electrons and holes at the BP side generates a depletion region, which is balanced by the electron influx and the resistance increase, due to the depletion width increase. Hence, an almost constant current of ~10 nA is maintained in our device. In the case of a silicon P-N junction, the current increases exponentially at forward bias.

On the other hand, when the drain bias is positive (reverse bias for a P-N junction), minority carriers from the BP and MoS2 layers start to flow to opposite sides, driven by the electric field. Moreover, depending on the drain bias, the tunneling component may also contribute to the drain current. The current flow, shown in Figure 2b, saturates at a high drain bias because the current flow is limited by the minority carrier supply. Unlike a P-N junction, where the diffusion of the majority carriers is the primary conduction mechanism, the drift of minority carriers is the primary conduction mechanism in this bias region. Many prior studies have correctly noted this difference; however, in our opinion, they did not carefully consider the off-current mechanism [13,14,16,18,19]. Most of prior works explained that the off state is due to the band shift closing the direct tunneling window, but they did not consider that the gate bias region—causing extremely low off current—did not match the gate bias region of the direct tunneling current.

Figure 3a shows the transfer characteristics of a MoS2/BP FET with a small positive drain bias. In this case, the current level is already in the 10 to 100 nA range at VG = 0 V and VD = 500 mV, as shown in Figures 2b and 3a. Thus, to turn off this device, a strong negative gate bias should be applied. Many prior studies described band structures similar to Figure 2d to explain the off-state, where the tunneling current does not flow because the carriers in the BP valence band cannot be transferred to the MoS2 conduction band. However, as indicated in Figure 2c, the minority carriers from the BP can be injected into the MoS2 (and vice versa) when VG is approximately −3 V and the current level is approximately 10 nA. Thus, the reduction of the tunneling component cannot explain the turn off mechanism of our device; i.e., the prior explanation is obviously wrong.

**Figure 3.** Electrical properties and current flow mechanism of a MoS2/BP heterojunction FET at VD = 500 mV. (**a**) Transfer characteristics and transconductance (Gm). (**b**) Band diagrams showing the states at different gate bias regions.

Thus, the extremely low off current at strong negative gate bias needs to be explained with another mechanism. If we think about the carrier conduction at very negative VG, only the electrons from MoS2 can be drifted into the BP region and hole drift is blocked by the high barrier as shown in Figure 3b. Then, electrons injected into the BP region recombines holes and form a depletion region. As the

VG becomes more negative, the width of depletion region increases further and the drain current decreases rapidly, until the hole diffusion current starts to increase at VG < −4.5 V. Thus, in our opinion, the off-current of the MoS2/BP FET can be better explained with the formation of a depletion region in the BP layer.

If we explain the device operation from the negative VG side, it is easier to understand the operating mechanism. The drain current does not flow at −4.5 V because of the large depletion width. As VG increases to the positive bias side, the depletion region decreases, and suddenly, the minority carriers start drifting to other materials. Then, as the MoS2 energy band moves further downward, the tunneling current starts to flow at −3.2 V. Since the tunneling current is added to the drift current, due to the minority carrier injection, the drain current shows a hump at −3.2 V in our device. Most tunnel FETs reported in the literature show this kind of hump in the transfer characteristics, confirming our model.

The transfer characteristics measured at different drain biases and temperatures support our operation mechanism model further. When VD is small, the drift current decreases, but the turn-on behavior is not strongly affected because it is more closely related to how the depletion region is formed by the initial band alignment at the BP and MoS2 interface. To detect the bias where the BTBT current starts to contribute, the second derivative of the transfer curves is calculated, as shown in Figure 4a. The starting point of the abrupt curvature change marked with an arrow indicates the point of the BTBT current initiation, and the peak position indicates the maximum tunneling current. As the drain bias increases, the band alignment approaches the state shown in Figure 2f. Thus, a higher negative VG should be applied to turn off the tunneling current by pushing the MoS2 energy band upward. The temperature dependence also shows an interesting characteristic. The position of tunneling current initiation has not changed significantly, but the peak height increased, indicating that the BTBT current increased due to the increase of thermally activated carriers in the valence band of BP. The drift current increase can be attributed to the increased minority carrier density at the higher temperature.

**Figure 4.** (**a**) Transfer characteristics of a MoS2/BP heterojunction FET for different drain voltages (50 mV, 100 mV, and 500 mV). Normalized second derivative of transfer curves are shown to note the initiation points of band-to-band tunneling (BTBT). (**b**) Temperature-dependent transfer characteristic at 273 and 300 K, VD = 50 mV. (**c**) Subthreshold swing (SS) versus drain current at VD = 500 mV, 300 K.

Finally, we would like to emphasize that our device also shows a sub-60 mV/dec subthreshold swing in some regions of the transfer curve, as shown in Figure 4c. In previous reports (Table 1), swing values below 60 mV/dec often suggest a tunneling mechanism [12,13,16–19]. However, 60 mV/dec is the limit set by the diffusion mechanism. Since we have proposed that the turn-on behavior of our device is governed by the formation of the depletion region at the MoS2/BP interface, the turn-on mechanism is closer to a junction FET, where the drain current starts to flow once a small current path is formed by the reduction of the depletion width. Thus, the swing that is smaller than 60 mV/dec is more closely related to geometric factors and the carrier profile at the BP region, which affect the shape of depletion region.


**Table 1.** Comparison of the performance of the 2D/2D tunneling FETs reported in the literature.

#### **4. Conclusions**

In conclusion, we demonstrated the MoS2/BP heterojunction FET and analyzed the device operation mechanism. We found that the BTBT is not the primary mechanism determining the on-off characteristics of the MoS2/BP heterojunction FET, but it contributes to the formation of the hump in the transfer curve. In addition, the rapid turn-on and extremely low off-current are explained by the depletion region formation. Our results can be applied to general 2D/2D heterojunction devices.

**Author Contributions:** S.K.L. designed and conducted the experiments and S.C.K. supported the electrical measurement and analysis. T.J.Y. set up the experiment system. S.K.L. and H.J.H supported the process of experiments and the analysis of data. B.H.L. supported and guided the experiment and the results. B.H.L. conceived and advised the publication of the paper.

**Funding:** This work was partially supported by the Nano Materials Technology Development Program (2016M3A7B4909942) and by the Creative Materials Discovery Program of the Creative Multilevel Research Center (2015M3D1A1068062) through the National Research Foundation (NRF) of Korea funded by the Ministry of Science and ICT.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Communication* **Exfoliation and Characterization of V2Se9 Atomic Crystals**

**Bum Jun Kim 1,†, Byung Joo Jeong 2,†, Seungbae OH 2, Sudong Chae 2, Kyung Hwan Choi 1, Tuqeer Nasir <sup>1</sup> , Sang Hoon Lee 2, Kwan-Woo Kim 2, Hyung Kyu Lim 2, Ik Jun Choi 2, Ji-Yun Moon 3, Hak Ki Yu 3, Jae-Hyun Lee 3,\* and Jae-Young Choi 1,2,\***


Received: 31 August 2018; Accepted: 18 September 2018; Published: 18 September 2018

**Abstract:** Mass production of one-dimensional, V2Se9 crystals, was successfully synthesized using the solid-state reaction of vanadium and selenium. Through the mechanical exfoliation method, the bulk V2Se9 crystal was easily separated to nanoribbon structure and we have confirmed that as-grown V2Se9 crystals consist of innumerable single V2Se9 chains linked by van der Waals interaction. The exfoliated V2Se9 flakes can be controlled thickness by the repeated-peeling method. In addition, atomic thick nanoribbon structure of V2Se9 was also obtained on a 300 nm SiO2/Si substrate. Scanning Kelvin probe microscopy analysis was used to explore the variation of work function depending on the thickness of V2Se9 flakes. We believe that these observations will be of great help in selecting suitable metal contacts for V2Se9 and that a V2Se9 crystal is expected to have an important role in future nano-electronic devices.

**Keywords:** V2Se9; atomic crystal; mechanical exfoliation; scanning Kelvin probe microscopy

#### **1. Introduction**

To overcome the high-density integration of electronic technology, which faces physical limitations (e.g., fabrication process and reduction in charge carrier mobility), researchers have been intensively trying to develop a new device architecture or novel materials [1–4]. A range of diverse candidate materials have been proposed since the 2000s. Among them, graphene, which is a single layer of carbon atoms arranged in a hexagonal lattice, is considered to be a promising solution for future electronic devices because of its superior physical properties such as high carrier mobility and excellent chemical stability; however, it has the fatal disadvantage in that it has difficulty forming a band gap [1,2,5–7]. Therefore, the development of applications for graphene-based electronic devices, the most promising field, does not meet public' expectation yet. Graphene nanoribbons (GNRs) are presented as the effective way to open the bandgap of graphene but it is difficult to produce a uniform width in large area [6,7]. In addition, the transport behavior of GNRs and newly introduced two-dimensional (2D) materials (e.g., transition metal dichalcogenides (TMDCs) and black phosphorous), with appropriate bandgaps, are reduced dramatically because of dangling bonds at the side edges and domain boundaries [8–11]. Unfortunately, most of the studies of the

2D material-based electronic devices thus far contain an etching process to define the conducting channel. Thus, the discovery of one-dimensional (1D) nanomaterials, which are free from edge and grain boundary scattering, is a key solution in the development of nano-electronic device.

Carbon nanotubes (CNTs), which exhibit high carrier mobility, ultimate mechanical strength, and chemical stability, have been considered as representative building blocks for next-generation transistors, chemical sensors, and nanocomposites [12–14]. However, the wide range of electronic structures that arise from the different chirality of the CNTs curtails the reliability of the manufacturing process of the nano-electronic devices [15]. Therefore, separation of single-chirality CNTs from the bulk CNTs or control of the chirality during the growth of the CNTs is required. Recently, studies on the synthesis and application of a new family of 1D nanomaterials in the form of three-dimensional (3D) bundles of numerous single-molecular chains coupled by weak van der Waals interactions have been reported [16–21]. For example, extensive studies on bulk synthesis and atomic-scale dispersion of the bio-compatible Mo6S9−xIx have been reported [21–23]. In addition, Sb2S3 was developed as an optoelectronic device by effectively reducing exciton decay due to the absence of dangling bonds [24]. Moreover, VS4 was utilized for an electrochemical energy storage device by using the van der Waals gap between the chains [25,26]. However, in the majority of studies on these materials, they have been utilized only as a thin-film structure, although the benefits of the layered characteristics can be exploited. In addition, the crystal structure of Mo6S9−xIx is not well defined because the position of the sulfur and iodine atoms bridged to the molybdenum atoms may vary even for the same stoichiometric composition.

In this study, we succeeded in mass producing 1D semiconductor V2Se9 crystals via a simple transport method. Through the mechanical exfoliation method, we confirmed that as-grown V2Se9 crystals consist of innumerable single V2Se9 chains linked via the van der Waals interaction, like graphite. In addition, a nanoribbons structure of V2Se9 which is capable of thickness control was obtained through repetitive mechanical exfoliation of the V2Se9 crystals. Lastly, the change in work function according to the thickness change of the V2Se9 flakes was analyzed by scanning Kelvin probe microscopy (SKPM) measurement.

#### **2. Materials and Methods**

**Synthesis:** V2Se9 was synthesized using V (Powder, −325 mesh, 99.5%, Sigma-Aldrich, St. Louis, MO, USA) and Se (powder, 99+%, Alfa Aesar, Haverhill, MA, USA). The mixture of V (0.2038 g) and Se (1.4213 or 1.9898 g) with a V to Se ratio of 2:9 or 2:12.6 was pelletized and then sealed in a 10 cm-long evacuated quartz tube. The quartz ampoule was heated for 120 h at a temperature of 300–400 ◦C (at 5.5 ◦C/h) and then cooled (at 10 ◦C/h). The resulting material was a dark gray sintered powder. The unreacted Se was sublimated by heat treatment in a tube furnace at 250 ◦C under Ar atmosphere for 24 h.

**Mechanical exfoliation:** The bulk V2Se9 was placed on wafer dicing tape (BT150EKL, Nitto Denko, Umeda, Osaka, Japan) and the materials were stuck several times to yield thinner-than-bulk materials. A substrate (300 nm SiO2/Si or bare Si) was cleaned by ultrasonication in acetone, ethanol, and DI water for 15 min, followed by heating at 100 ◦C in order to remove the moisture from the substrate. The polymer tape was adhered strongly to and pressed against the substrate. After adhesion, the polymer tape was removed from the substrate; this process was repeated for exfoliation.

**Characterization:** Powder X-ray diffraction (Mac Science, M18XHF22, Tokyo, Japan) was performed using Cu-Kα radiation (λ = 0.154 nm). Field emission-scanning electron microscopy (FE-SEM, Hitachi, S4300SE, Chiyoda, Tokyo, Japan) was operated at an acceleration voltage of 15 kV. Atomic force microscopy (AFM, Park systems, NX 10, Suwon, South Korea) was performed in a non-contact mode for the topographic analysis of the mechanically exfoliated V2Se9 on 300 nm Si/SiO2. The surface potentials of V2Se9 on Si substrate were measured by SKPM (Park systems, NX10, Suwon,

South Korea) measurement using Si tips coated with Cr-Pt (Multi75-G, Budget Sensors Inc., 1113 Sofia, Bulgaria) with resonance frequencies of 75 kHz, a scan rate of 0.3 Hz, and sample bias of ±1 V.

#### **3. Results and Discussion**

Since the transition metal vanadium has the outermost 3d orbital, it can produce various forms of compounds (e.g., V5Se4 to V2Se9) through a chemical reaction with selenium (see the phase diagram in Figure S1). Therefore, to synthesize V2Se9 crystals with a high-purity and high-crystallinity, the ratio of V:Se and the synthesis temperature should be considered carefully. For example, if the atomic mixing ratio of V and Se powder is adjusted precisely to 2:9 to synthesize V2Se9 crystals, unpredictable fluctuation occurs in the synthetic tube and VSe2, which is an undesirable impurity, is formed. We corrected these parameters experimentally, and as a result obtained pure V2Se9 crystals with an exact stoichiometry ratio of 2:9 by adding them in excess of Se, as shown in Figure 1a (V:Se atomic mixing ratio of 2:12.6). The crystallinity of the bulk V2Se9 crystal was verified by the X-ray diffraction (XRD) pattern (JCPDS 01-077-1589) (Figure 1b). The SEM images in Figure 1c,d clearly shows the 1D nanowire structures and the gaps generated during transfer of the sample onto the Si substrate.

**Figure 1.** (**a**) Photo-image of mass production of V2Se9 crystal. (**b**) XRD pattern of V2Se9 crystal. (**c**) Low- and (**d**) high-magnification SEM images of V2Se9 crystal. The inset shows an illustration of the crystal structure of V2Se9.

To investigate the structural characteristics of nanoscale V2Se9, the bulk V2Se9 crystal was mechanically exfoliated using the well-known tape method [1]. Although each single V2Se9 chains are linked by weak van der Waals interaction, we obtained a thin V2Se9 nanoribbon on a 300 nm SiO2/Si substrate (see in Figure 2). Unlike typical 2D materials, an exfoliated V2Se9 nanoribbon shows a rough surface.

**Figure 2.** (**a**) Atomic force microscopy (AFM) image of the 1D V2Se9 flake on 300 nm SiO2/Si substrate. (**b**) Line-profile of a V2Se9 flake as marked in Figure 2a.

We attempted a further delamination at the sample position using the tape, and found that some of them had been torn out (black dotted line) and that the thickness decreased from 90 to 20 nm (L1 to L1 ), and from 31 to 2 nm (L2 to L2 ) (see Figure 3).

**Figure 3.** (**a**) AFM image of exfoliated V2Se9 on 300 nm SiO2/Si substrate. (**b**) AFM image of additionally exfoliated V2Se9 on 300 nm SiO2/Si substrate. (**c**) Line-profile of 1D V2Se9 flakes on 300 nm SiO2/Si substrate before and after 2nd exfoliation.

Figure 4a shows the AFM image of an isolated V2Se9 nanoribbon on the 300 nm SiO2/Si substrate. The nanoribbon has an atomic scale thickness and a width of approximately 20 nm (Figure 4b). Since V2Se9 has a bundle structure in which single chains are bonded by van der Waals forces,

we expect that V2Se9 nanoribbons may exhibit ideal transport characteristics without degradation due to edge scattering.

**Figure 4.** (**a**) AFM image of the V2Se9 nanoribbon on the 300 nm SiO2/Si substrate. The inset shows an illustration of the V2Se9 nanoribbon. (**b**) Line-profiles of the V2Se9 nanoribbon as marked L1, L2, L3, and L4 in Figure 4a.

To investigate the electrical properties of V2Se9 flakes with a different number of layers, we performed an SKPM analysis, which is a non-destructive analytical tool that can investigate the local surface potential energy and work function by measuring the contact potential difference between the tip and the sample (*VCPD*) [27,28]. Because the V2Se9 nanoribbons were on the bare Si substrate, the work function of V2Se9 flakes can be calculated using the following equation:

$$V\_{CPD} = \frac{1}{\varepsilon} \left(\varphi\_t - \varphi\_f\right)\_{\prime} \tag{1}$$

$$\begin{split} \Delta V\_{\text{CPD}} &= V\_{\text{CPD}}(V\_2 \text{Se}\_9) - V\_{\text{CPD}}(\text{substate}) \\ &= \frac{1}{\varepsilon} \left( \boldsymbol{\varrho}\_t - \boldsymbol{\varrho}\_f \right) - \frac{1}{\varepsilon} (\boldsymbol{\varrho}\_t - \boldsymbol{\varrho}\_s) \\ &= \frac{1}{\varepsilon} \left( \boldsymbol{\varrho}\_s - \boldsymbol{\varrho}\_f \right) \end{split} \tag{2}$$

where *ϕt*, *ϕ<sup>s</sup>* and *ϕ<sup>f</sup>* represent the work functions of the tip, Si substrate, and V2Se9 flake, respectively.

As shown in Figure 5a, the surface potential energy varies with the thickness of the V2Se9 flakes. For example, the surface potential energy differences (Δ potential energy) between the V2Se9 flakes with thicknesses of 5 and 40 nm and the Si substrate were 38 and 60 mV, respectively (Figure 5b,c). A Statistical analysis of more than 27 samples shows that as the thickness of the V2Se9 flake is less than 25 nm, the surface potential energy difference and the work function become to decrease simultaneously (Figure 5d,e). These phenomena can be explained using an interlayer screening effect, which is also observed in typical 2D materials [27–29]. In general, the native Si oxide (e.g., SiOx), which forms naturally on the surface of the Si wafer, has a hydrophilic property, which caused a large number of charge-trapping sites owing to the moisture in the air. Therefore, it affected the charge transfer between the V2Se9 flakes and the Si substrate [27]. Since the effective area of the interlayer screening effects increases with decreasing flake thickness, the surface potential difference and the work function of 25 nm thick V2Se9 flakes decreased from that of the bulk V2Se9 (See in Figure S2).

**Figure 5.** (**a**) Scanning Kelvin probe microscopy (SKPM) image of exfoliated 1D V2Se9 flakes on the Si substrate. (**b**,**c**) Height and potential energy profiles of the V2Se9 flakes and Si substrate as labeled in Figure 5a. (**d**,**e**) Variation in potential energy difference and work function depending as a function of thickness of V2Se9 flakes.

#### **4. Conclusions**

In conclusion, the mass production of the high-purity and high-crystalline 1D material V2Se9 crystals was successfully demonstrated using the solid-state reaction of V and Se. Through the mechanical exfoliation method, we confirmed that as-grown V2Se9 crystals consist of innumerable covalently bonded V2Se9 chains linked by the van der Waals interaction. In addition, atomic nanoribbons structures of V2Se9 was obtained on the 300 nm SiO2/Si substrate. We used SKPM analysis to investigate the electrical characteristics of V2Se9 and established that the work function decreased with decreasing thickness of the V2Se9 flakes owing to the interlayer screening effect. These results will be of great help in selecting suitable metal contacts for V2Se9; these will have a significant influence on the overall performance. We believe that the 1D semiconductor V2Se9 crystal is expected to be a new family of 2D materials that will be considered essential in future device applications.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2079-4991/8/9/737/s1, Figure S1: Phase diagram of V-Se binary system, Figure S2: SKPM image of exfoliated 1D V2Se9 flake.

**Author Contributions:** J.-Y.C. designed the experiments, and B.J.K. and B.J.J. supported the elemental analysis. S.O. and S.C. performed the chemical reaction experiments and K.H.C., T.N. and S.H.L. support the chemical reaction experiment and K.-W.K., H.K.L., I.J.C. and J.-Y.M. supported structural analysis. H.K.Y., J.-H.L. and J.-Y.C. conceived and supervised this study and provided intellectual and technical guidance.

**Funding:** This work was supported by the Technology Innovation Program (or Industrial Strategic Technology Development Program) (10063400, Development of Growth and Transfer Technology for Defectless <sup>350</sup> <sup>×</sup> 350 mm2 Single Crystalline Graphene) funded By the Ministry of Trade, Industry and Energy (MOTIE, Korea). J.H.L. acknowledges support from the Presidential Postdoctoral Fellowship Program of the National Research Foundation in Korea (2014R1A6A3A04058169).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Feasible Route for a Large Area Few-Layer MoS2 with Magnetron Sputtering**

**Wei Zhong 1,† ID , Sunbin Deng 2,† ID , Kai Wang 2, Guijun Li 2, Guoyuan Li 1, Rongsheng Chen 1,2,\* and Hoi-Sing Kwok <sup>2</sup>**


Received: 20 June 2018; Accepted: 30 July 2018; Published: 3 August 2018

**Abstract:** In this article, we report continuous and large-area molybdenum disulfide (MoS2) growth on a SiO2/Si substrate by radio frequency magnetron sputtering (RFMS) combined with sulfurization. The MoS2 film was synthesized using a two-step method. In the first step, a thin MoS2 film was deposited by radio frequency (RF) magnetron sputtering at 400 ○C with different sputtering powers. Following, the as-sputtered MoS2 film was further subjected to the sulfurization process at 600 ○C for 60 min. Sputtering combined with sulfurization is a viable route for large-area few-layer MoS2 by controlling the radio-frequency magnetron sputtering power. A relatively simple growth strategy is demonstrated here that simultaneously enhances thin film quality physically and chemically. Few-layers of MoS2 are established using Raman spectroscopy, X-ray diffractometer, high-resolution field emission transmission electron microscope, and X-ray photoelectron spectroscopy measurements. Spectroscopic and microscopic results reveal that these MoS2 layers are of low disorder and well crystallized. Moreover, high quality few-layered MoS2 on a large-area can be achieved by controlling the radio-frequency magnetron sputtering power.

**Keywords:** few-layer MoS2; magnetron sputtering; magnetron sputtering power; raman spectroscopy; disorder

#### **1. Introduction**

The emergence of monolayer graphene [1,2] and transition metal dichalcogenides (TMDs) [3,4] has inspired a series of high-profile discoveries in the electronic and optoelectronic fields [5–7], and has initiated potentially new areas [8,9]. Thus, two-dimensional (2D) materials have recently been intensively studied. In the context of 2D TMDs, molybdenum disulfide (MoS2) is one of the attractive embodiments due to its stable form in few- and single-layers [10,11] as well as its desirable electrical and optical properties [12]. Few- and single-layer MoS2 is firstly obtained via top-down mechanical stripping techniques [13], which are commonly used for graphene exfoliation. Although this method of exfoliation and dip coating [14–17] has its own the advantages to achieve high-quality 2D materials, none of these are proper solutions for radical large-scale commercial manufacturing, where the mass-producible growth of large-area, continuous, and high-quality 2D MoS2 thin films on dielectrics is a pre-requisite. From this point of view, the bottom-up strategies for thin film growth, including chemical vapor deposition (CVD) and physical vapor deposition (PVD), are better choices when compared with the top-down methods mentioned above. CVD methods have already been successfully demonstrated [18–20], but the control of thin film thickness, purity and uniformity on a large scale needs to be further enhanced [21]. On the other hand, PVD methods, especially magnetron sputtering, have been broadly employed in large-scale commercial manufacturing at low cost and with easy control. However, the exploration of 2D MoS2 thin film growth using magnetron sputtering technology is quite insufficient [21]. In recent years, there have been few attempts at sputtering techniques for the growth of MOS2 thin films. Muratore et al. [22], Kaindl et al. [23], and Samassekou et al. [24] reported the synthesis of continuous few-layer MoS2 by sputtering method using a MoS2 target, and the sputtered MoS2 films was annealed in an argon atmosphere. Tao et al. [21] and Santoni et al. [25] reported MoS2 film using Mo target sputtered in vaporized sulfur ambient. However, the reported films either are relatively thick or have poor crystal quality and optical properties [21–25]. The main obstacles for large-area sputtered high quality MoS2 films are possibly located in the difficulty of disorder control during thin film deposition, and the lack of metrics to evaluate the deposited thin films.

Studies over the past decade have shown that Raman spectroscopy has historically played an important role in the structural characterization of graphitic materials [26–28] and has also become a powerful tool to understand the effects of process parameters on the production of high quality graphene by monitoring changes in disordered peaks [28–32]. Recently, Raman spectroscopy has also been used to study the effects of disorder on the MoS2 [33]. On the other hand, since many scholars have previously investigated the crystallinity of MoS2 thin film by high-temperature vulcanization or deposition on different substrates [21,34,35], few studies have investigated the effect of sputtering power on the MoS2 thin film. Therefore, in this work, we use a Raman spectroscopy approach to describe the quality and to study the effect of RF power on the deposition of the large-scale few-layer MoS2 films. We also investigated the crystalline structure of the films through an X-ray diffractometer (XRD) and a high-resolution field emission transmission electron microscope (HRTEM). The binding energies of Mo in the MoS2 film grown by radio frequency magnetron sputtering (RFMS) were further analyzed by X-ray photoelectron spectroscopy (XPS). For precise detection, the surface 5-nm-thick thin films were etched using Ar ions before XPS characterization.

#### **2. Materials and Methods**

The large-scale few layer MoS2 thin films were deposited using the RFMS technique. Firstly, both silicon substrates coated with thermally grown SiO2 and glass substrates (Eagle 2000, Corning) were ultrasonically cleaned in acetone and then isopropanol (IPA). After being rinsed in deionized (DI) water and dried, the substrates were loaded into the chamber of an RFMS system (AJA International Inc., Scituate, MA, USA) and heated to 400 ○C. The distance between the substrate and the MoS2 target (99.99%, 2 Inc., Plasmaterials Inc., Livermore, CA, USA) was 10 cm. In order to suppress the influence of oxygen and moisture on the deposited thin films, the base pressure of the chamber should be pumped down as low as possible. In this work, the value was 2.67 <sup>×</sup> <sup>10</sup>−<sup>5</sup> Pa. During deposition, Ar gas (99.999%) was allowed to flow into the chamber with a stable flow rate of 20 sccm, and the working pressure of the chamber was maintained at 3 <sup>×</sup> <sup>10</sup>−<sup>3</sup> Torr. The RF sputtering power applied to the MoS2 target was varied from 10 W to 150 W in order to investigate the relationship between the large-scale few layer MoS2 growth and RF power. The deposition time was dependent on the required thickness of the thin films and the thickness of the film is maintained at 15 nm, the thickness of the thin films is confirmed by HRTEM (High-Resolution Transmission Electron Microscopy). When the RFMS process was completed, the heater and the Ar gas pipeline were switched off, and the substrates were cooled down to room temperature naturally. After loading out of the chamber, the SiO2/Si substrates and glass substrates with as-deposited MoS2 thin film were immediately subjected to post-annealing in a sulfurization atmosphere for 60 min at 600 ○C to enhance thin film quality physically and chemically (Figure 1).

In order to identify the layered structure and obtain the disorder information, the deposited MoS2 thin films were analyzed using Raman spectroscopy (Renishaw invia RE04, 514 nm Ar laser with a 1 μm spot size, Renishaw plc, Gloucestershire, United Kingdom). For the crystalline phase characterizations of the thin films, an X-ray diffractometer (XRD, Empyrean, PANalytical, Almelo, The Netherlands) with Cu Kα radiation was used. It was operated in thin-film mode, and the angle between the X-ray and the thin film surface was fixed at 0.5 degrees. Furthermore, a high-resolution transmission electron microscopy (HRTEM, JEM-2010HR, JEOL, Tokyo, Japan) was applied to identify the number of layers and atomic structure of the MoS2 film. To analyze the material composition of the deposited thin films as well as the chemical environment of the atoms in the thin films, X-ray photoelectron spectroscopy (XPS) measurement was conducted on the Physical Electronics 5600 multi-technique system (Physical Electronics Inc., Chanhassen, MN, USA).

#### **3. Results and Discussion**

Figure 1 shows an image of MoS2 thin layer grown on glass substrates with different deposition time (15 s, 30 s, and 45 s). The deposited MoS2 layer was light gray, and after annealing under a sulfur atmosphere, the MoS2 layer was pale yellow and was found to have specular reflection of ambient light. This result is consistent with previous reports [21]. The size of the synthesized films is limited by the dimensions of our sample heating holder.

**Figure 1.** This post-deposition annealing treatment was performed to further enhance crystalline quality in as-sputtered MoS2 on glass substrates under sulfur environment.

Figure 2 shows the Raman spectra of the MoS2 thin films on SiO2/Si substrates deposited under different RF powers of 10 W, 80 W, 120 W and 150 W, respectively. From Figure 2, it can be seen that all of the thin films exhibit two specific Raman characteristic peaks of MoS2, namely the in-plane (*E*<sup>1</sup> 2*g*) peak at ~381 cm-1 and the out-of-plane (A1g) peak at ~405 cm<sup>−</sup>1. Their peak positions herein are quite consistent with the results in other reports [24,36,37] , which the *E*<sup>1</sup> <sup>2</sup>*<sup>g</sup>* and A1g mode peaks appear at ~381.5 cm−<sup>1</sup> and ~404.8 cm<sup>−</sup>1. Besides this, a relatively small LA(M) peak (a defect peak generated by LA phonons at the M point in the Brillouin zone) is also observed at ~227 cm<sup>−</sup>1. Theoretically, the *E*1 <sup>2</sup>*<sup>g</sup>* mode corresponds to the S and Mo atoms oscillating in antiphase parallel to the crystal plane, while the A1g mode corresponds to the S atoms oscillating in antiphase out-of-plane, as shown in the insets of Figure 2 [29,38]. In addition to the absolute positions of these two peaks, the frequency difference (Δk) between the A1g peak and the *E*<sup>1</sup> <sup>2</sup>*<sup>g</sup>* peak is a good indicator of the layer number in MoS2 thin films. In this work, Δk is around 24 cm-1, which is larger than that in monolayer MoS2 thin films (~18–19 cm<sup>−</sup>1) [19,33,39], but smaller than the typical value in bulk MoS2 (~26 cm<sup>−</sup>1) [37,38]. This indicates the existence of few-layer MoS2 [36]. Table 1 lists the peak positions corresponding to the *E*1 <sup>2</sup>*<sup>g</sup>* and A1g mode as well as the Δk values for all samples under various RF sputtering powers in this work. With the continuous increase of RF power from 10 W to 150 W, the Δk values remain in the vicinity of 24 cm<sup>−</sup>1. This means that all of the thin films under different RF powers are MoS2 with a

few layers [36,37]. Another indicator of film quality is the full width at half maxima (FWHM) of the observed vibration modes. FMHM values for the sputtered FL-MoS2 film with different RF powers is compared in Table 1. In general, higher FMHM values mean more disorder [24,25]. From the Table 1, it can be seen that the MoS2 films with RF powers of 120 W has the lowest FMHM values, meaning it has the least disorder. It also can be seen that the A1g peak and the *E*<sup>1</sup> <sup>2</sup>*<sup>g</sup>* peak all have a blue shift as the RF powers increased from 10 W to 120 W and a red shift as the RF powers increased from 120 W to 150 W. The red shift is attributed to the high RF power, resulting in an increase in the residual-stress of the film [40].While the blue shift is attributed to O2-doping of MoS2, which will be shown on the XPS result [41].

**Figure 2.** Raman spectra of the MoS2 thin films deposited on SiO2/Si substrates under different radio frequency (RF) powers. The insets illustrate the oscillating mode of the *E*<sup>1</sup> <sup>2</sup>*<sup>g</sup>* and A1g peak.

According to Raman fundamental selection rules, only phonons with wave vector q ≌ 0 are Raman active around the center of the Brillouin zone. However, this rule will be broken by defects, which cause the appearance of peaks away from the zone center [33,42]. Monitoring the evolution of disorder-related sub-peaks in the Raman spectrum enables us to understand the effects of process parameters and has allowed great strides to be made in the CVD preparation of high quality graphene [29,43,44]. Similarly, in Figure 2, apart from two major peaks with regard to the basic *E*1 <sup>2</sup>*<sup>g</sup>* and A1g vibration modes, several sub-peaks related to the defects can also be observed. Among them, the sub-peak at 227 cm−<sup>1</sup> is the most intense, which is attributed to the longitudinal phonon branch at the point M (LA (M)) of the Brillouin region [33,45]. Therefore, this sub-peak is able to form a very clear marker for disorder in the system, especially for the few- and single-layer MoS2. The intensity ratio of the LA(M) sub-peak to A1g peak is plotted as a function of RF power in Figure 3. It can be clearly observed that the intensity ratio reaches a minimum when the RF power climbs to 120 W, indicating the improvement of thin film quality [46]. In general, higher RF power could assist with the formation of crystalline films with lower disorder (namely, lower LA(M)/ A1g intensity ratio), but excessively high RF power is not welcomed. For instance, the LA(M)/A1g intensity ratio rises when the RF power increases from 120 W to 150 W. It reveals the increase of disorder in the deposited MoS2 thin films. A possible explanation for this phenomenon could be the increased defect generation caused by ion bombardment under over-high RF power.


**Table 1.** The *E*<sup>1</sup> <sup>2</sup>*g*- and A1g-related Raman peak information of MoS2 thin films deposited using radio frequency magnetron sputtering (RFMS) under various RF powers.

**Figure 3.** LA(M) to A1g peak intensity ratio of the MoS2 films with different RF powers deposited on SiO2/Si substrates.

The crystalline information of the samples was characterized using XRD. The X-ray diffraction patterns of MoS2 thin films on SiO2/Si substrates under different RF powers are shown in Figure 4. The MoS2 thin films deposited under different RF powers all exhibit three obvious diffraction peaks, which are located at 14.0○, 21.6○, and 51.5○ respectively. The two weak diffraction peaks at 21.6○ and 51.5○ correspond to the SiO2 (JCPDS: 27-0605) (111) and (400) planes, respectively. The strong diffraction peak at 14.0○ is an indicator of the MoS2 (JCPDS: 37-1492) (002) plane. For the MoS2 thin films deposited under an RF power of 10 W, the broad and weak diffraction peak at 14.0○ indicates the amorphous structure of the film. However, when the RF power increases to 80 W, 120 W and then 150 W, all of the MoS2 thin films exhibit a strong and narrow diffraction peak at 14.0○, which indicate the well crystallization of MoS2 thin films. Since the (002) plane is parallel to the surface of the substrates, the deposited MoS2 thin films using the RFMS technique under an RF power of 80 W, 120 W and 150 W could grow along the c-axis. Meanwhile, the exclusive diffraction peak also suggests the thin films are highly oriented. These properties are quite helpful for the formation of stacked microstructures in MoS2 thin films. Besides this, it should be noted that the intensity of the (002) peak for MoS2 thin films under an RF power of 150 W is lower compared to the other two samples. This is possibly related to the degradation of crystalline quality under such high RF powers [47]. At the same time, we also noticed that under the RF power of 150 W, the sample showed a weak diffraction peak at 28.4○, which corresponds to the Si (JCPDS: 27-1402) (100) plane.

**Figure 4.** X-ray diffraction patterns of the MoS2 thin films on SiO2/Si substrates under different RF sputtering powers.

To further elucidate the crystalline structure, the MoS2 thin film deposited by 120 W RF power was transferred onto a lacey copper grid for HRTEM characterizations. Figure 5a presents its cross-sectional HRTEM image, it can be seen that there are about 20 layers of the MoS2 thin films on the SiO2/Si substrate, and the interlayer spacing (0.68 nm) of the MoS2 thin films is consistent with the previous results [21]. Moreover, it is also verified that the deposited MoS2 thin films are stacked a few layers in parallel, which is in agreement with the results extracted from the Raman spectra above. As far as the typical high resolution TEM image in Figure 5b is concerned, the first-order diffraction spots of the FFT image (inset of Figure 5b) on a selected area are shown according to a regular hexagonal symmetry, thus indicating the presence of the MoS2 thin film made of single crystal domains (without Moireé patterns). Moreover, HRTEM images of the selected area in Figure 5b after FFT filtering are shown in Figure 5c. The IFFT-filtered image (Figure 5c) shows a regular honeycomb pattern due to the atomic arrangements of the Mo and S atoms. In Figure 5d, the calculated profile along the selected direction in Figure 5c is drawn, which reveals the (004) plane of MoS2 with a lattice spacing of 0.31 nm that is observed in Figure 5c. Figure 5e is the magnified image of the square-surrounded region in Figure 5c. The periodic atom arrangement for Mo confirms that the MoS2 thin films deposited under an RF power of 120 W own a crystalline structure. This is also consistent with the XRD results in Figure 4.

**Figure 5.** (**a**) Cross-sectional high-resolution field emission transmission electron microscope (HRTEM) image of samples deposited under an RF power of 120 W. (**b**) High resolution TEM image of the MoS2 film deposited by 120 W RF power transferred onto a lacey carbon grid. The inset shows the fast Fourier transformation (FFT) image corresponding to the TEM image selected area of a portion of (**b**), showing the hexagonal symmetry of the MoS2 structure. (**c**) Inverse FFT images corresponding to the TEM image selected area of a portion of (**b**). (**d**) Atomic spacing along the selected direction of the basal plane. (**e**) Zoom-in image of the area highlighted in (**c**). The hexagonal structure formed by Mo atoms is indicated.

XPS was used to analyze the chemical environment of Mo in the MoS2 thin films, and to detect any impurity (particularly oxygen) involvement during preparation. The high-resolution XPS spectra of the Mo 3d and S 2s region are shown in Figure 6. Since the peaks of Mo 3d and S 2s are too close to distinguish, in order to obtain the chemical environment information of the Mo species, both the S 2s and Mo 3d spectra are taken into account. Moreover, since the main Mo doublet peak signals of the thin films under an RF power of 10 W are so weak that they reach noise level, the analysis in terms of such samples is not reliable. Hence, only the data with an RF power of 80 W, 120 W and 150 W have been fitted by a 20% Lorentzian–Gaussian ration fit and the according results together with the Shirley background are presented in Figure 6b–d. According to [35], the two peaks at 229.1 eV and 232.2 eV for the MoS2 thin film are attributed to the doublet Mo 3d5/2 and Mo 3d3/2 orbitals, respectively. Meanwhile, the fitting shows that there is a second characteristic at the lower binding

energy. For species originating from lower binding energy (BE) Mo3d peaks, the lower BE Mo species is Mo still associated with the Mo-S lattice, or it reflects a single amorphous MoSx phase, where Mo has a different number of nearest neighbor S atoms [48,49]. According to [25], the lower BE Mo species can be assigned to zero-valent Mo (Mo(0)) occurring in small aggregates dispersed in a MoS2 matrix. From the principle of sputter coating, we can know that the presence of Mo(0) cannot be avoided, and the purpose of annealing under a sulfur atmosphere is to convert Mo(0) to Mo(IV), which can reduce disorder and defect formation. However, when the film is etched by an ion beam, there is also a chemical shift of its binding energy toward smaller values [25,35]. Therefore, compared to XPS characterization, the advantages of strong non-destructive characterization of Raman spectra are even more pronounced. For accurate detection, 5-nano-thin-thick-surface films should be etched with Ar ions prior to XPS characterization. However, this will lead to sample destruction and will have a certain impact on the test results. Although this changed the chemical state of the Mo atom, all the samples were processed in the same way, and the overall trend of its variation with sputtering power did not change. The relative ratio of Mo species calculated from the deconvolved Mo 3d spectrum results is shown in Figure 7. The Mo(IV) fraction is defined as the area under the Mo(IV) peak divided by the total Mo peak area. It can clearly be seen that the Mo (IV) fraction reaches a maximum when the RF power is increased to 120 W, indicating the improvement of thin film quality. In addition, smaller peaks at 230.2 eV and 233.3 eV are assigned to Mo5+ [50], indicating the presence of chemisorbed oxygen at sulfur vacancies [51,52] or sub-stoichiometric oxides MoOx [53]. The origin of MoOx is probably due to the sulfurization process. We believe that MoOx has probably formed during the deposition and/or during the sulfurization process at 600 ○C by reaction with the oxygen diffused from the SiO2/Si substrate. As shown in Figure 7, the Mo (V) fraction is around 0.1, 0.094 and 0.102 for an RF power of 80 W, 120 W, and 150 W, respectively. For different RF powers, the Mo(V) fraction is basically the same, indicating that the power change has no effect on the Mo(V) fraction.

**Figure 6.** High-resolution X-ray photoelectron spectroscopy (XPS) spectra of (**a**) Mo 3d/S 2s, and (**b**), (**c**), and (**d**) Mo3d core-level spectra for an RF power of 80 W, 120 W, and 150 W, respectively. The background is shown with a black line at the bottom. The black dots represent the raw data. The red line is the total least-squares fit. The orange lines indicate the Mo(V) 3d components. The blue lines are the Mo3d components linked to Mo (0) and Mo (IV), the Mo (0) are the lower BE doublet. The olive lines are the S2s components. (See text for more explanation).

**Figure 7.** Relative ratio of Mo species with various chemical states at different RF powers.

#### **4. Conclusions**

Few-layer MoS2 thin film deposition on large-area thermally oxidized silicon substrates was demonstrated using the RFMS technique. Raman analysis verified the achievement of few-layer MoS2 thin films. Meanwhile, it was proposed that the disorder inside the thin films could be monitored using Raman spectra, and controlled by adjusting the RF sputtering power. Furthermore, the XRD spectra and cross-sectional TEM images confirmed the high quality of few-layer MoS2 thin films, implying that RFMS was suitable for layered MoS2 growth. Additionally, the XPS characterizations on RFMS grown few-layer MoS2 thin films revealed the RF power has a great effect on the binding energies of Mo atoms. Our work illustrates that sputtering combined with sulfurization is a viable route for the high quality of large-area few-layer MoS2 by controlling the radio-frequency magnetron sputtering power.

**Author Contributions:** S.D. and K.W. conceived and designed the experiments; S.D. and K.W. performed the experiments; W.Z. and R.C. analyzed the data; S.D., K.W., W.Z. and G.L. (Guijun Li). contributed materials/analysis tools; W.Z., S.D. and R.C. wrote the paper; G.L. (Guoyuan Li) and H.-S.K. provided advice about the content and the structure of this work.

**Funding:** This research was funded by the National Natural Science Foundation of China under Grant 61604057, in part by the Partner State Key Laboratory on Advanced Displays and Optoelectronics Technologies under Grant ITC-PSKL12EG02, in part by the Science and Technology Program of Guangdong Province under Grant 2017A010101010, and in part by the Science and Technology Program of Guangdong Province under Grant No. 201807010098.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
