Applied Science and Convergence Technology 2019; 28(5): 159-163
Published online September 30, 2019
Copyright © The Korean Vacuum Society.
Seon Kyeong Kanga,b and Hyun Seok Leea,b,*
aDepartment of Physics, Chungbuk National University, Cheongju 28644, Republic of Korea
bDepartment of Scientific Instrumentation and Development, Chungbuk National University, Cheongju 28644, Republic of Korea
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-CommercialLicense (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution,and reproduction in any medium, provided the original work is properly cited.
Two-dimensional (2D) materials have received a considerable amount of attention owing to their unique properties compared with those of bulk materials, such as ultralow thickness with no dangling bonds and direct bandgap for monolayers. Monolayer MoS2 with a direct bandgap of ~1.8 eV is representative of 2D semiconductors, which can facilitate a high on/off ratio in field-effect transistors and can have various optoelectronic applications in the visible range. Large area MoS2 monolayers are generally synthesized using chemical vapor deposition (CVD) methods, where various types of metal precursors are utilized. Here, we report the effects of process parameters on the growth mode of MoS2 monolayers prepared by CVD using solution-based metal precursors. We investigated various growth modes of MoS2 flakes depending on the growth temperature and time, Mo-precursor concentration in precursor solutions, and carrier gas amount. The synthesized MoS2 monolayers revealed typical n-type semiconductor behaviors, which were investigated by optical microscopy, confocal photoluminescence and Raman spectroscopy, atomic force microscopy, and I–V characterization of field-effect transistors. The n-doping effect was reduced by removing unreacted precursors after transferring the MoS2 layer on new SiO2/Si substrates.
Keywords: Growth mode, MoS2 monolayers, Chemical vapor deposition, Solution-based metal precursor, Transition metal dichalcogenides
Transition metal dichalcogenides (TMDs) have received attention as van der Waals layered semiconductors because of their unexplored physical properties at an atomic level thickness, compared to those of conventional bulk semiconductors. As a representative of TMDs, MoS2 monolayers have been intensively studied with regard to their optical and electrical properties. A direct bandgap in the visible range allows for potential applications in several areas such as nanoelectronics, nanophotonics, valleytronics, and soft-electronics [1–6]. For such applications, TMDs synthesized via chemical vapor deposition (CVD) methods are widely used [2,7,8]. Sulfurization after spin-casting a metal precursor solution on substrates is one of the promising CVD methods [9–11]. Although the growth mode of monolayers is strongly correlated to various process parameters, the effects of such parameters and processes for their optimization have not yet been reported. In this study, we investigated different growth modes of MoS2 monolayers synthesized by CVD as functions of the growth temperature and time, solution concentration for metal precursor, and carrier gas amount, where each growth mode was monitored by optical microscopy (OM). After optimizing the growth condition, surface topographies of the as-grown MoS2 monolayers were studied via atomic force microscopy (AFM). The optical properties and doping effects of the as-grown and transferred MoS2 monolayers were investigated using micro-photoluminescence (PL) and Raman spectroscopy; field-effect transistor (FET) characteristics were also studied.
MoS2 crystals were grown on a SiO2/Si substrate by using two-zone thermal CVD with a 2-inch quartz tube . Prior to the growth process, SiO2/Si substrates were pre-treated with O2 plasma (100 W, 2 min) to obtain hydrophilic surfaces. Next, a liquid precursor solution was coated on the SiO2/Si substrate by using a spin-coater at 3000 rpm for 30 s. The liquid precursor solution consisted of two Mo-precursors, sodium molybdate dihydrate (SMD, Sigma-Aldrich) and ammonium molybdate tetrahydrate (AMT, Sigma-Aldrich), NaOH as a promoter, iodixanol (Opti, Sigma-Aldrich) as a surfactant, and deionized (DI) water for dilution . We mixed these materials to achieve the desired growth condition. We used a fixed solution ratio of NaOH : SMD : Opti = 4 : 0.5 : 0.5 as the reference condition. S-precursor source (S powder, Sigma-Aldrich) in a ceramic boat was placed at zone 1, and the liquid-solution-coated substrate was placed at zone 2. For synthesis, zone 1 and zone 2 were heated to 250 and 800 °C, respectively. The sublimated S at zone 1 (210–250 °C) was carried to zone 2 by N2 carrier gas (99.9999 %). The carried S atoms reacted with Mo-precursors at zone 2 (700–800 °C), resulting in the synthesis of MoS2, when the flow rate of N2 gas was fixed at 200 sccm. Na ions from NaOH promoted the nucleation and growth of MoS2 crystals . After completion of the growth process, the temperature of the furnace was naturally quenched below 80 °C.
Figure 1 shows a schematic of our optimized CVD process for the fabrication of triangular MoS2 monolayers. The inset shows the programed temperature setting for the growth process.
To characterize the electrical properties, the MoS2 layers were transferred onto new SiO2 (300 nm)/Si substrates. The transfer process consisted of the following steps: i) Polymethyl methacrylate (PMMA) was coated on MoS2/SiO2/Si by using a spin-coater at 1000 rpm for 60 s. ii) The PMMA/MoS2 layer was detached from the SiO2/Si substrate in DI water. iii) The floating PMMA/MoS2 layer on the DI water surface was transferred onto a new SiO2/Si substrate (treated with O2 plasma) and dried using a N2 blowgun. iv) The PMMA/MoS2 layer on the SiO2/Si substrate was heated using a hot plate at 80 °C for 5 min. The PMMA supporting layer was removed using acetone and sequentially cleaned using isopropyl alcohol.
The growth modes and shapes of the MoS2 flakes for the as-grown samples were observed by OM. The surface morphologies and residues of the as-grown samples were measured by AFM. We identified residues such as the Mo-precursor, promoter, and crater at the surface of the as-grown samples. The PL and Raman spectra for the as-grown and transferred samples were characterized using a micro-PL/Raman mapping system equipped with an objective lens of 0.75 N.A. and a laser beam with a wavelength of 532 nm (XperRam200VN, NANOBASE, Inc.).
To characterize the doping properties, we fabricated FETs by using the In micro-welding method, which is known as a polymer-residue-free process for mentalizations [12,13]. In micro-electrodes were connected to the flake edge of the transferred and annealed MoS2 monolayers on SiO2 (300 nm)/Si. Drain current (ID) curves as a function of the Si back gate bias (VG) were obtained for various values of the drain bias (VD) by using a Keithley 2614B Source measure unite.
Figure 2(a) shows the OM images of the synthesized MoS2 flakes on the substrate under temperature variation of zone 1 and zone 2, where the temperatures range from 210 to 250 °C for zone 1 and 700 to 800 °C for zone 2. At a fixed temperature of 700 °C in zone 2, the MoS2 flakes enlarge as the temperature in zone 1 increases. While small crystal flakes start to form at 230 °C and enlarge at 250 °C in zone 1, no crystal shapes are observed at 210°C in zone 1. We attribute the failure of MoS2 formation at a low temperature in zone 1 to the lack of sublimated S flux for MoS2 synthesis. This is because the size of the MoS2 flake is proportional to the sublimation temperature of S in zone 1 under a fixed temperature of the synthetic zone (700 °C). MoS2 flakes with a starfish shape, indicating off-stoichiometry , as well as solution residues (Mo-source and promoter) on the facets of MoS2 flakes and the substance are frequently observed. This is because the sublimation of Mo-precursors at a temperature of 700 °C in zone 2 is not sufficient for the MoS2 formation reaction. However, at 750 °C in zone 2, the MoS2 flakes exhibit a curved triangular shape for all zone 1 temperatures and lesser residues are observed. Eventually, MoS2 flakes with a near triangular shape and size exceeding 60 μm are synthesized at 800 °C in zone 2, indicating the exact stoichiometry of the Mo and S composition for MoS2. While multilayers form at a lower temperature (210 °C) in zone 2, uniform monolayers are synthesized at higher temperatures (230 and 250 °C). These growth modes according to the synthesis temperatures are schematically summarized in Fig. 2(b).
Figure 3(a) shows the OM images of the synthesized MoS2 flakes as a function of the AMT ratio in the Mo-precursor solution (top) and corresponding schematics depicting the growth mode (bottom). As the AMT content increases under the reference growth condition, the flake size increases to 100 μm for an AMT ratio of 0.5 owing to the sufficient amount of the Mo source. Moreover, the curved facet of the triangular shape changes to a flat facet, indicating the exact stoichiometry for MoS2 . While large-area growth of MoS2 can be achieved using a sufficient amount of AMT, an excess amount of AMT will promote multilayer growth. The effects of carrier gas amount are displayed in the OM images and schematics shown in Fig. 3(b). While the crystals are not formed at a smaller gas amount of 100 sccm owing to a lack of S vapor flux, triangular MoS2 monolayers are synthesized at 200 sccm. However, when the gas amount reaches 300 sccm, enlarged and curved triangles with multilayers are observed owing to an overly high flux of S vapor. This suggests that an optimum amount of carrier gas is essential for forming homogeneous monolayers. Figure 3(c) displays the effects of the deposition (or growth) time on the growth mode of MoS2. As the deposition time increases from 18 to 22 min, the size of MoS2 flakes increases and multilayers form at 22 min because of sufficient nucleation time for multilayer formation.
Under the reference condition optimized through the above processes, the surface morphology of the as-grown MoS2 was investigated using a tapping mode AFM, and the acquired topography image is shown in Fig. 4(a). The height of the MoS2 flake with triangular area is higher than the substrate area. While unreacted promoters of a residual particle shape are observed on the entire sample, they are selectively accumulated at the facet edges of the MoS2 flakes, which is typically observed in MoS2 prepared by CVD using solution-based metal precursors . Figure 4(b) shows the height profile of line A in Fig. 4(a). The height difference between the substrate and MoS2 layer reaches ~0.7 nm, indicating the typical monolayer thickness of MoS2 [8,10]. Figure 4(c) presents a magnified topography image of Fig. 4(a). Typical craters are clearly observed on the samples, which is attributed to craters originating from the unreacted Na2MoO4 . The height profile of line B in Fig. 4(c) is displayed in Fig. 4(d), where the measured crater depth reaches ~5 nm and the accumulated height of promoter residues at the facet boundary is over 17 nm. Such unreacted promoters and craters can be easily removed by undertaking transfer processes in water .
The effects of the residual promoters and craters on the optical properties of MoS2 were investigated using PL and Raman mapping for the as-grown MoS2 and transferred MoS2 on new SiO2/Si substrates (Fig. 5). The PL intensity map of A exciton peak for the as-grown and transferred MoS2 are shown in Figs. 5(a) and 5(b), respectively. The overall PL intensity distributions for both cases are homogeneous. However, the PL and Raman spectra exhibit prominent differences. In the comparative PL spectra shown in Fig. 5(c), while the A peak position for the as-grown MoS2 is in the range of 1.82 eV, that for transferred MoS2 is blue-shifted to ~1.87 eV with an enhanced PL intensity , indicating a p-doping effect [15,16]. This result agrees well with the comparative Raman spectra shown in Fig. 5(d). The doping-related A1g peak of the Raman spectra for the transferred MoS2 is slightly blue-shifted (~1 cm−1) owing to the p-doping effect, in comparison to that for the as-grown MoS2 . When the residual promoters, possibly acting as n-dopants in the as-grown MoS2, are removed by water during the transfer process , the transferred MoS2 is in a relatively de-doped condition. Therefore, the de-doping effect for the transferred MoS2 is attributed to the reduction in n-doping sources via the transfer process.
To confirm the de-doping effect of the transferred MoS2 monolayers, we measured the ID–VG transfer curves with varying VD for the MoS2 FET. Figure 6(a) illustrates the MoS2 FET with In electrode contacts, and Fig. 6(b) shows the measured ID–VG characteristics of the FET for various VD. These ID–VG curves exhibit bi-polar characteristic, and the on/off ratio reaches ~105 for a drain voltage of 0.7 V. Notably, while MoS2 FETs have been reported to typically exhibit n-type behavior for various metal electrodes [1,2,6,9], our FET exhibits bi-polar behavior because of the reduced n-doping effect and a good compatibility between MoS2 and In contact . Moreover, threshold voltages in the n-type region decrease with reduced VD and become nearly zero for a VD of 0.7 eV, in contrast to the fixed current level and threshold voltage in the p-type region. A change in the VD-induced threshold voltage is exhibited prominently between In and MoS2, which is attributed to the low work function (~4.0 eV) of In, because the VD dependence of thermionic effect is more sensitive for a small Schottky barrier .
We have studied the effects of process parameters on the growth modes for monolayer MoS2 flakes synthesized by CVD using solution-based Mo-precursors. Based on the variation in growth temperature, Mo-precursor ratio in the solution, carrier gas amount, and growth time, we have optimized the growth mode for triangular flakes with the exact stoichiometry of MoS2. As the most important process parameter, the growth temperature influenced the flake size and shape as well as the number of layers. As the AMT ratio was increased in the liquid precursor solution, the coverage of MoS2 flakes also improved. As the carrier gas amount and the deposition time increased, the coverage of MoS2 flakes and the number of layers also increased. The AFM morphology of the as-grown MoS2 surface revealed a considerable amount of precursor and promoter residues, which accumulated at the interfaces of the flake facets. The PL and Raman mapping images displayed an uniform contrast for intensities in a single flake; however, the blue-shift of the A exciton peak and A1g peak for the transferred samples indicated de-doping effects resulting from precursor residue removal during the transfer process using DI water. The ID–VG transfer curves of the MoS2 FET with In electrodes exhibited an on/off ratio of ~105 for a drain voltage of 0.7 V. Moreover, the FET displayed typical bi-polar transistor characteristics owing to the de-doping effect of transferred MoS2 and reduced Schottky barrier effect from In electrodes.
This work was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2018R1D1A1A02046206).