This is an open access article published under an ACS AuthorChoice License, which permits

copying and redistribution of the article or any adaptations for non-commercial purposes.

Letter

pubs.acs.org/NanoLett

Suppressing Nucleation in Metal−Organic Chemical Vapor

Deposition of MoS2 Monolayers by Alkali Metal Halides
HoKwon Kim,†,‡ Dmitry Ovchinnikov,†,‡ Davide Deiana,§ Dmitrii Unuchek,†,‡ and Andras Kis*,†,‡
†
Electrical Engineering Institute, École Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland
‡
Institute of Materials Science and Engineering, École Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland
§
Interdisciplinary Center for Electron Microscopy (CIME), École Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne,
Switzerland
*
S Supporting Information
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

ABSTRACT: Toward the large-area deposition of MoS2

layers, we employ metal−organic precursors of Mo and S for
a facile and reproducible van der Waals epitaxy on c-plane
sapphire. Exposing c-sapphire substrates to alkali metal halide
Downloaded via 94.64.111.217 on July 7, 2023 at 19:28:56 (UTC).

salts such as KI or NaCl together with the Mo precursor prior

to the start of the growth process results in increasing the
lateral dimensions of single crystalline domains by more than 2
orders of magnitude. The MoS2 grown this way exhibits high
crystallinity and optoelectronic quality comparable to single-
crystal MoS2 produced by conventional chemical vapor deposition methods. The presence of alkali metal halides suppresses the
nucleation and enhances enlargement of domains while resulting in chemically pure MoS2 after transfer. Field-effect
measurements in polymer electrolyte-gated devices result in promising electron mobility values close to 100 cm2 V−1 s−1 at
cryogenic temperatures.
KEYWORDS: Chemical vapor deposition, two-dimensional transition metal dichalcogenides, nucleation and growth,
microstructure engineering, FET devices

T he chemical vapor deposition of two-dimensional

materials is a highly promising method to produce
atomically thin layers at a large scale for harnessing their
gases within the gaseous mixture that is transported to the
substrate by controlling the evaporation rates of the solid
precursor and mass flow rates. An extensive vapor phase
attractive properties. Monolayer MoS 2 is a model 2D thermodynamics study performed by Kumar et al.19 has shown
semiconductor that can be used to realize field-effect transistors that growth temperatures above 850 °C at atmospheric
with high current on/off ratios.1 It is a naturally occurring pressure lead to layer-by-layer growth of MoS2 without
material with a good chemical stability that exhibits a wide
extraneous deposition of carbon contaminants resulting from
range of attractive properties such as a spin−orbit coupling-
induced band splitting,2,3 a mechanically tunable bandgap,4−8 the organic ligands of the precursor.
and a low temperature superconductivity.9−13 Although previous studies have paved the way to growing
Toward the large-scale synthesis of MoS2 thin films, a atomically thin films of MoS2 by MOCVD, the underlying
conventional chemical vapor deposition method of producing MOCVD growth mechanism and ways to control the critical
MoS2 monolayers typically involves low vapor pressure solid parameters that determine the quality of the deposited film
powder precursors such as MoO3 and sulfur. It has been such as coverage and domain sizes are still lacking. In this work,
investigated for centimeter-scale deposition of polycrystalline we show that pre-exposing the substrate before the growth
monolayer MoS2 with grain sizes of nanometer to micrometer stage to nucleation suppressors such as alkali metal halides,
and with controllable coverage.14,15 However, low vapor together with the Mo precursor, is an important factor in
pressures of the solid precursors require them to be loaded decreasing the nucleation density and enabling large-area
inside a heated zone of the reactor chamber leading to a limited growth of MoS2 domains on the order of 10 μm within a
control over the vapor phase composition and deposition rate.
short growth time of 30 min. This is a significant improvement
Thus, this synthesis approach heavily undermines the ability to
control the nucleation density, thickness, and coverage. over the growth morphology obtained without the use of the
Here, we aim to address this issue by employing well-known alkali metal halides and offers a reproducible and cost-effective
metal−organic precursors of molybdenum, Mo(CO)6, which is
a high vapor pressure solid, and of sulfur, H2S in gas phase.16−19 Received: June 1, 2017
This metal-organic chemical vapor deposition (MOCVD) Revised: July 7, 2017
approach allows reliably setting the concentration of precursor Published: July 12, 2017

© 2017 American Chemical Society 5056 DOI: 10.1021/acs.nanolett.7b02311

Nano Lett. 2017, 17, 5056−5063
Nano Letters Letter

Figure 1. (a) SEM image of MoS2 nuclei on c-sapphire obtained from an alkali metal halide-free growth. The dark, round-shaped domains are MoS2
islands that randomly nucleated on the surface of sapphire, which appears as a bright background in the image. (b) Optical microscope image of
MoS2 grown on sapphire during NaCl-assisted growth. Note that optical microscopy was used in panel b instead of SEM to give a larger view of the
sample and that MoS2 layers appear to be brighter than the substrate. (c) Raman spectra of MoS2 samples produced in the alkali−metal halide free
process (bottom) and NaCl-assisted growth (top). (d) Absorbance spectra of the samples produced from the alkali−metal halide free process
(black) and NaCl-assisted growth (blue). (e, f) Background-subtracted XPS core-level Mo 3d and S 2s spectra of the samples produced from the
alkali-metal halide free process (e) and NaCl-assisted growth (f). Gauss-Lorenztian mixed peak fitting analysis was performed on the spectra to
determine the chemical states of the deposited film.

way for mass-scale production of microstructure-engineered activated processes that lead to random nucleation of MoS2
two-dimensional transition metal dichalcogenides (TMDCs). arising from the collective crystallization of surface adatoms.20
Results and Discussion. We first employed a single- Raman and UV−vis absorbance spectra (Figure 1c,d)
growth-step procedure similar to Kumar et al.19 in a quartz measured over multiple grains on sapphire indicate that MoS2
tube, with a hot-wall furnace setup (Figure S1), releasing both is largely of monolayer nature and has good optical quality. The
precursors at the same growth stage at a fixed growth peak parameters and relative distances of E′ and A1′ ́ peaks of
temperature and without using any alkali metal halides. This Raman spectra correspond to those of monolayer MoS2.21 A
growth procedure leads to a deposition of finely grained, representative absorbance spectra for the growth temperature
of 950 °C is shown in Figure 2d where clear features of A, B, C,
circularly shaped MoS2 nuclei on sapphire with a high
and D excitons can been seen.22−25,15 Furthermore, the core-
nucleation density and an average grain-to-grain distance of
level XPS spectra of Mo 3d (Figure 1e) indicate that the
less than 100 nm for a wide range of growth temperatures MOCVD growth results in MoS2 with a small amount of peaks
(Figure 1a and Supporting Information, SI, Figure S2). Figure arising from Mo6+ oxidation state akin to that of MoO3, which
S2 shows the evolution of growth morphology of MOCVD could be originating from a slight oxidation of the sample
MoS2 on sapphire as a function of the growth temperature. The around the grain boundaries or other defects.26
higher growth temperature leads to reduced nucleation density To reduce the density of nuclei and improve their lateral size,
and thus larger grain sizes. This nucleation density versus Tgrowth we placed alkali metal halides (mainly KI or NaCl) upstream as
behavior (Figure S2b) is exponential in nature due to thermally first employed by Kang et al.18 It has been suggested that the
5057 DOI: 10.1021/acs.nanolett.7b02311
Nano Lett. 2017, 17, 5056−5063
Nano Letters Letter

Figure 2. (a) Optical microscope image of an MoS2 island grown on sapphire after KI-assisted MOCVD. (b) Room-temperature PL spectra of as-
grown MoS2 on sapphire, transferred onto SiO2/Si, and free-standing MoS2 transferred onto a TEM grid (normalized to as-grown on sapphire
spectra). (c) E′ Raman peak position map of the island constructed after multipeak fitting of the MoS2 Raman spectral region for each pixel. (d) PL
maximum peak position map of the MoS2 island. Scale bars, 5 μm (c and d).

presence of the alkali halides would reduce the concentration of inhomogeneity of the island, Raman and PL mapping was
water in the vapor phase environment of the MOCVD system performed on the same area. Especially, the positions of the E′
encouraging the two-dimensional growth of nuclei while Raman peaks and excitonic PL peak maxima appear to shift
suppressing nucleation. We found that a pre-exposure of the most prominently as illustrated in Figure 2c and d. Both of the
sapphire substrate to Mo(CO)6 together with a very small maps follow a similar pattern with band-like regions of red-
amount of KI (<1 mg) placed in front of the substrate prior to shifted peaks appearing to be radiating outward from the center
the growth stage is enough to significantly boost the domain of the island. This variation in peak positions can be attributed
sizes of MoS2 beyond 10 μm (Figure 1b and Figure S3), which to a nonuniform strain field within the island where parts of the
is more than a 2 orders of magnitude improvement over the island pinned to the substrate are in tensile strain due to the
one-growth-step approach on the nuclei size (∼100 nm vs ∼50 difference in the thermal expansion coefficient while some parts
μm at Tgrowth = 950 °C) and similar to that of conventional are relaxed during the sample cooling.8,15 On the basis of the
CVD involving MoO3(s) and S(s) precursors. Large-area Raman distributions of the E′ peak position and PL max. position
(Figure 1c) and absorbance (Figure 1d) measurements values, a tensile strain difference of ∼0.3% was estimated, which
demonstrate that the resulting film also exhibits good crystalline is consistent with the E′ position shift of a sample after a
characteristics.15 transfer onto a c-sapphire substrate (Figure S4a), and what is
The nucleation density and coverage appear to be limited by expected from the thermal expansion coefficient mismatch
the amount of Mo as shown in Figure S3. The growth time between bulk MoS2 and sapphire.8,15
beyond 30 min (Figure S3g,h) did not significantly influence For chemical state analysis, micro-XPS was performed on the
the nucleation density and coverage, suggesting a quick growth islands before and after transferring the entire film onto the
rate and Mo-limited growth. Further increasing further the SiO2/Si substrate. From the survey spectra (Figure 3a), the
exposure to Mo increased the nucleation density and coverage, main contaminant on the as-grown MoS2 appears to be alkali
however, it led to the decrease in the domain size (Figure S3f). metals originating from the alkali metal halides used for the
This significantly improved grain size allows us to carefully growth. The core-level spectra indicate that, while there are
investigate a single, individual island of MoS2 at a larger spatial alkali metals on top of the island or at the interface between the
scale. For instance, room temperature PL spectra (Figure 2b) of island and substrate, the oxidation states of Mo appear to
the as-grown island on sapphire exhibit a strong emission correspond to those of MoS2 islands (Mo4+) where the small
coming from the A exciton, characteristic of the direct band gap deviation is likely to be related to oxidized Mo states at the
of monolayer MoS2.22,25 The fwhm of ∼64 meV is comparable interface.
to as-grown monolayer MoS2 samples on sapphire produced by After transfer, the MoS2 film no longer contains a significant
conventional CVD method and suggests a high quality with a amount of alkali metals (Figure 3b,c). It is apparent that the
relatively small defect concentration. To probe the spatial alkali metals were washed away in the solvents used during the
5058 DOI: 10.1021/acs.nanolett.7b02311
Nano Lett. 2017, 17, 5056−5063
Nano Letters Letter

Figure 3. (a) XPS survey spectra of MoS2 samples that were grown in the presence of NaCl and KI. (b) XPS core-level spectra of the Na 1s region
for as-grown MoS2 sapphire (top) obtained by NaCl-assisted growth and the transferred sample on SiO2/Si substrate (bottom). (c) XPS core-level
spectra of K 2p region containing K 2p3/2 and K 2p1/2 peaks for as-grown MoS2 on sapphire (top) obtained by KI-assisted growth and sample
transferred sample onto SiO2/Si substrate (bottom).

Figure 4. (a) High-resolution HAADF-STEM image of a monolayer MoS2 imaged at 80 kV. The array of Mo atoms appears to be much brighter
than sulfur atoms due to the Z contrast in HAADF-STEM mode. The white arrows indicate substitutional unintentional dopant atoms heavier than
Mo. (b) High-resolution HAADF-STEM image of 2L MoS2 area grown near the center of the island and (c) corresponding FFT image of the 2L
superstructure of b. Scale bars, 2 nm (a, b); 4 nm−1 (c).

transfer process. Indeed, the examination of the sapphire as they still show a significant contribution from Mo5+ oxidation
substrate after the complete delamination shows no presence of states which can be linked to a oxidized form of Mo at the
potassium (Figure S5a−d). Additionally, in a separate experi- interface between MoS2 and sapphire that possibly participated
ment, we observe the removal of alkali metal after DI water in the growth process of MoS2.
washing of sapphire substrates that had been annealed in the To take a closer look at the atomic structure, high-resolution
atmosphere rich of alkali metal halides suggesting that the alkali HAADF-STEM imaging was carried out on freestanding MoS2
metals interact only weakly with the substrate (Figure S5e−g). layers transferred onto a TEM grid. Figure 4 illustrates a well-
Furthermore, the XPS core-level measurements (Figure S5b) ordered MoS2 atomic lattice apparently exhibiting an ideal 2H-
show that residual Mo are bound more strongly to the sapphire phase lattice structure with a low defect and impurity
5059 DOI: 10.1021/acs.nanolett.7b02311
Nano Lett. 2017, 17, 5056−5063
Nano Letters Letter

Figure 5. EDLT based on monolayer MOCVD MoS2. (a) Transfer characteristics of a 4-probe electric double-layer transistor, recorded with a sweep
rate of 5 mV/s. (Inset) Optical microscope image of a four-probe transistor device before electrolyte deposition. (b) Sheet conductivity at 280 K as a
function of VPE. The black arrows show the direction of gate voltage sweep with a 10 mV/s rate. (c) Temperature dependence of the four-probe field
effect electron mobility values for VPE = 1.3 V.

concentration. On the basis of the FFT analysis, the lattice gating to build electric double-layer transistors (EDLTs) based
spacing between the Mo atoms, which appear much brighter on monolayer MoS2 and explore a wide range of carrier
than sulfur atoms due to the Z contrast,27 was calculated to be densities. [EMIM][TFSI] in PS−PMMA−PS polymer ma-
3.2 Å, matching closely with the known values of the MoS2 trix35−37 was spin-coated onto our samples followed by baking
lattice parameter.25,28 Occasionally, bright substitutional of the samples in high vacuum for 2 h at 70 °C. Current Is as a
dopants could be observed in place of Mo atoms in the atomic function of polymer electrolyte voltage VPE is shown in Figure
scale images. This heavier element is likely tungsten, which was 5a for one sweep. We observe a sharp turn on with a
the most significant impurity in the Mo precursor used subthreshold slope (SS) value of 125 mV/dec at room
according to the impurity analysis provided by the supplier temperature, in line with measurements performed at 240 K
(lot no. MKBX7702 V, Sigma-Aldrich) as also detected by the in monolayer CVD MoS2 grown from MoO3 and solid sulfur
large-area STEM-EDS measurements (Figure S6). Again, no precursors.32 The dome-like shape at high VPE values is related
significant alkali metals (Na or K) and halides (Cl or I) were to electrolyte-induced disorder.37−39 Hysteresis is present in
observed in the MoS2 area, consistent with the XPS our measurements, but overall we found good reproducibility in
observation. Indeed, the PL measurements on the free-standing three devices that were measured. In Figure 5b we show sheet
monolayer MoS2 (Figure 3b) show a strong exciton A peak resistance extracted at a lower temperature (T = 280 K). For
with a minimum trion peak contribution with the fwhm of ∼45 further insights into transport properties, we performed
meV, similar to the best values of A exciton fwhm (∼47−50 temperature-dependent measurements (Figure 5c) applying
meV22,25,29−31) for room-temperature PL of free-standing fixed VPE at 280 K, lowering temperature below freezing point
monolayer MoS2 found in the literature. Moreover, low- of PE (220−240 K) and turning on the back gate. We found
temperature photoluminescence spectra of a transferred sample low contact resistances with PE gating down to low
taken at 4 K (as shown in Figure S4b) exhibit no significant temperatures (Figure S7b). Band-like transport could be
bound exciton peak at lower energies, which are normally achieved in contrast to solid-gated samples (e.g., Figure S7c),
found for CVD-grown MoS2 monolayers and likely arises from after a metal−insulator-transition above the critical conductivity
trapped charges or defects.32,33 of e2/h, to yield a four-probe electron mobility, μ4C ∼ 95 cm2/
Sometimes, stacks of bi- to few-layer MoS2 could be observed (V s) at T = 12 K. This behavior is in agreement with early
at the centers of individual islands. This is a common feature reports of band-like transport in exfoliated40 and CVD-grown
for significantly lowered nucleation density resulting in well- monolayer MoS2.41 We noticed that the low temperature
isolated domains (internucleus distance ≫ surface diffusion mobility is limited by electrolyte-induced disorder.39 A similar
length of the active species)34 at a high growth temperature and trend could be observed for example comparing PE-gated38 and
large amount of salts. The high-resolution STEM image indeed h-BN encapsulated42 monolayer WSe2 devices. Overall, we have
shows that a bilayer found near the center of an island can have demonstrated good electrical performance of our devices.
a random twisting angle as evident from a FFT analysis (e.g., We now turn our attention to the growth mechanism and the
Figure 4c). This unusual stacking and weak interaction between effect of the alkali metal halides on domain size enlargement. It
the layers were also evidenced by Raman and photo- is not likely that the alkali metal halides actually lower the
luminescence spectra in Figures S4c and d. amount of residual water vapor, which could affect the two-
For the electrical characterization of our material, MOCVD dimensional growth,43 as originally suggested by Kang et al. for
MoS2 was transferred onto degenerately doped Si substrates a low-temperature, low-pressure MOCVD process,18 since very
with a 270 nm thick, thermally grown SiO2 layer. First, we small amounts (0.1−1 mg) were used so that most of the salt
performed characterization of our samples using the solid back- evaporated or decomposed before the actual growth stage.
gate in vacuum after in situ annealing, which established the n- Furthermore, when a large quantity of salt was placed upstream,
type charge conduction. The sheet resistivity and contact away from the heating zone of the furnace preventing the
resistance were not low enough to observe band-like transport evaporation (Figure S8a), nucleation was not significantly
(Figure S7a,b) possibly due to the intrinsically low doping level suppressed. We have also investigated the effects of other ionic
of our sample. To this end, we used polymer electrolyte (PE) salts such as CaCl2 and CaO which are known to be more
5060 DOI: 10.1021/acs.nanolett.7b02311
Nano Lett. 2017, 17, 5056−5063
Nano Letters Letter

effective desiccating agents than the alkali metal halides.44,45 the pre-exposure stage and suppressing nucleation in the
However, we did not observe significant suppression of the growth stage. Further work is underway to expand the
nucleation density (Figure S8b,c). parameter space and optimize the coverage and thickness
Instead of acting as desiccants, the alkali metal halides are uniformity.
more likely to act in a similar way as planar, cyclic seeding Conclusion. We have developed a facile and reproducible
promoters such as perylene-3,4,9,10-tetracarboxylic acid method toward large-area growth of mono- to few-layer MoS2
tetrapotassium salt (PTAS), 3,4,9,10-perylene tetracarboxylic on c-sapphire using metal−organic precursors of Mo and S.
acid dianhydride (PTCDA), and copper hexadecafluoro- The pre-exposure of growth substrates to alkali metal halides
phthalocyanine (F16CuPc) that are known to suppress and the Mo precursor before the growth stage appears to play a
nucleation and enhance two-dimensional growth in the crucial role in suppressing nucleation and increasing the lateral
conventional chemical vapor deposition of TMDCs.14,46−50 size of single-crystal domains from a few hundred nanometers
However, the underlying mechanism is not well-understood to tens of microns. The MoS2 film grown this way exhibits high
partly due to the complexity of the chemical vapor deposition crystallinity and good optoelectronic properties with a narrow
process. Many factors such as nuclei surface/interfacial/edge A exciton and an electron mobility near 100 cm2V−1s−1 at
energy, surface diffusion rate, active species attachment rate, cryogenic temperatures in an electric double-layer transistor
and density of preferential nucleation sites are expected to configuration. The XPS and STEM analysis has shown that the
greatly influence the nucleation density on a substrate. One alkali metal halides do not bind strongly to the MoS2 resulting
possible explanation is that a planar molecule can act as a in chemically pure MoS2 after transfer. Our work is a major step
seeding promoter because it can stabilize the two-dimensional toward advancing the microstructure engineering and large-area
nuclei due to its geometry.48 Recently, an atomic additive such production of layered transition metal dichalcogenides.
as Cu was also shown to promote the two-dimensional growth Materials and Methods. Metal−Organic Chemical
by attaching itself preferentially on top of the hexagonal lattice, Vapor Deposition System. The metal−organic chemical
rather than at edge sites, blocking the three-dimensional growth vapor deposition system consists of a typical quartz tube
during CVD of TMDCs.51 We believe that alkali metal halides three-zone hot wall furnace (Carbolite HZS 12/600) with gas
can also act in a similar way to enhance two-dimensional flow rates controlled by mass flow controllers. High-purity H2S
growth versus nucleation. and Ar gases were provided by compressed gas cylinders
Additionally, a small intentional amount of oxygen has been (Carbagaz). Mo(CO)6 solid precursor (Sigma-Aldrich 577766,
shown to suppress nucleation of graphene52,53 and TMDCs in >99.9% purity) was loaded inside a stainless steel cylindrical
CVD by rendering potential nucleation sites less active or by bubbler (Strem Chemicals, Inc.) in a temperature controlled
increasing the size of the critical nucleus. In the case of alkali thermostat (Lauda Eco Silver) with a reverse bubbler
metal halides, the excessive amount of alkali metal (≫10 mg) configuration packed within silica beads 2 mm in diameter to
indeed appears to decrease the coverage as well as nucleation of facilitate the evaporation of Mo(CO)6 by increased surface
MoS2 layers (e.g., Figure S8a). The shapes of nuclei changed area.55
from circular to more faceted (Figure 1a, b), assuming a Before a typical growth step, the c-sapphire substrates
triangular or hexagonal shape with straight edges. In such case, (Al2O3(0001)) purchased from MTI Corp. were preannealed in
the nucleation rate must have sufficiently slowed down so that air at 1000 °C for 1 h to induce clean, atomically smooth
the density of nucleation would not reach the point of steps.56 The MOCVD process was performed in several stages.
saturation even toward the end of the growth stage.20,54 This For an alkali-metal salt-free, single-growth-step MOCVD
suggests that the alkali metal salts have likely induced a surface process, the substrates placed in the central zone of the furnace
poisoning effect that would limit the nucleation and coverage.34 were first heated to the growth temperature of 700−1020 °C
To further elucidate the effect of alkali metal halides, we under Ar atmosphere. We then introduced 1−3 sccm of H2S
interrupted the MOCVD process during a KI-assisted growth at into the quartz tube to create a sulfur-rich atmosphere in which
the end of the Mo(CO)6 exposure stage without introducing the substrates were annealed for 15 min. After this, Mo(CO)6
H2S. XPS analysis shows that Mo and K were adsorbed on the was introduced into the quartz tube using Ar carrier gas with an
surface after the exposure step while iodine is no longer present Ar flow rate of 1.5 sccm and a bubbler temperature of 5−20 °C
(Figure S5b). This adsorption of Mo could not be achieved which lead to the volumetric flow of 3.35 × 10−5 to 1.84 × 10−4
under the same amount of Mo(CO)6 exposure without the use sccm of Mo(CO)6 estimated using the reference values of
of the alkali metal halide. Furthermore, in a separate temperature-dependent vapor pressure values and ideal gas
experiment, Mo(CO)6 was introduced at a temperature above law.57 After 30−600 min of growth time (total Mo(CO)6
900 °C after evaporating away KI from the substrates where exposure of 3.9−7.48 μg leading to the overall Mo(CO)6 to
only K remains absorbed on the substrate. This did not lead to H2S molar ratio of ∼1:10 000 for the growth stage which
a significant deposition of Mo hinting us that both the alkali ensures S-rich environment), Mo(CO)6 flow was shut-off
metal and halide have to be present. This is consistent with our abruptly and the furnace allowed to cool to ambient
observation that directly attempting a single-step approach with temperature under the constant flow of H2S and Ar in order
alkali metal salts did not lead to the enhancement of two- to grow MoS2 in a sulfur-rich environment.
dimensional growth (Figure S9) likely due to the evaporation For the alkali metal halide-assisted growth, 10 to 0.5 mg of
of salts at typical growth temperature ranges (850−1000 °C). KI or NaCl was placed upstream of the sapphire substrate
In addition, we also tried the two-step approaches without any support and loaded together with c-sapphire substrates (Figure
alkali metal halides or with a halogen-only environment by S1a). The Mo(CO)6 precursor was first introduced into the
placing a boat or ampule of solid iodine upstream and repeated chamber when the temperature reached 600 °C, and the
the growth experiment, but this lead to poor morphologies substrate was exposed to the precursor for 30 min (Figure S1b)
(e.g., Figure S8d). Therefore, these results show that the alkali for optimal adsorption of Mo. The substrate was heated to the
metal halides play a dual role in facilitating adsorption of Mo in growth temperature of 800−1050 °C for large-domain size
5061 DOI: 10.1021/acs.nanolett.7b02311
Nano Lett. 2017, 17, 5056−5063
Nano Letters Letter

growth in Ar wherein H2S was introduced to initiate the via e-beam lithography. Au was deposited using an e-beam
sulfurization of the adsorbed Mo species. All other experimental evaporator with a thickness of 100 nm. [EMIM][TFSI] in the
configurations were kept the same as for the alkali metal halide- PS−PMMA−PS matrix was chosen as polymer electrolyte (PE)
free growth. for gating. The solution was spin-coated on top of our devices,
Raman, Photoluminescence (PL), and Absorbance Meas- and a large area gate electrode was used to apply PE voltage,
urements. Raman and PL measurements were performed using VPE. Prior to measurements, the sample was baked in high
a Renishaw inVia confocal Raman microscope with an objective vacuum for 2 h at 70 °C. For DC electrical measurements, we
lens of ×100 and laser wavelength of 532 nm with a laser power used Keithley 2400 SMU and Keithley 2000 DMM. Cryogenic
(<0.3 mW) carefully adjusted for a minimal damage by measurements were performed in a Janis closed-cycle cryogen-
checking to ensure that there are no peak shifts in the sample free cryostat. Mobility measurements were performed in four-
over repeated number of measurements. Diffraction gratings of contact geometry, and a geometric back gate capacitance of 270
1200 g/mm and 600 g/mm were used for Raman and PL nm SiO2 substrate 12.7 nF/cm2 was used for mobility
measurements, respectively, for suitable spectral ranges and extraction below the freezing point of electrolyte (220−240
resolution. Absorbance measurements were performed on an K). Interplay with the electrolyte gate could be excluded since
area approximately 1 cm in diameter using a Varian Cary UV/ the geometric capacitance was confirmed by Hall effect
vis spectrophotometer. measurements in similar devices based on ReS2.37
X-ray Photoelectron Spectroscopy (XPS). For survey and
core-level XPS analysis, ULVC-PHI Versa Probe II equipped
with laboratory Al K-alpha X-ray source (hv = 1486.6 eV) was
■
*
ASSOCIATED CONTENT
S Supporting Information
used with a typical beam size of 200 μm for a survey and The Supporting Information is available free of charge on the
average-area analysis while a beam size of 20 μm was used for ACS Publications website at DOI: 10.1021/acs.nano-
single-island analysis. A dual-beam charge neutralizer (low lett.7b02311.
energy electron and He ion) was used on the insulating
samples. The adventitious C 1s peak at the binding energy of Supporting figures and electrical properties of monolayer
284.8 eV was taken as a reference for offsetting the raw spectra. MOCVD MoS2 (PDF)

■
Atomic Force Microscopy (AFM) and Scanning Electron
Microscopy (SEM) Imaging Techniques. AFM topographical AUTHOR INFORMATION
imaging was performed with the Cypher Asylum Research
atomic force microscope in amplitude modulation mode. SEM Corresponding Author
images were obtained by a field emission SEM Zeiss Merlin *E-mail: andras.kis@epfl.ch.
microscope with a charge neutralizer to image the as-grown ORCID
sample on sapphire directly. Andras Kis: 0000-0002-3426-7702
Scanning Transmission Electron Microscopy (STEM). High- Notes
resolution high-angle annular dark field (HAADF) scanning The authors declare no competing financial interest.

■
transmission electron microscopy (STEM) was performed in a
double Cs-corrected FEI Titan Themis microscope equipped
ACKNOWLEDGMENTS
with a high brightness Schottky X-FEG electron source, a
monochromator, and a SuperX energy-dispersive X-ray We thank P. Mettraux (CIME, EPFL), A. Margez (MCP,
spectrometer (EDS), operated at 80 kV accelerating voltage EPFL), Z. Benes (CMI, EPFL), and F. Bobard (CIME, EPFL)
and low electron beam current values (below 50 pA). The for their training and technical support on XPS, Raman, e-beam
specimen for the STEM investigation was transferred on Si3N4 lithography, and SEM systems, respectively. We thank D.
TEM windows (Norcada, Canada) with arrays of 2 μm Dumcenco for insightful discussion on the CVD processes.
diameter holes. Images were processed by Wiener filtering in This project has received funding from the European Union’s
Gatan Digital Micrograph software. Horizon 2020 research and innovation programme under grant
Transfer Method and Device Characterization. For agreement no. 696656 (Graphene Flagship), European Union’s
transferring as-grown MoS2 on sapphire, the sample was spin- Seventh Framework Programme FP7/2007-2013 under grant
coated with a polymer support of poly(methyl methacrylate) agreement no. 318804 (SNM), and Swiss SNF grant no.
(PMMA) A2 1500 rpm for 1 min and was baked at 75 °C for 153298 and was performed in the frames of the Marie Curie
30 min. The edges of the sample were scratched by a diamond ITN network “MoWSeS” (grant no. 317451).
scriber for an efficient delamination of MoS2/PMMA before it
was immersed in the solution of KOH (30 wt %) at 70 °C for
about 30 min. It was then taken out and gently placed onto DI
■ REFERENCES
(1) Radisavljevic, B.; Radenovic, A.; Brivio, J.; Giacometti, V.; Kis, A.
water allowing the MoS2/PMMA film to delaminate from the Nat. Nanotechnol. 2011, 6 (3), 147−150.
substrate and float on the DI water. The floating piece of the (2) Zhu, Z. Y.; Cheng, Y. C.; Schwingenschlögl, U. Phys. Rev. B:
MoS2/PMMA was transferred to three beakers of DI water to Condens. Matter Mater. Phys. 2011, 84 (15), 153402.
be fished out onto a target substrate [typically SiO2 (270 nm)/ (3) Liu, G.-B.; Shan, W.-Y.; Yao, Y.; Yao, W.; Xiao, D. Phys. Rev. B:
Si for back-gated field effect measurements] which had been Condens. Matter Mater. Phys. 2013, 88 (8), 085433.
(4) J. Chem. Phys. 2014, 140 (17), 174707.
treated by a KOH solution (30 wt %) to render the surface
(5) Yue, Q.; Chang, S.; Kang, J.; Zhang, X.; Shao, Z.; Qin, S.; Li, J. J.
hydrophilic to improve the adhesion. The transferred sample Phys.: Condens. Matter 2012, 24 (33), 335501.
was left in acetone overnight and then annealed in a gaseous (6) Manzeli, S.; Allain, A.; Ghadimi, A.; Kis, A. Nano Lett. 2015, 15
mixture of H2 (10 v.%)/Ar at 380 °C for 4 h to remove PMMA. (8), 5330−5335.
The single crystalline area was chosen and etched in a (7) Pan, H.; Zhang, Y.-W. J. Phys. Chem. C 2012, 116 (21), 11752−
rectangular Hall bar geometry for four contact measurements 11757.

5062 DOI: 10.1021/acs.nanolett.7b02311

Nano Lett. 2017, 17, 5056−5063
Nano Letters Letter

(8) Lloyd, D.; Liu, X.; Christopher, J. W.; Cantley, L.; Wadehra, A.; (36) Yomogida, Y.; Pu, J.; Shimotani, H.; Ono, S.; Hotta, S.; Iwasa,
Kim, B. L.; Goldberg, B. B.; Swan, A. K.; Bunch, J. S. Nano Lett. 2016, Y.; Takenobu, T. Adv. Mater. 2012, 24 (32), 4392−4397.
16 (9), 5836−5841. (37) Ovchinnikov, D.; Gargiulo, F.; Allain, A.; Pasquier, D. J.;
(9) Costanzo, D.; Jo, S.; Berger, H.; Morpurgo, A. F. Nat. Dumcenco, D.; Ho, C.-H.; Yazyev, O. V.; Kis, A. Nat. Commun. 2016,
Nanotechnol. 2016, 11 (4), 339−344. 7, 12391.
(10) Saito, Y.; Nakamura, Y.; Bahramy, M. S.; Kohama, Y.; Ye, J.; (38) Allain, A.; Kis, A. ACS Nano 2014, 8 (7), 7180−7185.
Kasahara, Y.; Nakagawa, Y.; Onga, M.; Tokunaga, M.; Nojima, T.; (39) Gallagher, P.; Lee, M.; Petach, T. A.; Stanwyck, S. W.; Williams,
Yanase, Y.; Iwasa, Y. Nat. Phys. 2015, 12 (2), 144−149. J. R.; Watanabe, K.; Taniguchi, T.; Goldhaber-Gordon, D. Nat.
(11) Lu, J. M.; Zheliuk, O.; Leermakers, I.; Yuan, N. F. Q.; Zeitler, Commun. 2015, 6, 6437.
U.; Law, K. T.; Ye, J. T. Science 2015, 350 (6266), 1353−1357. (40) Radisavljevic, B.; Kis, A. Nat. Mater. 2013, 12, 815−820.
(12) Roldán, R.; Cappelluti, E.; Guinea, F. Phys. Rev. B: Condens. (41) Schmidt, H.; Wang, S.; Chu, L.; Toh, M.; Kumar, R.; Zhao, W.;
Matter Mater. Phys. 2013, 88 (5), 054515. Castro Neto, A. H.; Martin, J.; Adam, S.; Ö zyilmaz, B.; Eda, G. Nano
(13) Ge, Y.; Liu, A. Y. Phys. Rev. B: Condens. Matter Mater. Phys. Lett. 2014, 14 (4), 1909−1913.
2013, 87 (24), 241408. (42) Fallahazad, B.; Movva, H. C. P.; Kim, K.; Larentis, S.; Taniguchi,
(14) Lee, Y.-H.; Zhang, X.-Q.; Zhang, W.; Chang, M.-T.; Lin, C.-T.; T.; Watanabe, K.; Banerjee, S. K.; Tutuc, E. Phys. Rev. Lett. 2016, 116
Chang, K.-D.; Yu, Y.-C.; Wang, J. T.-W.; Chang, C.-S.; Li, L.-J.; Lin, T.- (8), 086601.
W. Adv. Mater. 2012, 24 (17), 2320−2325. (43) Kastl, C.; Chen, C. T.; Kuykendall, T.; Shevitski, B.; Darlington,
(15) Dumcenco, D.; Ovchinnikov, D.; Marinov, K.; Lazić, P.; T. P.; Borys, N. J.; Krayev, A.; Schuck, P. J.; Aloni, S.; Schwartzberg, A.
Gibertini, M.; Marzari, N.; Sanchez, O. L.; Kung, Y.-C.; Krasnozhon, M. 2D Mater. 2017, 4 (2), 021024.
D.; Chen, M.-W.; Bertolazzi, S.; Gillet, P.; Fontcuberta i Morral, A.; (44) Trusell, F.; Diehl, H. Anal. Chem. 1963, 35 (6), 674−677.
Radenovic, A.; Kis, A. ACS Nano 2015, 9 (4), 4611−4620. (45) Bower, J. H. Bur. Stand. J. Res. 1934, 12, 241.
(16) Dhar, S.; Kumar, V. K.; Choudhury, T. H.; Shivashankar, S. A.; (46) Ling, X.; Lee, Y.-H.; Lin, Y.; Fang, W.; Yu, L.; Dresselhaus, M.
Raghavan, S. Phys. Chem. Chem. Phys. 2016, 18 (22), 14918−14926. S.; Kong, J. Nano Lett. 2014, 14 (2), 464−472.
(17) Dumcenco, D.; Ovchinnikov, D.; Sanchez, O. L.; Gillet, P.; (47) Zhang, X.-Q.; Lin, C.-H.; Tseng, Y.-W.; Huang, K.-H.; Lee, Y.-
Alexander, D. T. L.; Lazar, S.; Radenovic, A.; Kis, A. 2D Mater. 2015, 2 H. Nano Lett. 2015, 15 (1), 410−415.
(4), 044005. (48) Yu, L.; El-Damak, D.; Radhakrishna, U.; Ling, X.; Zubair, A.;
(18) Kang, K.; Xie, S.; Huang, L.; Han, Y.; Huang, P. Y.; Mak, K. F.; Lin, Y.; Zhang, Y.; Chuang, M.-H.; Lee, Y.-H.; Antoniadis, D.; Kong, J.;
Kim, C.-J.; Muller, D.; Park, J. Nature 2015, 520 (7549), 656−660. Chandrakasan, A.; Palacios, T. Nano Lett. 2016, 16 (10), 6349−6356.
(19) Kumar, V. K.; Dhar, S.; Choudhury, T. H.; Shivashankar, S. A.; (49) Smithe, K. K. H.; English, C. D.; Suryavanshi, S. V.; Pop, E. 2D
Raghavan, S. Nanoscale 2015, 7 (17), 7802−7810. Mater. 2017, 4 (1), 011009.
(20) Robinson, V. N. E.; Robins, J. L. Thin Solid Films 1974, 20 (1), (50) Ling, X.; Lin, Y.; Ma, Q.; Wang, Z.; Song, Y.; Yu, L.; Huang, S.;
155−175. Fang, W.; Zhang, X.; Hsu, A. L.; Bie, Y.; Lee, Y.-H.; Zhu, Y.; Wu, L.;
(21) Lee, C.; Yan, H.; Brus, L. E.; Heinz, T. F.; Hone, J.; Ryu, S. ACS Li, J.; Jarillo-Herrero, P.; Dresselhaus, M.; Palacios, T.; Kong, J. Adv.
Nano 2010, 4 (5), 2695−2700. Mater. 2016, 28 (12), 2322−2329.
(22) Splendiani, A.; Sun, L.; Zhang, Y.; Li, T.; Kim, J.; Chim, C.-Y.; (51) Liu, J.; Zeng, M.; Wang, L.; Chen, Y.; Xing, Z.; Zhang, T.; Liu,
Galli, G.; Wang, F. Nano Lett. 2010, 10 (4), 1271−1275. Z.; Zuo, J.; Nan, F.; Mendes, R. G.; Chen, S.; Ren, F.; Wang, Q.;
(23) Britnell, L.; Ribeiro, R. M.; Eckmann, A.; Jalil, R.; Belle, B. D.; Rümmeli, M. H.; Fu, L. Small 2016, 12 (41), 5741−5749.
Mishchenko, A.; Kim, Y.-J.; Gorbachev, R. V.; Georgiou, T.; Morozov, (52) Hao, Y.; Bharathi, M. S.; Wang, L.; Liu, Y.; Chen, H.; Nie, S.;
S. V.; Grigorenko, A. N.; Geim, A. K.; Casiraghi, C.; Neto, A. H. C.; Wang, X.; Chou, H.; Tan, C.; Fallahazad, B.; Ramanarayan, H.;
Novoselov, K. S. Science 2013, 340 (6138), 1311−1314. Magnuson, C. W.; Tutuc, E.; Yakobson, B. I.; McCarty, K. F.; Zhang,
(24) Klots, A. R.; Newaz, A. K. M.; Wang, B.; Prasai, D.; Y.-W.; Kim, P.; Hone, J.; Colombo, L.; Ruoff, R. S. Science 2013, 342
Krzyzanowska, H.; Lin, J.; Caudel, D.; Ghimire, N. J.; Yan, J.; (6159), 720−723.
Ivanov, B. L.; Velizhanin, K. A.; Burger, A.; Mandrus, D. G.; Tolk, N. (53) Chen, W.; Zhao, J.; Zhang, J.; Gu, L.; Yang, Z.; Li, X.; Yu, H.;
H.; Pantelides, S. T.; Bolotin, K. I. Sci. Rep. 2015, 4, 6608. Zhu, X.; Yang, R.; Shi, D.; Lin, X.; Guo, J.; Bai, X.; Zhang, G. J. Am.
(25) Mak, K. F.; Lee, C.; Hone, J.; Shan, J.; Heinz, T. F. Phys. Rev. Chem. Soc. 2015, 137 (50), 15632−15635.
Lett. 2010, 105 (13), 136805. (54) Kim, H.; Mattevi, C.; Calvo, M. R.; Oberg, J. C.; Artiglia, L.;
(26) Kim, H.; Dumcenco, D.; Frégnaux, M.; Benayad, A.; Chen, M.- Agnoli, S.; Hirjibehedin, C. F.; Chhowalla, M.; Saiz, E. ACS Nano
W.; Kung, Y.-C.; Kis, A.; Renault, O. Phys. Rev. B: Condens. Matter 2012, 6 (4), 3614−3623.
Mater. Phys. 2016, 94 (8), 081401. (55) Fannin, L. W.; Webb, D. W.; Pearce, R. H. J. Cryst. Growth
(27) Krivanek, O. L.; Chisholm, M. F.; Nicolosi, V.; Pennycook, T. J.; 1992, 124 (1), 307−310.
(56) Yoshimoto, M.; Maeda, T.; Ohnishi, T.; Koinuma, H.; Ishiyama,
Corbin, G. J.; Dellby, N.; Murfitt, M. F.; Own, C. S.; Szilagyi, Z. S.;
O.; Shinohara, M.; Kubo, M.; Miura, R.; Miyamoto, A. Appl. Phys. Lett.
Oxley, M. P.; Pantelides, S. T.; Pennycook, S. J. Nature 2010, 464
1995, 67 (18), 2615−2617.
(7288), 571−574.
(57) Boxhoorn, G.; Ernsting, J. M.; Stufkens, D. J.; Oskam, A.
(28) Cheng, Y. C.; Zhu, Z. Y.; Tahir, M.; Schwingenschlögl, U. EPL
Thermochim. Acta 1980, 42 (3), 315−321.
Europhys. Lett. 2013, 102 (5), 57001.
(29) Mouri, S.; Miyauchi, Y.; Matsuda, K. Nano Lett. 2013, 13 (12),
5944−5948.
(30) Scheuschner, N.; Ochedowski, O.; Kaulitz, A.-M.; Gillen, R.;
Schleberger, M.; Maultzsch, J. Phys. Rev. B: Condens. Matter Mater.
Phys. 2014, 89 (12), 125406.
(31) Shin, B. G.; Han, G. H.; Yun, S. J.; Oh, H. M.; Bae, J. J.; Song, Y.
J.; Park, C.-Y.; Lee, Y. H. Adv. Mater. 2016, 28 (42), 9378−9384.
(32) Ponomarev, E.; Gutiérrez-Lezama, I.; Ubrig, N.; Morpurgo, A. F.
Nano Lett. 2015, 15 (12), 8289−8294.
(33) Zafar, A.; Nan, H.; Zafar, Z.; Wu, Z.; Jiang, J.; You, Y.; Ni, Z.
Nano Res. 2017, 10 (5), 1608−1617.
(34) Markov, I.; Stoyanov, S. Contemp. Phys. 1987, 28 (3), 267−320.
(35) Pu, J.; Yomogida, Y.; Liu, K.-K.; Li, L.-J.; Iwasa, Y.; Takenobu, T.
Nano Lett. 2012, 12 (8), 4013−4017.

5063 DOI: 10.1021/acs.nanolett.7b02311

Nano Lett. 2017, 17, 5056−5063