reviews www.advancedsciencenews.

com

Electronic Devices

Progress of Large-Scale Synthesis and Electronic

Device Application of Two-Dimensional Transition Metal
Dichalcogenides
Xiongfei Song, Zhongxun Guo, Qiaochu Zhang, Peng Zhou,* Wenzhong Bao,*
and David Wei Zhang

From the Contents

1. Introduction................................................2
The recent exploration of semiconducting two-
2. Synthetic Methods of 2D TMDCs..................2 dimensional (2D) transition metal dichalcogenides
(TMDs) with atomic thickness has taken both the
3. Electronic Transport in TMD Devices.............8
scientific and technological communities by storm.
4. Conclusion................................................18 Extensively investigated TMD that are accessible by
large-scale synthetic methods materials are remarkably
stable, such as MoS2 and WSe2. They allow superior
gate control due to their 2D nature and favorable
electronic transport properties, thus suggesting a bright
future for digital and RF electronics. In this review, the
latest developments in the controlled synthesis of large
scale TMDs are firstly introduced by discussing various
approaches. The major obstacles that must be overcome
to achieve wafer-scale, uniform, and high-quality TMD
films for practical electronic applications are included.
Advances in the electronic transport studies of TMDs
are presented, such as doping, contact engineering, and
mobility improvement, which contribute to overall device
performance. A perspective and a look at the future for
this field is provided in closing.

1700098 (1 of 22) www.small-journal.com © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim small 2017, 13, 1700098
 www.advancedsciencenews.com

1. Introduction of TMD devices.[19,26–32,34] Thus, material growth and sub-

sequent device processing are still the main challenges for
The discovery of graphene, one of the most famous two- large-scale realization of TMD integrated electronic systems.
dimensional layered materials (2DLMs), has triggered a Despite the unsolved issues in this field, the promising
world-wide research thrust over the last decade.[1–6] Gra- potential of TMDs is indicated by the increasing number of
phene has extraordinary electrical-transport properties that industrial electronic companies now active in this field. While
are the main justification for awarding the 2010 Nobel Prize the ultimate goal of completely replacing Si-based comple-
in Physics to Andre Geim and Kostya Novoselov, who con- mentary metal-oxide-semiconductors (CMOS) by atomically
ducted pioneering electrical-transport experiments.[2] Since thin 2DLMs is unrealistic, TMDs, because of their unique
then both the scientific and technological communities have properties, can still replace conventional materials in several
held great expectations for graphene electronics. They even existing applications, which can also inspire new applications
consider graphene to be among the candidate materials for such as flexible, wearable electronics.[36–38]
post-Si electronics.[7] T of a bandgap and controlled doping In this review, the focus is on semiconducting TMDs
approaches make it difficult to use in digital electronic appli- (mainly based on Mo and W for M, and S and Se for X) and
cations.[4] The zero bandgap of graphene is responsible for their applications in electronic devices. Specifically presented
the low ION/IOFF in graphene field effect transistors (FETs), are advances in the areas of material synthesis (including
which leads to considerable static power dissipation, while bottom-up methods including CVD, metal organic CVD
the conventional complementary logic operation must be (MOCVD), and epitaxial growth) as well as several major
based on controlled and stable doping processes, the perfect issues that are key to tailoring the electrical properties of
2D nature of graphene makes it difficult to move to digital 2D-TMDs (including doping, contact engineering, and dif-
electronic applications. Therefore, it is not clear whether it ferent scattering mechanisms). A perspective and a look
could migrate from Si to graphene integrated circuits (ICs). at the future is also provided. Summaries of earlier work
Consequently, huge efforts have been devoted to the are not provided here and can be found elsewhere in the
search for semiconducting 2DLMs with bandgaps greater literature.[39–41]
than ≈1 eV (so that the high on/off-current ratio is more
easily achieved at room temperature). One group of 2DLMs
meeting this criterion are the transition metal dichalco- 2. Synthetic Methods of 2D TMDCs
genides (TMDs). Unlike graphene, which consists only of
carbon atoms in a hexagonal lattice, TMDs follow an MX2 To shift from fundamental studies to the practical exploration
structure, where M represents a transition metal (from group of 2D-TMDs, it is essential to develop production approaches
4 to group 10, such as Mo, W, Ti, Pt and Hf) and X is a chal- that are reliable and scalable with controlled physical dimen-
cogen (S, Se, or Te). Depending on how M and X are com- sions and chemical compositions of TMD layers. As with
bined, TMDs can exhibit rich electronic properties, from graphene, there are two primary strategies for preparing
semiconducting to metallic or even superconducting.[8–10] In 2D-TMDs: the top-down strategy, where the bulk forms
addition, with their two-dimensional nature preserved band- are exfoliated into few-layer or monolayer sheets; and the
gaps can be opened and tuned by control of thickness, strain, bottom-up strategy to chemically synthesize these materials
and stacking orders.[11–15] A more distinct degree-of-freedom using chemical vapor deposition (CVD), atomic layer deposi-
feature of 2DLMs is that the weak van der Waals (vdW) tion (ALD), molecular beam epitaxy (MBE), etc.
force between layers allows flexible stacked vdW hetero- Since Novoselov et al.[2] first obtained monolayer gra-
structures without the limitation of lattice matching.[16] All phene via mechanical exfoliation of natural bulk graphite,
these features enable versatile engineering of band structures their method has been one of the most achievable top-down
at the atomic scale, providing a wide range of new materials approaches to produce clean, highly crystalline, and atomi-
suitable for electronic applications.[17–19] cally thin sheets of 2DLMs. It remains the best method for
The fast pace of TMD material development shows no fundamental investigations of the inherent physical prop-
sign of abatement. The main bottleneck preventing the prac- erties of 2DLMs and proof-of-concept device demonstra-
tical application of TMDs in electronics remains the need for tions,[12,42–45] but it is generally recognized as not practical for
uniform growth with wafer-scale size. Compared to chemical large scale synthesis of TMDs because it gives rise to limited
vapor deposition (CVD) synthesized graphene with a car-
rier mobility ≈11,000 cm2 V−1 s−1 at room temperature[20] X. Song, Z. Guo, Q. Zhang, Prof. P. Zhou,
and a wafer-scale single-crystalline grain size,[21] existing syn- Prof. W. Bao, D. W. Zhang
thetic methods so far can only produce TMD films with lim- State Key Laboratory of ASIC and System
ited spatial uniformity and fair electrical performance.[22–25] School of Microelectronics
Fudan University
Moreover, reliable and repeatable device-processing tech-
Shanghai 200433, China
nologies are lacking, since methods developed for conven- E-mail: [email protected];
tional bulk-SCs (Si, Ge, III-V materials, etc.) cannot be used [email protected]
in TMD device fabrication.[19,26–35] For example, acceptable The ORCID identification number(s) for the author(s) of this article
doping methods that do not degrade the 2D nature of TMDs can be found under https://doi.org/10.1002/smll.201700098.
are still under investigation.[33,35] Achieving good contact and
high carrier mobility are also critical for good performance DOI: 10.1002/smll.201700098

small 2017, 13, 1700098 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.small-journal.com (2 of 22) 1700098
 reviews www.advancedsciencenews.com

sheet size, poor reproducibility, and relatively low yield.

Other than this method, liquid sonication,[46,47] ion interca- Peng Zhou received his BS (2001) and PhD
lation,[48–54] and laser thinning[55] have also been employed (2006) in Physics from Fudan University,
to produce ultrathin flakes of 2DLMs. Recently developed and now he is a full professor at the Fudan
liquid phase exfoliation methods[56–59] achieve an improved University. His research focuses on the
memory devices based on two-dimensional
yield. Most top-down methods have poor control of thickness,
layered materials.
lateral dimension, and cleanness of resultant flakes. However,
even with these drawbacks they could be further refined with
time to meet the requirements of particular applications.
On the other hand, the bottom-up synthetic approaches
exhibit a potential to grow uniform 2DLMs with a controlled
thickness and lateral size, which are well suited for future
wafer-scale TMD devices. The breakthrough technique in
2008 of growing CVD on copper[60] has been a milestone
Wenzhong Bao is a full professor in the
for the commercial application of 2DLMs. A recent report School of Microelectronics at the Fudan
showed successful formation of ≈1.5 inch-large single-crystal- University. He graduated from Nanjing
line graphene by a single nucleus on the entire substrate.[21] University with a BS in physics (2006). He
Synthesis of large-area ultrathin TMD layers has been also received his PhD from the University of
extensively investigated recently, thus pointing toward large- California, Riverside (2011), having studied
in the Department of Physics and Astronomy.
scale integration of devices. The bottom-up methods include
His current research is on two-dimensional
CVD, MOCVD, and direct epitaxial growth, each of which is
layered materials, advanced energy storage
discussed. and flexible electronics.

2.1. Chemical Vapor Deposition of 2D-TMDs

Among the various bottom-up synthetic methods, the CVD

method has the most promise to realize high-quality, wafer- Figure 1e–f. There is also a contradictory report[65] claiming
scale 2DLMs at a low cost. The performance of CVD-grown that seed promoters are not essential for initializing the lay-
graphene electronic devices has been proved comparable to ered growth of TMDs, by demonstration of triangular MoS2
that of mechanically exfoliated samples.[60,61] The CVD syn- flakes with domain size ≈10 µm successfully grown on bare
thesis of 2D-TMDs is slightly different from that of graphene, SiO2/Si substrates. A recent report also demonstrated that
and can be mainly summarized by the following two routes: the number of obtained TMD layers could be controlled
one-step CVD synthesis and two-step CVD synthesis. by treating the surfaces of SiO2 substrates with the oxygen
plasma process.[66] More recently, Yu et al. reported a revised
2.1.1. One-Step CVD Synthesis setup in which a small quartz tube sealed at one end was
used as the container for the precursors and the substrate, as
In one-step growth gaseous transition metal-based precur- shown in Figure 1b. This method has proven to be effective in
sors and chalcogenide elements are simultaneously brought synthesizing triangular WS2 monolayers with an edge length
into a quartz tube, where they react directly to form TMDs of up to 178 µm (Figure 1c–d).[67] Unlike mounting the sub-
on arbitrary substrates. Such an approach allows the growth strate face-down above the transition metal-based compound,
of single crystalline TMD flakes in a controllable manner and the substrate and the transition metal-based compound could
has been widely used for producing TMD materials. be successively placed in the quartz tube of an LPCVD,[68] as
MoO3 (WO3) is one of the most popular precursors. shown in Figure 1g. Zande et al. have fabricated near-trian-
During the CVD process MoO3 vapor can be initially gular MoS2 flakes with the biggest domain size about 123 µm
reduced by sulfur vapor to form a volatile MoO3-x species, by using ambient-pressure CVD (Figure 1h).[24] Further-
which is subsequently delivered downstream by the carrier more, Gong et al. reported MoSe2 triangles with edge lengths
gas; it reacts with the sulfur-rich environment to form MoS2 greater than 1 mm, and hexagons with diagonals larger than
on the substrate. Figure 1a depicts a typical experimental 1.6 mm were successfully grown by controlling the nucleation
setup, in which sulfur and MoO3 powders were successively density (Figure 1i).[69]
placed in a quartz tube, with a substrate mounted face-down The choice of precursors is also crucial and, in a different
above the MoO3.[61,62] The substrate was pretreated by spin- work, MoCl5 and S were used to achieve high-quality mono­
casted seeding promoters such as reduced graphene oxide layers and few-layer MoS2 films over an area of centim-
(rGo), perylene-3,4,9,10-tetracarboxylic acid tetrapotassium eters with unprecedented uniformity and controllability.[70]
salt (PTAS), and perylene-3,4,9,10-tetracarboxylic dianhy- Different substrates also affect the quality of CVD-grown
dride (PTCDA), which were preferable for inducing effective TMD layers. Yu et al. reported that when WS2 on graphite is
nucleation. Using this treatment a few layers of nearly trian- compared with SiO2/Si substrates, WS2 grown on graphite is
gular MoS2[62–64] and WS2[64] sheets with domain size larger less affected by charged impurities and structural defects.[71]
than 1 µm were successfully produced by CVD, as shown in Hence, although SiO2/Si is currently considered the preferred

1700098 (3 of 22) www.small-journal.com © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim small 2017, 13, 1700098
 www.advancedsciencenews.com

Figure 1. One-step CVD synthesis of 2D-TMDs. a) Schematic illustration of an experimental setup (left), and an optical image of obtained MoS2
flakes (right). Reproduced with permission.[63] Copyright 2012, John Wiley & Sons, Inc. b) Experimental setup for growing large domain WS2, in
which sulphur and WO3 powders are loaded into a quartz tube that is sealed at one end. c) Optical image of WS2 flakes with a maximum side
length of 178 µm. d) SEM image of the area shown in (b). Reproduced with permission.[67] Copyright 2014, John Wiley & Sons, Inc. e–f) Typical
growth results obtained with and without coating PTAS as a seeding promoter. The insets in (e) are an optical image of the triangular MoS2 flakes
and an AFM image of the continuous MoS2 monolayer film. The insets in (f) are an AFM image of the deposited MoS2 and the corresponding height
cross-section analysis. Reproduced with permission.[64] Copyright 2013, American Chemical Society. g) Experimental setup of the LPCVD system.
Reproduced with permission.[68] Copyright 2013, American Chemical Society. h) Optical images of MoS2 flakes with large grain size on an SiO2/Si
substrate (left) and a monolayer MoS2 triangle with 123 µm side length (right). Reprinted by permission.[24] Copyright 2013, Macmillan Publishers
Ltd. i) Morphology of millimeter-scale MoSe2 single crystals and centimeter-scale MoSe2 continuous film. Reprinted by permission.[69] Copyright
2016, John Wiley & Sons, Inc.

substrate for growing MX2 thin layers, some other insulating metals (Mo/W) pre-deposited on the substrate; ii) similar to
single-crystal substrates such as sapphire,[68] mica, and SrTiO3 (i) but replacing the transition metals with transition metal
have attracted attention due to their ultra-flat surfaces, excel- oxide precursors (MoO3/WO3); and iii) thermolysis of the
lent thermal stability, and possible lattice registry with MX2 pre-deposited single-precursor ammonium thiomolybdate
adlayers. It should also be noted that telluride compounds (NH4)2MoS4.
can decompose and lose Te atoms at high temperatures. In Figure 2a–b, a thin layer of transition metal layer is
Hence, it is difficult to grow atomically thin tellurides directly initially deposited onto SiO2/Si substrates using either an
by CVD compared to sulfides and selenides. MoTe2[72,73] electron-beam (E-beam) or sputtering evaporation, and then
CVD growth has been reported recently, confirming that Te reacted in the CVD system at high temperature.[74,79–81] The
deficiency in the resultant film indeed exists. reaction mechanism can be understood as a direct chemical
reaction M + nX = MXn, where M is the transition metal and
2.1.2. Two-Step CVD Synthesis the X is the chalcogen. In this way, the thickness and size of
the TMD thin film is determined by the thickness of the pre-
An alternative attempt is the two-step growth route, in which deposited transit ion metal film and by the size of the growth
transition metal-based precursors are initially deposited on substrate.
the substrate and then converted into 2D-TMDs. Several Figure 2c–d show a schematic illustration for the synthesis
thermal CVD processes have been reported using different of TMD layers by metal-based precursor chalcogenization.
precursors: i) Direct chalcogenization of pure transition First, a layer of metal-based precursor is deposited on the

small 2017, 13, 1700098 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.small-journal.com (4 of 22) 1700098
 reviews www.advancedsciencenews.com

Figure 2. Two-step CVD synthesis of 2D-TMDs. a) Typical setup of 2D-MoS2 preparation by sulfurizing a pre-deposited Mo layer. b) Experimental
setup of the sulfurization furnace. Reproduced with permission.[74] Copyright 2014, John Wiley & Sons, Inc. c) Schematic of the setup for the synthesis
of MoS2 layers by MoO3 sulfurization. A thin layer of MoO3 was thermally evaporated on the sapphire substrate. Reproduced by permission.[75]
Copyright 2012, Royal Society of Chemistry. d) Schematic illustration of the synthetic procedure for the ALD-based WS2 nanosheets. Reproduced
by permission.[76] Copyright 2013, American Chemical Society. e) Schematic of the synthesis MoS2 via layer-by-layer surface sulfurization of MoO2
microplates, followed by further peel-off. Reproduced by permission.[77] Copyright 2013, American Chemical Society. f) Synthesis of MoS2 film by
the decomposition of thiomolybdates (NH4)2MoS4. Reproduced by permission.[78] Copyright 2012, American Chemical Society.

substrate using thermal evaporation,[75,82] PVD,[83] ALD,[76,84] at 1000 °C to obtain MoS2 films.[78] This thermolysis method
or other methods, forming a uniformly thin layer suitable produces high quality MoS2 thin films. However, due to the
for chalcogenization. Then, the transition metal-based pre- technical limitations of uniform and ultra-thin (NH4)2MoS4
cursor is converted to a TMD. In the second step the tran- film preparation, it is hard to achieve the desired thickness
sition metal-based precursor could directly react chemically for large-area MoS2 films using this method, which limits the
with chalcogenide to form TMDs.[82,83] As shown in Figure 2c, wafer-scale uniformity and its further application.
a two-step thermal process is an alternative approach, as
MoO3 converts primarily into MoO2, then MoO2 reacts with 2.1.3. Comparison of the Two CVD Synthetic Methods
sulfur to form MoS2.[75] Crystalline MoO2 has also been
used to directly react with sulfur to yield MoS2,[77] as shown Huge efforts have been devoted to the two CVD methods
in Figure 2e, with varied numbers of layers, depending on to achieve wafer-scale, continuous, and uniform TMDs with a
the annealing duration. Using this synthetic method, wafer- large crystalline grain size. Although both routes have proven
scale TMD thin films are successfully prepared on substrates, effective for producing large area TMDs, there is still debate
where the film thickness depends on the thickness of the about which route can yield better TMDs with high quality
deposited precursor. for practical applications. The one-step method allows effi-
TMD ultrathin film synthesis via the decomposition of cient chalcogenization in the gas phase for single crystalline
pre-deposited TMD-based precursor layers is an effective TMDs grown directly on arbitrary substrates. Hence, this
method of preparing large-area TMD thin layers. As shown method has been most widely applied to produce synthetic
in Figure 2f, (NH4)2MoS4 was initially dip-coated on the sub- TMD monolayers. However, a wide grain-size distribution
strate and converted to MoS2 by annealing at a high tem- and random localized nucleation centers usually occur during
perature in an Ar/H2 flow to remove the residual solvents, the CVD synthesis, which indicate that the morphology
NH3 molecules, and other byproducts dissociated from the of substrates and control of nucleation density are rather
precursors. This was followed by sulfurization in sulfur vapor crucial.[69] In the two-step method direct chalcogenization

1700098 (5 of 22) www.small-journal.com © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim small 2017, 13, 1700098
 www.advancedsciencenews.com

provides a fast and simple way to obtain atomically thin next-generation ICs from 2D-TMDs. The resulting TMD films
TMD layers, and in ideal situations the lateral size and thick- display moderate electrical properties, with MoS2 and WS2
ness of the pre-deposited metal/metal-based films determine room temperature mobility as high as 30 and 18 cm2 V−1 s−1,
the thickness and size of the resultant TMD layer. However, respectively. However, the grain sizes of these continuous
it is difficult to form continuous single-layer TMDs using this films are still limited to several micrometers, which under-
method. When targeting multiple (more than five) layers the mine their performance and applications in electronics
result may be relatively uniform; however, the obtained thin because of the scattering of carriers at grain boundaries.
films show a metallic transport property with a low on/off Complicated techniques, high cost, and the risks of handling
current ratio when the TMD layer becomes too thick.[19,85] hazardous gas also limit their broader implementation.
This is likely due to the presence of an unreacted metal layer.
Also, the TMD layers are mostly polycrystalline with a small
crystal grain size, which is limited by its reaction kinetics. 2.3. Direct Epitaxial Growth
Therefore, to exert a steady control over the material
growth process requires accurate control over the growth Different from the direct chemical reaction using CVD, epi-
temperature, flow rate, and super-saturation of the vapor taxial growth is an important route for achieving uniform
phase reactant, as well as uniform spatial distribution. All TMD films. The dangling bonds on the substrate surface, how-
these are difficult to achieve using simple thermal CVD pro- ever, can only be connected to well-lattice-matched materials
cesses. Besides, a high temperature is normally necessary due to the unique epitaxial growth mechanism. As a result,
during the CVD process, and this hampers its broader appli- to obtain a high-quality TMD ultrathin layer using heteroepi-
cation in flexible electronics or next generation thin film tran- taxy, the main challenge is the lattice-matching between the
sistors (TFTs). Therefore, CVD synthesis of ideal 2D-TMDs substrate and the target TMD material. However, when clean
in a controllable manner has been a longtime challenge. surfaces of freshly cleaved 2DLMs serve as growth substrates
without dangling bonds, the lattice-matching requirement for
heteroepitaxy growth is then relaxed due to a much weaker
2.2. MOCVD vdW interaction (and termed as vdW epitaxy).[87] This vdW
epitaxy has been demonstrated even with a lattice mismatch
Recently, homogeneous and wafer-scale growth of MoS2 and as large as 50%.[88] Recently, different epitaxial growth
WS2 have been demonstrated by Kang et al. through the use methods of TMDs for both monolayer and multi-layers have
of MOCVD,[86] as shown in Figure 3a–c. A well-designed been demonstrated. These use disparate approaches ranging
MOCVD process can yield highly uniform MoS2 and WS2 from CVD[89,90] to MBE[91,92] on different epitaxial substrates
monolayers in a four-inch scale, which is a breakthrough for including graphene,[89,93] highly oriented pyrolytic graphite

Figure 3. a) MOCVD growth of continuous MoS2 and WS2 films. b–c) Photographs of monolayer MoS2 and WS2 grown on 4-inch fused silica
substrates. Reproduced with permission.[86] Copyright 2015, Nature Publishing Group. d) FFT patterns of the monolayer MoS2 film grown on a
graphene membrane. The arrows indicate the diffraction spots originating from the graphene and the MoS2 layers. Reproduced with permission.[89]
Copyright 2012, American Chemical Society. e) A cross-sectional TEM image of grown HfSe2 on HOPG. f) High-angle annular dark field (HADDF)
STEM image of grown HfSe2 on MoS2 showing abrupt interfaces and layered crystalline films. Reproduced with permission.[91] Copyright 2015,
American Chemical Society. g) Five layers of MoS2 stacked on top of a SnS2 crystal shown by a cross-sectional TEM image of such heterostructures
viewed from zone axis (ZA) of SnS2; inset is the corresponding SAED pattern. h) A high-resolution TEM image along [001] zone axis showing both
lattice fringes and Moiré fringes of MoS2−SnS2 heterostructures. Reproduced with permission.[95] Copyright 2014, American Chemical Society.
i) Photograph of continuous MoS2 monolayer with full coverage on mica. Reproduced with permission.[90] Copyright 2013, American Chemical
Society.

small 2017, 13, 1700098 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.small-journal.com (6 of 22) 1700098
 reviews www.advancedsciencenews.com

Table 1. Summary of the large scale TMD synthesis methods.

Synthetic methods TMDs Key Preparation conditions Precursors Growth conditions Crystal size Doping On/off ratio &
type Mobility
One-step CVD synthesis MoS2 Pre-treated by seeding Sulfur and MoO3 Atmosphere; >2 µm N ≈104;
promoter; Substrate faced powder 650 °C ≈0.02 cm2/V·s
down. Ref. [63]
Substrate faced up. Ref. [65] Sulfur and MoO3 Atmosphere; ≈10 µm N ≈105–6 × 106;
powder 850 °C ≈0.2–18 cm2/V·s
Substrates were cleaned in Sulfur and MoO3 Atmosphere; ≈123 µm N ≈105–107;
H2SO4/H2O2 (3:1) for 2 h powder 700 °C ≈1–8 cm2/V·s
and 5 min for O2 plasma;
Substrate faced down.
Ref. [24]
Substrate faced up. Ref. [70] Sulfur and MoCl5 2 Torr; 850 °C / N ≈0.003–0.03 cm2/V·s
powder
WS2 Pre-treated by seeding Sulfur and WO3 Atmosphere; ≈10 µm N >105;
promoter; Substrate faced powder 650 °C ≈0.01 cm2/V·s
down. Ref. [64]
Substrate faced down. Sulfur and WO3 550 to 750 °C ≈178 µm / /
Ref. [67] powder
Substrate faced up. Ref. [68] Sulfur and WO3 225 mTorr; 900 °C ≈50 µm Bipolar ≈102;
powder ≈0.28–0.46 cm2/V·s
MoTe2 Substrate faced down. Selenium and Atmosphere; ≈1.3–1.6 mm N ≈105–107;
Ref. [69] MoO3 powder 750 °C ≈42 ± 19 cm2/V·s
Two-step CVD synthesis MoS2 Mo thin layer pre-deposited Sulfur powder Atmosphere; / N ≈0.004–0.04 cm2/V·s
on substrate. Ref. [25] 500–750 °C
MoO3 thin layer pre-deposited Sulfur powder 1 to 600 Torr; 500 / N ≈105;
on sapphire. Ref. [75] to 1000 °C ≈0.8 cm2/V·s
MoO2 microplates by reduc- Sulfur powder Atmosphere; 850 ≈5–20 µm N ≈104–106;
tion of MoO3. Ref. [77] to 950 °C ≈0.1–0.7 cm2/V·s
(NH4)2MoS4 was dip-coated (NH4 )2 MoS4 and 1 to 500 Torr; ≈160 nm N ≈105
on SiO2/Si or sapphire sub- sulfur powder 1000 °C
strate. Ref. [78]
≈6 cm2/V·s
WS2 WO3 thin layer pre-deposited Sulfur powder High vacuum; ≈5–15 µm / /
on substrate. Ref. [82] 800 °C
WO3 thin layer pre-deposited H2S gas Atmosphere; 470 / ≈3.9 cm2/V·s
by ALD. Ref. [76] to 1000 °C
MOCVD MoS2 Grown on sapphire and Mo(CO)6 gas, 7.5 Torr; 550 °C ≈100 nm–10 µm N ≈106;
SiO2. Ref. [86] (C2H5)2S gas ≈10–40 cm2/V·s
WS2 Grown on sapphire. Ref. [86] W(CO)6 gas, 7.5 Torr; 550 °C ≈100 nm–10 µm N ≈106;
(C2H5)2S gas ≈5–18 cm2/V·s
Direct Epitaxial Growth MoS2 Graphene-covered Cu foil (NH4 )2 MoS4 10 mTorr to ≈100 nm–10 µm / /
as the epitaxial substrate. atmosphere;
Ref. [89] 400 °C
Mica as the epitaxial Sulfur and MoO3 ≈30 pa; ≈0.32 to 3 µm / /
substrate. Ref. [90] powder 530–700 °C
HfSe2 HOPG and MoS2 as epitaxial Hf and Se 1 × 109 mbar; / / /
substrates. Ref. [91] 400–550 °C
MoTe2 Bulk MoS2 as the epitaxial Mo and Te 25 to 350 °C / / /
substrate. Ref. [92]

(HOPG),[91] mica,[90] MoS2,[91,92], GaSe[94] and SnS2,[95] as appropriately chosen substrate, has also been utilized to syn-
shown in Figure 3d–i. Through these methods high quality thesize MoS2 films thicker than 100 nm.
films of MoS2, MoTe2,[92] WS2,[95] WSe2,[95] and HfSe2[91] Common methods for TMD large-scale synthesis have
have been prepared by epitaxial growth] Pulsed laser evap- been summarized in Table 1 in order to show a clear com-
oration (PLD),[96] combined with the orienting effect of an parison among all the methods, including key preparation

1700098 (7 of 22) www.small-journal.com © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim small 2017, 13, 1700098
 www.advancedsciencenews.com

conditions, precursors, growth temperature, crystal quality, technologies, as compared to the case of conventional bulk-
and carrier mobilities. Most of these methods are general and SCs (Si, Ge, III-V materials, etc.). For example, substitution
can be readily extended to the growth of similar 2D-TMDs. doping, a common strategy adopted to access CMOS con-
However, large-scale synthesis that meets industrial stand- figuration in bulk-SCs, is not suitable here because it con-
ards is still challenging, and includes two unsolved issues: currently modifies the structure and degrades the TMDs.
1) It is necessary to improve the grain size, film continuity, Achieving good contact is another critical issue for achieving
uniformity, and thickness control, preferably synthesized at high-performance 2D-TMD devices. Clear mechanisms and
a low temperature; and 2) A specific substrate is normally effective approaches for this remain under debate. Contact
necessary to support the obtained TMDs, thus a transfer resistance also obscures the carrier mobility that is a critical
process is required to peel them off onto desired substrates. factor in device performance. Therefore, contact problems
However, the complicated transfer process is incompatible and the carrier scattering both limit the carrier mobility of
with the standard CMOS process, and inevitably introduces 2D-TMDs, which is now relatively modest compared to that
impurities, ruptures, and ripples. Before all these problems of graphene and III-V SCs.
are solved, a compromise is necessary to achieve acceptable In the following, the current understanding of several
quality of the TMD film. At present, the only way to achieve major issues that are key to tailoring the electrical proper-
ideal wafer-scale TMDs is still likely to be one of the previ- ties of 2D-TMDs and realizing high performance devices are
ously mentioned scalable methods. summarized. The scientific and technological communities
must work together on these issues to push TMD (as well as
2DLMs) electronics forward.
3. Electronic Transport in TMD Devices
The ultra-high carrier mobility and capability of size scaling 3.1. Doping
are more promising for graphene than for conventional Si.[97]
However, the extent to which we might improve the perfor- To apply TMDs to low-power, high-performance complemen-
mance of graphene-based FETs (G-FETs) is still unclear tary logic applications requires the development of both n-
because it lacks the bandgap required for digital electronic and p-type FETs. Therefore, doping is one of the most critical
applications. This zero bandgap prevents depletion of charge technologies in TMDs. In traditional host SCs, ion implanta-
carriers in G-FETs, inevitably resulting in high static power tion is a versatile technique that achieves spatially controlled
dissipation and low voltage gain for logic operation[97–100] as incorporation of donor or acceptor impurities in a control-
compared to the traditional Si-based CMOS configuration. lable manner. However, it is incompatible with 2DLMs due
The emerging 2D-TMDs, on the other hand, perfectly satisfy to their ultrathin structure. Hence, alternative doping tech-
the requirements for logic circuits. This is due to their suit- niques, discussed below, have been developed for TMDs. For
able bandgap range and ultrathin channel thickness, which practical applications it is still a challenge to keep the doping
allow superior gate control, large switching ratios, and capa- stable and highly efficient, and novel strategies that allow
bility of suppressing short channel effects.[101] However, early stable doping against changes in environmental factors such
work by Novesolov et al.[102] identified a field effect mobility as temperature and humidity are needed.
for single-layer MoS2 that was at least three orders of mag-
nitude lower than that of graphene. Years later interest 3.1.1. Electrostatic Gating
in TMD FETs was revived by the work of Radisavljevic
et al. in 2011,[103] which displayed top-gated SL-MoS2 FETs Recent studies have also shown that some of these 2D-TMDs
with moderate mobilities (≈60–70 cm2 V−1 s−1), large on/off show ambipolar behavior.[45,85,111–116] Thus local control of
ratios (≈108), and low subthreshold swings (74 mV dec−1) carrier types and densities of such devices can be electrostati-
at room temperature, creating huge opportunities in future cally tuned by both globally and locally patterned gate elec-
nanoelectronics. trodes to create a spatial modulation of the doping profile
Based on their unique 2D nature, 2D TMD devices have via solid gate dielectrics.[117–119] As shown in Figure 4a–b,[118]
electrical properties that are known to be highly sensitive to lateral P-N, P-P, and N-N junctions can be easily obtained
various effects that arise from surfaces and interfaces, thus by varying gate voltages of two independently patterned
providing versatile knobs for tuning these devices. Various back gate electrodes. Requiring additional voltage to local
modification methods have been applied to further physically gate structures limits the practical application in electronics.
or chemically tune 2DLMs based on their unique intrinsic However, such gate structures can still be regarded as a con-
properties. To date these include dimensional sizing (ver- venient experimental platform to explore electronic trans-
tical[44,103–106] and horizontal direction),[63,78,89] ion-intercala- port and optoelectronic properties of 2D materials without
tion,[48–54] application of an external field,[107–110] tuning the complex lateral PN junction formation by CVD or selective
stacking order,[15] and strain engineering,[13,14] etc. These fun- doping.
damental modification methods provide useful engineering Ion-gating is another straightforward and effective way
tools to improve TMD electronic devices. However, the to manipulate 2D-TMD electronic characteristics. The high
perfect 2D structure of TMDs is a double-edged sword that capacitance induced by formation of an electric double
brings extra challenges in practical application. This is mainly layer (EDL) on the material surface can tune the carrier
due to the absence of reliable and repeatable processing density up to 1015 cm−2 with a small gate bias, significantly

small 2017, 13, 1700098 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.small-journal.com (8 of 22) 1700098
 reviews www.advancedsciencenews.com

Figure 4. a–b) Local electrostatic gating of WSe2 transistor. a) Optical micrograph of a monolayer WSe2 device controlled by two local gates.
Bottom: schematic side view of the device including electrical connections. b) Current–voltage (Ids − Vds) curve under four doping configurations:
NN, PP, PN, and NP. Both gates were set to 10 V for the NN configuration and −10 V for PP. Vlg was set to +10 V and Vrg to +10 V for PN/NP. The NN
and PP configurations (yellow and black curves, respectively) are Ohmic at low Vds, while the PN and NP configurations (blue and green curves,
respectively) strongly rectify current in opposite directions. Reproduced with permission.[118] Copyright 2014, Nature Publishing Group. c–d) MoS2-
based EDLT device and its transistor properties. c) Double-gate device and measurement configuration. d) Transfer curve of transistor operations
by accumulating carriers by EDL top liquid gate (red: ramping VLG up; blue: ramping VLG down) and HfO2 bottom solid gate (green), both at 220 K.
c–d) Reproduced with permission.[122] Copyright 2012, AAAS.

larger than the limit of a solid dielectric gating (≈1013 cm−2). through interaction with adlayers, atoms, or molecules, as dis-
A typical experiment is shown in Figure 4c–d. Because the cussed below.
local Fermi level modulation has superior control capa- The air stable n-type dopants include SiNx[132] on WSe2
bility of the carrier density, it can be extended from the and Cs2CO3[133] on MoS2, which are inorganic coating mate-
conduction band to the valance band even in an intrinsic rials. Solution immersion of polyethyleneimine (PEI)[134]
n-type TMD FET (e.g., MoS2) and thus exhibit ambi- or benzyl viologen (BV) (Figure 5a)[127] have also proved
polar behavior.[108,120–122] Besides enhanced gate control, effective on few-layer MoS2; The p-type dopants on MoS2
the surface electric dipole layer near the electrodes can include MoO3 (Figure 5b),[128] gold nanoparticles (AuCl3)
reduce contact resistance by thinning the Schottky barrier (Figure 5c),[129] and tetracyanoquinodimethane (TCNQ).[135]
width.[108] The superiority of ion-gating has been exam- A detailed list of dopants can be found in a recent review
ined among a variety of 2D-TMDs, including WSe2,[109,110] and references therein.[136] Doping of TMD sensitively also
SnS2,[123] MoTe2[124,125] and SnSe2.[126] depends on the gaseous environment. High susceptibility
can be achieved for MoS2 by various vapors, including tri-
3.1.2. Charge Transfer Doping ethylamine (TEA), tetrahydrofuran (THF), and all polar
molecules such as acetone.[137] Gaseous NO2[138] is also
Charge transfer can lead to either electron or hole doping, favorable for inducing the n-type, while NH3 (Figure 5d)[130]
depending on the reduction potential of the doping species is favorable for p-type doping (Figure 5d). Other gaseous
with respect to the chemical potential of 2D-TMDs. Thus, the dopants include NO[139] (on MoS2) and O2[140] (on MoS2,
manipulation of optoelectronics and electrical properties is WSe2 and MoSe2).
correlated to the doping level enabled in 2D-TMDs. Charge For ambipolar TMCs with a low density of intrinsic sur-
transfer doping has been widely applied in 2D-TMDs mainly face defects (i.e., low Fermi pinning factor)[141] manipulation

1700098 (9 of 22) www.small-journal.com © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim small 2017, 13, 1700098
 www.advancedsciencenews.com

Figure 5. a–d) Transfer characteristics for different doping materials on TMDs. a) Few-layer MoS2 top-gate device before and after BV doping,
at bias Vds = 50 mV and 0.1 V. Reproduced with permission.[127] Copyright 2014, American Chemical Society. b) Back-gated few-layer MoS2
transistor before and after 0.1 nm MoO3 deposition, with bias Vds = 0.1V. Reproduced with permission.[128] Copyright 2013, AIP publishing LLC.
c) Few-layer MoS2 EDLT before and after Au nanoparticle treatment. Reproduced with permission.[129] Copyright 2013, Nature Publishing Group.
d) Monolayer CVD-MoS2 back-gated transistor with different NH3 exposure concentration, at bias Vds = 5V. Reproduced with permission.[130]
Copyright 2014, America Chemical Society. e) Schematic of MoS2 -Li micro-battery for in situ electrical measurement during intercalation.
f) Resistance versus time of MoS2 flake undergoing electrochemical lithiation at constant current. Insets show a schematic of the measurement
setup (upper right), and optical images of lithiated MoS2 at different intercalation stages. Reproduced with permission.[131] Copyright 2015,
John Wiley & Sons, Inc.

of electron/hole injection by contact metals with dif- 3.1.3. Substrate Effect and Doping
ferent work functions can also be regarded as a convenient
charge-transfer way to “dope” material and obtain n- or Ultrathin 2DLMs, are more sensitive to the substrate-mate-
p-type electrical characteristics. Besides, intercalated species rial interface, since the charge carriers cannot completely
also tend to induce charge transfers in the host 2D-TMDs screen the extrinsic potential disturbance. Previous studies
(Figure 5e–f).[131] Note that the doping capability is signifi- on graphene have demonstrated that its electrical proper-
cantly greater than those of the above-mentioned methods, ties can be tuned by interactions with different substrate
even greater than the doping level via electron-double-layer materials.[142–145] For TMDs studies have found that the
(EDL) gating with an ionic electrolyte or liquid. performance of MoS2 devices is greatly influenced by the
Doping stability varies among these approaches. In underlying species.[85,146,147] The self-assembled monolayer
the case of surface-adsorbed gas molecules, the resulting (SAM),[148] a bunch of molecules with different functional
doping effect is only temporary or “volatile,” and TMDs groups, provides a new platform for tuning electrical prop-
become immediately undoped once the gas molecules are erties of TMDs. Inserting SAM between TMD and the tra-
detached. It is also noted that the adsorbed small molecules ditional substrate (e.g., SiO2) is recognized as an effective
or coated metals on TMDs also tend to desorb from sur- substrate doping method. Its most competitive aspect is the
faces known to be sensitive to moisture and oxygen. In fact, flexibility to choose among various functional groups that
most reports show that the doping effects change with time will serve as the buffering layer to screen the scattering from
in the long term under ambient conditions (this is mainly traditional substrates. For example, SAMs with a functional
due to the pristine surface of a 2DLM that lacks dangling group of amine (NH3) and fluoro (CF3) can transfer
bonds). This makes it difficult to form strong interface electrons and holes to MoS2, respectively.[149] Other func-
bonds with dopants. Thus, the reversible physisorption pro- tional groups, including methyl (CH3), thiol (SH), have
cesses (even ion-intercalation,[54] which is not based on the also been systematically investigated.[150] Besides the doping
covalent chemical modification of the TMDs surfaces) are effect, SAM also effectively enables reduction of interface
usually not stable enough to ensure reliability in long-term trap states and SO phonon scattering, thus improving carrier
operation. mobility.[150]

small 2017, 13, 1700098 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.small-journal.com (10 of 22) 1700098
 reviews www.advancedsciencenews.com

When compared to organic SAMs, inorganic solid-state n- and p-type doping without changing the basic band
ultrathin films are more stable and compatible with semicon- structure. Other transition metals, such as group 5 element
ductor processing. For inorganic films the more important niobium and group 7 element rhenium, can also be incorpo-
factors affecting TMD devices are the Fermi level difference rated into composition during growth to obtain n- and p-type
and lattice mismatch that correspond to charge transfer and 2D-TMDs, respectively.[162] However, the alloying method
interface strain, respectively. Systematic studies have been still faces a spatial selective incorporation problem inher-
conducted to compare various substrate doping effects by ently limited by CVD or PVT growth. Other problems, such
depositing monolayer MoS2 on LaAlO3, SrTiO3, SiO2[151] and as defect-induced quality degradation, remain to be solved to
graphene,[152] and then using Raman and photoluminescence realize better device performance.
(PL) spectroscopy to identify structural strain and charge
transfer. The results indicate that charge transfer plays an
important role, and the PL intensity can be drastically altered 3.2. Contact Engineering
by different substrates. Such substrate dependence was also
studied in other films such as ZnO[153] and VO2.[154] Electron Achieving good contact is yet another key aspect that deter-
injection or extraction can effectively tune the exciton charge mines on-state current Ion and high-frequency performance
state that is correlated to modification of PL spectroscopy.[135] of TMD devices. Previous studies have shown that high
contact resistance stems from the special interface between
3.1.4. Alloying TMDs and 3D metal contact.[19,29,163] When the channel resist-
ance is relatively small compared to the contact resistance,
Alloying is an alternative method for stable doping of TMDs mobility and Ion are mainly governed by the contact rather
achieved by incorporating similar-element atoms into target than the channel itself. However, the interface between the
materials during the CVD or physical vapor transport (PVT) 2DLMs and metallic electrodes, whence the contact resist-
growth process. It provides a powerful tool to modulate the ance is mainly derived, is not only attributed to the formation
composition of a material’s chemical and physical properties of a Schottky barrier,[164] but also to other factors, including
(e.g., bandgap, carrier density) and is regarded as an effec- the existence of van der Waals (vdW) gaps, Fermi level pin-
tive substitute doping method in the 2DLM field.[155–158] For ning, hybridization between contacts and 2DLMs, etc.[19,26,27]
some TMDs with similar lattice structures, such as hexagonal
MoS2 and WS2, we can precisely control growth conditions 3.2.1. Interface between TMDs and Contact Electrodes
to form continuous van der Waals alloys, e.g., MoxW1-xS2
and MoS2(1-x) Se2x.[159–161] Since MoS2 and WS2 band edges Contact resistance is mainly formed at the interface
mainly rely on orbit superposition from d-orbitals of Mo and between TMDs and the 3D contact metal, and can be
W atoms, it is possible to alter the position of band edges divided into two types: top contact and edge contact, as
simply by tuning their compositions to further modulate shown in Figure 6a–b.[27] Top contact fabrication is facile

Figure 6. Schematic of a) Top-contact structure; b) edge contact structure. c) typical metal/bulk-SC contact interface with formation of covalence
bonding. d) Metal/2DSC contact interface with vdW gap. b) Representation of different injection mechanisms of typical metal/bulk-SC contacts.
From top to bottom: thermionic emission, thermionic field emission, and tunneling. d) Representation of injection mechanism of metal/2DSC
contact; only thermionic emission is available. a–f) Reproduced with permission.[27] Copyright 2015, Nature Publishing Group.

1700098 (11 of 22) www.small-journal.com © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim small 2017, 13, 1700098
 www.advancedsciencenews.com

and compatible with the traditional CMOS fabrication 3.2.2. Fermi Pinning Effect
process. However, traditional contact interface geometry
(Figure 6c,e) no longer applies to TMD devices due to their In an ideal situation, when the interface between
perfect 2D structure. When 2D-TMDs contact 3D-metal, 2D-TMDs and metal contacts is clean enough (without any
only the vdW gap forms between them due to the absence of defects or surface state), the Schottky barrier height (SBH)
covalent bonds on the pristine surfaces of TMDs, as shown ∅b is determined by the Schottky-Mott formula, which
in Figure 6d,f.[165] Charge carriers must overcome such a includes the metal’s work function Wm and electron affinity
“vdW gap barrier” between TMDs and the electrodes; and χ of TMDs[167]: ∅b = Wm − χ. However, recent reports have
the existence of additional such barriers can scatter some shown that the Schottky-Mott formula usually does not
transporting carriers, leading to high contact resistance. As apply to a TMD-metal interface,[19] primarily due to the
for edge contact, the metal atoms and edges of TMDs are Fermi pinning effect that is introduced by the surface state
stitched together by covalent bonds in the horizontal direc- and described by the Bardeen theory.[168,169] In the case of
tion, which allows carrier transport between the metal and bulk-SC when the surface states are filled by flowing elec-
TMDs without a “vdW gap barrier,” showing much lower trons, a depletion region forms near the surface as a surface
contact resistance.[166] dipole layer. Before direct contact the system, including
Another important difference between top contact and metal, bulk-SC, and the surface dipole layer, maintains
edge contact is the width of the Schottky barrier. When the electron neutrality with different Fermi levels of metal, sur-
carriers transport from a channel to a metal contact, a thin face state, and semiconductor EFM, EFSC, EFSS (Figure 7a),
Schottky barrier facilitates the tunneling through of carriers respectively. While the metal approaches the bulk-SC,
and their injection into metal contacts, decreasing contact the electrons will flow within the system to achieve ther-
resistance. Basic semiconductor (SC) physics tells us that the modynamic equilibrium (Figure 7b–c). With the existence
width of the Schottky barrier is proportional to the length of of a surface state, the final Fermi level is not determined
the depletion region on the carrier transport path. For the by metal and bulk-SC only, but instead is described by
top contact on bulk-SC, the carriers can be injected into con- Bardeen. In Bardeen theory, the SBH (q∅b) is determined
tacts from a non-depleted region and “choose” the minimum by Equation (1):
depletion region length (Figure 6c,e). On the contrary, due
to the atomic thickness of TMDs, the region under the metal
q∅ b = S (Wm − Ws′ ) + (Ws′ − χ ), (1)
contact is fully depleted, which means the carriers go through
a longer depletion region under metal contact, leading to a
thicker Schottky barrier (Figure 6d,f). As for edge contact, where S and W′s are the Fermi pinning factor and surface
a depleted region only appears at the metal contact-channel electron system work function, respectively. The Fermi pin-
interface, which is much thinner than that of the top contact. ning factor is given by Equation (2):

Figure 7. a–c): Schematic of the Fermi level alignment at the semiconductor-metal interface. a) Metal, surface state (S.S.), and semiconductor
(S.C.) are three independent electron systems maintaining electron neutrality. b) Electrons flow between S.S. and S.C. to achieve local equilibrium
and form a depletion region. c) When metal approaches the S.C. surface, electrons flow between metal and S.S to form surface potential difference
ΔV with electrical field depth δ. d) Schematic of MIGS, where metal electron wave functions “invade” the SC bandgap. The detailed mechanism
can be found in ref. [170] Reproduced with permission.[170] Copyright 2001, American Physical Society.

small 2017, 13, 1700098 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.small-journal.com (12 of 22) 1700098
 reviews www.advancedsciencenews.com

d∅ b 1 barrier height. When the applied bias qV ≫ kBT, the term

S = = ,
d∅ m e 2δ Dg (2)  qV 
exp  − can be ignored, and the expression for the current
1+
εi  k BT 
density simplifies to Equation (4):

where e is the electron charge, δ is the surface electrical field  q∅ ESB 

J ds = AT 2 exp  − (4)
depth, and Dg is the surface electron state density. For S = 0,  k BT 
the Schottky barrier height (q∅b) is pinned, and only depends
on W′s of the surface electron system and χ of TMDs, and not By using an Arrhenius plot (Figure 8c) the slope of the
on the Wm. For S = 1, the formula reverts to the Schottky- curve can be extracted. This slope gives the value of ∅ESB.
Mott relation, which is the ideal case. In Figure 8d ∅ESB is extracted under different gate voltages.
However, in the case of atomic thin TMD layers, the For a large reverse gate bias the source-to-channel current
definitions of “surface” and “bulk” become blurry. When is composed only of thermionic emission current, which cor-
surface states exist in the bandgap of TMDs, the depletion responds to the linear region. When the gate doping level is
region diffuses laterally instead of just in the vertical direc- larger than the flat band condition, ∅ESB deviates from lin-
tion, thus increasing the effective SBH. The surface states earity, because the Schottky barrier is thin enough to assist
are mainly induced by two factors: i) Surface defects. For quantum tunneling transport to further lower ∅ESB. Fol-
an ideal 2D atomic lattice that is free of dangling bonds, no lowing this principle, the flat band condition can be identified
surface states exist in the band structure. However, for some in Figure 8d, at which ∅ESB corresponds to the true SBH.
conditions (e.g., atom vacancy, surface contamination), the It should be noted that in this Schottky barrier analysis,
surface defects have been experimentally observed,[171–173] any temperature-dependent scattering mechanisms in the
thus introducing extra surface states. ii) Metal-induced gap channel are ignored, because the power law dependence on
state (MIGS) (Figure 7d).[174,175] The electron wave function temperature does not affect the exponential dependence of
of a metal contact can “invade” the bandgap of TMDs at the the Schottky barrier relation. Thus, the extraction of the slope
metal-TMD interface, further introducing electron states in of the Arrhenius plot can be regarded as a believable meas-
the bandgap, which can be charged by electrons to achieve urement of the Schottky barrier height.[19]
thermodynamic equilibrium.[165]
3.2.4. How to Achieve “Ohmic Contact”
3.2.3. How to Extract the Schottky Barrier Height
Experiments have been conducted to reduce contact resist-
Charge carriers must overcome various energy barriers ance with TMD channels, including chemical absorption
to transport from metal contact to the TMD channel. At doping, metal contact engineering, phase engineering, and
the metal-TMD interface, both thermionic emission and tunneling contact, etc.
quantum tunneling transport contribute to the carrier injec- Due to the atomically thin nature of TMDs, a substitu-
tion, and the ratio of the two transport processes depends on tion ion-implantation doping introduces tremendous damage
the gate doping level. As shown in Figure 8a, at a reversed to a TMD’s lattice, and the doping concentration is difficult
and subthreshold gate doping level, the main current com- to precisely control.[180] Metal contact engineering is a facile
ponent is a thermionic emission current, and the quantum and feasible method to ideally lower the Schottky barrier to
tunneling current is overwhelmed by a large Schottky bar- improve carrier injection. Many metals have been adopted
rier width. In this situation the quantum tunneling current is in TMD transistor fabrication, including Pt,[181] Au,[181,182]
only considered when the gate voltage is higher than the flat- Ni,[19,183] Sc,[19] In,[184] Al,[184] Ti,[185] Cr,[186] and Mo.[31] As
band voltage (VFB). When increasing the gate doping level, analyzed above, the Fermi pinning effect is inevitable, and
the quantum tunneling current enhances rapidly due to the thus the TMD-metal Schottky barrier deviates from the ideal
attenuation of Schottky barrier thickness and increase of lineup of the metal work function and 2DLM’s energy band
the density of states (DOS) in the channel. To calculate the (Figure 9a–b). Some attempts to introduce bonding between
Schottky barrier height, it is essential to identify VFB by the TMDs and metal contacts have been reported using Ti[182]
current-temperature relation in an Arrhenius plot to evaluate and Mo,[31] showing comparably low contact resistance of
the ratio of the two current components to the variation of 0.8 kΩμm and 2 kΩμm, respectively. However, a Ti contact
the gate doping level.[19] is only effective for MoS2; and a Mo contact is susceptible to
Assuming thermionic emission current can be described oxidation. In general it is difficult to achieve “Ohmic contact”
by classic thermionic emission theory, the current going by simply relying on a metal contact.
through the Schottky barrier is given by Equation (3):[176] Inspired by the mechanism of a spin injection layer,
inserting a buffering layer (including semimetal and insu-
 q∅ ESB    qV   lator) between the metal and TMDs can solve the dilemma
J ds = AT 2 exp  − 1 − exp  − (3)
 k BT    k BT   encountered in direct contact, as discussed above. When com-
pared to a traditional metal contact, a 2D “semimetal” family
where A is Richardson’s constant, kB is the Boltzmann con- (such as graphene, a kind of Dirac semimetal), because of
stant, T is temperature, V is the applied bias across the source a tunable work function and zero bandgap, is preferred to
and channel interface, and ∅ESB is the effective Schottky achieve low contact resistance. Chuang et al. have developed

1700098 (13 of 22) www.small-journal.com © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim small 2017, 13, 1700098
 www.advancedsciencenews.com

Figure 8. a) Schematic of injection mechanisms contributing to the working current with varying gating levels. The energy diagrams from top to
bottom show: in the reverse bias region, the gate doping level is below the flat band condition, where quantum tunneling current can be ignored;
The gate doping level is larger than the flat band condition, where thermionic emission and quantum tunneling current both contribute to the total
current. Reproduced with permission.[27] Copyright 2015, Nature Publishing Group. b) Transfer curve of back-gate MoS2 transistor with Ni contacts at
different temperatures used for Arrhenius-type plot. c) Arrhenius-type plot under different back-gate voltages. The effective Schottky barrier height
can be extracted by fitting the slope. d) The diagram of effective Schottky barrier height ∅ESB with different gate doping levels ∅GS. The point where
∅ESB begins to deviate from linearity is the flat band condition. b–d) Reproduced with permission.[19] Copyright 2013, American Chemical Society.

a few-layer WSe2 transistor with graphene electrodes and other hand, improved contact by deposition of an insulator
successfully achieved low resistance contact for both elec- has been experimentally demonstrated using ALD-grown
tron and hole injection by gate tuning the work function of Ta2O5 (Figure 9c)[32] and CVD-grown BN (Figure 9d)[190]
graphene, i.e., a gate controlled Schottky barrier.[187] Roy as tunneling layers on MoS2; the results are shown in
et al. first fabricated MoS2-based “all-2D” devices using gra- Figure 9c–d. The improved performance can be explained by
phene as source/drain electrodes to improve contact resist- MIGS surface density attenuation and defects in the insu-
ance.[188] A similar method has been used in MoS2 FETs lator that facilitate the electron tunneling process.
to achieve a low resistance value of 3.7 kΩμm.[189] In these As mentioned in Section 3.1, chemical adsorption
cases low resistance can be achieved for three main reasons: doping is another effective way to reduce the Schottky bar-
i) the work function of graphene can be effectively tuned by rier without introducing defects to a TMD surface. Similar
a gate voltage that minimizes the Schottky barrier height; to the Si-based self-aligned doping process that uses a top-
ii) WSe2 is an ambipolar TMD with low density of intrinsic gated structure to localize the doping region near the con-
surface defects; and iii) an MIGS surface state is totally tacts and realize selective treatment, Fang et al. conducted
eliminated by the graphene-electrode structure where WSe2 potassium[177] (Figure 9e–f) and NO2[138] doping on TMDs
is no longer directly underneath the metal contact. On the with ZrO2 as a top gate dielectric, with both n- and p-type

small 2017, 13, 1700098 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.small-journal.com (14 of 22) 1700098
 reviews www.advancedsciencenews.com

Figure 9. a) Lineup of metal Fermi level with the electronic bands of MoS2, considering the electron affinity of MoS2 and the work function of
corresponding metals. Reprinted with permission from ref. [173] Copyright 2014, American Chemical Society. b) Transfer curve of back-gate MoS2
transistor with different electrodes. Reproduced with permission.[19] Copyright 2013, American Chemical Society. c) Measured contact resistance
as a function of Ta2O3 thickness. Reproduced with permission.[32] Copyright 2015, American Chemical Society. d) Calculated Schottky barrier height
for 1≈2 layers BN. e–f) Reproduced with permission.[177] Copyright 2016, John Wiley & Sons, Inc. e) Schematic of a top-gated few-layer WSe2 n-FET
with chemically n-doped S/D contacts by K exposure. f) Transfer characteristics of a device before and after NO2 patterned doping of the S/D
contacts. e–f) Reproduced with permission.[177] Copyright 2012, American Chemical Society. g) Schematic models of single-layered MoS2 with 2H
(left) and 1T (right) phases in basal plane and cross-section views. Reproduced with permission.[178] Copyright 2014, Nature Publishing Group. h)
Electrostatic force microscopy phase image of a monolayer MoS2 nanosheet showing the difference between locally patterned 2H (bright color)
and 1T (dark color) phases. i) Transfer characteristics of top-gated devices, measured at Vds = 1V (logarithmic scale on the left and linear scale on
the right). Blue curves represent devices with 1T phase electrodes and black curves are with Au on the 2H phase. Reproduced with permission.[179]
Copyright 2014, Nature Publishing Group.

performance observed, respectively. All of this shows evi- Until now phase engineering has been the most effec-
dence of degenerate doping level in the S/D region and tive method, stemming from the fact that TMDs can exist
reduction of contact resistance. Some air-stable doping mate- in several different structure phases. The phase engineering
rials have also been demonstrated. Benzyl viologen (BV), a experiment has been conducted for MoS2 by means of the
promising candidate of ambient-stable n-doping material lithium-intercalation chemical method, which induces a tran-
for carbon nanotube and graphene transistors, can create a sition from the semiconducting 2H phase to a metastable
degenerate n-type doping of MoS2 and reduce contact resist- 1T metallic phase.[193,194] The atomic structure difference
ance by a factor greater than 3.[127] Some inorganic materials is shown in Figure 9g. The 1T phase is transformed from
have also been used, including amorphous titanium suboxide the 2H phase by structure destabilization due to the inter-
(ATO)[191] and chloride (Cl) molecules.[192] However, as dis- calated lithium ions.[195] The first application of phase engi-
cussed above, the TMD doping strategy is still immature, and neering that realized a low-contact-resistance exfoliated and
it is useful to explore stable, effective, and convenient doping CVD MoS2 transistor was demonstrated by R. Kappera and
methods for future TMD applications. his co-workers.[26,196] Metal contact was deposited on a 1T

1700098 (15 of 22) www.small-journal.com © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim small 2017, 13, 1700098
 www.advancedsciencenews.com

phase MoS2 (Figure 9h), achieving ultra-low contact resist- creating image charges in the surrounding dielectric materials.
ance (≈200–300 Ωμm) and enhancing device performance This is much different than the bulk-SC case (Figure 10b–c).
(Figure 9i) for both few-layer exfoliation and monolayer Theoretical analysis of monolayer MoS2 has shown that CIs
CVD MoS2 at zero gate bias. The ultra-low contact resistance located inside MoS2 (z = 0) can induce infinite image-charge
originates from the atomically sharp, coherent boundary of arrays at point zn = na, where n is an integer, with a charge
1T/2H phases without obvious defects,[136,196,197] thus facili- magnitude of eγ |n|, where γ = (εs − εe)/(εs + εe).[209,210] Thus
tating carrier injection from channel to contact region. the unscreened Coulomb potential of CI is determined by
Other methods, such as ultra-high vacuum (UHV) depo- both the relative dielectric of MoS2 (εs) and the surrounding
sition of contacts, have been proposed.[198] The UHV method materials (εe). By considering the electron screening effect in
achieved a resistance of Au contact ≈700 Ωμm at room tem- MoS2, the induced charge density can be determined by the
perature without doping. This work uncovered the influence boundary condition of the interface between MoS2 and sur-
of adsorbates (impurities) trapped at the metal-MoS2 inter- rounding material using the Gauss theorem:
face, showing the importance of obtaining a clean interface in
∂φ ind ∂φ ind (5)
the fabrication of a TMD transistor. σ ind (r ) = Dz (0 + ) − Dz (0 − ) = − ε s1 + ε s2
∂z ∂z
In conclusion, the contact resistance of TMDs is intensely 0+ 0−

affected by the material surface environment that include

where εs1 and εs2 are dielectric constants of the top and
intrinsic (lattice defects) and extrinsic (adsorbates) factors,
bottom material, respectively. The screening function of
while the work function of the metal contact is less important
MoS2 can be written in a form similar to that of bulk-SC:
than it is for bulk semiconductors.
e2
ε 2 d (q, ω ) = 1 − Π(q, ω ),[205] where the difference here
s1 (ε + ε )q
s2

is the use of the 2D wave vector q instead of the 3D wave

3.3. Scattering Mechanisms vector Q. Then, intrinsic CI scattering can be derived from
the Fermi Golden theory, where the CI scattering rate  1 
As promising materials for post CMOS electronics, TMDs  τ CI 
have undergone intensive studies that focused on the physical has a positive correlation with the 2D Fourier transform of
understanding of carrier transport in order to achieve high Coulomb potential: V (q, z) scr = V (q , z) bare /ε 2 d (q, ω ) . That is
performance electronic devices. Well investigated scattering why using high-k dielectric material can reduce CI scattering
sources in graphene, including charge impurity (CI)[199,200] scat- and enhance device mobility.[103,211,212]
tering, surface optical (SO) phonon scattering,[142,201] and sur- Extrinsic CI scattering is also called remote CI scat-
face roughness[202] from the substrate, play equivalent roles in tering, because the CI source is outside the material itself.
TMD devices. For example, experimentally achieved mobility The remote CI interacts with carriers through potential fluc-
of MoS2 is much lower than the theoretical prediction in the tuation induced by the short-range Coulomb force, which
intrinsic phonon-limited transport regime, which can be up to alters the local energy band, leading to carrier localization
410 cm2 V−1 s−1.[203,204] How these scattering sources can tailor puddles that have been observed in graphene.[202] In a low
device performance will be discussed below in detail. gate doping level or with thickness that is less than the
Thomas-Fermi screening depth, the carrier localization can
3.3.1. Charge Impurity Scattering greatly impact mobility, because the carriers transport only
by hopping (Figure 10d).[171] As the gate doping or thick-
Charge impurity (CI) represents static charges or dipoles
ness increases, high carrier density will screen the influence
that generate a constant electrical field. The sources of CI
of remote CI and smooth the energy band edge to increase
include impurity defects in the crystal lattice (e.g., sulfur
mobility. On the other hand, if the remote CI is arrayed
vacancy in MoS2[172]), namely intrinsic CI sources; dangling
continuously and conformably, the carrier density can be
bonds of substrate or dielectric layer (e.g., SiO2[171]); and
enhanced by a strong doping effect without encumbering
adsorbed polar molecules (e.g., H2O[198]) that are classified as
mobility. It has already been demonstrated that inserting a
extrinsic CI sources. In bulk-SCs, when point charges such as
polar molecule layer (PMMA, CYTOP, PVP) can modify
donor or acceptor atoms locate inside the material, the Cou-
carrier density and threshold voltage without lowering
lomb potential can be completely screened by surrounding
mobility.[85,213]
electrons so that the electrical field of CI can be restricted
inside the SCs (Figure 10a). According to electron screening
Remote Phonon Scattering
theory, the Fourier transform of the screened CI poten-
tial can be described as V (q, z) scr = V (q, z) bare/ε 3d (q, ω ),
Surface optical (SO) phonon scattering, which is also
and the dielectric function of bulk-SC can be given as
e2 called remote phonon scattering, is induced by oscilla-
ε 3d (Q, ω ) = 1 − Π(Q, ω ),[208] where Q is a three-dimensional
ε 0ε bQ 2 tion of dipoles from the substrate and dielectric layer. This
wave vector, εb is the relative dielectric constant of bulk-SC, type of scattering generates an electromagnetic field and
and Π(Q, ω) is the polarization function of the electron gas. interacts with moving carriers in a TMD channel. The role
This formula indicates that the screened Coulomb potential of SO phonon scattering in thin bulk-SC film[214] and gra-
of point charges inside a bulk-SC is lowered only by its own phene[142,215] has been investigated and found to be a non-
dielectric constant. For atomic-thickness TMDs, the Coulomb negligible influence on carrier transport, thus degrading
field of CI is not totally restricted to within the TMDs, thus carrier mobility. Recently, experimental exploration of SO

small 2017, 13, 1700098 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.small-journal.com (16 of 22) 1700098
 reviews www.advancedsciencenews.com

Figure 10. Comparison of electron screening between the 3D and 2D cases, a) in bulk-SC (bulk MoS2), the electrical field of a charge impurity (CI)
can be screened by surrounding electrons, and b) in 2DSC (monolayer MoS2), the electrical field of a CI cannot be totally screened in the vertical
direction. c) For monolayer MoS2, the Coulomb field of a CI is not completely restricted inside TMDs, creating image charges in the surrounding
dielectric materials. Reproduced with permission.[205] Copyright 2013, American Physical Society. d) In low carrier density or with thickness thinner
than Thomas-Fermi screening depth, a remote CI can modify the bandgap edge, inducing carrier localization and influencing mobility. Reproduced
with permission.[171] Copyright 2012, American Chemical Society. e–f) Demonstration of room temperature phonon-limited transport. e) Four-probe
conductivity as a function of Vg for monolayer MoS2 device on HfO2 substrate (device H1). Inset shows device structure. f) Field-effect mobility as a
function of temperature for three devices on SiO2 (black), Al2O3 (green), and HfO2 (red), respectively, under n = 7.1 × 1012 cm−2. Solid lines are the
modeling results. Reproduced with permission.[206] Copyright 2016, John Wiley & Sons, Inc. g) Electron mobility as a function of an environmental
dielectric constant at high temperature (300K, red line) and relatively low temperature (100K, dark line). h) The dielectric-dependent electron
mobility at different temperatures with impurity and carrier density of 1013 cm−2. G–h) Reproduced with permission.[207] Copyright 2014, American
Physical Society.

phonon scattering in room temperature transport behavior can be calculated. The result is shown in Figure 10g, where
of MoS2 has been conducted, as shown in Figure 10e–f.[206] the dashed lines show mobility without SO phonon scat-
By adopting thiol-treated MoS2 (minimizing sulfur vacancy) tering, enhanced by dielectric screening of CI scattering. The
and a high-k dielectric, CI scattering is suppressed to a very solid lines indicate mobility obtained by considering both CI
low level. Thus room temperature phonon-limited transport and SO phonon scattering. In Figure 10h the temperature
can be observed by fabricating a monolayer MoS2 transistor dependence of net mobility that considers both CI scattering
with a room-temperature mobility around 150 cm2 V−1 s−1. and SO phonon scattering is shown as black curves, with
Experiment shows that the surface optical (SO) phonons red and blue lines standing for mobility that considers sepa-
adjacent to high-κ dielectrics play a crucial role in limiting rately the SO phonon and CI scattering, respectively, indi-
charge transport in atomically thin TMDs. A corresponding cating that SO phonon scattering is the main factor limiting
theoretical investigation of SO scattering in TMDs has mobility at room temperature.
been conducted by combining density functional theory
(DFT)[203,204] and relaxation-time approximation (RTA) of 3.3.2. Grain Boundary Scattering
the Boltzmann transport equation.[207] Theoretical calcula-
tions show that SO phonon scattering is the most significant The role of the grain boundary (GB), a collective 1-D line
factor at high temperatures, and phonon-limited mobility defect linked with differently oriented single crystals, is

1700098 (17 of 22) www.small-journal.com © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim small 2017, 13, 1700098
 www.advancedsciencenews.com

essential for the electronic transport behavior of TMDs. of problems related to TMD-based electronic device
Studies on the morphology of graphene GBs and its impact application.
on transport properties have been extensively carried
out.[216–219] GBs of CVD grown TMDs have also been exper-
imentally observed, including MoS2,[220–224] MoSe2,[225,226]
WSe2[227,228] and WS2,[229–233] whereas a systematic inves-
tigation of their roles in electronic transport property is Acknowledgements
lacking. An early experimental study on MoS2 found that
the mirror twin and tilt boundaries show opposite contri- X.-F.S. and Z.-X.G. contributed equally to this work. This work was
butions to the electrical conductivity.[234] By performing a supported by the National Key Research and Development Program
temperature-dependent study on the two different types (2016YFA0203900) and National Natural Science Foundation of
of GBs of MoS2, the charge transport mechanism can be China (61376093, 61622401).
unveiled by adopting the Anderson model, with the dis-
order effect originating from the accumulation of point or
other structural defects at these GBs.[235] More detailed Conflict of Interest
experimental investigations on MoS2 GBs brought out
that the misorientation-angle-dependent carrier mobility The authors declare no conflict of interest.
across GBs arises from the angle-dependent-variation of
the potential barrier through GBs, which is supported by
a DFT calculation.[236] This is also consistent with pre-
[1] A. K. Geim, K. S. Novoselov, Nat. Mater. 2007, 6, 183.
vious high-resolution STM and spectroscopy measurement
[2] K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang,
results.[237] S. V. Dubonos, I. V. Grigorieva, A. A. Firsov, Science 2004, 306,
666.
[3] A. K. Geim, Science 2009, 324, 1530.
4. Conclusion [4] F. Schwierz, Nature Nanotechnol. Nat. Nanotechnol. 2010, 5,
487.
Recent advances in various semiconducting TMDs have led [5] J.-H. Lee, E. K. Lee, W.-J. Joo, Y. Jang, B.-S. Kim, J. Y. Lim,
S.-H. Choi, S. J. Ahn, J. R. Ahn, M.-H. Park, C.-W. Yang, B. L. Choi,
to promising electronics applications. However, most pro-
S.-W. Hwang, D. Whang, Science 2014, 344, 286.
gress has still occurred in the laboratory, and engineering [6] F. Bonaccorso, L. Colombo, G. Yu, M. Stoller, V. Tozzini,
problems urgently require a solution for practical applica- A. C. Ferrari, R. S. Ruoff, V. Pellegrini, Science 2015, 347,
tions. In this review we summarized and compared var- 1246501.
ious preparation methods of TMD materials, such as CVD [7] ITRS International Technology Roadmap for Semiconductors,
methods that are suitable for wafer-scale fabrication. To date http://www.itrs.net (accessed: July 2016).
[8] S. Bhattacharyya, A. K. Singh, Phys. Rev. Phys. Rev. B 2012, 86,
studies indicate there is still significant room to improve
075454.
growth parameters or explore new synthetic methods. [9] H. Terrones, M. Terrones, 2D Mater. 2014, 1, 011003.
Besides preparing the material, electronic transport prop- [10] J. Lu, O. Zheliuk, I. Leermakers, N. F. Yuan, U. Zeitler, K. T. Law,
erties of TMDs that are crucial for practical applications in J. Ye, Science 2015, 350, 1353.
electronic devices can be further tailored by band structure [11] W. Zhao, Z. Ghorannevis, L. Chu, M. Toh, C. Kloc, P.-H. Tan,
engineering or device processing procedures that include G. Eda, ACS Nano 2012, 7, 791.
engineering of doping, contact, and scattering. All of these [12] K. F. Mak, C. Lee, J. Hone, J. Shan, T. F. Heinz, Phys. Rev. Lett.
2010, 105, 136805.
are regarded as primary challenges for future practical elec-
[13] M. Ghorbani-Asl, S. Borini, A. Kuc, T. Heine, Phys. Rev. Phys.
tronics application of TMDs. Thus, to clarify opportunities Rev. B 2013, 87, 235434.
and challenges for future TMD-based electronic devices, it [14] Y. Wang, C. Cong, W. Yang, J. Shang, N. Peimyoo, Y. Chen,
is of great importance to develop a comprehensive under- J. Kang, J. Wang, W. Huang, T. Yu, Nano Res. 2015, 8, 2562.
standing of the large-scale growth and novel electronic [15] J. He, K. Hummer, C. Franchini, Phys. Rev. Phys. Rev. B 2014, 89,
transport properties of TMDs, as well as the correlation 075409.
between them. This will bridge the gap between fundamental [16] Y. Liu, N. O. Weiss, X. Duan, H.-C. Cheng, Y. Huang, X. Duan, Nat.
Rev. Mater. 2016, 1, 16042.
research and engineering applications. Going forward, much
[17] D. Sarkar, X. Xie, W. Liu, W. Cao, J. Kang, Y. Gong, S. Kraemer,
more work is needed to attack many issues related to engi- P. M. Ajayan, K. Banerjee, Nature 2015, 526, 91.
neering application, which includes improving the synthesis [18] F. Withers, O. Del Pozo-Zamudio, A. Mishchenko, A. Rooney,
and transfer of 2D-TMD films on arbitrary substrates; opti- A. Gholinia, K. Watanabe, T. Taniguchi, S. Haigh, A. Geim,
mizing device structures together with improving processing A. Tartakovskii, K. S. Novoselov, Nat. Mater. 2015, 14, 301.
techniques; as well as exploring appropriate applications [19] S. Das, H.-Y. Chen, A. V. Penumatcha, J. Appenzeller, Nano Lett.
2013, 13, 100.
such as flexible and low-power electronics that make more
[20] Y. Hao, M. Bharathi, L. Wang, Y. Liu, H. Chen, S. Nie,
direct use of benefits from the “atomically thin nature”. A X. Wang, H. Chou, C. Tan, B. Fallahazad, H. W. Ramanarayan,
singular breakthrough in one of these areas is certainly inad- C. Magnuson, E. I. Tutuc, B. F. Yakobson, K. McCarty,
equate and hopefully, more comprehensive progress will be Y.-W. Zhang, P. Kim, J. Hone, L. S. Colombo, R. Ruoff, Science
achieved in the coming years, providing a thorough solution 2013, 342, 72.

small 2017, 13, 1700098 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.small-journal.com (18 of 22) 1700098
 reviews www.advancedsciencenews.com

[21] T. Wu, X. Zhang, Q. Yuan, J. Xue, G. Lu, Z. Liu, H. Wang, H. Wang, [55] A. Castellanos-Gomez, M. Barkelid, A. M. Goossens,
F. Ding, Q. Yu, Nat. Mater. 2016, 15, 43. V. E. Calado, V. D. Z. Hs, G. A. Steele, Nano Lett. 2012, 12, 3187.
[22] Q. Ji, Y. Zhang, Y. Zhang, Z. Liu, Chem. Soc. Rev. 2015, 44, 2587. [56] Z. Rafiei-Sarmazdeh, S. H. Jafari, S. J. Ahmadi,
[23] Y. Shi, H. Li, L.-J. Li, Chem. Soc. Rev. 2015, 44, 2744. S. M. Zahedi-Dizaji, J. Mater. Sci. 2016, 51, 3162.
[24] A. M. V. D. Zande, P. Y. Huang, D. A. Chenet, T. C. Berkelbach, [57] S. Lin, Y. Chui, Y. Li, S. P. Lau, FlatChem 2017, 2, 15.
Y. M. You, G. H. Lee, T. F. Heinz, D. R. Reichman, D. A. Muller, [58] E. O. Ortiz-Quiles, C. R. Cabrera, FlatChem 2017, 2, 1.
J. C. Hone, Nat. Mater. 2013, 12, 554. [59] J. Zhu, H. Wang, J. Liu, L. Ouyang, M. Zhu, Nanotechnology
[25] Y. Zhan, Z. Liu, S. Najmaei, P. M. Ajayan, J. Lou, Small 2012, 8, 2017, 28, 115604.
966. [60] X. Li, W. Cai, J. An, S. Kim, J. Nah, D. Yang, R. Piner,
[26] R. Kappera, D. Voiry, S. E. Yalcin, B. Branch, G. Gupta, A. Velamakanni, I. Jung, E. Tutuc, Science 2009, 324, 1312.
A. D. Mohite, M. Chhowalla, Nat. Mater. 2014, 13, 1128. [61] C. Mattevi, H. Kim, M. Chhowalla, J. Mater. Chem. 2010, 21,
[27] A. Allain, J. Kang, K. Banerjee, A. Kis, Nat. Mater. 2015, 14, 415.
1195. [62] X. Ling, Y.-H. Lee, Y. Lin, W. Fang, L. Yu, M. S. Dresselhaus,
[28] I. Popov, G. Seifert, D. Tománek, Phys. Rev. Lett. 2012, 108, J. Kong, Nano Lett. 2014, 14, 464.
489. [63] Y. H. Lee, X. Q. Zhang, W. Zhang, M. T. Chang, C. T. Lin,
[29] H. Liu, M. Si, S. Najmaei, A. T. Neal, Y. M. Du, P. Ajayan, J. Lou, K. D. Chang, Y. C. Yu, T. W. Wang, C. S. Chang, L. J. Li, Adv. Mater.
D. P. Ye, Nano Lett. 2013, 13, 2640. 2012, 24, 2320.
[30] J. R. Chen, P. M. Odenthal, A. G. Swartz, G. C. Floyd, H. Wen, [64] Y.-H. Lee, L. Yu, H. Wang, W. Fang, X. Ling, Y. Shi, C.-T. Lin,
K. Y. Luo, R. K. Kawakami, Nano Lett. 2013, 13, 3106. J.-K. Huang, M.-T. Chang, C.-S. Chang, M. Dresselhaus,
[31] J. Kang, W. Liu, K. Banerjee, Appl. Phys. Lett. 2014, 104, T. Palacios, L. J. Li, J. Kong, Nano Lett. 2013, 13, 1852.
093106. [65] S. Najmaei, Z. Liu, W. Zhou, X. Zou, G. Shi, S. Lei, B. I. Yakobson,
[32] S. Lee, A. Tang, S. Aloni, H. S. Philip Wong, Nano Lett. 2016, 16, J. C. Idrobo, P. M. Ajayan, J. Lou, Nat. Mater. 2013, 12, 754.
276. [66] J. Jeon, S. K. Jang, S. M. Jeon, G. Yoo, Y. H. Jang, J. H. Park,
[33] C. H. Chen, C. L. Wu, J. Pu, M. H. Chiu, P. Kumar, T. Takenobu, S. Lee, Nanoscale 2015, 7, 1688.
L. J. Li, 2D Mater. 2014, 1, 034001. [67] C. Cong, J. Shang, X. Wu, B. Cao, N. Peimyoo, C. Qiu, L. Sun,
[34] S. Chuang, C. Battaglia, A. Azcatl, S. McDonnell, J. S. Kang, T. Yu, Adv. Opt. Mater. 2014, 2, 131.
X. Yin, M. Tosun, R. Kapadia, H. Fang, R. M. Wallace, Nano Lett. [68] Y. Zhang, Y. Zhang, Q. Ji, J. Ju, H. Yuan, J. Shi, T. Gao, D. Ma,
2014, 14, 1337. M. Liu, Y. Chen, ACS Nano 2013, 7, 8963.
[35] M. R. Laskar, D. N. Nath, L. Ma, E. W. LeeII, C. H. Lee, T. Kent, [69] Y. Gong, G. Ye, S. Lei, G. Shi, Y. He, J. Lin, X. Zhang, R. Vajtai,
Z. Yang, R. Mishra, M. A. Roldan, J.-C. Idrobo, Appl. Phys. Lett. S. T. Pantelides, W. Zhou, Adv. Funct. Mater. 2016, 26, 2009.
2014, 104, 092104. [70] Y. Yu, C. Li, Y. Liu, L. Su, Y. Zhang, L. Cao, Sci. Rep. 2012, 3,
[36] D. Akinwande, N. Petrone, J. Hone, Nature Commun. 2014, 5, 1866.
5678. [71] Y. Kobayashi, S. Sasaki, S. Mori, H. Hibino, Z. Liu, K. Watanabe,
[37] R. Cheng, S. Jiang, Y. Chen, Y. Liu, N. Weiss, H.-C. Cheng, H. Wu, T. Taniguchi, K. Suenaga, Y. Maniwa, Y. Miyata, ACS Nano 2015,
Y. Huang, X. Duan, Nat. Commun. 2014, 5, 5143. 9, 4056.
[38] X. Peng, L. Peng, C. Wu, Y. Xie, Chem. Soc. Rev. 2014, 43, 3303. [72] L. Zhou, K. Xu, A. Zubair, A. D. Liao, W. Fang, F. Ouyang,
[39] M. Bosi, RSC Adv. 2015, 5, 75500. Y. H. Lee, K. Ueno, R. Saito, T. Palacios, J. Am. Chem. Soc. 2015,
[40] X. Duan, C. Wang, A. Pan, R. Yu, X. Duan, Chem. Soc. Rev. 2015, 137, 11892.
44, 8859. [73] J. C. Park, S. J. Yun, H. Kim, J. H. Park, S. H. Chae, S. J. An,
[41] D. Voiry, A. Mohite, M. Chhowalla, Chem. Soc. Rev. 2015, 44, J. G. Kim, S. M. Kim, K. K. Kim, Y. H. Lee, ACS Nano 2015, 9,
2702. 6548.
[42] A. Splendiani, L. Sun, Y. Zhang, T. Li, J. Kim, C. Y. Chim, G. Galli, [74] A. Tarasov, P. M. Campbell, M. Y. Tsai, Z. R. Hesabi, J. Feirer,
F. Wang, Nano Lett. 2010, 10, 1271. S. Graham, W. J. Ready, E. M. Vogel, Adv. Funct. Mater. 2014, 24,
[43] S. Bertolazzi, J. Brivio, A. Kis, ACS Nano 2011, 5, 9703. 6389.
[44] Z. Yin, H. Li, H. Li, L. Jiang, Y. Shi, Y. Sun, G. Lu, Q. Zhang, [75] Y. C. Lin, W. Zhang, J. K. Huang, K. K. Liu, Y. H. Lee, C. T. Liang,
X. Chen, H. Zhang, ACS Nano 2012, 6, 74. C. W. Chu, L. J. Li, Nanoscale 2012, 4, 6637.
[45] Y. Zhang, J. Ye, Y. Matsuhashi, Y. Iwasa, Nano Lett. 2012, 12, [76] J. G. Song, J. Park, W. Lee, T. Choi, H. Jung, C. W. Lee,
1136. S. H. Hwang, J. M. Myoung, J. H. Jung, S. H. Kim, ACS Nano
[46] J. N. Coleman, M. Lotya, A. O’Neill, S. D. Bergin, P. J. King, 2013, 7, 11333.
U. Khan, K. Young, A. Gaucher, S. De, R. J. Smith, Science 2011, [77] X. Wang, H. Feng, Y. Wu, L. Jiao, J. Am. Chem. Soc. 2013, 135,
331, 568. 5304.
[47] R. J. Smith, P. J. King, M. Lotya, C. Wirtz, U. Khan, S. De, [78] K. K. Liu, W. Zhang, Y. H. Lee, Y. C. Lin, M. T. Chang, C. Y. Su,
A. O’Neill, G. S. Duesberg, J. C. Grunlan, G. Moriarty, Adv. Mater. C. S. Chang, H. Li, Y. Shi, H. Zhang, Nano Lett. 2012, 12, 1538.
2011, 23, 3944. [79] Y. Zhan, Z. Liu, S. Najmaei, P. M. Ajayan, J. Lou, Small 2012, 8,
[48] M. Osada, T. Sasaki, J. Mater. Chem. 2009, 19, 2503. 966.
[49] G. Eda, H. Yamaguchi, D. Voiry, T. Fujita, M. Chen, M. Chhowalla, [80] C. M. Orofeo, S. Suzuki, Y. Sekine, H. Hibino, Appl. Phys. Lett.
Nano Lett. 2011, 11, 5111. 2014, 105, 083112.
[50] Z. Zeng, Z. Yin, X. Huang, H. Li, Q. He, G. Lu, F. Boey, H. Zhang, [81] G. Tai, T. Zeng, J. Yu, J. Zhou, Y. You, X. Wang, H. Wu, X. Sun,
Angew. Chem. 2011, 50, 11093. T. Hu, W. Guo, Nanoscale 2015, 8, 2234.
[51] Z. Zeng, T. Sun, J. Zhu, X. Huang, Z. Yin, G. Lu, Z. Fan, Q. Yan, [82] H. R. Gutiérrez, N. Perea-López, A. L. Elías, A. Berkdemir,
H. H. Hng, H. Zhang, Angew. Chem. Int. Ed. 2012, 51, 9052. B. Wang, R. Lv, F. López-Urías, V. H. Crespi, H. Terrones,
[52] Z. Jian, Z. Han, S. Dong, Y. Liu, T. N. Chang, H. S. Shin, Y. J. Hu, M. Terrones, Nano Lett. 2013, 13, 3447.
L. Bo, K. P. Loh, Nat. Commun. 2014, 5, 149. [83] S. Hussain, M. A. Shehzad, D. Vikraman, M. F. Khan, J. Singh,
[53] N. Liu, P. Kim, J. H. Kim, J. H. Ye, S. Kim, C. J. Lee, ACS Nano D. C. Choi, Y. Seo, J. Eom, W. G. Lee, J. Jung, Nanoscale 2016, 8,
2014, 8, 6902. 1.
[54] J. Wan, S. D. Lacey, J. Dai, W. Bao, M. S. Fuhrer, L. Hu, Chem. [84] K. Y. Ko, J. G. Song, Y. Kim, T. Choi, S. Shin, C. W. Lee, K. Lee,
Soc. Rev. 2016, 45, 6742. J. Koo, H. Lee, J. Kim, ACS Nano 2016, 10, 9287.

1700098 (19 of 22) www.small-journal.com © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim small 2017, 13, 1700098
 www.advancedsciencenews.com

[85] W. Bao, X. Cai, D. Kim, K. Sridhara, M. S. Fuhrer, Appl. Phys. [117] J. S. Ross, P. Klement, A. M. Jones, N. J. Ghimire, J. Yan,
Lett. 2012, 102, 042104. D. Mandrus, T. Taniguchi, K. Watanabe, K. Kitamura, W. Yao,
[86] K. Kang, S. Xie, L. Huang, Y. Han, P. Y. Huang, K. F. Mak, Nature Nanotechnol. 2014, 9, 268.
C. J. Kim, D. Muller, J. Park, Nature 2015, 520, 656. [118] B. W. Baugher, H. O. Churchill, Y. Yang, P. Jarillo-Herrero, Nature
[87] A. Koma, K. Sunouchi, T. Miyajima, J. Vac. Sci. Technol., Nanotechnol. 2014, 9, 262.
B: Microelectron. Process. Phenom. 1985, 3, 724. [119] A. Pospischil, M. M. Furchi, T. Mueller, Nature Nanotechnol.
[88] A. Koma, J. Cryst. Growth 1999, 201, 236. 2014, 9, 257.
[89] Y. Shi, W. Zhou, A.-Y. Lu, W. Fang, Y.-H. Lee, A. L. Hsu, S. M. Kim, [120] Y. Zhang, J. Ye, Y. Yomogida, T. Takenobu, Y. Iwasa, Nano Lett.
K. K. Kim, H. Y. Yang, L.-J. Li, Nano Lett. 2012, 12, 2784. 2013, 13, 3023.
[90] Q. Ji, Y. Zhang, T. Gao, Y. Zhang, D. Ma, M. Liu, Y. Chen, X. Qiao, [121] L. Chu, H. Schmidt, J. Pu, S. Wang, B. Özyilmaz, T. Takenobu,
P.-H. Tan, M. Kan, Nano Lett. 2013, 13, 3870. G. Eda, Sci. Rep. 2013, 4, 7293.
[91] R. Yue, A. T. Barton, H. Zhu, A. Azcatl, L. F. Pena, J. Wang, [122] J. Ye, Y. Zhang, R. Akashi, M. Bahramy, R. Arita, Y. Iwasa, Science
X. Peng, N. Lu, L. Cheng, R. Addou, ACS Nano 2014, 9, 474. 2012, 338, 1193.
[92] H. C. Diaz, R. Chaghi, Y. Ma, M. Batzill, 2D Mater. 2015, 2, [123] H. Yuan, M. Toh, K. Morimoto, W. Tan, F. Wei, H. Shimotani,
044010. C. Kloc, Y. Iwasa, Appl. Phys. Lett. 2011, 98, 012102.
[93] X. Liu, I. Balla, H. Bergeron, G. P. Campbell, M. J. Bedzyk, [124] W. Shi, J. Ye, Y. Zhang, R. Suzuki, M. Yoshida, J. Miyazaki,
M. C. Hersam, ACS Nano 2016, 10, 1067. N. Inoue, Y. Saito, Y. Iwasa, Sci. Rep. 2015, 5, 12534.
[94] Z. Ben Aziza, H. Henck, D. Pierucci, M. G. Silly, E. Lhuillier, [125] H. Xu, S. Fathipour, E. W. Kinder, A. C. Seabaugh,
G. Patriarche, F. Sirotti, M. Eddrief, A. Ouerghi, ACS Nano 2016, S. K. Fullerton-Shirey, ACS Nano 2015, 9, 4900.
10, 9679. [126] T. Pei, L. Bao, G. Wang, R. Ma, H. Yang, J. Li, C. Gu, S. Pantelides,
[95] X. Zhang, F. Meng, J. R. Christianson, C. Arroyo-Torres, S. Du, H. Gao, Appl. Phys. Lett. 2016, 108, 053506.
M. A. Lukowski, D. Liang, J. R. Schmidt, S. Jin, Nano Lett. 2014, [127] D. Kiriya, M. Tosun, P. Zhao, J. S. Kang, A. Javey, J. Am. Chem.
14, 3047. Soc. 2014, 136, 7853.
[96] M. Donley, N. McDevitt, T. Haas, P. Murray, J. Grant, Thin Solid [128] J. Lin, J. Zhong, S. Zhong, H. Li, H. Zhang, W. Chen, Appl. Phys.
Films 1989, 168, 335. Lett. 2013, 103, 063109.
[97] K. S. Novoselov, V. Fal, L. Colombo, P. Gellert, M. Schwab, [129] Y. Shi, J.-K. Huang, L. Jin, Y.-T. Hsu, S. F. Yu, L.-J. Li, H. Y. Yang,
K. Kim, Nature 2012, 490, 192. Sci. Rep. 2013, 3, 1839.
[98] D. Jariwala, V. K. Sangwan, L. J. Lauhon, T. J. Marks, [130] B. Liu, L. Chen, G. Liu, A. N. Abbas, M. Fathi, C. Zhou, ACS Nano
M. C. Hersam, Chem. Soc. Rev. 2013, 42, 2824. 2014, 8, 5304.
[99] D. Jariwala, A. Srivastava, P. M. Ajayan, J. Nanosci. Nanotechnol. [131] J. Wan, W. Bao, Y. Liu, J. Dai, F. Shen, L. Zhou, X. Cai, D. Urban,
2011, 11, 6621. Y. Li, K. Jungjohann, Adv. Energy Mater. 2015, 5, 1401742.
[100] A. H. C. Neto, F. Guinea, N. M. R. Peres, K. S. Novoselov, [132] K. Chen, D. Kiriya, M. Hettick, M. Tosun, T.-J. Ha,
A. K. Geim, Rev. Mod. Phys. 2009, 81, 109. S. R. Madhvapathy, S. Desai, A. Sachid, A. Javey, APL Mater.
[101] Y. Yoon, K. Ganapathi, S. Salahuddin, Nano Lett. 2011, 11, 3768. 2014, 2, 092504.
[102] K. Novoselov, D. Jiang, F. Schedin, T. Booth, V. Khotkevich, [133] J. D. Lin, C. Han, F. Wang, R. Wang, D. Xiang, S. Qin, X.-A. Zhang,
S. Morozov, A. Geim, Proc. Natl. Acad. Sci. U. S. A. 2005, 102, L. Wang, H. Zhang, A. T. S. Wee, ACS Nano 2014, 8, 5323.
10451. [134] Y. Du, H. Liu, A. T. Neal, M. Si, D. Y. Peide, IEEE Electron Device
[103] B. Radisavljevic, A. Radenovic, J. Brivio, i. V. Giacometti, A. Kis, Lett. 2013, 34, 1328.
Nature Nanotechnol. Nanotechnol. 2011, 6, 147. [135] S. Mouri, Y. Miyauchi, K. Matsuda, Nano Lett. 2013, 13, 5944.
[104] T. Cao, G. Wang, W. Han, H. Ye, C. Zhu, J. Shi, Q. Niu, P. Tan, [136] H. Schmidt, F. Giustiniano, G. Eda, Chem. Soc. Rev. 2015, 44,
E. Wang, B. Liu, Nat. Commun. 2012, 3, 887. 7715.
[105] K. F. Mak, K. He, J. Shan, T. F. Heinz, Nature Nanotechnol. 2012, [137] F. K. Perkins, A. L. Friedman, E. Cobas, P. Campbell, G. Jernigan,
7, 494. B. T. Jonker, Nano Lett. 2013, 13, 668.
[106] H. Zeng, J. Dai, W. Yao, D. Xiao, X. Cui, Nature Nanotechnol. [138] H. Fang, S. Chuang, T. C. Chang, K. Takei, T. Takahashi, A. Javey,
2012, 7, 490. Nano Lett. 2012, 12, 3788.
[107] J. Pu, Y. Yomogida, K.-K. Liu, L.-J. Li, Y. Iwasa, T. Takenobu, Nano [139] H. Li, Z. Yin, Q. He, H. Li, X. Huang, G. Lu, D. W. H. Fam,
Lett. 2012, 12, 4013. A. I. Y. Tok, Q. Zhang, H. Zhang, Small 2012, 8, 63.
[108] M. M. Perera, M.-W. Lin, H.-J. Chuang, B. P. Chamlagain, [140] S. Tongay, J. Zhou, C. Ataca, J. Liu, J. S. Kang, T. S. Matthews,
C. Wang, X. Tan, M. M.-C. Cheng, D. Tománek, Z. Zhou, ACS L. You, J. Li, J. C. Grossman, J. Wu, Nano Lett. 2013, 13, 2831.
Nano 2013, 7, 4449. [141] S. Xu, Z. Wu, H. Lu, Y. Han, G. Long, X. Chen, T. Han, W. Ye,
[109] H. Yuan, X. Wang, B. Lian, H. Zhang, X. Fang, B. Shen, G. Xu, Y. Wu, J. Lin, 2D Mater. 2016, 3, 021007.
Y. Xu, S.-C. Zhang, H. Y. Hwang, Nature Nanotechnol. 2014, 9, [142] J. H. Chen, C. Jang, S. Xiao, M. Ishigami, M. S. Fuhrer, Nature
851. Nanotechnol. Nat. Nanotechnol. 2008, 3, 206.
[110] A. Allain, A. Kis, ACS Nano 2014, 8, 7180. [143] B. Lee, Y. Chen, F. Duerr, D. Mastrogiovanni, E. Garfunkel,
[111] V. Podzorov, M. Gershenson, C. Kloc, R. Zeis, E. Bucher, Appl. E. Y. Andrei, V. Podzorov, Nano Lett. 2010, 10, 2427.
Phys. Lett. 2004, 84, 3301. [144] J. C. Meyer, A. K. Geim, M. I. Katsnelson, K. S. Novoselov,
[112] H. Fang, S. Chuang, T. C. Chang, K. Takei, T. Takahashi, A. Javay, T. J. Booth, S. Roth, Nature 2007, 446, 60.
Nano Lett. 2012, 12, 3788. [145] Z. Yan, Z. Sun, W. Lu, J. Yao, Y. Zhu, J. M. Tour, ACS Nano 2011,
[113] D. Braga, I. Gutiérrez Lezama, H. Berger, A. F. Morpurgo, Nano 5, 1535.
Lett. 2012, 12, 5218. [146] A. Ayari, E. Cobas, O. Ogundadegbe, M. S. Furthrer, J. Appl.
[114] S. Das, J. Appenzeller, Appl. Phys. Lett. 2013, 103, 103501. Phys. 2007, 101, 014507.
[115] N. R. Pradhan, D. Rhodes, Y. Xin, S. Memaran, L. Bhaskaran, [147] K. Dolui, I. Rungger, S. Sanvito, Phys. Rev. Phys. Rev. B 2013,
M. Siddiq, S. M. Hill, P. Ajayan, L. Balicas, ACS Nano 2014, 8, 87, 165402.
7923. [148] T. Kondo, R. Yamada, K. Uosaki, in Organized Organic Ultrathin
[116] Y. F. Lin, Y. Xu, S. T. Wang, S. L. Li, M. Yamamoto, Films, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany
A. Aparecido-Ferreira, W. Li, H. Sun, S. Nakaharai, W. B. Jian, 2013, pp 7–42.
Adv. Mater. 2014, 26, 3263. [149] Y. Li, C.-Y. Xu, P. Hu, L. Zhen, ACS Nano 2013, 7, 7795.

small 2017, 13, 1700098 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.small-journal.com (20 of 22) 1700098
 reviews www.advancedsciencenews.com

[150] S. Najmaei, X. Zou, D. Er, J. Li, Z. Jin, W. Gao, Q. Zhang, S. Park, [184] W. Liu, J. Kang, D. Sarkar, Y. Khatami, D. Jena, K. Banerjee, Nano
L. Ge, S. Lei, J. Kono, V. B. Shenoy, B. I. Yakobson, A. George, Lett. 2013, 13, 1983.
P. M. Ajayan, J. Lou, Nano Lett. 2014, 14, 1354. [185] H. Liu, M. Si, Y. Deng, A. T. Neal, Y. Du, S. Najmaei, P. M. Ajayan,
[151] Y. Li, Z. Qi, M. Liu, Y. Wang, X. Cheng, G. Zhang, L. Sheng, J. Lou, P. D. Ye, ACS Nano 2014, 8, 1031.
Nanoscale 2014, 6, 15248. [186] H. Wang, L. Yu, Y.-H. Lee, Y. Shi, A. Hsu, M. L. Chin, L.-J. Li,
[152] D. Pierucci, H. Henck, C. H. Naylor, H. Sediri, E. Lhuillier, M. Dubey, J. Kong, T. Palacios, Nano Lett. 2012, 12,
A. Balan, J. E. Rault, Y. J. Dappe, F. Bertran, P. L. Fèvre, 4674.
A. T. C. Johnson, A. Ouerghi, Sci. Rep. 2016, 6, 26656. [187] H.-J. Chuang, X. Tan, N. J. Ghimire, M. M. Perera, B. Chamlagain,
[153] M. S. Kim, S. Roy, J. Lee, B. G. Kim, H. Kim, J.-H. Park, S. J. Yun, M. M.-C. Cheng, J. Yan, D. Mandrus, D. Tománek, Z. Zhou, Nano
G. H. Han, J.-Y. Leem, J. Kim, ACS Appl. Mater. Interfaces 2016, Lett. 2014, 14, 3594.
8, 28809. [188] T. Roy, M. Tosun, J. S. Kang, A. B. Sachid, S. B. Desai, M. Hettick,
[154] J. Hou, X. Wang, D. Fu, C. Ko, Y. Chen, Y. Sun, S. Lee, K. X. Wang, C. C. Hu, A. Javey, ACS Nano 2014, 8, 6259.
K. Dong, Y. Sun, S. Tongay, L. Jiao, J. Yao, K. Liu, J. Wu, Small [189] Y. Du, L. Yang, J. Zhang, H. Liu, K. Majumdar, P. D. Kirsch,
2016, 12, 3976. P. D. Ye, IEEE Electron Device Lett. 2014, 35, 599.
[155] A. Kutana, E. S. Penev, B. I. Yakobson, Nanoscale 2014, 6, 5820. [190] J. Wang, Q. Yao, C. W. Huang, X. Zou, L. Liao, S. Chen, Z. Fan,
[156] J. Mann, Q. Ma, P. M. Odenthal, M. Isarraraz, D. Le, E. Preciado, K. Zhang, W. Wu, X. Xiao, Adv. Mater. 2016, 28, 8302.
D. Barroso, K. Yamaguchi, G. von Son Palacio, A. Nguyen, Adv. [191] M. Montazeri, A. Wade, M. Fickenscher, H. E. Jackson,
Mater. 2014, 26, 1399. L. M. Smith, J. M. Yarrison-Rice, Q. Gao, H. H. Tan, C. Jagadish,
[157] Y. C. Lin, D. O. Dumcenco, H. P. Komsa, Y. Niimi, Nano Lett. 2011, 11, 4329.
A. V. Krasheninnikov, Y. S. Huang, K. Suenaga, Adv. Mater. 2014, [192] L. Yang, K. Majumdar, H. Liu, Y. Du, H. Wu, M. Hatzistergos,
26, 2857. P. Hung, R. Tieckelmann, W. Tsai, C. Hobbs, Nano Lett. 2014, 14,
[158] I. S. Kim, V. K. Sangwan, D. Jariwala, J. D. Wood, S. Park, 6275.
K.-S. Chen, F. Shi, F. Ruiz-Zepeda, A. Ponce, M. Jose-Yacaman, [193] S. J. Sandoval, D. Yang, R. Frindt, J. Irwin, Phys. Rev. Phys. Rev.
ACS Nano 2014, 8, 10551. B 1991, 44, 3955.
[159] D. Dumcenco, K. Chen, Y. Wang, Y. Huang, K. Tiong, J. Alloys [194] P. J. Mulhern, Can. J. Phys. 1989, 67, 1049.
Compd. 2010, 506, 940. [195] M. Chhowalla, H. S. Shin, G. Eda, L.-J. Li, K. P. Loh, H. Zhang,
[160] M. Zhang, J. Wu, Y. Zhu, D. O. Dumcenco, J. Hong, N. Mao, Nature chemistry 2013, 5, 263.
S. Deng, Y. Chen, Y. Yang, C. Jin, ACS Nano 2014, 8, 7130. [196] R. Kappera, D. Voiry, S. E. Yalcin, W. Jen, M. Acerce, S. Torrel,
[161] Y. Chen, J. Xi, D. O. Dumcenco, Z. Liu, K. Suenaga, D. Wang, B. Branch, S. Lei, W. Chen, S. Najmaei, APL Mater. 2014, 2,
Z. Shuai, Y.-S. Huang, L. Xie, ACS Nano 2013, 7, 4610. 092516.
[162] J. Wilson, A. Yoffe, Adv. Phys. 1969, 18, 193. [197] G. Eda, T. Fujita, H. Yamaguchi, D. Voiry, M. Chen, M. Chhowalla,
[163] H. Liu, A. T. Neal, P. D. Ye, ACS Nano 2012, 6, 8563. ACS Nano 2012, 6, 7311.
[164] D. Jena, K. Banerjee, G. H. Xing, Nat. Mater. 2014, 13, 1076. [198] C. D. English, G. Shine, V. E. Dorgan, K. C. Saraswat, E. Pop,
[165] J. Kang, W. Liu, D. Sarkar, D. Jena, K. Banerjee, Phys. Rev. Phys. Nano Lett. 2016, 16, 3824.
Rev. X 2014, 4, 031005. [199] J.-H. Chen, C. Jang, S. Adam, M. Fuhrer, E. Williams, M. Ishigami,
[166] L. Wang, I. Meric, P. Huang, Q. Gao, Y. Gao, H. Tran, T. Taniguchi, Nat. Phys. 2008, 4, 377.
K. Watanabe, L. Campos, D. Muller, Science 2013, 342, 614. [200] E. Hwang, S. Adam, S. D. Sarma, Phys. Rev. Lett. 2007, 98,
[167] S. M. Sze, K. K. Ng, Physics of Semiconductor Devices, 3rd ed., 186806.
John Wiley & Sons, Hoboken, NJ, USA 2007. [201] S. Fratini, F. Guinea, Phys. Rev. Phys. Rev. B 2008, 77, 195415.
[168] J. Bardeen, Phys. Rev. 1947, 71, 717. [202] J. Martin, N. Akerman, G. Ulbricht, T. LOhmann, J. V. Smet, K. Von
[169] A. Cowley, S. Sze, J. Appl. Phys. 1965, 36, 3212. Klitzing, A. Yacoby, Nat. Phys. 2008, 4, 144.
[170] R. T. Tung, Phys. Rev. B 2001, 64, 205310. [203] X. Li, J. T. Mullen, Z. Jin, K. M. Borysenko, M. B. Nardelli,
[171] S. Ghatak, A. N. Pal, A. Ghosh, ACS Nano 2011, 5, 7707. K. W. Kim, Phys. Rev. Phys. Rev. B 2013, 87, 115418.
[172] J. Hong, Z. Hu, M. Probert, K. Li, D. Lv, X. Yang, L. Gu, N. Mao, [204] K. Kaasbjerg, K. S. Thygesen, K. W. Jacobsen, Phys. Rev. Phys.
Q. Feng, L. Xie, Nat. Commun. 2015, 6, 6293. Rev. B 2012, 85, 115317.
[173] S. McDonnell, R. Addou, C. Buie, R. M. Wallace, C. L. Hinkle, ACS [205] Z.-Y. Ong, M. V. Fischetti, Phys. Rev. Phys. Rev. B 2013, 88,
Nano 2014, 8, 2880. 165316.
[174] V. Heine, Phys. Rev. 1965, 138, 1689. [206] Z. Yu, Z. Y. Ong, Y. Pan, Y. Cui, R. Xin, Y. Shi, B. Wang, Y. Wu,
[175] J. Tersoff, Phys. Rev. Lett. 1984, 52, 465. T. Chen, Y. W. Zhang, Adv. Mater. 2016, 28, 547.
[176] A. Anwar, B. Nabet, J. Culp, F. Castro, J. Appl. Phys. 1999, 85, [207] N. Ma, D. Jena, Phys. Rev. Phys. Rev. X 2014, 4, 011043.
2663. [208] J. H. Davies, The Physics of Low-Dimensional Semiconductors:
[177] H. Fang, M. Tosun, G. Seol, T. C. Chang, K. Takei, J. Guo, A. Javey, An Introduction, 1st ed., Cambridge University Press,
Nano Lett. 2013, 13, 1991. Cambridge, UK 1997.
[178] Y.-C. Lin, D. O. Dumcenco, Y.-S. Huang, K. Suenaga, Nature [209] D. Jena, A. Konar, Phys. Rev. Lett. 2007, 98, 136805.
Nanotechnol. 2014, 9, 391. [210] A. Konar, D. Jena, J. Appl. Phys. 2007, 102, 123705.
[179] T. Li, B. Wan, G. Du, B. Zhang, Z. Zeng, AIP Advances 2015, 5, [211] S. Kim, A. Konar, W.-S. Hwang, J. H. Lee, J. Lee, J. Yang, C. Jung,
057102. H. Kim, J.-B. Yoo, J.-Y. Choi, Nat. Commun. 2012, 3, 1011.
[180] L. Pelaz, R. Duffy, M. Aboy, L. Marques, P. Lopez, I. Santos, [212] D. Lembke, A. Kis, ACS Nano 2012, 6, 10070.
B. Pawlak, M. Van Dal, B. Duriez, T. Merelle, in IEEE Int. Electron [213] G. Yoo, S. L. Choi, S. Lee, B. Yoo, S. Kim, M. S. Oh, Appl. Phys.
Devices Meet., IEEE, San Francisco, CA, USA 2008, pp 1–4. Lett. 2016, 108, 263106.
[181] S. Walia, S. Balendhran, Y. Wang, R. Ab Kadir, A. S. Zoolfakar, [214] K. Hess, P. Vogl, Solid State Commun. 1979, 30, 797.
P. Atkin, J. Z. Ou, S. Sriram, K. Kalantar-zadeh, M. Bhaskaran, [215] A. Konar, T. Fang, D. Jena, Phys. Rev. Phys. Rev. B 2010, 82,
Appl. Phys. Lett. 2013, 103, 232105. 115452.
[182] W. Liu, J. Kang, W. Cao, D. Sarkar, Y. Khatami, D. Jena, [216] X. Zou, Y. Liu, B. I. Yakobson, Nano Lett. 2013, 13, 253.
K. Banerjee, in IEEE Int. Electron Devices Meet., (Ed: W. Liu), [217] O. V. Yazyev, S. G. Louie, Nat Mater 2010, 9, 806.
IEEE, Washington, DC, USA 2013; pp 19.4.1–19.4.4. [218] A. W. Tsen, L. Brown, M. P. Levendorf, F. Ghahari, P. Y. Huang,
[183] S. Larentis, B. Fallahazad, E. Tutuc, Appl. Phys. Lett. 2012, 101, R. W. Havener, C. S. Ruizvargas, D. A. Muller, P. Kim, J. Park,
223104. Science 2012, 336, 1143.

1700098 (21 of 22) www.small-journal.com © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim small 2017, 13, 1700098
 www.advancedsciencenews.com

[219] X. Li, C. W. Magnuson, A. Venugopal, J. An, J. W. Suk, B. Han, [229] X. Zou, B. I. Yakobson, Small 2015, 11, 4503.
M. Borysiak, W. Cai, A. Velamakanni, Y. Zhu, Nano Lett. 2014, [230] Y. Zhang, Y. Zhang, Q. Ji, J. Ju, H. Yuan, J. Shi, T. Gao, D. Ma,
10, 4328. M. Liu, Y. Chen, ACS Nano 2013, 7, 8963.
[220] A. E. Yore, K. K. H. Smithe, W. Crumrine, A. Miller, J. A. Tuck, [231] Y. Rong, Y. Sheng, M. Pacios, X. Wang, Z. He, H. Bhaskaran,
B. Redd, E. Pop, B. Wang, A. K. Newaz, J. Phys. Chem. C 2016, J. H. Warner, ACS Nano 2016, 10, 1093.
120, 24080. [232] Y. Rong, K. He, M. Pacios, A. W. Robertson, H. Bhaskaran,
[221] J. Shi, Y. Yang, Y. Zhang, D. Ma, W. Wei, Q. Ji, Y. Zhang, X. Song, J. H. Warner, ACS Nano 2015, 9, 3695.
T. Gao, C. Li, Adv. Funct. Mater. 2015, 25, 842. [233] A. Azizi, X. Zou, P. Ercius, Z. Zhang, G. Stone, M. Terrones,
[222] M. C. Prado, R. Nascimento, B. E. Faria, M. J. Matos, H. Chacham, B. I. Yakobson, N. Alem, Nat. Commun. 2014, 5,
B. R. Neves, Nanotechnology 2015, 26, 475702. 4867.
[223] S. Najmaei, Z. Liu, W. Zhou, X. Zou, G. Shi, S. Lei, B. I. Yakobson, [234] V. D. Z. Am, P. Y. Huang, D. A. Chenet, T. C. Berkelbach, Y. You,
J. C. Idrobo, P. M. Ajayan, J. Lou, Nat. Mater. 2013, 12, 754. G. H. Lee, T. F. Heinz, D. R. Reichman, D. A. Muller, J. C. Hone,
[224] C. Jingxin, J. Tao, J. Qingqing, Z. Yu, L. Zhiming, S. Yuwei, Nature Mater. 2013, 12, 554.
Z. Yanfeng, G. Xingao, L. Weitao, W. Shiwei, Adv. Mater. 2015, [235] S. Najmaei, M. Amani, M. L. Chin, Z. Liu, A. G. Birdwell,
27, 4069. T. P. O’Regan, P. M. Ajayan, M. Dubey, J. Lou, ACS Nano 2014, 8,
[225] Y. Ma, H. C. Diaz, J. Avila, C. Chen, V. Kalappattil, R. Das, 7930.
M. H. Phan, T. Cadez, J. M. Carmelo, M. C. Asensio, Nat. [236] L. T. Hue, D. J. Perello, J. Zhao, Q. Deng, K. Hyun, G. H. Han,
Commun. 2017, 8, 14231. C. S. Hoon, J. H. Yun, L. Y. Hee, Nat. Commun. 2016, 7,
[226] L. Jiao, H. J. Liu, J. L. Chen, Y. Yi, W. G. Chen, Y. Cai, J. N. Wang, 10426.
X. Q. Dai, N. Wang, W. K. Ho, New J. Phys. 2015, 17, 053023. [237] Y. L. Huang, Y. Chen, W. Zhang, S. Y. Quek, C. H. Chen, L. J. Li,
[227] Y. L. Huang, Z. Ding, W. Zhang, Y. H. Chang, Y. Shi, L. J. Li, W. T. Hsu, W. H. Chang, Y. J. Zheng, W. Chen, Nat. Commun.
Z. Song, Y. J. Zheng, D. Chi, Y. Q. Su, Nano Lett. 2016, 16, 2015, 6, 6298.
3682.
[228] H. Ago, S. Fukamachi, H. Endo, P. Solísfernández, R. M. Yunus, Received: January 9, 2017
Y. Uchida, V. Panchal, O. Kazakova, M. Tsuji, ACS Nano 2016, 10, Revised: May 13, 2017
3233. Published online: July 19, 2017

small 2017, 13, 1700098 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.small-journal.com (22 of 22) 1700098