REVIEW

2D Layered Materials www.advmat.de

Recent Progress in CVD Growth of 2D Transition Metal

Dichalcogenides and Related Heterostructures
Yu Zhang, Yuyu Yao, Marshet Getaye Sendeku, Lei Yin, Xueying Zhan, Feng Wang,
Zhenxing Wang, and Jun He*

about many breakthroughs due to the

In recent years, 2D layered materials have received considerable research abundant electronic, physical and mechan-
interest on account of their substantial material systems and unique physico- ical properties within their 2D planes.[4–12]
chemical properties. Among them, 2D layered transition metal dichalcogenides As one striking example of 2D atomic-
(TMDs), a star family member, have already been explored over the last few layer crystals, transition metal dichalco-
years and have exhibited excellent performance in electronics, catalysis, and genides (TMDs) have been explored for a
few years because of their tunable band-
other related fields. However, to fulfill the requirement for practical application, gaps from ≈1–2 eV, blooming a series of
the batch production of 2D TMDs is essential. Recently, the chemical vapor dep- excellent properties and performances
osition (CVD) technique was considered as an elegant alternative for success- in electronic devices and energy conver-
fully growing 2D TMDs and their heterostructures. The latest research advances sion.[7,13–20] The in-plane of each layer in
in the controllable synthesis of 2D TMDs and related heterostructures/superlat- TMDs is held together by strong chemical
bonds, while adjacent layers are coupled
tices via the CVD approach are illustrated here. The controlled growth behavior,
by a weak van der Waals (vdW) interac-
preparation strategies, and breakthroughs on the synthesis of new 2D TMDs tion.[21] This available structure allows
and their heterostructures, as well as their unique physical phenomena, are TMD materials to be easily exfoliated
also discussed. Recent progress on the application of CVD-grown 2D materials into monolayers or few layers. Monolayer
is revealed with particular attention to electronics/optoelectronic devices and TMDs possess remarkable properties,
catalysts. Finally, the challenges and future prospects are considered regarding and molybdenum disulfide (MoS2) can
be considered as the best example among
the current development of 2D TMDs and related heterostructures. TMD family members. For instance, the
bulk MoS2 crystal possesses an indirect
bandgap. However, a transition from
1. Introduction indirect bandgap to direct bandgap surprisingly occurs in the
monolayer and is accompanied by a strong photoluminescence
Since the advent of graphene,[1,2] 2D atomic crystals[3] have (PL) emission.[22–25] Moreover, monolayer MoS2 has a bandgap
gained significant attention over past decades and have brought structure that is tunable with external strain and a strong light–
matter interaction,[15,26] which promote its excellent electronic
and optoelectronic performance in 2D devices.[27–30] Further-
Dr. Y. Zhang, Y. Yao, M. G. Sendeku, L. Yin, X. Zhan, Dr. F. Wang,
Prof. Z. Wang, Prof. J. He more, the proper bandgap and metallic edges of MoS2 also
CAS Center for Excellence in Nanoscience inspire its application in hydrogen evolution reaction (HER) as
CAS Key Laboratory of Nanosystem and Hierarchical Fabrication an efficient electrocatalyst or photocatalyst.[31–36] On the other
National Center for Nanoscience and Technology hand, the broken inversion symmetry and strong spin–orbit
Beijing 100190, China
E-mail: [email protected] coupling in monolayer MoS2 have brought a new research
Y. Yao opportunity to the field of valleytronics.[37–41] Generally, all these
Sino-Danish College unique physicochemical properties, combined with appreci-
University of Chinese Academy of Science able charge mobility, have made 2D TMD materials promising
Beijing 100049, China candidates for deep investigations into fundamental physical
Y. Yao, M. G. Sendeku, L. Yin issues and high-performance electronics. Consequently, few-
Center of Materials Science and Optoelectronics Engineering
layer MoS2 has been obtained by the exfoliation method and
University of Chinese Academy of Science
Beijing 100049, China fabricated into field-effect transistor (FET) devices, exhibiting
Prof. J. He a significant electronic performance,[42–45] that forecast a vital
School of Physics and Technology potential application of 2D TMDs. On top of that, a plethora
Wuhan University of investigations devoted to exfoliated 2D TMD materials have
Wuhan 430072, China been communicated,[46–52] successfully receiving some impor-
The ORCID identification number(s) for the author(s) of this article tant breakthroughs. However, the exfoliation method always
can be found under https://doi.org/10.1002/adma.201901694.
affords 2D TMD flakes of a random thickness and domain size,
DOI: 10.1002/adma.201901694 and the technique is also time-consuming, making its mass

Adv. Mater. 2019, 1901694 1901694 (1 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

production challenging. Since the discovery of the chemical

vapor deposition (CVD) method as a successful approach for Yu Zhang received her
the growth of graphene crystals,[53,54] the method has also been B.S. degree in the School
expanded to the preparation of other 2D layered materials. Cur- of Physical Science and
rently, this approach has been extended to rationally design var- Technology from the Central
ious 2D TMD materials, such as MoS2 nanosheets,[55–58] MoSe2 China Normal University,
nanosheets,[59,60] WS2 nanosheets,[61,62] ReS2 nanosheets,[63] and in 2011. Then, she received
some heterostructures,[64,65] which have sparkled several inter- her Ph.D. degree from the
esting chemical and electronic properties. Although various College of Engineering (COE)
2D TMDs have been successfully grown using a facile CVD from the Peking University,
method, the intrinsic growth mechanism and the batch produc- China, in 2016. Currently, she
tion of wafer-scale films, conducive to the high-performance is working at the National
in electronics/optoelectronics and other fields, still need more Center for Nanoscience and
exploration. Very recently, some important progress on the syn- Technology (NCNST) as an assistant professor. Her major
thesis of 2D TMDs, multi-heterostructures and superlattices research interests include 2D nanomaterials synthesis,
has been realized. characterization, and their applications.
Herein, we concentrate on the recent progress in the syn-
thesis of 2D TMDs and related heterostructures via the CVD
method. We start with the effects of some key factors such as Yuyu Yao received her B.S.
temperature, carrier gas, and the ratio of precursors on the syn- degree in material physics
thesis of 2D TMDs via CVD technique. Then we summarize from the University of Jinan
the importance of the growth substrates and the growth system in 2017. Presently, she is a
to large-scale synthesis. Hence, large-scale 2D TMDs have been Master candidate in Prof. Jun
obtained on the basis of properly designed CVD growth. Fur- He’s group at NCNST. Her
thermore, some recent breakthroughs in the growth of new current research interests
kinds of semiconducting and metallic 2D TMDs via the CVD mainly focus on the con-
technique are described. The fundamental electronic properties trolled growth of 2D advanced
and unique physical phenomena emanating from high quality materials via chemical vapor
CVD-grown TMDs are also reviewed. Finally, the controlled deposition (CVD) and their
growth of alloys with tunable band structures, different multi- performance studies.
heterostructures with atomically sharp interfaces and superlat-
tices from 2D TMD crystals is analyzed, which can offer further Jun He received his Ph.D.
step for the synthesis of 2D crystals and provides significantly in semiconductor physics
more material systems. Lastly, we emphasize a summary of this from the Institute of
review and the future prospects of these materials. Semiconductors, Chinese
Academy of Sciences (CAS),
in 2003. Then, he worked
2. 2D Semiconducting TMDs successively at the Applied
Physics Department of
2.1. Growth Control and System Design Technische Universiteit
Eindhoven, Netherlands, the
2.1.1. Key Factors in the CVD Growth Process Material Department of the
University of California, Santa
Recently, the CVD method has been widely regarded as a tech- Barbara, and the California NanoSystem Insitute (CNSI),
nique available for growing 2D atomic crystals. Various atomi- University of California, Los Angeles, USA. He joined the
cally thin layered MX2 semiconductors, such as MoS2,[55,66,67] “100-Talents” Program of CAS in November 2010 and
WS2,[68] MoSe2,[69] MoTe2,[70,71] etc., have been successfully became a full-time professor of NCNST since then. His
achieved by the CVD method. Through in-depth research on main research interests are the synthesis, characterization,
the growth dynamics in the CVD system, several key factors and devices of low-dimensional semiconductor materials.
in the growth process for obtaining a large-scale, uniform
monolayer MX2 have been explored. For example, trian-
gular WS2 flakes were synthesized on a sapphire substrate
via a CVD system, and a detailed analysis of the controllable When the other growth parameters (growth temperature, car-
growth behavior was given.[61] The growth system is main- rier gas and growth time) remain constant, a small Dss value
tained at a very low pressure during the growth process via always produces few-layer WS2 flakes with a large domain size
an operating pump, which provides a uniform flow of vapor (Figure 1a), while a larger Dss value always results in mono­
precursors and is beneficial to the uniform growth. First, the layer WS2 flakes with a smaller domain size (Figure 1b).
distance between the source and substrate (Dss) was discov- This dependence on Dss manifests in the synthesis of WS2
ered to be a vital factor in the control of the WS2 thickness. domains being highly dependent on the density of vaporized

Adv. Mater. 2019, 1901694 1901694 (2 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 1. a,b) SEM images of WS2 flakes with a Dss of ≈9 and 10 cm. c,d) SEM images of synthesized WS2 flakes under pure Ar carrier gas and
mixed carrier gas (Ar + H2), respectively. a–d) Reproduced with permission.[61] Copyright 2014, American Chemical Society. e) Time-dependent edge
length of grown WSe2 triangles. Insets are the optical microscopy (OM) images of WSe2 triangles obtained at 10 and 20 s. f) The FRSe-dependent
growth rate of WSe2. g) The growth temperature (T) dependence of the growth rate. e–g) Reproduced with permission.[74] Copyright 2017, Wiley-VCH.
h–j) OM images of as-grown MoS2 crystals under different S supplies. Scale bar: h,i) 10 µm and j) 100 µm. k) The formation energies of different edges
varying with µS. l) A schematic view of the edge evolution during growth. m) An evolution of barrier energy for different edges with respect to µS. h–m)
Reproduced with permission.[75] Copyright 2018, Wiley-VCH.

precursors. The result demonstrates that varying the Dss could powder, which is beneficial to WS2 growth. Moreover, a suf-
efficiently control the thickness distribution and domain sizes. ficient concentration of precursors is also required for the
Second, the carrier gas is also essential to the CVD growth. formation of a regular shape.[31,72,73] Hence, the introduction
When H2 is mixed with Ar as the carrier gas, rather than of H2 gas can effectively adjust the concentration of the pre-
pure Ar gas, the WS2 domain with saw-tooth edge converts cursors and optimize the flake shape. Furthermore, the depo-
to a sharp-edged triangle (Figure 1c,d). The fact is that H2 (or sition temperature has also been mentioned as an important
formed H2S) can efficiently accelerate the reduction of WO3 factor, a relatively high deposition temperature is favorable for

Adv. Mater. 2019, 1901694 1901694 (3 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

the synthesis of high-quality crystals. Thus, there are sev- 2.1.2. Reaction System Design
eral key parameters during the CVD growth process, such as
growth Dss, carrier gas and growth temperature, that should To optimize the growth result in a controllable way, besides
be strictly controlled to realize the controllable growth of 2D the typical growth parameters, the proper design of a growth
materials. Recently, the synthesis of monolayer WSe2 on Au system also plays an important role. Lu et al.[77] reported a
foil was demonstrated by Ren et al.[74] Figure 1e shows the special synthesis of MoS2 crystals in a designed system. Addi-
growth-time dependence of the WSe2 domain size, and a tionally, Chi et al.[78] demonstrated the synthesis of uniform
millimeter-sized triangular WSe2 domain could be obtained monolayer MoS2 via introducing a nickel oxide foam as a bar-
within 30 s (as an inset in Figure 1e). This result indicates the rier during the growth process. Figure 2a presents a schematic
growth time also significantly affect the domain size of WSe2 view of the growth system. MoO3 powder was covered by a piece
flakes obtained in the growth process. At a proper nucleation of NiO foam and the foam was covered by a sapphire substrate.
density, a prolonged growth time can enlarge the crystal size Therein, the deposition of vaporized MoO3 could be effectively
by providing continuous precursors. Additionally, the feeding obstructed by the NiO foam, which led to a low nucleation rate
rate of the precursors and the growth temperature also signifi­ and large domain size of monolayer MoS2 on sapphire sub-
cantly influence the domain size and growth rate of monolayer strates. Figure 2b exhibits an optical microscopy (OM) image of
WSe2 triangles. Figure 1g shows that the growth rate of WSe2 CVD-synthesized MoS2, with the length of an isolated MoS2 tri-
triangles remarkably increases with the feeding rate of Se angle being up to 170 µm. Moreover, a scanning transmission
(FRSe) at 800 °C, and the maximum growth rate is ≈8.3 µm s−1, electronic microscopy (STEM) image shows the good single-
which no longer increases. Interestingly, the growth rate crystalline quality without any vacancies, as shown in Figure 2c.
clearly increases with the temperature (Figure 1h). Hence, X-ray photoelectron spectroscopy (XPS) was utilized to further
a high feeding rate of Se combined with a high growth tem- investigate the growth mechanism behind introducing the NiO
perature can offer the synthesis of a large-scale WSe2 mono­ foam. Before the growth, the characteristic Ni 2p3/2 and 2p1/2
layer with an ultrafast growth rate. Afterward, Loh et al.[75] peaks for NiO were located at 853.5 and 872.3 ± 0.7 eV, respec-
developed a detailed exploration of the morphology evolution tively. However, after the growth process, an evident shift in the
by controlling the growth precursor. Here, the Mo supply was peaks exists, while the multiple split peaks are absent, which
in excess during the growth process, which is due to the use is due to the formation of NiMoO4. Hence, the Mo precur-
of Mo foil as the Mo source. With a decreasing S supply, the sors are partially adsorbed on the NiO foam and emerged as a
morphology of as-grown MoS2 dramatically changes from a stable NiMoO4 compound via using the NiO foam. Therefore,
triangle with straight edges to a dendritic shape (Figure 1h–j). the nucleation densities can be controlled and reduced, and
With a different sulfur chemical potential (µS), different eventually, large-domain MoS2 crystals, together with a well-
edges (S edge or Mo edge) exhibit distinct growth rates, oriented continuous MoS2 film can be achieved. The electronic
which induce the morphological changes. Figure 1k shows characterization of the MoS2-based device shows a respectable
an apparent µS dependence on the formation energy for dif- mobility of 0.5–2.0 cm2 V−1 s−1, indicating the high quality of
ferent edges. Figure 1l further schematically illustrates that the synthesized MoS2 (Figure 2e). The results reveal that intro-
the growth rate of each edge is related to the energy barrier ducing a chemical barrier is beneficial for controlling the den-
ΔE1 and ΔE2. ΔE1 represents the energy barrier for zigzag sity of reactive precursors and improving the growth coverage
S-edge propagation, while ΔE2 indicates the energy barrier for and single-crystal size. Later, Dresselhaus et al.[79] realized the
the evolution of zigzag Mo-edges during the growth process. large-scale growth of MoTe2 layers in a controllable way via a
Figure 1m reveals that both ΔE1 and ΔE2 are highly dependent designed growth system. Figure 2f schematically illustrates
on µS, thus the growth rates of S and Mo edges can efficiently the growth process. Mo film on SiO2/Si deposited by electron
modulate the morphology of MoS2 nanosheets and the type beam evaporation is first oxidized in air to form MoO3 and then
of edges.[76] For example, by increasing the supply of S, the further reacted with Te powder. Notably, molecular sieves are
Mo edge grows substantially faster than the S edge and finally added to the crucible before growth that can adsorb byprod-
disappears, thus, MoS2 domain terminated with S edges is ucts efficiently during the growth and release Te vapor in a
developed. Therefore, to facilitate a deep insight into the prep- controllable and uniform way. Hence, a uniform MoTe2 film
aration of 2D materials by the CVD method, some crucial fac- with a scratch line showing the bare SiO2 substrate is obtained
tors affecting the growth results have been found, which in (Figure 2g). The atomic force microscopy (AFM) image (inset
turn can offer direction for the controllable growth of other in Figure 2g) of the edge indicates a uniform thickness of
2D materials. 3.1 nm. Further, the high-resolution transmission electron
To summarize, the above representative works demonstrate that microscopy (HRTEM) image (Figure 2h) shows a perfect hex-
several key factors modulating the 2D growth have been explored, agonal atomic arrangement, verifying its high crystal quality
such as the growth temperature, carrier gas and the source– and 2H phase. Hence, the proper design of the reactive system
substrate distance, as well as the supply of precursors. These fac- is very crucial for a controllable growth. To satisfy the require-
tors can efficiently control the nucleation density, vapor transport ments of practical applications, the batch production of 2D
rate, the growth rate, and the morphology. During the CVD growth materials is integral. Zhang et al.[80] reported the synthesis of
process, even though each parameter plays its main role, all of wafer-scale monolayer MoS2 films on sapphire wafers by the
them still restrict mutually and exhibit synergistic action. Hence, CVD method. Different from the other reported CVD systems,
a proper consideration and design of all the growth parameters is this growth system here introduces separate pathways for the
desirable for the controllable growth of 2D materials. S and MoO3 sources with different carrier gases, as shown in

Adv. Mater. 2019, 1901694 1901694 (4 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 2. a) Reaction system for MoS2 growth. b) An OM image of the synthesized MoS2 nanosheets. c) A STEM image obtained at 60 kV. d) XPS
spectra of the Ni 2p orbital before and after MoS2 growth. e) Transfer curves of an MoS2-based device. The inset is an OM image of the FET device.
a–e) Reproduced with permission.[78] Copyright 2016, American Chemical Society. f) The growth process of MoTe2 layers. g) An OM image of a uniform
MoTe2 film with a scratched line showing the substrate. The inset is the AFM image of the MoTe2 with a thickness of ≈3.1 nm. h) An FFT-filtered HRTEM
image of the transferred MoTe2. f–h) Reproduced with permission.[79] Copyright 2015, American Chemical Society. i) The modified CVD system for
MoS2 growth. j) A photograph of a 2 in. MoS2 film grown on sapphire substrates. k) The field-effect mobilities extracted from 24 devices with different
channel lengths. i–k) Reproduced with permission.[80] Copyright 2017, American Chemical Society.

Figure 2i. Using two pathways for the S and MoO3 sources, the the growth system plays a key role, and its proper design is
carrier gas can be controlled separately. Moreover, the evapo- essential.
rated S and volatile MoO3−x only meets and reacts at the growth Subsequently, space-confined growth systems were explored
substrate region, which efficiently promotes the formation of for growing various 2D TMDs. Yu et al.[81] demonstrated a
monolayer MoS2. Figure 2j shows as-grown monolayer MoS2 large-area synthesis of a WS2 monolayer via the CVD method.
film on a series of sapphire wafers. An electronic property of Different from the reported CVD growth process, the SiO2/Si
the monolayer MoS2 films was also measured by fabricating substrate was placed face-down above another piece of SiO2/Si
back-gated field-effect transistor (FET) devices. Figure 2k shows substrate covered by WO3 powder, as shown in Figure 3a. The
that the average mobility of 24 devices is ≈40 cm2 V−1 s−1, indi- reaction between WO3 and S occurs in the confined space
cating the high crystalline quality of the as-grown wafer-scale between the two substrates, which is conductive to the adequate
MoS2. It is well known that, wafer-scale MoS2 films always con- reaction. Figure 3b shows a typical OM image of the synthesized
sist of several merged single-crystal domains, which often leads WS2 with the edge length approaching 178 µm. Additionally, the
to many grain boundaries in the films. In addition, the grain corresponding HRTEM image in Figure 3c reveals the perfect
boundaries following some defects or vacancies will impede the lattice arrangement, validating the high quality crystal. Recently,
high performance of electronic devices. Hence, the controllable Lu et al.[77] demonstrated the synthesis of ultrathin MoS2 crys-
growth of wafer-scale single crystals is still a challenge. In brief, tals by a space-confined growth strategy. Figure 3d schematically
for the batch production of large-scale 2D layered materials, shows the sandwich-structured reaction system with an Mo

Adv. Mater. 2019, 1901694 1901694 (5 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 3. a) CVD system for WS2 growth. b) An OM image of WS2 triangles. c) An HRTEM image of a WS2 monolayer. a–c) Reproduced with permis-
sion.[84] Copyright 2014, American Chemical Society. d) A schematic view of the MoS2 growth. e) OM images of as-grown MoS2 flakes. f) A schematic
illustration of the proposed growth mechanism. d–f) Reproduced with permission.[77] Copyright 2018, American Chemical Society. g) A diagram of the
HfS2 growth with an inner quartz tube. h) An AFM image of a typical synthesized HfS2 flake. i) The thickness evolution of HfS2 flakes with respect to
different d values. g–i) Reproduced with permission.[82] Copyright 2017, Wiley-VCH.

film and S plate as the reaction precursors. Interestingly, MoS2 as-grown HfS2 flakes with different d values. Apparently, a
flakes of different thicknesses were synthesized on the SiO2/Si smaller d leads to the formation of thin flakes with irregular
substrate (Figure 3e). Figure 3f schematically illustrates the shapes, which is attributed to the low transport efficiency. The
growth mechanism in the space-confined sulfurization reaction. transport in a confined space is revealed to be limited and con-
First, enough sulfur was physically adsorbed on the substrate to siderably slower than that without confined space because of the
provide excessive vaporized S precursor. Then, the Mo film was limited space. Hence, a confined space is of great importance
vulcanized and converted to triangular MoS2 islands. During in the synthesis of HfS2 flakes by varying the reaction environ-
the growth process, the top cover in the sandwich reactor effi- ment. The proper design of the growth system renders conveni-
ciently restricts the S to the confined space ensuring an exces- ence for the preparation of 2D TMDs in a controllable way.
sive amount of sulfur vapor, which then increases the reaction In brief, the growth system also plays an important role in the
between Mo and S atoms. Generally, the confined reaction CVD growth of 2D TMDs. For example, introducing a NiO foam
system can be applied for the synthesis of various 2D materials. or a molecular sieves can effectively control the reaction rate. And
Accordingly, a useful design of the growth system was utilized introducing separate pathways for each precursor can promote
by Li et al.[82] for realizing the growth of ultrathin HfS2 flakes. the direct formation of MoS2 on substrates. Additionally, a space-
Figure 3g depicts the CVD system for the HfS2 growth using confined growth system can maintain a more stable transport and
HfCl4 and S as precursors. Differently, the mica substrates constant supply of precursors, which is beneficial to a uniform
are inserted into an additional tube, and such a design pro- growth. Hence, apart from the representative growth parameters,
vides a more stable transport of the carrier gas, maintaining the design of the growth system also acts as an vital role for an
a constant supply of precursors during the growth process. efficient growth, which provides a significant guidance.
The AFM image (Figure 3h) shows the thickness of a hexagonal
HfS2 crystal as 1.2 nm with d = 1.5 mm (d denotes the distance
between the mica and the bottom of the inner tube). To inves- 2.1.3. Effect of Growth Substrates
tigate the influence of the confined space on the synthesis of
HfS2 flakes, a series of growth experiments was conducted For a deep insight into 2D material synthesis via the CVD
by tuning d. Figure 3i illustrates the statistical thicknesses of method, several key ingredients can be further studied to result

Adv. Mater. 2019, 1901694 1901694 (6 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 4. a) Map of the surface growth. b) An SEM image of dendritic MoS2 flakes. c) A false-colored dark-field TEM image. d) Polarization curves of
several samples with different coverages and morphologies. In (d), curve a: A relatively compact MoS2/Au sample with a coverage of 60%; curve b: A rela-
tively fractal MoS2/Au sample with a coverage of 12%; and curve c: Absolute dendritic MoS2/Au with a coverage of 12%. a–d) Reproduced with permis-
sion.[84] Copyright 2014, American Chemical Society. e) The adjacent twin domains of a LAO (100) substrate. f,g) SEM images of synthesized MoS2 flakes.
h) Calculated exchange current densities. e–h) Reproduced with permission.[85] Copyright 2016, IOP Publishing. i) Statistics for 30 dendritic MoS2 flakes
(left inset: an SEM image of a typical MoS2 flake; right inset: a box-counting estimation of the fractal dimension). j) The adsorption configuration for a
MoS3 monomer on STO (001). Green, red, gray, yellow, and cyan balls denote the Sr, O, Ti, S, and Mo atoms, respectively. k) An illustration of the prob-
able diffusion pathways. l) Calculated energy barriers for the diffusion pathways in (k); i–k) Reproduced with permission.[86] Copyright 2016, Wiley-VCH.

in a controllable growth. Apart from the fundamental growth electron diffraction (SAED) pattern all indicate a single-crys-
parameters (temperature, carrier gas, and distance) and the talline feature. Moreover, the transferred fractal MoS2 reveals
growth system mentioned above, the growth substrate plays an interesting application in the HER as an electrocatalyst
an indispensable role in the CVD growth. Liu et al.[83] demon- (Figure 4d). Even at a rather low coverage, the dendritic MoS2
strated that insulating single crystals (mica, sapphire, etc.), in flakes exhibited a rather low onset potential (≈0.1 V, a high
view of their atomically flat surfaces, good stability, and perfect exchange current density ≈24.5 mA cm−2). The results reveal
lattice matching, were recognized as more appropriate sub- that the dendritic MoS2, with abundant edges, provides a
strates for growing 2D materials than polycrystalline SiO2/Si. plethora of active sites for the HER, which leads to an excel-
In fact, various 2D materials with large grain sizes and contin- lent HER activity. Furthermore, Zhang et al.[85] chose another
uous films have been achieved on these substrates. Recently, single crystal—LaAlO3 (LAO(100))—as a growth substrate
Zhang et al.[84] first communicated the achievement of den- and obtained fractal MoS2 on the corrugated LAO(100) with
dritic MoS2 on unique SrTiO3 (STO (100)) crystals revealing twin crystal domains induced by a second order phase transi-
their high performance in the hydrogen evolution reaction tion. Figure 4e schematically illustrates the formation of the
(HER). Figure 4a schematically depicts the chemical reaction twin crystal, which is attributed to the deformation taking
for synthesizing monolayer MoS2 on STO (100). The scan- place along different directions in two adjacent lattices. The
ning electron microscopy (SEM) image in Figure 4b shows SEM image (Figure 4f) manifests the overall appearance of as-
the typical fractal morphologies of as-grown MoS2 with abun- grown fractal MoS2 on the undulated surface of the LAO(100)
dant edges. Interestingly, the dendritic MoS2 flakes maintain a substrate. Furthermore, the image exhibits a distinct fractal
highly crystalline quality as evidenced by the TEM characteriza- morphology originating from the central nucleation site of
tion. Figure 4c demonstrates a false-color dark-field TEM image each MoS2 flake, which may be attributed to the strong inter-
with uniform contrast and its corresponding selected area action between the substrate (LAO(100)) and MoS2 (Figure 4g).

Adv. Mater. 2019, 1901694 1901694 (7 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Similarly, the exposure of abundant active edge sites and a of the schematic view in Figure 5a, during the growth process
high nucleation density result in the dendritic MoS2 being a at an increased temperature, the solid glass melts and pro-
good electrocatalyst, as is reflected by a high catalytic activity duces a clean atomically flat liquid surface. Importantly, the
(Figure 4h). Hence, different from SiO2/Si, mica and sapphire, flat liquid surface with few defects can decrease the nuclea-
as well as STO (100) and LAO(100) substrates, possess strong tion density,[91] which promotes the synthesis of MoSe2 crystals
interactions with MoS2 and induce a distinct fractal growth with large domain size. The OM image in Figure 5b presents
behavior. In other words, the growth results can be modulated triangular MoSe2 crystals with domain sizes up to a millim-
by utilizing different growth substrates. Therefore, for a deep eter on a molten glass. The SAED characterization was also
exploration of the formation mechanism, Zhang et al.[86] inves- utilized to probe the crystal structure and lattice orientation
tigated the growth behavior of dendritic MoS2 on STO(100) via of diverse regions in a triangle, revealing the single-crystal
experimental characterizations and theoretical calculations.[87,88] nature. Additionally, the STEM Z-contrast image (Figure 5c)
Through a standard box-counting evolution, Figure 4i presents shows the perfect atomic arrangement of the MoSe2 without
statistics on the fractal dimensions (D) of a series dendritic any defects, which further confirms the high quality of the
flakes with an average D of 1.65 ± 0.08 (30 flakes; the left inset crystal. The electronic transport studies based on as-grown
is a typical SEM image, and a linear fitting of the fractal dimen- MoSe2 crystals exhibit excellent device properties, further
sions is shown in the right inset of Figure 4i). This result is in highlighting the perfect crystalline quality. This study provides
accordance with that of fractal Au islands grown on Ru(0001) new insights into substrate engineering for the realization of
substrates, which followed a classical diffusion-limited aggre- large-scale synthesis of 2D materials. Recently, Au foil was
gation (DLA) mechanism. Thus, the DLA mechanism can also also selected as a growth substrate for WS2 and WSe2 mono­
apply to the dendritic growth of MoS2 on STO(100). For a fur- layers via the CVD method.[57,74,92,93] Ren et al.[62] reported the
ther investigation, the diffusion behavior of a MoSx monomer synthesis of millimeter-sized WS2 single crystals on Au foils.
on the surface of STO (001) was deduced by DFT calculations. The OM image in Figure 5d displays triangular WS2 domains
Figure 4j shows that Sr sites are the favorable adsorption sites. a millimeter in size on Au foil. The DFT calculations demon-
According to the theoretical simulation, two possible pathways strate the energy barrier for the sulfurization of WO3 by sulfur
for the MoS3 monomer hopping among these Sr sites are dem- atoms on Au is lower than the direct sulfurization of WO3 in
onstrated in Figure 4k, represented by red arrows (Pathway 1) a sulfur atmosphere, which leads to a large-area synthesis of
and blue arrows (Pathway 2). Interestingly, the energy barrier WS2 on Au. Moreover, the authors also synthesized millimeter-
for the diffusion along Pathway 1 (<0.3 eV) is substantially size WSe2 domains on reusable Au substrates at a very high
smaller than that along Pathway 2 (>3 eV) (Figure 4l). Therefore, growth rate. Figure 5e shows a the typical OM image of an
the monomer precursors exhibit a clear diffusion anisotropy on as-grown WSe2 domain obtained with a growth time of 10 s.
the surface of STO (001). In short, the diffusion anisotropy and The inset of Figure 5e exhibits the perfect atomic arrangement
distinct lattice symmetry induce the DLA growth of dendritic with no defects or vacancies, emphasizing the high quality of
MoS2. Later, Gu et al. also reported the strong influence of sub- WSe2 crystals. DFT calculations were utilized to deduce the
strate symmetry on the CVD growth of MoS2 fractals on SiC.[89] ultrafast growth behavior on Au foil. The calculation results
Hence, these results reveal that the growth substrates plays a show that the attachments of both Se atoms and W2 dimers on
considerable role in the 2D growth. The diffusion barrier and the the edge of a WSe2 cluster are exothermic processes, resulting
interface interaction on different substrates are quite diverse, as in energy decreases of 0.88 and 3.12 eV, respectively (Figure 5f
well as the lattice symmetry, which strongly impact the diffusion and g). Hence, the growth rate of WSe2 on Au foil is not domi-
of precursors and the formation of 2D materials on substrates. 2D nated by the diffusion of Se and W2. Therefore, large-scale
growth is quite susceptible to the localized concentration of pre- WSe2 crystals can be generated by quickly supplying the W
cursors which depends on the surface diffusion. Thus, the proper and Se precursors. Additionally, FETs were fabricated on all
choice of growth substrate can satisfy the specific requirement, the obtained CVD-grown TMDs and exhibited comparable per-
for example, the dendritic shape is better for electrocatalysis, and formance. For example, Figure 5h demonstrates the mobilities
a large domain size is expected for device applications. and on/off ratios of the 19 MoSe2-based FETs, showing a high
Currently, various 2D layered materials have been applied performance (mobility: 5–95 cm2 V−1 s−1; on/off ratio: 104–107).
to electronics and optoelectronics, exhibiting remarkable appli- In addition, the carrier mobility and on/off current ratio of
cation prospects on account of their abundant properties. the WS2-based device is 1.7 cm2 V−1 s−1 and 107, respectively
Hence, the batch production of 2D TMD layers with a large (Figure 5i). Similarly, the mobility of CVD-grown WSe2 FETs
grain size plays an important role in their practical application. is ≈139 cm2 V−1 s−1 with an on/off ratio of 9.2 × 106. All these
According to the previous research conducted on CVD growth, results are equal to those of mechanically exfoliated samples,
the growth substrate can efficiently influence the growth result. which proves the high crystal quality of the CVD-synthesized
Recently, some valuable progress on the CVD growth by ingen- 2D TMDs and forecasts their application potentials.
ious designs has been reported. Loh et al.,[90] for the first time, In short, both molten glass and Au foil are unique growth
successfully achieved the synthesis of mono­layer millimeter- substrates for the 2D growth. Molten glass can provide a
scale MoSe2 crystals via the CVD route by using molten glass flat liquid surface and lower the nucleation density, which
as a substrate. Figure 5a shows the experimental setup and efficiently promote the synthesis of large domain crystals.
CVD reaction process. MoO3 and Se powders were the reac- Lower energy barrier on Au foils drive the ultrafast growth
tion precursors and the molten glass was held by a piece of of high-quality WSe2 crystals. Thus different substrates can
Mo foil supported by SiO2/Si. As described by the bottom part induce different growth behavior, a feasible selection of growth

Adv. Mater. 2019, 1901694 1901694 (8 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 5. a) Schematic of the CVD process. b) An OM image of as-grown MoSe2 crystals. c) A STEM image of MoSe2 crystals. Scale bar: 1 nm.
a–c) Reproduced with permission.[90] Copyright 2016, American Chemical Society. d) An OM image of synthesized WS2 triangles. Reproduced
under the terms of the CC-BY Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/).[62]
Copyright 2015, Springer Nature Publishing AG. e) An OM image of a typical WSe2 triangle grown for 30 s. The inset is the HAADF-STEM
image. f) The calculated energy evolution for the attachment of a Se atom. g) The calculated energy evolution for the attachment of a W2 dimer.
e–g) Reproduced with permission.[74] Copyright 2017, Wiley-VCH. h) The mobilities and on/off ratios of various MoSe2 FETs. The inset is an OM image
of a device. Reproduced with permission.[90] Copyright 2016, American Chemical Society. i) Transfer characteristics of a WS2-based device. The inset is
an OM image of the device. Reproduced under the terms of the CC-BY Creative Commons Attribution 4.0 International License (http://creativecom-
mons.org/licenses/by/4.0/).[62] Copyright 2015, Springer Nature Publishing AG. j) Transfer curves of a monolayer WSe2 BG-FET. The inset is an OM
image of the device. Reproduced with permission.[74] Copyright 2017, Wiley-VCH.

substrates, considering the surface diffusion barrier and lattice which is markedly lower than the melting point of Mo foils.
symmetry, is desirable for the controllable growth. Figure 6a presents a typical OM image of as-grown MoS2 tri-
angles. In this regard, a 14 × 6 cm2 uniform, full-covered MoS2
film on a glass substrate was finally obtained by the merging of
2.1.4. Other Agents Promoting the CVD Growth composite domains for only 8 min (in the inset of Figure 6a).
Notably, this growth rate of MoS2 on glass is significantly
Though the growth substrate plays an important role in the CVD higher than those of commonly used substrates. To investigate
growth process, the practicable variety of substrates is still lim- the growth mechanism of MoS2 on soda-lime glass, the MoS2
ited. Then, some new agents, such as sodium chloride (NaCl) growth behavior on both sola–lime glass and quartz substrates
and potassium chloride (KCl) crystals, with low melting points, was compared in detail. Combined with DFT calculations, using
have also been added and utilized during the growth process, elemental Na in the glass substrate resulted in a profound effect
exhibiting an extraordinary performance and good universality facilitating the growth of MoS2. Figure 6b,c displays the calcu-
in the controllable growth of 2D TMDs and . Recently, Zhang lated energies for the MoS2 growth. With the incorporation of
et al.[94] reported the growth of a 6 in. uniform MoS2 monolayer Na, the highest energy barrier was reduced from 0.53 to 0.29 eV,
on glass. In the low-pressure CVD growth process, Mo foil and which consequently promotes the growth rate of MoS2. Hence,
S powder were used as precursors and the soda-lime glass was the soda-lime glass is crucial for the rapid growth of monolayer
placed directly under the Mo foil. The Mo foil is oxidized into MoS2. In addition, Liu et al.[95] group very recently realized the
MoO3−x (x = 2, 3) by the introduced O2 carrier gas, and then synthesis of 47 compounds based on TMDs via a salt-assisted
reacts with S. Notably, the growth temperature is ≈500 °C, CVD method. As can be visualized from Figure 6d, the presence

Adv. Mater. 2019, 1901694 1901694 (9 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 6. a) OM image of monolayer MoS2 at a growth time of ≈5 min. Scale bar: 0.1 mm. The inset is a photograph of a uniform MoS2 film.
b) DFT calculations for the MoS2 growth without Na. c) DFT calculations for the MoS2 growth with Na. a–c) Reproduced under the terms of the
CC-BY Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/).[94] Copyright 2018, The Authors,
published by Springer Nature Publishing AG. d) An illustration of the reaction process and OM images of the synthesized materials. Reproduced
with permission.[95] Copyright 2018, Springer Nature Publishing AG.

of the salt can promote the evaporation of the precursors and placed between the SiO2/Si substrate and precursors. Figure 7b
increase the reaction rate, which helps to produce millimeter- schematically displays the WSe2 growth process. The typical
sized single-crystal 2D TMD materials over a relatively short morphology of a monolayer WSe2 triangle is demonstrated
growth time and at a low temperature. Step (8) in Figure 6d in Figure 7c. From the HRTEM image shown in Figure 7d, a
shows some typical OM images of various TMDs synthesized perfect atomic arrangement with no apparent defects indicates
with large domain sizes. These studies indicate the introduction a high-quality crystal. Furthermore, DFT calculations were con-
of Na or NaCl could efficiently promote the growth of 2D TMDs. ducted to further study the effect of Cu during the growth pro-
Apart from the agents mentioned above, the liquid metal, cess. As shown in Figure 7e, the absorption energy of Cu on the
serving as a catalyst, maybe conducive to uniform synthesis on surface of WSe2 is 2.05 eV, but the value for the absorption of W
nonmetal substrates, such as insulating substrates or polycrys- on WSe2 is strikingly increased to 7.21 eV. The adsorption force
talline substrates. Commonly used nonmetal substrates, for of Cu atoms on WSe2 is appreciably stronger than that of W
example, SiO2/Si, have no catalytic effect and always induce an atoms. Hence, during the growth process, Cu atoms fully occupy
uncontrollable growth with random thickness and size. Intrigu- the sites on the WSe2 surface and W atoms can only attach to the
ingly, Fu et al.[96] proved this effect in the growth of 2D TMDs edges of WSe2 domains, consequently the domain size increases
and reported the synthesis of monolayer WSe2 on an SiO2/Si constantly. Thus, the large-domain-size growth of monolayer
substrate assisted by Cu foil. Notably, Cu foils have been often WSe2 single crystals can be obtained via introducing a Cu foil.
applied to the growth of graphene as a substrate, which also The transfer curve of the WSe2 FET (Figure 7f) exhibits a typical
effectively promote the splitting of precursors.[53,97] Figure 7a p-type conduction behavior, showing a mobility of 45 cm2 V−1 s−1.
illustrates the reaction setup, in which a piece of Cu roll is Additionally, the performance of a WSe2 photodetector was also

Adv. Mater. 2019, 1901694 1901694 (10 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 7. a) The setup for the WSe2 growth assisted by Cu foil. b) A schematic of the WSe2 synthesis. c) An OM image of a representative WSe2
triangle. d) An HRTEM image of a WSe2 crystal. e) Calculated absorption energies. f) The transfer curve of a WSe2 FET. The inset is an SEM image. g) The
photocurrent curves for the device. h) The time-dependent photoresponse of the device. a–h) Reproduced with permission.[96] Copyright 2016, Wiley-VCH.

studied. Figure 7g,h demonstrates a prominent current increase and substrates, wafer-scale uniform 2D semiconducting
with the illumination of light as well as the high responsivity TMDs with a large domain size were achieved. The latest
of the WSe2-based devices. This result reveals the achievement progress in the CVD growth of some semiconductor TMDs,
of high-quality WSe2 crystals via a Cu-assisted growth process, such as MoS2,[96] WS2,[62] MoSe2,[90] and WSe2,[74] is presented
which may be extended to the growth of various 2D materials. in Table 1. Based on a series of studies on the system design
To summarize, the introducing of other agents also can and growth mechanism, millimeter-sized 2D TMD domains
modulate the growth behavior. The introducing of sodium and uniform film have been realized in a controllable way. Fur-
chloride (NaCl) or potassium chloride (KCl) crystals with low thermore, these CVD-grown TMDs have also been applied to
melting point can lower the energy barrier, which consequently electronic/optoelectronic devices and demonstrated excellent
promotes the growth rate of MoS2. Additionally, the introduced electronic properties compared to previous results (see also
Cu atoms largely occupy the sites on the WSe2 surface due Table 1).
to its small adsorption energy, which consequently promote
the expanding of the lateral domain size. And the design in
above works opens new ways for the large-scale growth and 2.2. New Kinds of Semiconducting TMDs
these agents also can be extended to the synthesis for other
TMD materials. Apart from those typical TMDs discussed in the previous sec-
Hence, some key factors for the CVD growth have been tion, some new kinds of semiconducting TMDs have subse-
explored, such as temperature, carrier gas, the ratio of precursors quently been explored. ZrX2 and HfX2 (X = S, Se, and Te) layers
and growth substrates, as well as the growth system. And a valid have rarely been studied, including their preparation, optical
control of them is needed to realize the batch production. Via properties and physical properties. However, these materials
controlling growth parameters, designing the growth systems have been proposed to exhibit electronic properties superior to

Adv. Mater. 2019, 1901694 1901694 (11 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Table 1. The latest progress in the CVD growth and applications of four typical semiconductor TMDs.

Material Growth precursors and method Growth substrate Thickness Uniformity Domain size Mobility On–off ratio

MoS2 Mo foil + S powder LPCVD Glass Monolayer 6 in. uniform Edge length > 6.3–11.4 cm 2
105–106
monolayer film 400 µm V−1 s−1

WS2 MoO3 + S powder APCVD Au foil Monolayer Large-area films Edge length > 1–2 cm2 V−1 s−1 4 × 106–5 × 107
500 µm

MoSe2 MoO3 + Se powder APCVD Molten glass Monolayer – Domain size ≈95 cm2 V−1 s−1 107
reaching ≈2.5 mm

WSe2 WO3 + Se powder APCVD Au foil Monolayer Large-area con- Millimeter-sized ≈143 cm2 V−1 s−1 9 × 106
tinuous films in triangular WSe2
≈60s domain

those of MoS2.[98–102] Recently, Xie et al.[103] reported the growth The ZrS2 flakes were selectively deposited on BN instead of
of ZrS2 nanosheets via a CVD method. A typical SEM image on bare SiO2/Si regions, which could be due to the atomically
of ZrS2 hexagons on a BN substrate is shown in Figure 8a. flat surface of BN flakes. The AFM image (Figure 8b) exhibits

Figure 8. a) SEM image of synthesized ZrS2 flakes on BN. b) An AFM image of a ZrS2. c) The thickness and vapor pressure varying with the evapora-
tion temperature of ZrCl4. d) The Ids–Vds curves for different gate voltages. a–d) Reproduced with permission.[103] Copyright 2015, ACS Publications.
e) An OM image of as-obtained HfS2 crystals on mica. f) An AFM image of an HfS2 crystal. g) Transfer curves of HfS2 devices. h) The responsivity and
photogain of HfS2 phototransistor at Vg = 80 V, −80 V. e–h) Reproduced with permission.[104] Copyright 2017, Wiley-VCH. i) A field-emission scanning
electron microscopy (FESEM) image of PtSe2 flakes. j) Transfer curves of a PtSe2-based device. The inset shows the I–V curves for various gate volt-
ages. i,j) Reproduced with permission.[111] Copyright 2016, Wiley-VCH. k,l) OM images of synthesized HfSe2 and PtS2 crystals. k,l) Reproduced with
permission.[95] Copyright 2018, Springer Nature Publishing AG.

Adv. Mater. 2019, 1901694 1901694 (12 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

the morphology of monolayer ZrS2 crystals. Furthermore, the the 2D family members and these materials exhibits superior
thickness of the as-grown ZrS2 crystals was correlated with the electronic properties, providing an abundance of promising
evaporation temperature of ZrCl4 (Figure 8c), demonstrating a candidates for electronics and optoelectronics.
clear relationship between the thickness and evaporation tem-
perature. The ZrS2-based FET exhibited an n-type transport
behavior, as shown in Figure 8d. The mobility was extracted to 3. 2D Metallic TMDs
be ≈1.1 cm2 V−1 s−1, which is considerably lower than the theo-
retical mobility[102] on account of the poor electrode contacts and Over the years, 2D semiconductor TMD materials have attracted
small domain size. However, the synthesis of ZrS2 broadens considerable interest and generated abundant research due to
the available material systems and provides a new opportunity their various bandgaps and physicochemical properties. These
for potential applications in nanoelectronics. Subsequently, the 2D materials have enabled unprecedented developments in
growth of atomically layered HfS2 crystals on mica was reported electronic devices, valleytronics, and electrocatalysis. Recently,
by Liu et al. via vdW epitaxial growth.[104] Accordingly, mono­ metallic TMDs, such as TiSe2,[112–115] NbSe2,[116–118] TaSe2,[119–121]
layer HfS2 crystals were synthesized by a CVD route using and TaS2,[122–125] have gained considerable attention on account
HfCl4 and S powders as precursors. Figure 8e demonstrates of their unique characteristics that include magnetism,[126–128]
an OM image of CVD-synthesized HfS2 triangles, and the cor- superconductivity[129] and charge-density waves (CDWs).[130]
responding AFM image (Figure 8f) reveals their monolayer 2D metallic TMD materials represent an essential component
nature with a thickness of 0.68 nm. The CVD-grown HfS2 crys- of the 2D family. Moreover, metallic TMD materials with no
tals were transferred onto SiO2/Si and fabricated into FETs. The bandgap are also forecast to hold potential applications in elec-
transfer characteristics of HfS2 crystals with different thick- tronics and energy conversion/storage depending on their elec-
nesses (Figure 8g) reveal an n-type behavior and an on/off cur- trical conductivity. Initially, the superconductivity phenomenon
rent ratio of ≈108.[105] Additionally, a mobility of 7.6 cm2 V−1 s−1 and CDW transitions were realized in ultrathin metallic TMD
was recorded for a few-layer HfS2 crystal, while the mobility flakes, which are mainly obtained by exfoliation methods. How-
of an HfS2 monolayer ranges from 1.5 to 2.1 cm2 V−1 s−1. ever, the exfoliated metallic TMD flakes also possess relatively
The device mobility generally increases with the thickness small sizes with random thickness, which greatly hinders the
of the HfS2 crystal (Figure 8g). This results indicate that the investigation of their intrinsic magnetism and further applica-
electrons transport through different layers of HfS2 is inde- tion. Hence, the controllable growth of 2D metallic TMDs pos-
pendent, which is similar to that of MoS2-based devices.[106,107] sesses significant value for further research and applications. To
Importantly, the HfS2-based phototransistor exhibits prominent tackle the above issues, recently, breakthroughs regarding the
photoelectric properties, as presented in Figure 8h. It demon- controllable growth of 2D metallic TMD flakes via an ambient-
strates a prominent responsivity of 3.08 × 105 A W−1 (808 nm pressure CVD method have been reported, which paves the way
incident laser).[108] Generally, ZrX2 and HfX2 layers have also to the intensive study of 2D metallic TMDs.
demonstrated excellent properties, thus the uniform growth of Vanadium disulfide (VS2) is a representative member with a
large-size monolayers is a prerequisite. Recently, Liu et al.[97] layered structure crystallizing in the 1T phase. Zhang et al.[131]
demonstrated a molten-salt-assisted CVD system for growing demonstrated the achievement of few-layer VS2 nanosheets
various monolayer TMDs. The controllable synthesis of mono­ via an ambient pressure chemical vapor deposition (APCVD)
layer ZrX2 and HfX2 (X = S, Se, and Te) were realized, and growth method. Vanadium trichloride (VCl3) and sulfur (S) pow-
their crystal structures and physical properties were further ders were used as the growth precursors, and VS2 flakes were
investigated. Apparently, apart from these semiconductor obtained at a growth temperature of 600 °C. Figure 9a demon-
TMDs, precious-metal-based 2D TMDs (such as PtSe2) have strates a typical OM image of VS2 nanosheets obtained under a
been considered as another promising class of candidates mixed carrier gas (Ar and H2). Interestingly, the average edge
for electronics.[109,110] Furthermore, Dong et al.[111] achieved length of VS2 domains could by efficiently tuned by varying the
ultrathin PtSe2 nanosheets by the CVD method using hexa- flow rate of the carrier gas (Figure 9b). Evidently, the domain
chloroplatinic acid (H2PtCl6) and selenium (Se) as the source size increases strikingly with decreasing H2. However, in the
materials. Figure 8i shows an SEM image of as-grown PtSe2 absence of H2 as a carrier gas, no VS2 flakes appears on sub-
hexagons and the related HRTEM image (inset of Figure 8i) strates, which confers the necessity of H2 gas during the growth
reveals a perfect hexagonal arrangement and high-quality process. The hydrogen flow is thus expected to promote the
crystal. In this exciting work, back-gated FETs were fabricated reaction rate of vaporized VCl3 and S, which continuously mod-
based on PtSe2 nanosheets. Figure 8j depicts typical transfer ulates the nucleation density and growth rate of VS2 nanosheets.
and output characteristics of PtSe2-based transistors, indicating Hence, the type of carrier gas plays a vital role in the CVD
a p-type transport behavior. The hole mobility can be calculated growth. Accordingly, an ≈6.8 nm thick VS2 domain is obtained
as 7 cm2 V−1 s−1. Furthermore, Liu et al. successfully achieved (Figure 9c). In addition, an STEM image (Figure 9d) reveals the
atomically thin HfX2, and PtX2 (X = S, Se, and Te) with a large presence of two coexisting atomic patterns. Furthermore, the
domain size in a controllable way.[97] Figure 8k,l demonstrates corresponding fast Fourier transformation (FFT) reveals two
typical OM images of ultrathin HfSe2 and PtS2 domains with kinds of patterns, which are identified as red and yellow regions.
uniform contrast. This result indicates the formation of a superlattice in VS2
In brief, the facile CVD method has been successfully flakes. Afterward, Zhang et al. realized the synthesis of several
extended to growing new 2D semiconducting TMDs. And the metallic TMDs on different substrates. For example, the authors
synthesis of ZrX2, HfX2 and PtX2 (X = S, Se, and Te) enlarge grew few-layer VSe2 and TaS2 crystals using the CVD technique

Adv. Mater. 2019, 1901694 1901694 (13 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 9. a) OM image of the VS2 flakes (100 sccm Ar and 2 sccm H2). b) VS2 edge lengths varying with the H2 flow rate. c) An AFM image
of a transferred VS2 flake. d) An HAADF-STEM image and the FFTs of the selected regions. a–d) Reproduced with permission.[131] Copyright
2017, ACS Publications. e) An AFM image of a 4.9 nm thick VSe2 flake. f) The thickness of VSe2 flakes varying with the Ar flow rate. g) A false-
color coded HAADF-STEM image. e–g) Reproduced with permission.[136] Copyright 2017, Wiley-VCH. h) An SEM image of TaS2 flakes on Au
foil. Scale bar: 20 µm; the inset is an AFM image. Scale bar: 50 µm. i) An HRTEM image of the transferred TaS2 sample. Scale bar: 1 nm. h,i)
Reproduced under the terms of the CC-BY Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/
by/4.0/).[133] Copyright 2017, The Authors, published by Springer Nature Publishing AG. j) An OM image of NbSe2 crystals on SiO2/Si. Scale
bar: 40 µm. The inset is an AFM image. Scale bar: 1 µm. k) An ADF-STEM image of the NbSe2 crystal. The inset shows the structural model.
Scale bar: 2 nm. j,k) Reproduced under the terms of the CC-BY Creative Commons Attribution 4.0 International License (http://creativecommons.org/
licenses/by/4.0/).[134] Copyright 2017, The Authors, published by Springer Nature Publishing AG. l) An SEM image of uniform NbS2 crystals. Scale bar:
50 µm. The inset shows an NbS2 crystal with an edge length of 0.1 mm. m) An AFM image of an NbS2 flake. Scale bar: 5 µm. n) An HRTEM image of
an NbS2 crystal. Scale bar: 5 nm. l–n) Reproduced with permission.[135] Copyright 2018, Wiley-VCH.

and the as-grown materials exhibit high electrical conductivi- NbX2 (X = S, Se, and Te), has also been subsequently explored.
ties and unique thickness-dependent CDW phase transitions, as Liu et al.[134] communicated the first experimental report on the
well as demonstrating the excellent performances in electroca- CVD growth of monolayer metallic NbSe2 and its supercon-
talysis.[132,133] As shown in Figure 9e, the VSe2 nanosheets were ductivity. Accordingly, they synthesize monolayer NbSe2 by an
synthesized on mica, and the evolution of the average thickness APCVD method using partially oxidized niobium powder (NbOx
of VSe2 nanosheets is displayed in Figure 9f, showing a reduc- (x ≤ 2.5)) and selenium (Se) powder as precursor. A representa-
tion in thickness with increasing gas flow. Figure 9g shows a tive OM image of the hexagonal NbSe2 crystals is presented in
false-color STEM image, and the V and Se atoms can be distin- Figure 9j, showing an apparent OM contrast. The corresponding
guished as green and yellow balls, respectively, which confirms AFM image reveals the thickness of an NbSe2 crystal is ≈1.1 nm,
their perfect atomic arrangement in a relatively high-quality indicating its monolayer nature (inset of Figure 9j). The high-
crystal. Similarly, 2D metallic TaS2 nanosheets have been dem- magnification angle annular dark (ADF)-STEM image of a
onstrated via both LPCVD and APCVD methods using tantalic monolayer NbSe2 crystal (Figure 9k) displays a perfect lattice
chloride (TaCl5) and S as precursors. An SEM image exhibits the arrangement, which corresponds to Se and Nb atomic columns.
synthesized TaS2 domains on an Au foil via an LPCVD method As demonstrated by the atomic model, the result confirms the
(Figure 9h). The inset of Figure 9 reveals the thickness of the existence of the hexagonal phase and a good crystal quality.
TaS2 nanosheet is ≈3 nm. Furthermore, Figure 9i exhibits a Recently, monolayer metallic NbS2 was also synthesized with a
perfect atomic arrangement, confirming the high quality of large grain size using an APCVD system.[135] Here, the authors
CVD-grown TaS2. Thus far, progressive report on the synthesis obtain edge lengths as large as 100 µm for as-grown NbS2 trian-
of TaSe2 and TaTe2 crystals via the CVD method were reported gles deposited on sapphire (Figure 9l and the inset OM image
by Liu et al.[97] Meanwhile, another class of metallic TMDs, of Figure 9l). An AFM image (Figure 9m) further demonstrates

Adv. Mater. 2019, 1901694 1901694 (14 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 10. a) Temperature-dependent resistivity of VSe2 nanosheets. b) The Rxx–T curve (black line) and dR/dT–T curve (blue dots) of a VSe2 nanosheet
(10 nm thick). a,b) Reproduced with permission.[136] Copyright 2017, Wiley-VCH. c) Positions of all the Raman peaks for a 2H-TaS2 flake as a function of the
temperature. d) The thickness-dependent phase transition temperature of CVD-grown 2H-TaS2. c,d) Reproduced under the terms of the CC-BY Creative Com-
mons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/).[133] Copyright 2017 , The Authors, published by Springer Nature
Publishing AG. e) The temperature-dependent Raman results for a 1.4 nm thick TiSe2 flake. f) The thickness-dependent phase transition temperature for
CVD-grown TiSe2 samples. e,f) Reproduced with permission.[114] Copyright 2018, Wiley-VCH. g) The temperature-dependent resistance (Rxx) of a monolayer
NbSe2 device. Upper-left inset: an OM image of the NbSe2 device. Lower-right inset: the Rxx evolution for different thickness. h) Superconductivity in an
NbSe2 crystal for different magnetic fields. g,h) Reproduced under the terms of the CC-BY Creative Commons Attribution 4.0 International License (http://
creativecommons.org/licenses/by/4.0/).[134] Copyright 2017, The Authors, published by Springer Nature Publishing AG. i) Transfer curves of a VS2-contacted
1L-MoS2 device. The inset is an OM image of the device. j) Cyclic voltammograms (CVs) in H2SO4. i,j) Reproduced with permission.[131] Copyright 2017, ACS
Publications. k) ΔGH* for different edges. l) Tafel slopes of different TaS2 samples. k,l) Reproduced under the terms of the CC-BY Creative Commons Attribu-
tion 4.0 International License (http://creativecommons.org/licenses/by/4.0/).[133] Copyright 2017, The Authors, published by Springer Nature Publishing AG.

the thickness of an NbS2 triangle is ≈1.0 nm, demonstrating properties of CVD-grown VSe2 crystals. Accordingly, six-ter-
its monolayer feature. Additionally, the perfect atomic arrange- minal Hall-bar devices were fabricated on several CVD-grown
ment of NbS2 can be supported by the result obtained from the 1T-VSe2 crystals. The resistivity values of several VSe2 crys-
HRTEM image (Figure 9n). By and large, a series of 2D metallic tals vary with the temperature, as illustrated in Figure 10a.
TMD crystals has been successfully synthesized by utilizing a Interestingly, the electronic resistance decreases dramatically
facile CVD route, and the components of which exhibits signifi- with the temperature, suggesting the metallic behavior of the
cant properties, such as extra-high electrical conductivity, excel- synthesized atomically thin VSe2 crystals. Additionally, the
lent HER catalysis, and superconductivity phenomena. VSe2 crystals demonstrate an ultrahigh electronic conduc-
Metallic TMDs in the bulk state have demonstrated a variety tivity approaching to 106 S m−1. More importantly, when below
of intriguing properties, such as magnetism, CDWs, and super- 100 K, the resistivity values of all VSe2 crystals show a sharp
conductivity. More interestingly, these unique phenomena drop, which is attributed to the existence of a CDW transition. As
have also been observed in the 2D limit. Moreover, the CVD- seen in Figure 10b, the CDW transition appears at 63 K for a 10
grown 2D metallic TMDs have also displayed excellent per- nm VSe2 crystal. Hence, below 100 K, CVD-grown VSe2 crystals
formances in electronic devices and electrocatalysis. In 2017, with different thicknesses are displayed. Recently, a temperature-
Zhang et al.[136] explored the intrinsic physical and electronic dependent Raman technique was performed on CVD-grown

Adv. Mater. 2019, 1901694 1901694 (15 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

2H-TaS2 to investigate the CDW phase transition.[133] Figure 10c a smaller contact resistance across the VS2/MoS2 interface. Addi-
depicts the positions of all Raman peaks of a 20 nm thick TaS2 tionally, 2D VS2 layers realized a capacitance of 8.6 × 102 F g−1,
flakes varying with the temperature. When the temperature is as shown in Figure 10j. Apart from their successful employ-
below 150 K, some new peaks can be viewed below 100 cm−1, ment in device applications, 2D metallic materials have also been
which indicates a CDW phase transition occurred during the applied to electrocatalysis. In a recent study conducted by Zhang
cooling process. In addition, the warming process shows a transi- et al.,[133] CVD-grown metallic TaS2 flakes were used as an elec-
tion temperature of ≈210 K. Furthermore, temperature-dependent trocatalyst and e­ xhibited an excellent catalytic performance. The
Raman spectra were obtained from a series of TaS2 flakes with DFT calculations demonstrates the Gibbs free energies (ΔGH*)
different thicknesses. The extracted nearly commensurate charge- for the Ta-edge, S-edge, and basal-plane of the TaS2 nanosheet
density-wave (NCCDW)/CCDW phase transition temperatures are −0.04, −0.10, and 0.15 eV, respectively (Figure 10k). This cal-
are plotted in Figure 10d, which reveals that the phase transition culated result suggests that both the edges and basal-planes of
temperature observably decreases with the thickness. Hence, the 2H-TaS2 can provide active sites during the catalytic reaction. As
obtained results match well with the electrical transport result of shown in Figure 10l, the Tafel slope for 2H-TaS2 is ≈33 mV dec−1,
the exfoliated TaS2 layers.[137] These results further prove the reli- which confirms the prominent catalytic properties of 2H-TaS2.
ability of CVD-grown samples for exploring CDW transitions and Therefore, all these results promote the exploration of new phys-
other interesting physical properties. Furthermore, for the as- ical properties in the 2D metallic TMD materials, as well as the
synthesized TiSe2 samples, the thickness-dependent CDW phase expansion of the material systems and the exploitation of their
transition was also observed in temperature-dependent Raman potential toward versatile applications in different fields.
spectra (Figure 10e,f).[114] The CDW transition temperature in 2D metallic TMDs is an essential building block for 2D family,
the few-layer TiSe2 samples is ≈212.5 K, which also indicates the showing lots of intriguing physical properties, such as supercon-
high-quality crystals of the CVD TiSe2 films. Recently, monolayer ductivity, magnetism, and CDW. The properties of 2D metallic
superconductors have gained significant interest and will pro- TMDs is quite different from that of 2D semiconducting TMDs,
vide ideal platforms for exploring superconductivity phenomena which should be an important supplement. And the combination
in the 2D limit, as well as topological physics.[138,139] Inspired by of 2D metallic and semiconducting TMDs will create more pos-
such unique properties, Liu et al.[134] investigated the supercon- sibilities for further applications. The discovery and synthesis of
ductivity of a CVD-grown monolayer NbSe2 crystal. Six-terminal metallic TMDs could be a significant advance for enriching the
Hall-bar devices were fabricated on the CVD-grown NbSe2 2D family, and provides more material system for scientists to
crystal (Figure 10g), which was first-coated by a continuous explore next-generation electronic and energy-related applications.
monolayer graphene film before the fabrication. The tempera-
ture-dependent resistance (Rxx) of an NbSe2 monolayer device
at zero magnetic field (Figure 10g), shows metallic behavior at 4. 2D TMD Alloys
high temperatures. Notably, the Rxx begins to decrease sharply
from 1.5 K and drop to zero at 0.8 K, showing the occurrence of 2D TMD materials have aroused significant interest in the
superconductivity. Additionally, the thickness of an NbSe2 crystal fields of electronic/optoelectronics owing to their layer-
can efficiently modulate the transition temperature (the lower dependent bandgap and excellent electronic mobility. Recently,
right inset of Figure 10g). These data also supports the forma- 2D semiconducting alloy materials based on TMDs have exhib-
tion of a high-quality CVD-derived NbSe2 monolayer. The tem- ited outstanding characteristics in modulating the bandgap
perature and magnetic field dependencies of Rxx for the NbSe2- and carrier type owing to their tunable components.[140–144]
based device are demonstrated in Figure 10h. It is evident that Among the various approaches, the CVD method for growing
the superconducting transition temperature shifts to lower tem- large-scale monolayer ternary TMD alloys[145,146] is a promising
peratures as the strength of the magnetic field increases. Finally, candidate. For example, Xie et al.[147] first reported the prepara-
the superconductivity is completely suppressed with a magnetic tion of 2D MoS2(1−x)Se2x (0 ≤ x ≤ 0.40) semiconductor alloys
field of ≈1.5 T at 0.3 K. All of these results reveal that CVD tech- with a range of bandgaps (1.86–1.73 eV) that exhibited good
nology can provide not only an excellent platform for the inves- device performances. Subsequently, several kinds of 2H TMD
tigation of intrinsic superconductivity of monolayer NbSe2 but alloys have been achieved and show a tunable bandgap and dif-
also opportunities for further applications of 2D metallic TMDs ferent photoluminescence (PL) signals. However, the control
in electronics and electrocatalysis. Recently, CVD-grown 2D of the carrier type via constructing alloy structures has rarely
metallic TMDs have gained important attention from researchers been reported. Recently, Xu et al. realized the CVD-grown of
because of their fascinating characteristics and enhanced perfor- monolayer 1T′-ReS2xSe2(1−x) (x = 0–1.0) alloys with different
mance. For example, CVD-grown VS2 crystals[131] were reported bandgaps and carrier types.[148] During the synthesis process,
to possess a very high electrical conductivity (≈3 × 103 S cm−1). the precursors are rhenium trioxide (ReO3), S and Se powders
Interestingly, the grown VS2 crystals proved to have potential use (see the schematic illustration in Figure 11a). By varying the
in electronics applications and supercapacitors. Furthermore, an ratio of S and Se in different preparation batches, ReS2xSe2(1−x)
electronic device was fabricated on a monolayer of MoS2 while domains with different composition could be obtained and
VS2 nanosheets acted as contact electrodes for the device per- which demonstrate different bandgaps and carrier types.
formance evaluation (the inset OM image of Figure 10i). In this Furthermore, Figure 11b portrays a synthesized hexagonal
work, the mobility of the VS2-contacted monolayer MoS2 device ReS2xSe2(1−x) flake on a mica substrate. The TEM-EDS ele-
is ≈7.8 cm2 V−1 s−1 (Figure 10i), showing a higher mobility than mental maps from a ReS2xSe2(1−x) crystal (Figure 11c) reveal the
MoS2 devices using pure Ni/Au electrodes. This result indicates uniform distribution of all elements, indicating the formation

Adv. Mater. 2019, 1901694 1901694 (16 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 11. a) The schematic view of the synthesis system and the structure. Left part: the experiment set-up and reaction process; the right part:
different ratio of S and Se in ReS2xSe2(1–x) lead to bandgap and different conduction type. b) A typical OM image of a transferred alloy on SiO2/Si. c)
An ADF-TEM image and the corresponding element maps. d) PL results for different x. e) Transfer curves of the ReS0.84Se1.16 alloy transistors at dif-
ferent Vds. f) Transfer curves of ReS2xSe2(1−x)-based devices. g) Hole and electron mobilities of ReS2xSe2(1−x) alloys for different x. a–g) Reproduced with
permission.[147] Copyright 2017, Wiley-VCH.

of uniformity. Then, the PL spectra were utilized to evaluate corresponding SEM image (Figure 12b) demonstrates the for-
the composition-dependent bandgap (Figure 11d). The PL peak mation of a dendritic WS2(1−x)Se2x flake. The HRTEM image
position shifts clearly from 940 to 765 nm with the change reveals a perfect hexagonal lattice, suggesting the good crystal
in x. In other words, the corresponding band energy (Eg) shows quality (Figure 12c). Furthermore, the TEM-EDS elemental
an evident dependence on x, which presents a large range of maps indicate the uniform distribution of all elements, explic-
potential bandgaps. Hence, the alloy structures can largely itly confirming the formation of the alloy structure (Figure 12d).
develop the material system and satisfy the requirements for Interestingly, the PL peak position (≈622.6 nm) (Figure 12e)
electronic applications. Furthermore, back-gated FETs based on shows a clear redshift compared to that of pristine WS2
the obtained alloys were fabricated to investigate the electrical flakes.[153] For HER experiments, the samples were first trans-
properties with different components. The transfer curves of ferred onto Au foils and then served as an electrocatalyst. Using
the ReS0.84Se1.16-based device are presented in Figure 11e and linear sweep voltammetry (LSV) measurement, the polarization
indicate a typical n-type conduction behavior. Then Figure 11f curves of three kinds of samples (bare Au foil, dendritic WS2
demonstrates the electronic properties of a series of alloys. The and dendritic WS2(1−x)Se2x) were obtained (Figure 12f). Appar-
transfer curves indicate that the ReS2xSe2(1−x)-based devices ently, the dendritic WS2(1−x)Se2x displayed a very low overpoten-
(x = 0.6–1.0) exhibit an n-type conduction behavior, whereas a tial (156 mV at 10 mA cm−2) when compared to that of dendritic
p-type semiconducting property is observed in the ReS2xSe2(1−x) WS2 (≈310 mV at 10 mA cm−2). Figure 12g reveals a higher
alloys with x = 0–0.3. Interestingly, the ReS2xSe2(1−x)-based FETs exchange current density of ≈ 50.1 µA cm−2 for the WS2(1−x)Se2x
(x = 0.3–0.6) behave as a bipolar semiconductor. Figure 11g samples, which should be attributed to the extra active sites in
shows the hole and electron mobilities of several devices with the surface induced by the Se incorporation. Thus, both the
different conduction types, which confirms the excellent elec- alloy structure and fractal morphology are mainly responsible
tronic properties.[149] The electronic transport results reveal for the observed outstanding HER activity due to the abundant
the possibility of systematically modulating bandgaps and car- active sites. Furthermore, the smaller charge-transfer resist-
rier types by turning the x value in the 1T′-ReS2xSe2(1−x) alloy. ances (Rct) and series resistance (Rs) of dendritic WS2(1−x)Se2x
Hence, 1T′ ReS2xSe2(1−x) alloys differ from commonly 2H-TMD manifest an effective charge transfer after the Se incorporation
alloys by being able to modulate the conduction type and bring (Figure 12h).
about a variety of electronic properties. 1T′-TMD alloys with Briefly, 2D TMD alloys indeed demonstrate a valuable appli-
unique electronic and optical properties offer more possibilities cation potential for the use in electronic/optoelectronics and
for 2D TMD materials in electronic/optoelectronic applications. energy conversion due to their tunable bandgap and complex
Additionally, 2D TMD alloys have been widely utilized in atomic structures. This class of materials shows intriguing sys-
electrocatalysis owing to their unsaturated active sites at the tematic modulation in their electronic properties with slight
edges and basal plane,[150,151] which is different from that of variations in compositions, which endows more convenience
binary TMDs. For example, dendritic WS2(1−x)Se2x alloys were for satisfying the requirement of versatile applications, as well
obtained on an STO(100) substrate.[152] In this work, tung- as exploring novel physical phenomena. The CVD-growth of
sten trioxide (WO3), S and Se powders were used as precur- 2D TMD alloys greatly enriches the 2D family and provides
sors (see the related chemical reaction in Figure 12a). The abundant material systems for various applications.

Adv. Mater. 2019, 1901694 1901694 (17 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 12. a) CVD growth of WS2(1−x)Se2x nanosheets on STO(100). b) An SEM image of a typical as-grown flake. c) An HRTEM image. The inset
shows the corresponding SEAD pattern. d) EDX maps of all elements. e) PL results of different samples on SiO2/Si. f) Polarization curves of bare Au
foil, dendritic WS2 and dendritic WS2(1−x)Se2x. g) Extracted exchange current densities. h) Nyquist plots. a–h) Reproduced with permission.[152] Copyright
2017, Royal Society of Chemistry.

5. Heterostructures Based on TMD 5.1. Vertical Heterostructures Based on TMDs

Layered Materials
Recently, WS2/MoS2, and WSe2/MoS2 vertical heterostruc-
2D heterostructures based on diverse 2D layered materials tures have been reported by one-step and two-step CVD
have shown tremendous potential for electronic and opto- routes, and show interesting optical and electronic properties.
electronic applications due to their tunable band alignment Ajayan et al.[164] communicated the preparation of vertical
and atomically sharp interfaces.[154–158] Currently, the 2D WS2/MoS2 heterostructures for the first time and explored
family contains abundant material systems showing distinct their stacking structures and interesting optical properties.
properties. These atomic layers can also be combined to Via a facile CVD system, vertical WS2/MoS2 heterostructures
build a series of heteromaterials by different stacking strate- were obtained at a relatively high growth temperature using
gies, realizing several new structures, and exhibiting novel W, MoO3, and S powder as reaction precursors. During the
physicochemical properties. VdW heterostructures based one-step growth process, a MoS2 monolayer and WS2 mono­
on 2D TMDs were originally obtained via mechanical tech- layer were obtained successively due to the distinct evapora-
niques (mechanical exfoliating and stacking).[159–163] How- tion temperature of MoO3 and W powders. Meanwhile, for
ever, this transfer method is time-consuming, affords little the synthesis of the WS2 monolayer, a little tellurium powder
control over the stacking ways, and is even incompatible was added to the W powder, which is an ingenious design.
with batch production. Hence the technique is arduous and Figure 13a presents a typical OM image of the synthesized ver-
clearly unscalable for practical applications. Furthermore, tical WS2/MoS2 heterostructure, showing WS2/MoS2 bilayer
the preparation of lateral heterostructures without alloy triangles with a clear optical contrast. Then PL spectroscopy
formation remains challenging by mechanically in-plane was utilized to investigate the optical properties of the vertical
stitching. Hence, the a direct-growth technique for control- WS2/MoS2 heterostructure. The PL spectra (Figure 13b) from
lable and scalable production is urgently desired, especially points 1 and 2 in Figure 13a shows only one strong PL peak
for synthesizing lateral heterostructures. The CVD method at 680 nm, corresponding to monolayer MoS2. Different from
has recently been utilized to synthesize various 2D TMD this, the PL spectra from points 3 and 4 exhibit three promi-
materials, which forecast the preparation of heterostructures nent peaks at 630, 680, and 875 nm. Evidently, the peaks at 630
by the CVD method. In 2014, Ajayan et al.[164] reported the and 680 nm originate from monolayer WS2 and MoS2, respec-
synthesis of WS2/MoS2 heterostructures via a facile CVD tively. Most interestingly, the peak at 850 nm is observed at the
method. Subsequently, the growth of several heterostructures bilayer region for the first time and is attributes to the strong
have been communicated.[165–169] In the following sections, coupling between MoS2 and WS2. The PL result indicates a
direct-growth methods for vertical and lateral heterostruc- strong coupling between monolayer MoS2 and monolayer
tures/superlattices based on layered semiconducting TMDs, WS2. Furthermore, to study the atomic structures, aberra-
as well as their optical properties and electric/optoelectronic tion-corrected STEM was performed on the vertically stacked
performances, are discussed in detail. WS2/MoS2 heterostructure. The STEM image in Figure 13c

Adv. Mater. 2019, 1901694 1901694 (18 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 13. a) OM image of an as-grown WS2/MoS2 vertical heterostructure. b) PL results from different regions in (a). c) A STEM image of the stacked
WS2/MoS2 heterostructure and the corresponding intensity profile. d) Transfer curves of different devices. a–d) Reproduced with permission.[164] Copy-
right 2014, Springer Nature. e) A schematic illustration of the two-step growth of WSe2/SnS2 vertical p–n junctions. f) OM image of as-grown WSe2/
SnS2 vertical heterostructure. g) A PL map of the heterostructure. Scale bar: 60 µm. The inset is the OM image. Scale bar: 100 µm. h) An OM image
of a device based on a WSe2/SnS2 vertical heterostructure. i) Ids–Vbg curves across the p–n heterojunction region. j) Ids–Vds output curves across the
WSe2/SnS2 heterojunction. k) Ids–Vds curves. e–k) Reproduced under the terms of the CC-BY Creative Commons Attribution 4.0 International License
(http://creativecommons.org/licenses/by/4.0/).[175] Copyright 2017, The Authors, published by Springer Nature Publishing AG.

shows a perfect hexgonal lattice arrangement, which indicates WSe2 powder was used as the growth precursor. Then, the
the 2H stacking behavior of the as-grown WS2/MoS2 vertical het- authors used S and SnO2 powders to form SnS2 on WSe2/SiO2.
erostructure. The as-grown WS2/MoS2 vertical heterostructure An OM image (Figure 13f) demonstrates the typical morphology
predominantly exhibit a 2H stacking behavior, which suggests of the as-grown large-scale WSe2/SnS2 vertical heterostructure.
a uniform stacking of heterostructure can be achieved by CVD Furthermore, the as-grown WSe2 triangle was partially covered
growth. Back-gated FETs were fabricated on the CVD-grown by the SnS2 domain after the second growth. The PL mapping
WS2/MoS2 vertical heterostructure, and Figure 13d reveals of a WSe2/SnS2 heterostructure shows that the monolayer
that the mobility is ≈15–34 cm2 V−1 s−1 with a large on/off WSe2 region exhibits very strong PL intensity at 766 nm, which
ratio (>106). These results suggest heterostructures synthesized corresponds to direct excitons from monolayer WSe2, while the
by the CVD method and showing a distinct interface provide stacked WSe2/SnS2 region shows conspicuous PL quenching,
substantial opportunities for use in electronic/optoelectronic which indicates an energy transfer between monolayer SnS2
devices. Eventually, different research groups reported a series and monolayer WSe2 induced by a strong interface interaction
of vertical heterostructures with large domain size and high- (Figure 13g). Multielectrode back-gated FETs were fabricated on
quality crystals. Since then, not only n-n junctions but also synthesized WSe2/SnS2 vertical heterostructures (Figure 13h).
p-n junctions, have been synthesized by the CVD method, and The transfer curves of the heterostructure region (Figure 13i)
exhibit intriguing optical/physical properties and excellent per- show a p-type conduction behavior with an on/off ratio of 107.
formances.[169–174] Recently, Pan et al.[175] reported a two-step Additionally, the output curves show a clear current rectification
growth of vertical WSe2/SnS2 heterostructures (Figure 13e). behavior by tuning the gate voltages (Figure 13j), indicating the
Briefly, the authors first employed a physical vapor deposition formation of a p–n junction across the heterostructure region.
(PVD) method to grow a WSe2 monolayer on SiO2/Si, in which Furthermore, the CVD-grown WSe2/SnS2 p–n heterojunction

Adv. Mater. 2019, 1901694 1901694 (19 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 14. a) Schematic view of the synthesis process. b) An SEM image of the as-grown vertical heterostructures. Scale bar: 40 µm. c,d) Raman maps
at 160.5 cm−1 (the E2g mode of ReS2) and 351.3 cm−1 (the E2g mode of WS2), respectively. Scale bar: 5 µm. e,f) TEM images of the twinned ReS2/WS2
vertical heterostructures. g) The corresponding fast Fourier transform (FFT). The inset image reveals a rotation angle of 5.6°. h) The corresponding
model for the twinned ReS2/WS2 vertical heterostructure with a rotation angle of 5.6°. i) Theoretical calculations confirming the twinned growth mecha-
nism. a–h) Reproduced under the terms of the CC-BY Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/
by/4.0/).[179] Copyright 2016, The Authors, published by Springer Nature Publishing AG.

also exhibits excellent photoconductive properties. As shown in image in Figure 14e shows a corner of the ReS2/WS2 flake.
Figure 13k, Isc remarkably increases with the laser power. The Figure 14f shows a periodic moiré pattern in the stacked bilayer.
response speed of the WSe2/SnS2 heterojunction photodetector Additionally, the corresponding FFT extracted from the HRTEM
(≈500 µs) is markedly faster than that of most CVD-grown verti- image (Figure 14g) reveals that the two sets of hexagonal lattices
cally stacked 2D heterostructures.[173,174,176–178] from the ReS2/WS2 vertical heterostructure present a very small
However, maintaining the growth behavior of heterostructures rotation of 5.6°. Furthermore, similar moiré pattern can be seen
in a completely controllable way is still a challenge because of the through the simulation of the atomic structure, as provided in
complex nucleation processes and the inhomogeneous growth Figure 14h. This result indicates there is negligible strain between
rate. Hence, the synthesis of vertical heterostructures in a con- the ReS2 and WS2, suggesting a good lattice match between two
trolled manner is desirable and useful for further applications in constituent layers. Additionally, DFT calculations were conducted
electronics. Subsequently, Fu et al.[179] realized the twinned syn- to investigate the twinned growth mechanism in detail. Step (1)
thesis of ReS2/WS2 vertical heterostructures using Au foil and W– of Figure 14i shows the reaction process. Further first-principles
Re alloy foil as a growth substrate and growth source, respectively. calculations were applied to calculate the absorption energies.
Figure 14a schematically illustrates the synthesis of twinned As shown in (2)–(4) of Figure 14i, the adsorption of W atoms on
ReS2/WS2 vertical heterostructures. Additionally, the SEM image Au(111) needs an energy of 2.77 eV, while the adsorption energy
of as-grown ReS2/WS2 taken from the Au foil (Figure 14b) dem- for Re atoms on WS2 is 3.04 eV. Differently, the adsorption of Re
onstrates a high yield of uniform vertically stacked the ReS2/WS2 atoms on Au (111) needs a substantially weaker input of 1.52 eV.
with domain sizes reaching 600 µm2. Furthermore, Raman map- Thus, the Re atoms on Au foil could desorb easily at 900 °C,
ping was utilized to confirm the overlapped structure of ReS2/ which would prevent the formation of ReS2 on Au (111). Then
WS2 bilayer, as shown in Figure 14c,d. The Raman intensity after the formation of WS2 on the Au foil, the synthesis of ReS2
maps using the E2g mode of ReS2 (160.5 cm−1) and E2g mode of on WS2 surface could be subsequently realized, finally obtaining
WS2 (51.3 cm−1) show uniform contrasts, which reveals the for- the growth of twinned ReS2/WS2 crystals. This study indicates
mation of a twinned ReS2/WS2 bilayer. Furthermore, an HRTEM that the growth substrate is very important in the synthesis of
characterization was conducted to investigate the crystal quality twinned ReS2/WS2 crystals. Furthermore, the selected growth
and stacking behavior of ReS2/WS2 heterostructures. The TEM precursor (W–Re alloy foil) also plays a crucial role. Hence, for

Adv. Mater. 2019, 1901694 1901694 (20 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

the synthesis of twinned ReS2/WS2 crystals, the CVD growth which forms a WSe2–MoS2 concentric triangle. The separate
technique can be employed by using only W–Re alloy foil as the regions were clearly discriminated according to their distinct
precursor and adjusting the growth temperature to 900 °C. contrasts. Furthermore, the STEM image can clearly distinguish
Generally, such 2D atomically layered vertical heterostructures, the W, Mo, Se, and S atoms by different intensities, as shown
especially p–n junctions are essential for applications in elec- in Figure 15c. This work presents an efficient route to obtain
tronics. Recently, vertical heterostructures have exhibited excel- lateral heterostructures. However, recent studies have mostly
lent physical and electrical properties by integrating the properties been limited to the simplest heterostructures with only two
from single material via vertical stacking. Through constructing distinct materials. Whereas, multi-heterostructures or superlat-
vertical heterostructures using different TMD materials can tices have attracted an abundance of interest due to their unique
induce diverse interfaces and distinct physical phenomena. Mean- structures and electronic properties. In contrast, the growth of
while, the stacking angles or ways also can lead to different prop- these structures has rarely been reported and is limited by more
erties, which widens the application potentials. Then 2D vertical complicated growth processes and an elaborate control of the
heterostructures are essential platforms for realizing new struc- growth conditions. Surprisingly, Duan et al.[182] demonstrated
tures and developing new performance. At present, CVD method the synthesis of multi-heterostructures and superlattices based
provides an facile way for synthesis 2D heterostructures in a con- on 2D TMDs. In a typical synthesis procedure, a reverse carrier
trollable way. Designing the growth system and precursors are gas flow was applied during the sequential growth processes, as
vital for the synthesis of 2D large-scale heterostructures. shown in Figure 15d. The reverse gas flow prevents the exist-
ence of unnecessary precursors, thus, enabling a highly robust
epitaxial growth. The typical OM image in Figure 15e demon-
5.2. In-Plane Heterostructures Based on TMDs strates the successful synthesis of a superlattice on SiO2/Si.
Additionally, the corresponding Raman maps further confirm
5.2.1. Semiconductor–Semiconductor Lateral Heterostructures a well-defined spatial distribution with alternating WS2 and
WSe2 regions. Via this reliable strategy, a series of multi-heter-
Vertical heterostructures based on TMDs exhibit various ostructures has also been synthesized by precisely controlling
stacking structures and interesting properties. Nevertheless, the spatial modulation. Figure 15g,h exhibits an OM image
lateral heterostructures with their edges stitched would make and STEM image of a synthesized WS2–MoSe2–WSe2 triangle
it easier to precisely tune the band offset because the mate- with distinct contrasts. Hence, by the proper design and con-
rials within the lateral heterostructure are more spatially sep- trol of the carrier gas flow during the step-by-step growth, this
arated. However, the synthesis of lateral heterostructures still work realized the robust synthesis of multi-heterostructures
remains challenging because it is easier to form TMD alloys and superlattices, which still need a multistep growth process.
according to a thermodynamically stable manner. Recently, Consequently, Gutiérrez et al.[183] reported the growth of multi-
MoS2–MoSe2,[65] WS2–MoS2,[164] and WSe2–MoSe2[64,180] lat- heterostructures via a uniquely designed gas environment. The
eral heterostructures were reported and exhibited interesting authors used a single heterogeneous solid source as the pre-
optical and electrical properties. However, most of these lat- cursor, and the different compositions of the carrier gas can
eral heterostructures present a smooth interface with an alloy realize the synthesis of different TMD materials at each step.
region due to the co-existence of precursors in the vapor phase In brief, the formation of MoX2 occurs using N2 + H2O(g) as
during the reaction process. The alloy regions greatly influence the carrier gas, whereas a carrier gas of Ar + H2 (5%) benefits
the band structure of the heterostructure and impede investi- the synthesis of WX2 instead of MoX2. Eventually multijunc-
gations into intrinsic lateral heterostructures, hindering the tion lateral heterostructures can be obtained by cyclically var-
material from being used in different applications. Therefore, ying the carrier gas. Figure 15i shows a typical OM image of
the synthesis of lateral heterostructures without an alloy region the multi-heterostructures with a large domain size of up to
is very important, especially for p–n junctions. Following this, 200 µm. Furthermore, monolayer MoSe2 and WSe2 have distinct
lateral WSe2–MoS2 heterostructures with a clear interface were PL peaks. Then, the PL maps (Figure 15j) reveal the alternating
reported by Li et al.[181] via two-step epitaxial growth for the constituents of the multi-heterostructure domains, containing
first time. As schematically depicted in Figure 15a, the lateral MoSe2 and WSe2. Figure 15k demonstrates the evolution of the
WSe2–MoS2 heterostructures were synthesized by sequential PL intensity along with the position across the heterostructure.
CVD-growth of WSe2 and MoS2 on a sapphire substrate. First, Junctions 2 and 4 exhibit a dramatic jump, which suggests the
single-crystalline monolayer WSe2 triangles were prepared at a formation of junctions at sharp interfaces. Moreover, junctions
higher growth temperature (925 °C) using WO3 and S powders 2 and 4 behave substantially sharper than the junctions 3 and
as precursors. Subsequently, the MoS2 monolayer was obtained 5, which represent “smoother” interfaces. These results can
at 755 °C using a separate furnace, which can avoid introducing be clearly verified by high-angle annular dark-field (HAADF)-
other impurities. Notably, during the second growth process, an STEM images (Figure 15l and m). The results show that the
excess of the Mo precursor induced the vertical growth of MoS2 transition from MoSe2 to WSe2 produces a smooth interface
on WS2, while an excess of the S precursor resulted in the pro- with an alloy region, while a clear interface arises at the switch
duction of redundant WS2. Hence, the control of MoO3 and S from WSe2 to MoSe2. Notably, the change in carrier gas gives
vapors is the key for achieving a lateral heterostructure with rise to the formation of different interfaces in the multijunc-
clean interface. From the OM image of an as-grown WSe2– tion heterostructures. Furthermore, electronic devices were
MoS2 lateral heterostructures (Figure 15b), it can be seen that fabricated on the single-junction WSe2–MoSe2 heterostructure.
the centric WSe2 triangle is surrounded by monolayer MoS2, Figure 15n shows the OM image of the fabricated devices with

Adv. Mater. 2019, 1901694 1901694 (21 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 15. a) Two-step growth process. b) An OM image of an as-grown lateral heterostructure. c) A STEM image of the transferred WSe2–MoS2.
a–c) Reproduced with permission.[181] Copyright 2015, American Association for the Advancement of Science. d) A CVD system for the synthesis of
lateral heterostructure. e) An OM image of an as-grown superlattice. Scale bar: 5 µm. f) Raman intensity maps at 250 and 350 cm−1. Scale bar: 5 µm.
g) An OM image of the WS2–WSe2–MoS2 multi-heterostructure on SiO2/Si. Scale bar: 5 µm. h) A STEM image. Scale bar: 100 nm. d–h) Reproduced
with permission.[182] Copyright 2017, The Authors, published by American Association for the Advancement of Science; exclusive licensee AAAS. i) An
OM image of as-grown multijunction heterostructures. j) PL maps of an MoSe2–WSe2 multijunction heterostructure. k) PL intensities with the position
across the heterostructures. l,m) STEM images at different junctions. n) An OM image of the fabricated device based on an MoSe2–WSe2 heterostruc-
ture. o) Ids−Vbg curves of separate MoSe2 and WSe2 domains. p) Ids–Vds curves across the MoSe2–WSe2 junction. i–p) Reproduced with permission.[183]
Copyright 2018, Macmillan Publishers Limited, Springer Nature Publishing AG.

several electrodes based on the WSe2–MoSe2 heterostructure. indicating the realization of a p–n junction. This work pro-
Proven by their transport properties (Figure 15o), WSe2 shows vides a different synthetic method to realize the formation of
a p-type conducting behavior, while MoSe2 exhibited an n-type multijunctions or superlattices, which is versatile and scalable.
conducting behavior. Additionally, the I–V curves across the Although 2D heterostructures or multi-heterostructures have
junction (Figure 15p) display a typical rectification behavior, been synthesized by CVD methods, the synthesis of coherent

Adv. Mater. 2019, 1901694 1901694 (22 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 16. a) Schematic view of the modulated growth of a TMD superlattice. b) SEM images of the monolayer superlattices. Scale bars: 200 nm.
c) An ADF-STEM image at the heterointerface between WS2 and WSe2. d) A SAED pattern taken from an area with a diameter of 280 nm. e) A diagram
depicting the incoherent and coherent epitaxy evolution during the growth. f) The strain engineering of a WS2/WSe2 superlattice with different width
of WS2 and WSe2. g) PL spectra of intrinsic WS2 (dashed line) and WS2 from superlattices I to V in (f). The inset is a representative PL spectrum of a
WS2/WSe2 superlattice. h) An AFM image of a WS2/WSe2 superlattice. Scale bar: 500 nm. a–h) Reproduced with permission.[184] Copyright 2018, The
Authors, published by American Association for the Advancement of Science; exclusive licensee AAAS.

TMD superlattices remains a challenge. Such coherent het- concentration of each precursor was precisely controlled, which
erostructures or superlattices without lattice dislocations will finally induce the direct modulation of the supercell dimen-
establish the possibility of advanced scientific and technological sions. Figure 16b shows that dark and bright regions corre-
applications in high-efficiency light-emitting diodes. sponding to WS2 and WSe2 monolayers were clearly observed.
Very recently, Park et al.[184] employed a modulated metal– Apart from this, every triangular unit of WS2 and WSe2 is highly
organic CVD (MOCVD) process to synthesize a coherent symmetrical, and the widths of each unit can be as narrow as
monolayer WS2–WSe2 superlattices with precisely controlled 20 nm. The ADF-STEM image in Figure 16c, which is obtained
supercell dimensions and lattice coherence. In this study, the from an interface between WS2 and WSe2, shows a perfect
MOCVD system plays a vital role in the controllable growth, atomic arrangement with no misfit dislocations. Additionally,
which can be accurately controlled by tuning the gaseous pre- the SAED result taken from a typical superlattices displays
cursors. As schematically presented in Figure 16a, the spatial one hexagonal pattern (Figure 16d), which indicates the single-
distribution of WS2–WSe2 superlattices can be switched by crystal feature of the heterostructure. As schematically depicted
simply changing the precursors for a different growth. The in Figure 16e, the growth behavior here belongs to a coherent

Adv. Mater. 2019, 1901694 1901694 (23 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

epitaxy mechanism and the superlattices present coherent het- 2D NbS2–WS2 lateral metal–semiconductor heterostructures
erointerfaces. An interesting phenomenon observed in this deposited on a sapphire (0001) substrate using a CVD method
study is that the lattice coherence results in a tensile (com- (Figure 17).[135] The growth process is schematically presented
pressive) strain within the WS2 (WSe2) region in the superlat- in Figure 17a. Briefly, a WS2 monolayer was first synthesized by
tices, and depends on the supercell dimensions. As shown in an APCVD method using WO3 (mixed with a little NaCl) and S
Figure 16f, a narrower of WS2 region (dWS2) increases the ten- powders as precursors. Then, the WS2 monolayer supported by
sile strain in WS2. Additionally, a wider WSe2 (dWSe2) decreases the sapphire was used as the growth substrate for the second
the compressive strain in WSe2. Furthermore, the band struc- growth. A typical SEM image of a monolayer WS2 triangle
ture of WS2 and WSe2 are sensitive to the strain. The inset obtained from the first growth step is shown in Figure 17b, and
in Figure 16g exhibits two PL peaks that, correspond to WS2 the SEM image in Figure 17c demonstrates a triangular NbS2–
and WSe2. Compared to its intrinsic PL peak, the grown WS2 WS2 lateral heterostructure after the second growth step. The
shows a shift in its PL peak as large as 250 meV (Figure 16h), Raman spectra (Figure 17d) from the centric part (black one)
which is induced by the strong epitaxial strain. In addition, the and peripheral part (red one) exhibit totally distinct Raman
strain-dependent band structure brings about a modulation of peaks, well matched to those of WS2 and NbS2, respectively.
the electronic properties with the superlattice design, which The Raman results further indicate the formation of NbS2–WS2
may extend the practical applications of the structure. Basically, lateral heterostructures without alloy formation. Figure 17e
the relaxation of strain in the 2D films always induces out-of- shows the PL spectra collected from different regions within
plane deformations such as wrinkles and ripples. These defects the heterostructure with distinct characteristics, which also con-
may lead the film to have a nonflat surface with curved edges. firms the formation of NbS2–WS2 lateral heterostructures. The
Though their edges are highly strained, being subjected to Raman maps (378 cm−1 from NbS2; 352 cm−1 from WS2) in
alternating compressive and tensile stresses during the growth Figure 17f–g with clear boundaries demonstrate individual NbS2
process, the synthesized ultrathin WS2–WSe2 superlattices with and WS2 regions, verifying the formation of NbS2–WS2 lateral
lattice coherence and symmetry still maintained their 2D flat heterostructures with no alloy formation at the interface. STEM
structure. This result is attributed to the strong vdW interac- characterization was also conducted to investigate the atomic
tions between the adlayer and the growth substrates. How- structure and chemical distribution at the nanoscale dimension
ever, cooling the sample from its original growth temperature (Figure 17h,i). The STEM images taken from both sides away
(600 °C) induces ripples in the as-synthesized superlattice. The from the interface demonstrate the high-quality crystallinity of
AFM height image of a representative WS2–WSe2 superlattice the pure WS2 and NbS2 regions. Furthermore, the corresponding
in Figure 16h demonstrates clear out-of-plane ripples in WSe2, FFT pattern also confirms the presence of individual chemical
which again confirms the presence of a strong epitaxial strain component of WS2 and NbS2, as well as the formation of lateral
during the growth. Briefly, this work precisely describes the NbS2–WS2 heterostructures. The electronic transport properties
coherent growth of 2D superlattices, and creates a new avenue and the crystal quality of the as-synthesized NbS2–WS2 lateral
for the epitaxial growth of 2D materials and the modulation of heterostructures were studied by fabricating a series of FETs on
their optical properties. Meanwhile, the findings clearly indicate SiO2/Si. The characteristic transfer curves of this device reveal
the advent of a new interdisciplinary research direction that an n-type conduction type (Figure 17j). Importantly, Figure 17k
widens the scope for electronic and optoelectronic applications. demonstrates a well-defined rectification behavior. Even though
Different from vertical heterostructures, lateral TMD het- the device performance here is not dramatic due to the compli-
erostructures is more favorable to building the p–n junction cated device fabrication processes, the finding not only pushes
within a monolayer thickness, which provides grand value forward the synthesis of new metal–semiconductor heterostruc-
for constructing 2D ultrathin electronic devices. Meanwhile, tures but also creates new possibilities for electronic devices.
the pristine excellent properties of each material can be well In short, the successful growth of metal–semiconductor
remained within a lateral heterostructures. Additionally, the NbS2–WS2 heterostructures paves the way for researchers to
atomic interface in a lateral heterostructure exhibits a plenty explore new heteromaterials via a facile CVD method. And the
of interesting phenomena, resulting lots of attention. For synthesis of 2D metal–semiconductor heterostructures not only
example, by a strictly control during the growth process, the makes an important advance toward the batch production for
superlattice can be realized by tuning the width of each mate- new heterostructure systems, but also creates new platform
rial, which looks like a new material. Unlike the synthesis of for scientist to solve the electrode contact issue and realize the
vertical heterostructures by both CVD method and mechanical comprehensive 2D electronic devices.
stacking technique, CVD method should be the most suitable
way for the synthesis of the lateral heterostructures with atomic
sharp interface, as well as superlattices. 6. Summary and Future Outlook
Herein, we have summarized the latest development in the syn-
5.2.2. Metal–Semiconductor Heterostructures Based on TMDs thesis of 2D TMD materials by using the facile CVD approach.
We have mainly focused on the recent progress made in
Though several p–n, p–p, and n–n junctions have been synthe- growing 2D materials, ranging from new materials and alloys
sized by facile CVD methods in recent years, metal–semicon- to heterostructures/superlattices (Figure 18). The controlled
ductor TMD heterostructures have not yet been directly realized growth of these materials via the CVD technique depends on
by the CVD approach. Recently, our group first introduced several key parameters, such as the temperature, carrier gas,

Adv. Mater. 2019, 1901694 1901694 (24 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

Figure 17. a) CVD growth process. b) A SEM image of a typical monolayer WS2 triangle. Scale bar: 25 µm. c) An SEM image of an NbS2–WS2 lateral
heterostructure. Scale bar: 20 µm. Inset: an AFM image shows the thickness of the NbS2 region is ≈10 nm. d) Raman analysis of the NbS2–WS2 lateral
heterostructure. Inset: SEM image of a NbS2–WS2 triangle. e) PL spectra from different regions of the NbS2–WS2 lateral heterostructure. f,g) 2D Raman
maps of the NbS2 region and WS2 region within the NbS2–WS2 lateral heterostructure, respectively. h,i) STEM images of a WS2 region and NbS2 region
away from the interface and the corresponding FFT images, respectively. j) The transfer curves of device fabricated with the NbS2–WS2 lateral hetero-
structure. k) The corresponding Ids–Vds output curves. a–k) Reproduced with permission.[135] Copyright 2018, Wiley-VCH.

growth time, and the ratio of the precursors. We have further behavior of 2D TMDs. Furthermore, some new kinds of semi-
summarized the effect of the aforementioned factors on the conductor TMDs have been explored, enriching the material
nucleation and growth of various 2D materials. Additionally, systems, and physicochemical properties of the 2D family.
experimental efforts have been enacted to prepare large crystals 2D metallic transition metal dichalcogenides, as another
of 2D TMDs using the facile CVD technique that can further important component of the TMD family, have been considered
offer the opportunity to use this class of materials in various as ideal archetypes for unearthing physics, such as magnetism,
applications, such as future nanoelectronics, optoelectronics superconductivity, CDW transitions, for various applications.
and energy conversion. To obtain wafer-scale monolayer TMD Very recently, the CVD method has been utilized to grow 2D
films and millimeter-sized TMD single crystals, a carefully metallic TMDs. Several kinds of ultrathin metallic TMD mate-
designed growth system and the appropriate choice of growth rials, such as VS2, TaS2, NbSe2, and NbS2, have been successfully
substrates are vital. By considering this basic principle, wafer- synthesized along with a large domain size and a high quality.
scale monolayer TMD film and millimeter-sized TMD single These metallic TMDs in their 2D limit exhibit excellent physical
crystals have been realized via the facile CVD method. Various properties, such as superconductivity and CDW phase transi-
2D materials fabricated via this technique have been reported tions. Additionally, benefitting from their excellent conductivity,
to exhibit excellent electronic/optoelectronic properties. Apart they serve as a good contact electrode in electronic devices and
from controlling the growth parameters, the design of the as efficient catalysts for the HER. Hence, further exploration
growth system and the growth substrate are also is crucial to of 2D metallic TMDs could promote significant advancements
the 2D growth. Through varying the design of a growth system, toward the batch production of several new 2D TMDs.
one can alter the reaction atmosphere and growth mechanism. The preparation of diverse TMD materials, alloys and het-
Meanwhile, different growth substrates possess different lattice erostructures/superlattices based on 2D TMD crystals has also
symmetries, surface roughness values, interface interaction, attracted immense interests due to their tunable band structures
and surface adsorption energies for TMD materials. Hence, and novel properties. Thus far, 2D TMD alloys with tunable
these properties could all fundamentally influence the growth bandgaps and complex structures have been successfully

Adv. Mater. 2019, 1901694 1901694 (25 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

electro-optical devices. In addition, making

2D magnetic heterostructures[190] by com-
bining magnetic material and semicon-
ducting TMDs has proven to be an effective
way to engineer highly functional materials
for spintronic applications. Nonetheless, at
present, the application of such an assembly
is rarely explored, let alone investigations
into the material/device preparation.
Despite the considerable advances that
have been achieved in the growth of 2D lay-
ered TMDs via the CVD method, there are
still some challenges that require further
attention to enhance their practical applica-
tions in future electronics/optoelectronics.
As outlined in this review, the intrinsic
growth mechanism, the batch production
of wafer-scale crystals, and the precise con-
struction of significantly more appropriate
vdW heterostructures/superlattices are all
indispensable. Furthermore, the explora-
tion of 2D magnetic materials and magnetic
heterostructures remains full of challenges.
Most notably, further progress in the syn-
thesis and device demonstration of 2D mate-
Figure 18. Schematic representation summarizing the research contents (growth behavior,
new materials/structures, properties and applications based on 2D TMDs synthesized via the
rials would efficiently put forward the prac-
CVD approach). tical performance for electronics/optoelec-
tronics in the future. Hence, we hope that
realized. This class of materials shows intriguing systematic this review has provided an understanding
modulations of their electronic properties with slight varia- of the development of 2D materials via the CVD technique
tions in compositions. Hence, they have been widely utilized and the factors affecting this process, as well as the means to
in electronic devices and electrocatalysis. Recently, heterostruc- broaden the application of 2D TMDs to meet current energy-
tures, multi-heterostructures, and superlattices with precisely related problems.
controlled spatial modulation have been achieved via a modi-
fied CVD approach, which represents a grand breakthrough in
2D growth. Additionally, FETs that are based on CVD-grown Acknowledgements
p–n junctions have been fabricated, and their excellent elec-
tronic properties have been recorded. This advancement This work was supported by the Ministry of Science and Technology
reveals a new bright path for the further development of 2D of China (No. 2016YFA0200700), the National Natural Science
Foundation of China (Nos. 61625401, 21703047, 61474033, 61574050,
devices. Intriguingly, a metal–semiconductor heterostructures and 11674072), the Strategic Priority Research Program of the Chinese
has also been obtained via a two-step CVD method, which Academy of Sciences (Grant No. XDA09040201), and the CAS Key
opens the a door for the creation of new heterostructures and Laboratory of Nanosystem and Hierarchical Fabrication. The authors
offers more possibilities for further applications. However, the also gratefully acknowledge the support of the Youth Innovation
robust growth of large-scale uniform metal–semiconductor Promotion Association CAS.
heterostructures in a controllable fashion remains a major
challenge.
On the other hand, 2D magnets such as Fe3GeTe2,[185] Conflict of Interest
CrGeTe3,[186] and CrI3,[187] have recently sparked enormous
The authors declare no conflict of interest.
attention, due to their unique magnetism in the 2D limit and
expressed potential for application in spintronic devices. How-
ever, most of these magnetic materials in the 2D limit are
unstable in air, which makes further investigations into their Keywords
intrinsic magnetism and practical applications difficult. Broadly
speaking, the family of 2D TMD materials contains abundant 2D layered materials, CVD approach, growth behavior, heterostructures,
transition metal dichalcogenides
compounds, which may provide opportunities for exploiting
some new magnetic materials. [188–190]
Hence, the control-
lable synthesis of 2D magnetic materials with a good stability Received: March 18, 2019
also deserve further research. Meanwhile, 2D heterostruc- Revised: May 20, 2019
tures based on TMDs have shown great potential for flexible Published online:

Adv. Mater. 2019, 1901694 1901694 (26 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

[1] A. K. Geim, K. S. Novoselov, Nat. Mater. 2007, 6, 183. [32] T. F. Jaramillo, K. P. Jorgensen, J. Bonde, J. H. Nielsen, S. Horch,
[2] A. K. Geim, Science 2009, 324, 1530. I. Chorkendorff, Science 2007, 317, 100.
[3] K. S. Novoselov, D. Jiang, F. Schedin, T. J. Booth, V. V. Khotkevich, [33] L. P. Hansen, Q. M. Ramasse, C. Kisielowski, M. Brorson,
S. V. Morozov, A. K. Geim, Proc. Natl. Acad. Sci. USA 2005, 102, E. Johnson, H. Topsoe, S. Helveg, Angew. Chem. 2011, 50, 10153.
10451. [34] C. Tsai, F. Abildpedersen, J. K. Norskov, Nano Lett. 2014, 14, 1381.
[4] K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, [35] M. Bollinger, J. V. Lauritsen, K. W. Jacobsen, J. K. Norskov,
M. I. Katsnelson, I. V. Grigorieva, S. V. Dubonos, A. Firsov, Nature S. Helveg, F. Besenbacher, Phys. Rev. Lett. 2001, 87, 196803.
2005, 438, 197. [36] B. Hinnemann, P. G. Moses, J. Bonde, K. P. Jorgensen,
[5] C. Lee, X. Wei, J. W. Kysar, J. Hone, Science 2008, 321, 385. J. H. Nielsen, S. Horch, I. Chorkendorff, J. K. Norskov, J. Am.
[6] Y. Zhang, Y. W. Tan, H. L. Stormer, P. Kim, Nature 2005, 438, 201. Chem. Soc. 2005, 127, 5308.
[7] B. Radisavljevic, A. Radenovic, J. Brivio, V. Giacometti, A. Kis, Nat. [37] H. Zeng, J. Dai, W. Yao, D. Xiao, X. Cui, Nat. Nanotechnol. 2012,
Nanotechnol. 2011, 6, 147. 7, 490.
[8] S. Z. Butler, S. M. Hollen, L. Cao, Y. Cui, J. Gupta, H. R. Gutierrez, [38] R. Suzuki, M. Sakano, Y. Zhang, R. Akashi, D. Morikawa,
T. F. Heinz, S. S. Hong, J. Huang, A. Ismach, E. J. Halperin, A. Harasawa, K. Yaji, K. Kuroda, K. Miyamoto, T. Okuda, Nat.
M. Kuno, V. V. Plashnitsa, R. D. Robinson, R. S. Ruoff, Nanotechnol. 2014, 9, 611.
S. Salahuddin, J. Shan, L. Shi, M. G. Spencer, M. Terrones, [39] S. X. Wu, J. Ross, G. Liu, G. Aivazian, A. M. Jones, Z. Fei, W. Zhu,
W. Windl, J. E. Goldberger, ACS Nano 2013, 7, 2898. D. Xiao, W. Yao, D. Cobden, Nat. Phys. 2013, 9, 149.
[9] Q. H. Wang, K. Kalantarzadeh, A. Kis, J. N. Coleman, M. S. Strano, [40] K. F. Mak, K. Mcgill, J. Park, P. L. Mceuen, Science 2014, 344, 1489.
Nat. Nanotechnol. 2012, 7, 699. [41] A. M. Srivastava, M. Sidler, A. Allain, D. Lembke, A. Kis,
[10] C. Dean, A. Young, I. Meric, C. Lee, L. Wang, S. Sorgenfrei, A. Imamoglu, Nat. Phys. 2015, 11, 141.
K. Watanabe, T. Taniguchi, P. Kim, K. L. Shepard, J. Hone, Nat. [42] S. B. Desai, S. R. Madhvapathy, A. B. Sachid, J. P. Llinas, Q. Wang,
Nanotechnol. 2010, 5, 722. G. H. Ahn, G. Pitner, M. J. Kim, J. Bokor, C. Hu, H. S. Wong,
[11] K. Watanabe, T. Taniguchi, H. Kanda, Nat. Mater. 2004, 3, 404. A. Javey, Science 2016, 354, 99.
[12] F. Xia, T. Mueller, Y. Lin, P. Avouris, Conf. Lasers Electro-Opt. Eur. [43] Y. Zhu, Y. Li, G. Arefe, R. A. Burke, C. Tan, Y. Hao, X. Liu, X. Liu,
2010, 4, 839. W. J. Yoo, M. Dubey, Q. Lin, J. C. Hone, Nano Lett. 2018, 18, 3807.
[13] M. Chhowalla, H. S. Shin, G. Eda, L. Li, K. P. Loh, H. Zhang, Nat. [44] E. Liu, Y. Fu, Y. Wang, Y. Feng, H. Liu, X. Wan, W. Zhou, B. Wang,
Chem. 2013, 5, 263. L. Shao, C. H. Ho, Y.-S. Huang, Z. Cao, L. Wang, A. Li, J. Zeng,
[14] X. Huang, Z. Zeng, H. Zhang, Chem. Soc. Rev. 2013, 42, 1934. F. Song, X. Wang, Y. Shi, H. Yuan, H. Y. Hwang, Y. Cui, F. Miao,
[15] L. Britnell, R. Ribeiro, A. Eckmann, R. Jalil, B. D. Belle, D. Xing, Nat. Commun. 2015, 6, 6991.
A. Mishchenko, Y. J. Kim, R. V. Gorbachev, T. Georgiou, [45] Y. Zhang, T. Oka, R. Suzuki, J. T. Ye, Y. Iwasa, Science 2014, 344, 725.
S. V. Morozov, Science 2013, 340, 1311. [46] H. Zeng, G. Liu, J. Dai, Y. Yan, B. Zhu, R. He, L. Xie, S. J. Xu,
[16] K. F. Mak, K. He, C. Lee, G. Lee, J. Hone, T. F. Heinz, J. Shan, Nat. X. Chen, W. Yao, Sci. Rep. 2013, 3, 1608.
Mater. 2013, 12, 207. [47] B. Sipos, A. F. Kusmartseva, A. Akrap, H. Berger, L. Forro, E. Tutis,
[17] D. Voiry, M. Salehi, R. Silva, T. Fujita, M. Chen, T. Asefa, Nat. Mater. 2008, 7, 960.
V. B. Shenoy, G. Eda, M. Chhowalla, Nano Lett. 2013, 13, 6222. [48] J. N. Coleman, M. Lotya, A. O. Neill, S. D. Bergin, P. J. King,
[18] M. A. Lukowski, A. S. Daniel, F. Meng, A. Forticaux, L. Li, S. Jin, U. Khan, K. Young, A. Gaucher, S. De, R. J. Smith, I. V. Shvets,
J. Am. Chem. Soc. 2013, 135, 10274. S. K. Arora, G. Stanton, H. Y. Kim, K. Lee, G. T. Kim,
[19] Q. Ji, Y. Zhang, J. Shi, J. Sun, Y. Zhang, Z. Liu, Adv. Mater. 2016, 28, G. S. Duesberg, T. Hallam, J. J. Boland, J. Wang, J. F. Donegan,
6207. J. C. Grunlan, G. P. Moriarty, A. Shmeliov, R. J. Nicholls,
[20] T. Yokoya, T. Kiss, A. Chainani, S. Shin, M. Nohara, H. Takagi, J. M. Perkins, E. M. Grieveson, K. Theuwissen, D. W. Mccomb,
Science 2001, 294, 2518. P. D. Nellist, V. Nicolosi, Science 2011, 331, 568.
[21] H. S. S. R. Matte, A. Gomathi, A. K. Manna, D. J. Late, R. Datta, [49] D. Voiry, H. Yamaguchi, J. Li, R. Silva, D. C. B. Alves, T. Fujita,
S. K. Pati, C. N. R. Rao, Angew. Chem. 2010, 49, 4059. M. Chen, T. Asefa, V. B. Shenoy, G. Eda, Nat. Mater. 2013, 12, 850.
[22] W. Zhao, Z. Ghorannevis, L. Chu, M. Toh, C. Kloc, P. Tan, G. Eda, [50] Z. Zeng, Z. Yin, X. Huang, H. Li, Q. He, G. Lu, F. Y. C. Boey,
ACS Nano 2013, 7, 791. H. Zhang, Angew. Chem. 2011, 50, 11093.
[23] A. Splendiani, L. Sun, Y. Zhang, T. Li, J. Kim, C. Chim, G. Galli, [51] G. Eda, H. Yamaguchi, D. Voiry, T. Fujita, M. Chen, M. Chhowalla,
F. Wang, Nano Lett. 2010, 10, 1271. Nano Lett. 2011, 11, 5111.
[24] K. F. Mak, C. Lee, J. Hone, J. Shan, T. F. Heinz, Phys. Rev. Lett. [52] S. S. Chou, B. Kaehr, J. Kim, B. M. Foley, M. De, P. E. Hopkins,
2010, 105, 136805. J. Huang, C. J. Brinker, V. P. Dravid, Angew. Chem. 2013, 52, 4160.
[25] H. R. Gutierrez, N. Perealopez, A. L. Elias, A. Berkdemir, B. Wang, [53] X. Li, W. Cai, J. An, S. Kim, J. Nah, D. Yang, R. D. Piner,
R. Lv, F. Lopezurias, V. H. Crespi, H. Terrones, M. Terrones, Nano A. Velamakanni, I. Jung, E. Tutuc, S. K. Banerjee, L. Colombo,
Lett. 2013, 13, 3447. R. S. Ruoff, Science 2009, 324, 1312.
[26] H. Shi, H. Pan, Y. Zhang, B. I. Yakobson, Phys. Rev. B 2013, 88, 205305. [54] Q. Yu, L. A. Jauregui, W. Wu, R. Colby, J. Tian, Z. Su, H. Cao,
[27] B. W. H. Baugher, H. O. H. Churchill, Y. Yang, P. Jarilloherrero, Z. Liu, D. Pandey, D. Wei, Nat. Mater. 2011, 10, 443.
Nat. Nanotechnol. 2014, 9, 262. [55] Y. Lee, X. Zhang, W. Zhang, M. Chang, C. Lin, K. Chang, Y. Yu,
[28] O. Lopezsanchez, D. Lembke, M. Kayci, A. Radenovic, A. Kis, Nat. J. T. Wang, C. Chang, L. Li, Adv. Mater. 2012, 24, 2320.
Nanotechnol. 2013, 8, 497. [56] Y. Shi, W. Zhou, A. Lu, W. Fang, Y. Lee, A. Hsu, S. M. Kim,
[29] K. Xu, D. Chen, F. Yang, Z. Wang, L. Yin, F. Wang, R. Cheng, K. K. Kim, H. Y. Yang, L. Li, Nano Lett. 2012, 12, 2784.
K. Liu, J. Xiong, Q. Liu, Nano Lett. 2017, 17, 1065. [57] Q. Ji, Y. Zhang, T. Gao, Y. Zhang, D. Ma, M. Liu, Y. Chen, X. Qiao,
[30] A. Nourbakhsh, A. Zubair, R. N. Sajjad, K. G. A. Tavakkoli, P. Tan, M. Kan, Nano Lett. 2013, 13, 3870.
W. Chen, S. Fang, X. Ling, J. Kong, M. S. Dresselhaus, E. Kaxiras, [58] S. Najmaei, Z. Liu, W. Zhou, X. Zou, G. Shi, S. Lei, B. I. Yakobson,
Nano Lett. 2016, 16, 7798. J. Idrobo, P. M. Ajayan, J. Lou, Nat. Mater. 2013, 12, 754.
[31] A. V. Der Zande, P. Y. Huang, D. Chenet, T. C. Berkelbach, Y. You, [59] X. Lu, M. I. B. Utama, J. Lin, X. Gong, J. Zhang, Y. Zhao,
G. Lee, T. F. Heinz, D. R. Reichman, D. A. Muller, J. Hone, Nat. S. T. Pantelides, J. Wang, Z. Dong, Z. Liu, Nano Lett. 2014, 14,
Mater. 2013, 12, 554. 2419.

Adv. Mater. 2019, 1901694 1901694 (27 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

[60] Y. H. Chang, W. Zhang, Y. Zhu, Y. Han, J. Pu, J. K. Chang, [85] C. Li, Z. Yu, Q. Ji, J. Shi, Z. Chen, X. Zhou, Q. Fang, Y. Zhang, 2D
W. T. Hsu, J. K. Huang, C. L. Hsu, M. H. Chiu, T. Takenobu, H. Li, Mater. 2016, 3, 035001.
C.-I. Wu, W.-H. Chang, A. T. S. Wee, L.-J. Li, ACS Nano 2014, 8, [86] Y. Zhang, Q. Ji, J. Wen, J. Li, C. Li, J. Shi, X. Zhou, K. Shi, H. Chen, Y. Li,
8582. S. Deng, N. Xu, Z. Liu, Y. Zhang, Adv. Funct. Mater. 2016, 26, 3299.
[61] Y. Zhang, Y. Zhang, Q. Ji, J. Ju, H. Yuan, J. Shi, T. Gao, D. Ma, [87] T. A. Witten, L. M. Sander, Phys. Rev. Lett. 1981, 47, 1400.
M. Liu, Y. Chen, ACS Nano 2013, 7, 8963. [88] H. Röder, K. Bromann, H. Brune, K. Kern, Phys. Rev. Lett. 1995, 74,
[62] Y. Gao, Z. Liu, D. Sun, L. Huang, L. Ma, L. Yin, T. Ma, Z. Zhang, 3217.
X. L. Ma, L. Peng, Nat. Commun. 2015, 6, 8569. [89] D. Wu, T. Min, J. Zhou, C. Li, G. Ma, G. Lu, M. Xia, Z. Gu, Sci. Rep.
[63] K. Keyshar, Y. Gong, G. Ye, G. Brunetto, W. Zhou, D. P. Cole, 2017, 7, 15166.
K. Hackenberg, Y. He, L. D. Machado, M. A. Kabbani, Adv. Mater. [90] J. Chen, X. Zhao, S. J. R. Tan, H. Xu, B. Wu, B. Liu, D. Fu, W. Fu,
2015, 27, 4640. D. Geng, Y. Liu, J. Am. Chem. Soc. 2017, 139, 1073.
[64] Y. Gong, S. Lei, G. Ye, B. Li, Y. He, K. Keyshar, X. Zhang, Q. Wang, [91] Y. Chen, J. Sun, J. Gao, F. Du, Q. Han, Y. Nie, Z. Chen,
J. Lou, Z. Liu, Nano Lett. 2015, 15, 6135. A. Bachmatiuk, M. K. Priydarshi, D. Ma, X. Song, X. Wu, C. Xiong,
[65] X. Duan, C. Wang, J. C. Shaw, R. Cheng, Y. Chen, H. Li, X. Wu, M. H. Rümmeli, F. Ding, Y. Zhang, Z. Liu, Adv. Mater. 2015, 27, 7839.
Y. Tang, Q. Zhang, A. Pan, Nat. Nanotechnol. 2014, 9, 1024. [92] J. Shi, D. Ma, G. Han, Y. Zhang, Q. Ji, T. Gao, J. Sun, X. Song,
[66] X. Ling, Y. Lee, Y. Lin, W. Fang, L. Yu, M. S. Dresselhaus, J. Kong, C. Li, Y. Zhang, ACS Nano 2014, 8, 10196.
Nano Lett. 2014, 14, 464. [93] J. Shi, Y. Yang, Y. Zhang, D. Ma, W. Wei, Q. Ji, Y. Zhang, X. Song,
[67] Y. Zhan, Z. Liu, S. Najmaei, P. M. Ajayan, J. Lou, Small 2012, 8, T. Gao, C. Li, X. Bao, Z. Liu, Q. Fu, Y. Zhang, Adv. Funct. Mater.
966. 2015, 25, 842.
[68] Y. Kobayashi, S. Sasaki, S. Mori, H. Hibino, Z. Liu, K. Watanabe, [94] P. Yang, X. Zou, Z. Zhang, M. Hong, J. Shi, S. Chen, J. Shu,
T. Taniguchi, K. Suenaga, Y. Maniwa, Y. Miyata, ACS Nano 2015, 9, L. Zhao, S. Jiang, X. Zhou, Nat. Commun. 2018, 9, 979.
4056. [95] J. Zhou, J. Lin, X. Huang, Y. Zhou, Y. Chen, J. Xia, H. Wang, Y. Xie,
[69] J. Rhyee, J. Kwon, P. Dak, J. H. Kim, S. M. Kim, J. Park, Y. K. Hong, H. Yu, J. Lei, D. Wu, F. Liu, Q. Fu, Q. Zeng, C.-H. Hsu, C. Yang,
W. G. Song, I. Omkaram, M. A. Alam, S. Kim, Adv. Mater. 2016, 28, L. Lu, T. Yu, Z. Shen, H. Lin, B. I. Yakobson, Q. Liu, K. Suenaga,
2316. G. Liu, Z. Liu, Nature 2018, 556, 355.
[70] C. H. Naylor, W. M. Parkin, J. Ping, Z. Gao, Y. R. Zhou, Y. M. Kim, [96] J. Liu, M. Zeng, L. Wang, Y. Chen, Z. Xing, T. Zhang, Z. Liu, J. Zuo,
F. Streller, R. W. Carpick, A. M. Rappe, M. Drndic, J. M. Kikkawa, F. Nan, R. G. Mendes, Small 2016, 12, 5741.
A. T. Johnson, Nano Lett. 2016, 16, 4297. [97] J. D. Wood, S. W. Schmucker, A. S. Lyons, E. Pop, J. W. Lyding,
[71] L. Zhou, A. Zubair, Z. Wang, X. Zhang, F. Ouyang, K. Xu, W. Fang, Nano Lett. 2011, 11, 4547.
K. Ueno, J. Li, T. Palacios, Adv. Mater. 2016, 28, 9526. [98] Y. Li, J. Kang, J. Li, RSC Adv. 2014, 4, 7396.
[72] E. Loginova, N. C. Bartelt, P. J. Feibelman, K. F. Mccarty, New J. [99] C. Kreis, S. Werth, R. Adelung, L. Kipp, M. Skibowski,
Phys. 2008, 10, 093026. E. E. Krasovskii, W. Schattke, Phys. Rev. B 2003, 68, 235331.
[73] J. V. Lauritsen, J. Kibsgaard, S. Helveg, H. Topsoe, B. S. Clausen, [100] F. Gianluca, B. Francesco, I. Giuseppe, P. Tomás, N. Daniel,
E. Laegsgaard, F. Besenbacher, Nat. Nanotechnol. 2007, 2, 53. S. Alan, S. K. Banerjee, C. Luigi, Nat. Nanotechnol. 2014, 9, 768.
[74] Y. Gao, Y. Hong, L. Yin, Z. Wu, Z. Yang, M. Chen, Z. Liu, T. Ma, [101] L. Li, X. Fang, T. Zhai, M. Liao, U. K. Gautam, X. Wu, Y. Koide,
D. Sun, Z. Ni, X. L. Ma, H. M. Cheng, W. Ren, Adv. Mater. 2017, Y. Bando, D. Golberg, Adv. Mater. 2010, 22, 4151.
29, 1700990. [102] W. Zhang, Z. Huang, W. Zhang, Y. Li, Nano Res. 2014, 7, 1731.
[75] J. Chen, X. Zhao, G. Grinblat, Z. Chen, S. J. R. Tan, W. Fu, Z. Ding, [103] M. Zhang, Y. Zhu, X. Wang, Q. Feng, S. Qiao, W. Wen, Y. Chen,
I. Abdelwahab, Y. Li, D. Geng, Y. Liu, K. Leng, B. Liu, W. Liu, M. Cui, J. Zhang, C. Cai, L. Xie, J. Am. Chem. Soc. 2015, 137, 7051.
W. Tang, S. A. Maier, S. J. Pennycook, K. P. Loh, Adv. Mater. 2018, [104] L. Fu, F. Wang, B. Wu, N. Wu, W. Huang, H. Wang, C. Jin,
30, 1704674. L. Zhuang, J. He, L. Fu, Adv. Mater. 2017, 29, 1700439.
[76] S. Wang, Y. Rong, Y. Fan, M. Pacios, H. Bhaskaran, K. He, [105] Y. H. Lee, X. Q. Zhang, W. Zhang, M. T. Chang, C. T. Lin,
J. H. Warner, Chem. Mater. 2014, 26, 6371. K. D. Chang, Y. C. Yu, J. T. Wang, C. S. Chang, L.-J. Li, Adv. Mater.
[77] D. Li, Z. Xiao, S. Mu, F. Wang, Y. Liu, J. Song, X. Huang, L. Jiang, 2012, 24, 2320.
J. Xiao, L. Liu, S. Ducharme, B. Cui, X. Hong, L. Jiang, J. F. Silvain, [106] S. Kim, A. Konar, W. S. Hwang, J. H. Lee, J. Lee, J. Yang, C. Jung,
Y. Lu, Nano Lett. 2018, 18, 2021. H. Kim, J. B. Yoo, J. Y. Choi, Y. W. Jin, S. Y. Lee, D. Jena, W. Choi,
[78] Y. Lim, K. Priyadarshi, F. Bussolotti, P. K. Gogoi, X. Cui, M. Yang, K. Kim, Nat. Commun. 2012, 3, 1011.
J. S. Pan, S. W. Tong, S. J. Wang, S. J. Pennycook, K. E. J. Goh, [107] Z. Yu, Y. Pan, Y. Shen, Z. Wang, O. ZY, T. Xu, R. Xin, L. Pan,
A. T. S. Wee, S. L. Wong, D. Z. Chi, ACS Nano 2018, 12, B. Wang, L. Sun, J. Wang, G. Zhang, Y. Zhang, Y. Shi, X. Wang,
1339. Nat. Commun. 2014, 5, 5290.
[79] L. Zhou, K. Xu, A. Zubair, A. D. Liao, W. Fang, F. Ouyang, Y. Lee, [108] W. Choi, M. Y. Cho, A. Konar, J. H. Lee, G. B. Cha, S. C. Hong,
K. Ueno, R. Saito, T. Palacios, J. Kong, M. S. Dresselhaus, J. Am. S. Kim, J. Kim, D. Jena, J. Joo, S. Kim, Adv. Mater. 2012, 24,
Chem. Soc. 2015, 137, 11892. 5832.
[80] H. Yu, M. Liao, W. Zhao, G. Liu, X. Zhou, Z. Wei, X. Xu, K. Liu, [109] Y. Zhao, J. Qiao, Z. Yu, P. Yu, K. Xu, S. P. Lau, W. Zhou, Z. Liu,
Z. Hu, K. Deng, S. Zhou, J. A. Shi, L. Gu, C. Shen, T. Zhang, L. Du, X. Wang, W. Ji, Y. Chai, Adv. Mater. 2017, 29, 1604230.
L. Xie, J. Zhu, W. Chen, R. Yang, D. Shi, G. Zhang, ACS Nano 2017, [110] C. Yim, K. Lee, N. Mcevoy, M. O. Brien, S. Riazimehr,
11, 12001. N. C. Berner, C. P. Cullen, J. Kotakoski, J. C. Meyer, M. C. Lemme,
[81] C. Cong, J. Shang, X. Wu, B. Cao, N. Peimyoo, C. Qiu, L. Sun, G. S. Duesberg, ACS Nano 2016, 10, 9558.
T. Yu, Adv. Opt. Mater. 2014, 2, 131. [111] Z. Wang, Q. Li, F. Besenbacher, M. Dong, Adv. Mater. 2016, 28,
[82] C. Yan, G. Lin, Z. Xing, J. Guo, Y. Li, Adv. Funct. Mater. 2017, 27, 10224.
1702918. [112] L. J. Li, E. C. O’Farrell, K. P. Loh, G. Eda, Ö. B, A. H. Castro Neto,
[83] Q. Ji, Y. Zhang, Y. Zhang, Z. Liu, Chem. Soc. Rev. 2015, 44, Nature 2016, 529, 185.
2587. [113] Y. I. Joe, X. M. Chen, P. Ghaemi, K. D. Finkelstein, G. A. D. L. Peña,
[84] Y. Zhang, Q. Ji, G. Han, J. Ju, J. Shi, D. Ma, J. Sun, Y. Zhang, M. Li, Y. Gan, J. C. T. Lee, S. Yuan, J. Geck, G. J. Macdougall, T. C. Chiang,
X. Lang, ACS Nano 2014, 8, 8617. S. L. Cooper, E. Fradkin, P. Abbamonte, Nat. Phys. 2014, 10, 421.

Adv. Mater. 2019, 1901694 1901694 (28 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

[114] J. Wang, H. Zheng, G. Xu, L. Sun, D. Hu, Z. Lu, L. Liu, J. Zheng, [141] Y. Y. Lv, L. Cao, X. Li, B. B. Zhang, K. Wang, B. P. B. Pang, L. Ma,
C. Tao, L. Jiao, J. Am. Chem. Soc. 2016, 138, 16216. D. Lin, S. H. Yao, J. Zhou, Y. B. Chen, S. Dong, W. Liu, M. Lu,
[115] P. Goli, J. M. Khan, D. Wickramaratne, R. K. Lake, A. A. Balandin, Y. Chen, Y. Chen, Sci. Rep. 2017, 7, 44587.
Nano Lett. 2012, 12, 5941. [142] I. Belopolski, D. S. Sanchez, Y. Ishida, X. Pan, P. Yu, S.-Y. Xu,
[116] M. Langer, M. Kisiel, R. Pawlak, F. Pellegrini, G. E. Santoro, G. Chang, T.-R. Chang, H. Zheng, N. Alidoust, G. Bian,
R. Buzio, A. Gerbi, G. Balakrishnan, A. Baratoff, E. Tosatti, M. Neupane, S.-M. Huang, C.-C. Lee, Y. Song, H. Bu, G. Wang,
E. Meyer, Nat. Mater. 2014, 13, 173. S. Li, G. Eda, H.-T. Jeng, T. Kondo, H. Lin, Z. Liu, F. Song, S. Shin,
[117] M. M. Ugeda, A. J. Bradley, Y. Zhang, S. Onishi, Y. Chen, R. Wei, M. Z. Hasan, Nat. Commun. 2016, 7, 13643.
C. Ojedaaristizabal, H. Ryu, M. T. Edmonds, H. Z. Tsai, A. Riss, [143] A. Lu, H. Zhu, J. Xiao, C. Chuu, Y. Han, M. Chiu, C. Cheng,
S.-K. Mo, D. Lee, A. Zettl, Z. Hussain, Z.-X. Shen, M. F. Crommie, C. Yang, K. Wei, Y. Yang, Y. Wang, D. Sokaras, D. Nordlund,
Nat. Phys. 2015, 12, 92. P. Yang, D. A. Muller, M. Y. Chou, X. Zhang, L.-J. Li, Nat. Nano-
[118] X. Xi, Z. Wang, W. Zhao, J. H. Park, K. T. Law, H. Berger, L. Forró, technol. 2017, 12, 744.
S. Jie, K. F. Mak, Nat. Phys. 2015, 12, 139. [144] P. Yu, J. Lin, L. Sun, Q. L. Le, X. Yu, G. Gao, C. H. Hsu, D. Wu,
[119] P. Hajiyev, C. Cong, C. Qiu, T. Yu, Sci. Rep. 2013, 3, 2593. T. R. Chang, Q. Zeng, F. Liu, Q. Wang, H.-T. Jeng, H. Lin,
[120] L. Lu, T. Sun, W. Liu, H. Cao, S. Yang, Phy. Rev. B 2015, 92, A. Trampert, Z. Shen, K. Suenaga, Z. Liu, Adv. Mater. 2017, 29,
224303. 1603991.
[121] R. Samnakay, D. Wickramaratne, T. R. Pope, R. K. Lake, [145] Q. Fu, L. Yang, W. Wang, A. Han, J. Huang, P. Du, Z. Fan, J. Zhang,
T. T. Salguero, A. A. Balandin, Nano Lett. 2015, 15, 2965. B. Xiang, Adv. Mater. 2015, 27, 4732.
[122] B. Sipos, A. F. Kusmartseva, A. Akrap, H. Berger, L. Forró, E. Tutis, [146] H. Li, X. Duan, X. Wu, X. Zhuang, H. Zhou, Q. Zhang, X. Zhu,
Nat. Mater. 2008, 7, 960. W. Hu, P. Ren, P. Guo, L. Ma, X. Fan, X. Wang, J. Xu, A. Pan,
[123] G. Liu, B. Debnath, T. R. Pope, T. T. Salguero, R. K. Lake, X. Duan, J. Am. Chem. Soc. 2015, 137, 5284.
A. A. Balandin, Nat. Nanotechnol. 2016, 11, 845. [147] Q. Feng, Y. Zhu, J. Hong, M. Zhang, W. Duan, N. Mao, J. Wu,
[124] Y. Yu, F. Yang, X. F. Lu, Y. J. Yan, Y. Cho, L. Ma, X. H. Niu, S. Kim, H. Xu, F. Dong, F. Lin, C. Jin, C. Wang, J. Zhang, L. Xie, Adv.
Y. Son, D. L. Feng, Nat. Nanotechnol. 2015, 10, 270. Mater. 2014, 26, 2763.
[125] E. Maximilian, S. F. Hanjo, K. Marina, B. Markus, D. Jure, [148] F. Cui, Q. Feng, J. Hong, R. Wang, Y. Bai, X. Li, D. Liu, Y. Zhou,
B. Helmuth, M. Gustavo, S. Germán, R. J. D. Miller, Nature 2010, X. Liang, X. He, Z. Zhang, S. Liu, Z. Lei, Z. Liu, T. Zhai, H. Xu, Adv.
468, 799. Mater. 2017, 29, 1705015.
[126] K. Xu, P. Chen, X. Li, C. Wu, Y. Guo, J. Zhao, X. Wu, Y. Xie, Angew. [149] A. Dathbun, Y. Kim, S. Kim, Y. Yoo, M. S. Kang, C. Lee, J. H. Cho,
Chem. 2013, 52, 10477. Nano Lett. 2017, 17, 2999.
[127] H. L. Zhuang, R. G. Hennig, Phys. Rev. B 2016, 93, 054429. [150] F. Wang, J. Li, W. Feng, T. A. Shifa, Z. Cheng, Z. Wang, X. Kai,
[128] Y. Zhou, Z. Wang, P. Yang, X. Zu, L. Yang, X. Sun, F. Gao, ACS X. Zhan, Q. Wang, J. He, Adv. Funct. Mater. 2015, 25, 6077.
Nano 2012, 6, 9727. [151] X. Kai, W. Fengmei, W. Zhenxing, Z. Xueying, W. Qisheng,
[129] X. Zhu, Y. Guo, H. Cheng, J. Dai, X. An, J. Zhao, K. Tian, S. Wei, C. Zhongzhou, S. Muhammad, J. He, ACS Nano 2014, 8,
X. C. Zeng, C. Wu, Nat. Commun. 2016, 7, 11210. 8468.
[130] F. Weber, S. Rosenkranz, J. P. Castellan, R. Osborn, R. Hott, [152] Y. Zhang, K. Liu, F. Wang, T. A. Shifa, Y. Wen, F. Wang, K. Xu,
R. Heid, K. P. Bohnen, T. Egami, A. H. Said, D. Reznik, Phys. Rev. Z. Wang, C. Jiang, J. He, Nanoscale 2017, 9, 5641.
Lett. 2011, 107, 107403. [153] T. A. Shifa, F. Wang, K. Liu, K. Xu, Z. Wang, X. Zhan, C. Jiang,
[131] Q. Ji, C. Li, J. Wang, J. Niu, Y. Gong, Z. Zhang, Q. Fang, Y. Zhang, J. He, Small 2016, 12, 3802.
J. Shi, L. Liao, X. Wu, L. Gu, Z. Liu, Y. Zhang, Nano Lett. 2017, 17, [154] A. K. Geim, I. V. Grigorieva, Nature 2013, 499, 419.
4908. [155] M. Chiu, X. M. Li, X. W. Zhang, W. Hsu, W. Chang, M. Terrones,
[132] Z. Zhang, J. Niu, P. Yang, Y. Gong, Q. Ji, J. Shi, Q. Fang, S. Jiang, H. Terrones, L. Li, ACS Nano 2014, 8, 9649.
H. Li, X. Zhou, Adv. Mater. 2017, 29, 1702359. [156] T. Roy, M. Tosun, X. Cao, H. Fang, D. Lien, P. Zhao, Y. Chen,
[133] J. Shi, X. Wang, S. Zhang, L. Xiao, Y. Huan, Y. Gong, Z. Zhang, Y. Chueh, J. Guo, A. Javey, ACS Nano 2015, 9, 2071.
Y. Li, X. Zhou, M. Hong, Q. Fang, Q. Zhang, X. Liu, L. Gu, Z. Liu, [157] H. Kroemer, Proc. IEEE 1982, 70, 13.
Y. Zhang, Nat. Commun. 2017, 8, 958. [158] Y. Ohno, D. K. Young, B. Beschoten, F. Matsukura, H. Ohno,
[134] H. Wang, X. Huang, J. Lin, J. Cui, Y. Chen, C. Zhu, F. Liu, Q. Zeng, D. D. Awschalom, Nature 1999, 402, 790.
J. Zhou, P. Yu, X. Wang, H. He, S. H. Tsang, W. Gao, K. Suenaga, [159] X. Hong, J. Kim, S.-F. Shi, Y. Zhang, C. Jin, Y. Sun, S. Tongay, J. Wu,
F. Ma, C. Yang, L. Lu, T. Yu, E. H. T. Teo, G. Liu, Z. Liu, Nat. Y. Zhang, F. Wang, Nat. Nanotechnol. 2014, 9, 682.
Commun. 2017, 8, 394. [160] R. Cheng, F. Wang, L. Yin, Z. Wang, Y. Wen, T. A. Shifa, J. He, Nat.
[135] Y. Zhang, L. Yin, J. Chu, T. A. Shifa, J. Xia, F. Wang, Electron. 2018, 1, 356.
Y. Wen, X. Zhan, Z. Wang, J. He, Adv. Mater. 2018, 30, [161] P. Rivera, J. R. Schaibley, A. M. Jones, J. Ross, S. X. Wu,
1803665. G. Aivazian, P. Klement, K. L. Seyler, G. Clark, N. Ghimire, J. Yan,
[136] Z. Zhang, J. Niu, P. Yang, Y. Gong, Q. Ji, J. Shi, Q. Fang, S. Jiang, D. Mandrus, W. Yao, X. Xu, Nat. Commun. 2015, 6, 6242.
H. Li, X. Zhou, L. Gu, X. Wu, Y. Zhang, Adv. Mater. 2017, 29, [162] A. F. Rigosi, H. M. Hill, Y. Li, A. Chernikov, T. F. Heinz, Nano Lett.
1702359. 2015, 15, 5033.
[137] F. Wei, C. Yu, J. Lin, X. Wang, Q. Zeng, J. Zhou, Z. Lu, W. Hong, [163] Y. Yu, S. Hu, L. Su, L. Huang, Y. Liu, Z. Jin, A. A. Purezky,
Y. He, H. He, Q. Fu, K. Suenaga, T. Yu, Z. Liu, Chem. Mater. 2016, D. B. Geohegan, K. W. Kim, Y. Zhang, L. Cao, Nano Lett. 2015, 15,
28, 7613. 486.
[138] C. K. Shih, J. Kim, Science 2009, 324, 1314. [164] Y. Gong, J. Lin, X. Wang, G. Shi, S. Lei, Z. Lin, X. Zou, G. Ye,
[139] S. Yu, K. Yuichi, Y. Jianting, I. Yoshihiro, N. Tsutomu, Science 2015, R. Vajtai, B. I. Yakobson, H. Terrones, M. Terrones, B. K. Tay,
350, 409. J. Lou, S. T. Pantelides, Z. Liu, W. Zhou, P. M. Ajayan, Nat. Mater.
[140] A. Apte, V. Kochat, P. Rajak, A. Krishnamoorthy, P. Manimunda, 2014, 13, 1135.
J. A. Hachtel, J. C. Idrobo, S. A. Syed Amanulla, P. Vashishta, [165] Y. Yu, H. Shi, L. Su, L. Huang, L. Yi, Z. Jin, A. A. Purezky,
A. Nakano, R. K. Kalia, C. S. Tiwary, P. M. Ajayan, ACS Nano 2018, D. B. Geohegan, K. W. Kim, Z. Yong, L. Cao, Nano lett. 2015, 15,
12, 3468. 486.

Adv. Mater. 2019, 1901694 1901694 (29 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.advmat.de

[166] C. Huang, S. X. Wu, A. M. Sanchez, J. J. P. Peters, R. Beanland, [178] W. Zhang, M.-H. Chiu, C.-H. Chen, W. Chen, L.-J. Li, A. T. S. Wee,
J. Ross, P. Rivera, W. Yao, D. Cobden, X. Xu, Nat. Mater. 2014, 13, ACS Nano 2014, 8, 8653.
1096. [179] T. Zhang, B. Jiang, Z. Xu, R. G. Mendes, Y. Xiao, L. Chen, L. Fang,
[167] X. Duan, C. Wang, J. C. Shaw, R. Cheng, Y. Chen, H. Li, X. Wu, T. Gemming, S. Chen, M. H. Rümmeli, L. Fu, Nat. Commun. 2016,
Y. Tang, Q. Zhang, A. Pan, J. Jiang, R. Yu, Y. Huang, X. Duan, Nat. 7, 13911.
Nanotechnol. 2014, 9, 1024. [180] X. Zhang, C. Lin, Y. Tseng, K. Huang, Y. Lee, Nano Lett. 2015, 15,
[168] X. Q. Zhang, C. H. Lin, Y. W. Tseng, K. H. Huang, Y. H. Lee, Nano 410.
Lett. 2014, 15, 410. [181] M. Y. Li, Y. Shi, C. C. Cheng, L. S. Lu, Y. C. Lin, H. Tang, M. L. Tsai,
[169] Q. Zhang, X. Xiao, R. Zhao, D. Lv, G. Xu, Z. Lu, L. Sun, S. Lin, C. Chu, K. Wei, J.-H. He, W.-H. Chang, K. Suenaga, L.-J. Li, Science
X. Gao, J. Zhou, C. Jin, F. Ding, L. Jiao, Angew. Chem., Int. Ed. 2015, 349, 524.
2015, 127, 9085. [182] Z. Zhang, P. Chen, X. Duan, K. Zang, J. Luo, X. Duan, Science
[170] L. Samad, S. M. Bladow, Q. Ding, J. Zhuo, R. M. Jacobberger, 2017, 357, 788.
M. S. Arnold, S. Jin, ACS Nano 2016, 10, 7039. [183] P. Sahoo, S. Memaran, Y. Xin, L. Balicas, H. R. Gutierrez, Nature
[171] Z. Xing, Z. Nan, L. Chao, H. Song, T. Zhai, 2D Mater. 2017, 4, 2018, 553, 63.
025048. [184] S. Xie, L. Tu, Y. Han, L. Huang, K. Kang, K. U. Lao, P. Poddar, C. Park,
[172] J. H. Yu, H. R. Lee, S. S. Hong, D. Kong, H. Lee, H. Wang, D. A. Muller, R. A. Distasio, J. Park, Science 2018, 359, 1131.
F. Xiong, S. Wang, Y. Cui, Nano Lett. 2015, 15, 1031. [185] Z. Yun, H. Lu, X. Zhu, S. Tan, F. Wei, L. Qin, Z. Wen, Q. Chen,
[173] Y. Gong, S. Lei, G. Ye, B. Li, Y. He, K. Keyshar, X. Zhang, Q. Wang, L. Yi, X. Luo, Sci. Adv. 2018, 4, eaao6791.
J. Lou, Z. Liu, R. Vajtai, W. Zhou, P. M. Ajayan, Nano Lett. 2015, [186] A. Stern, C. Gong, C. Wang, H. Ji, J. Xia, L. Li, R. Cava,
15, 6135. S. G. Louie, T. Cao, W. Bao, C. Wang, Y. Wang, Z. Q. Qiu, R. J. Cava,
[174] X. Li, M. W. Lin, J. Lin, B. Huang, A. A. Puretzky, C. Ma, K. Wang, S. G. Louie, J. Xia, X. Zhang, Nature 2017, 546, 265.
W. Zhou, S. T. Pantelides, M. Chi, I. I. Kravchenko, J. D. Fowlkes, [187] B. Huang, G. Clark, D. R. Klein, D. Macneill, E. Navarro-Moratalla,
C. M. Rouleau, D. B. Geohegan, K. Xiao, Sci. Adv. 2016, 2, K. L. Seyler, N. Wilson, M. A. Mcguire, D. H. Cobden, D. Xiao,
e1501882. W. Yao, P. J. Herrero, X. Xu, Nat. Nanotechnol. 2018, 13, 544.
[175] T. Yang, B. Zheng, Z. Wang, T. Xu, C. Pan, J. Zou, X. Zhang, Z. Qi, [188] J. Niu, Q. Ji, B. Yan, M. Li, G. Peng, Z. Liu, Y. Zhang, D. Yu, X. Wu,
H. Liu, Y. Feng, W. Hu, F. Miao, L. Sun, X. Duan, A. Pan, Nat. Phys. Rev. B 2017, 96, 075402.
Commun. 2017, 8, 1906. [189] N. Mounet, M. Gibertini, P. Schwaller, D. Campi, A. Merkys,
[176] B. Li, L. Huang, M. Zhong, Y. Li, Y. Wang, J. Li, Z. Wei, Adv. Elec- A. Marrazzo, T. Sohier, I. E. Castelli, A. Cepellotti, G. Pizzi,
tron. Mater. 2016, 2, 1600298. N. Marzari, Nat. Nanotechnol. 2018, 13, 246.
[177] C. Jianyi, L. Bo, L. Yanpeng, T. Wei, N. C. Tai, L. Linjun, Z. Jian, [190] D. Zhong, K. L. Seyler, X. Linpeng, R. Cheng, N. Sivadas, B. Huang,
G. Libo, Z. Yi, S. H. S. Shin, H. Y. Jeong, K. P. Loh, Adv. Mater. E. Schmidgall, T. Taniguchi, K. Watanabe, M. A. Mcguire, W. Yao,
2016, 27, 6722. D. Xiao, K.-M. C. Fu, X. Xu, Sci. Adv. 2017, 3, e1603113.

Adv. Mater. 2019, 1901694 1901694 (30 of 30) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim