Review

www.acsnano.org

Contacts between Two- and Three-

Dimensional Materials: Ohmic, Schottky, and
p−n Heterojunctions
Yang Xu,*,†,‡,⊥ Cheng Cheng,†,⊥ Sichao Du,† Jianyi Yang,† Bin Yu,† Jack Luo,† Wenyan Yin,† Erping Li,†
Shurong Dong,† Peide Ye,§ and Xiangfeng Duan‡
†
College of Information Science and Electronic Engineering, Zhejiang University, Hangzhou, Zhejiang 310027, China
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

‡
Department of Chemistry and Biochemistry, University of California, Los Angeles, California 90095, United States
§
School of Electrical and Computer Engineering, Purdue University, West Lafayette, Indiana 47906, United States
Downloaded via SLUB DRESDEN on June 21, 2021 at 09:35:34 (UTC).

ABSTRACT: After a decade of intensive research on two-dimensional

(2D) materials inspired by the discovery of graphene, the field of 2D
electronics has reached a stage with booming materials and device
architectures. However, the efficient integration of 2D functional layers
with three-dimensional (3D) systems remains a significant challenge,
limiting device performance and circuit design. In this review, we
investigate the experimental efforts in interfacing 2D layers with 3D
materials and analyze the properties of the heterojunctions formed
between them. The contact resistivity of metal on graphene and related 2D
materials deserves special attention, while the Schottky junctions formed
between metal/2D semiconductor or graphene/3D semiconductor call for
careful reconsideration of the physical models describing the junction
behavior. The combination of 2D and 3D semiconductors presents a form
of p−n junctions that have just marked their debut. For each type of the
heterojunctions, the potential applications are reviewed briefly.
KEYWORDS: 2D materials, 3D materials, graphene, transition metal dichalcogenides, metals, transfer length method, Ohmic contact,
Schottky, heterojunctions, sensors, detectors

W ith the advent of various two-dimensional (2D)

materials ranging from insulators, semiconductors,
to metals, enormous efforts have been put into the
investigation of their fundamental physical/chemical properties.
combination of electronic performance and mechanical
flexibility.17−20
While the “pure 2D” concept is revolutionary, interfacing 2D
materials with the three-dimensional (3D) world21 is never a
subject that could be circumvented but somehow has not
They are expected to bring about revolutionary opportunities in
nanoelectronics as the traditional silicon-based technology is received enough attention. The most prominent issue might be
reaching its physical limit.1,2 Though relatively young, the field the contact resistance.22 Most of the investigations on intrinsic
properties of 2D materials are carried out in the configuration
of 2D electronics has seen rapid development in a number of
of field-effect transistors (FETs). Without the careful selection
directions with high-impacting review articles already avail-
of the contact metal along with optimized integration protocols,
able,3−6 especially on graphene7−10 and 2D transition metal the channel properties of 2D materials could be masked easily
dichalcogenides (TMDs).11,12 In parallel with the studies of by the contact resistance.23,24 This artifact would be aggravated
intrinsic properties of monolayer 2D materials, the van der in the short-channel limit that is often necessary for high-
Waals (vdW) heterostructures have attracted special attention performance devices,25 which poses a critical challenge in
in the past 2−3 years.13−16 The endless possibilities of stacking device scaling. On the other hand, a considerable amount of
different 2D materials to create atomic scale heterostructures work has been put into the Schottky junction devices formed
and superlattices can open up exciting opportunities for between graphene (as transparent electrode) and conventional
creating designed device architectures with tunable electronic semiconductors or metal on 2D semiconductors, as will be
properties beyond the reach of single material. In addition, with
excellent electronic properties and exceptional mechanical Received: March 16, 2016
robustness, 2D materials have also attracted considerable Accepted: April 30, 2016
attention for a generation of highly flexible electronics with a Published: April 30, 2016

© 2016 American Chemical Society 4895 DOI: 10.1021/acsnano.6b01842

ACS Nano 2016, 10, 4895−4919
ACS Nano Review

discussed in the following sections. Despite the fact that these intrinsic transport properties of the channel material as well as
devices are different forms of Schottky junctions with charge improving the device performance. The electrical contact issue
carriers being transported from 2D (3D) to 3D (2D) material, calls for close attention especially when ideal Ohmic contact is
the conventional 3D Schottky junction transport model has difficult to achieve, as is the case for emerging 2D electronic
been commonly adapted to characterize the junction properties materials. Here, we start with the very basics of contact
and functions as a fundamental guideline for device resistance characterization in conventional semiconductor
optimization. devices and then review recent efforts in adapting these
In this review, we would like to point out that, in many cases, fundamental techniques into the nascent yet more complicated
it might no longer be appropriate to analyze the 2D/3D 2D cases.
interface based merely on conventional 3D models, as Though graphene has been intensively investigated ever since
commonly practiced (note that 2D−2D contact is out of the its discovery in 2004,26 it was not until around 2010 that the
scope of this review). There are four categories of 2D/3D issue of contact resistance to graphene had received proper
contacts depending on the material type, upon which we divide attention.27 As a critical limiting factor of the on-current in
our discussions into the following parts: (1) the metal contact graphene FETs, the contact resistance was characterized using
on graphene, (2) the metal contact on 2D semiconductors the TLM commonly implemented in conventional semi-
(TMDs, black phosphorus (BP), silicene, etc.), (3) graphene on conductor devices. Figure 2A illustrates the TLM configuration.
conventional semiconductors, and (4) 2D semiconductors on
conventional semiconductors (Figure 1). In the first section, we

Figure 1. Outline of the review. The review concerns specifically

the heterojunctions formed between a 2D and a 3D material.
According to the material type (metal or semiconductor, either in
2D or 3D), the contacts are Ohmic, Schottky, or p−n junctions. Figure 2. TLM for extracting the contact resistance RC and the
The contact resistance problem is discussed thoroughly in the transfer length LT in conventional semiconductor devices: (A) top
graphene/metal case, as well as in the 2D semiconductor/metal view of the measurement configuration; (B) side view to show the
contacts. resistive network for calculating LT, and (C) Rtotal(L) plot for
extracting RC and LT.
focus on the contact resistance (RC) problem on graphene.
After a brief review of the conventional 3D transfer length
method (TLM), we introduce the modified face-contact model
proposed for the 3D metal/2D graphene system and move on The total resistance (Rtotal) of the contact(C)−channel(Ch)−
to the edge-contact model studied more recently. Following the contact(C) structure is measured with two probes, and Rtotal =
work on contacting graphene, we discuss the efforts made to RChS(L/W) + 2RC, where RChS is the sheet resistance, L is the
contact 2D semiconductors in the second section. The third length, and W is the width of the channel. By varying L, RChS
section covers the graphene/conventional semiconductor can be extracted from the slope of Rtotal as a function of L, as
junctions. Until now, various semiconductors have been put shown in Figure 2C. Meanwhile, the contact resistance RC is
in contact with graphene to form the Schottky diode, and characterized by the intercept of the fitted line on the y-axis.
nearly all of the studies rely on a classical 3D Schottky junction The specific contact resistivity ρC is defined conventionally as
transport model to analyze their results. We introduce the most RCAC in the unit of Ω·cm2, where AC is the contact area. Figure
recent works concerning the modification of the Schottky diode 2B illustrates the resistive network model used for extracting
model, taking into account the 2D nature of graphene and its the transfer length LT, another critical parameter in contacts.
tunable Fermi level under bias. In the last section, we The current and the voltage along the x-direction can be solved
summarize the limited number of articles available studying using the equations
the p−n junction formed between a 2D semiconductor and a V (x )
3D semiconductor. This area of research is just emerging, but d I (x ) = −
ρC /(W dx) (1)
we expect to see a fast development here.

METAL ON GRAPHENE: OHMIC CONTACT S ⎛⎜

dx ⎞⎟
dV (x) = −I(x)R Ch
⎝W ⎠ (2)
Contact Resistance. For the development of any electronic
devices, making reliable electrical contact to the active layer is and the results are
the first step. The precise characterization of the Ohmic contact
or control over the Schottky contact is essential for probing the V (x) = V0e−x / L T (3)

4896 DOI: 10.1021/acsnano.6b01842

ACS Nano 2016, 10, 4895−4919
ACS Nano Review

Table 1. Contact Resistivity ρC on Graphene Measured by TLMa

metal (thickness) ρC (Ω·cm2) refs (year) graphene type
Ni (25, 60 nm) ∼10−5 27,28 (2010) mechanically exfoliated
Cr, Ti/Au (10/20 nm) 10−4−10−1 28 (2010) mechanically exfoliated
Ti/Au (10/50 nm) >10−5 29 (2011) epitaxial on SiC
Ti, Ni, Cu, Pt, Pd/Au + “ash” <10−7 29 (2011) epitaxial on SiC
Ti/Pt/Au ∼10−8−10−7 30 (2012) epitaxial on SiC
a
The Si industry requires ∼10−8 Ω·cm2, and the short-channel graphene devices would require ∼10−10 Ω·cm2.

1 W
I (x ) = S
V0e−x / L T
R Ch LT (4)

where LT = (ρC/RSCh)1/2. Because of the resistance at the metal/

channel interface and the channel sheet resistance, the current
flows preferably close to the contact edge and drops to 1/e of
the edge value at the length of LT, a phenomenon known as
“current crowding”. The effective contact area is therefore
defined as LTW. Using LT, the total resistance of the system can
be expressed as Rtotal = (RSCh/W)(L + 2LT). The intercept of the
fitted Rtotal(L) line with the x-axis gives LT directly.
The specific contact resistivity of graphene measured using
this TLM approach is summarized in Table 1. For contact
lengths larger than LT, current crowding is so significant that
the contact resistivity is described more effectively as RCW in
Ω·μm. The ρC is also affected by the metal type because the Figure 3. Transport at the metal/graphene contact proposed by Xia
Fermi level shift in graphene on contacting metals has different et al.32 (A) Tansmission probabilities: TMG, carrier injection from
work functions.31 Various efforts have been made to improve metal to graphene and TK, from graphene under the contact to the
contact to graphene devices, and low ρC values are reported. channel. (B,C) Band diagram at the metal−graphene interface and
Although the TLM is used regularly for characterizing the the graphene p−n junction, respectively.
contact resistance RC and the transfer length LT at the face
contact between graphene and metal, whether this 3D resistive where λ≪ λm, TMG = λ /λm ; in the ballistic limit (λ ≫ λm),
network model is valid for a 2D−3D contact is questionable.
Only through proper understanding of the electron transport TMG ∼ 1. The effective transfer length LT in this case is λλm .
mechanism and current flow at the contact can one improve ρC It is worth noting here that λ in graphene is significantly longer
effectively, which has recently been systematically addressed by than that in conventional semiconductors, leading to a majorly
Xia et al.32 In conventional TLM, there are two basic ballistic transport at the metal/graphene interface. Therefore,
assumptions: (1) the metal/semiconductor contact is diffusive the contact resistance shows a negative temperature depend-
due to a very short electron mean free path in doped ence, in contrary to conventional semiconductors. Based on
semiconductors, and (2) the sheet resistance of the semi- this model, to increase the contact conductance GC ∝ TM,
conductor is identical in the transport channel or under the either M could be increased by heavy doping of the graphene
contact since the current flows in the bulk. However, neither of under the metal or T could be enhanced by increasing the mean
these assumptions holds true for graphene. As a result, the free path λ in graphene and reducing the effective coupling
resistive network model is no longer appropriate for extracting length λm with the contact metal.
the transfer length LT, which characterizes the current crowding Current Flow between Contacts. In the contact
and is critical in contact design. Meanwhile, it is necessary to resistance models discussed above, the metal electrodes act
differentiate the sheet resistance in the channel and under the on the graphene layer in their immediate adjacency. On the
contact, as graphene is doped by the contact metal. other hand, as a layered anisotropic material (in-plane vs out-of-
Figure 3 illustrates the model proposed for replacing the plane), graphene has unique distribution of current in
resistive network model shown in Figure 2B. It focuses on the multilayer channels compared with 3D isotropic materials,
transmission probability T at the two contact junctions: (1) the which in return influences the contact and gating behavior in
metal−graphene junction, TMG, and (2) the p−n junction for devices. Sue and Appenzeller established the model of
graphene under the contact and in the channel, TK. According interlayer charge coupling and screening in multilayer
to Landauer’s approach, the conductance G ∝ TM, where M is graphene, which also provides insights to current flow in
the number of conduction modes in graphene, is determined by other 2D materials.33 They presented the systematic exper-
imental study on current distribution in multilayer graphene
the density of states (DOS) and its broadening due to
FETs by varying the channel layer number. As shown in Figure
inhomogeneities as well as coupling to the metal d states. When
4A, the on-state conductance of the channel Gon shows a
the electron is injected from graphene to metal, the coupling of
decreasing trend with increasing layer number (channel
graphene to metal induces the only scattering process in the
thickness d), contrary to intuitions, while the off-state
ballistic limit, characterized by the effective coupling length λm.
conductance Goff increases; the Ion/Ioff shows a 1/d dependence.
The transmission probability is TMG = λ /(λm + λ) , where λ The resistive network model illustrated in Figure 4B gives a
is the electron mean free path in graphene. In the diffusive limit satisfactory interpretation of the experimental data, composed
4897 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

1 ⎛ t + ri ⎞
Qi ∝ exp⎜ − ox ⎟
tox + ri ⎝ λ ⎠ (5)
N
∑ Q i = Q gate
i=1 (6)
where Qi is the charge in the ith layer, tox is the dielectric
thickness, and ri is the distance from the dielectric/channel
interface to the ith layer. Assuming the conductance in each
layer is proportional to the charge induced in it, we have
ci−1
Gon, i = N
Ggate + Gmin
∑j=1 c j−1 (7)
where
Qi ⎛ r − ri − 1 ⎞
c= ≈ exp⎜ − i ⎟
Q i−1 ⎝ λ ⎠ (8)
and Ggate corresponds to the conductance under total gate
charge.
In this model, the two critical parameters are the ratio Rint/
Roff, characterizing the interlayer coupling, and the screening
length λ. The fitting results shown in Figure 4A uses Rint =
0.05Roff and λ = 0.6 nm, suggesting a short depth of the current
flow from the top contact into the bulk of the multilayer
Figure 4. Resistive network model for current distribution in channel and rather limited gate control across the channel
multilayer graphene FET. (A) Experimental data of the current on/ thickness since λ extends only two graphene layers.
off ratio (inset), the on- and off-channel conductance, and the fitted Improving the Contact: Surface or Edge? Based on the
line using the model shown in (B); each layer has its resistance Ri, models discussed above, it is clear that the interlayer tunneling
and the interlayer resistance is Rint.33 of electrons limits the contact conductance in a metal/graphene
system significantly, due to the lack of bonding sites at the
surface of graphene. An alternative strategy is then to utilize the
of intralayer resistance R1 to RN for N layers of graphene and edge contact. Wang et al. reported the device topology in which
interlayer resistance R int . Each layer has a minimum the encapsulated 2D graphene channel makes pure one-
conductance Gmin regardless of the gating condition, and the dimensional (1D) edge contact to the 3D metal electrodes
ratio between Rint and the off-state single-layer resistance Roff is (Figure 5).34 The device shows 25% less specific contact
estimated to be Rint/Roff ∼ 0.02−0.2. When the FET is gated, resistance (Cr/Pd/Au 1/15/60 nm, ∼150 Ω·μm) compared
the total charge on the gate is Qgate, and the vertical (z) with that reported in ref 32 and other intrinsic face contacts and
Coulomb potential in the channel φ(z) ∝ (Q/z)exp(−z/λ), demonstrates almost ideal transport over length scales longer
where Q is the charge under Thomas−Fermi screening and λ is than 15 μm at low temperatures. The much improved contact is
the screening length. Therefore, we have attributed to the shorter bonding distance at the metal/

Figure 5. Edge contact to graphene. (A) Graphene monolayer is encapsulated between boron nitride to expose the edge, followed by regular
deposition of Cr/Pd/Au (1/15/60 nm) to form the edge contact. (B) Scanning transmission electron microscopy image (top panel) and
electron energy loss spectroscopy map (bottom panel) confirming the edge-contact formation.34

4898 DOI: 10.1021/acsnano.6b01842

ACS Nano 2016, 10, 4895−4919
ACS Nano Review

Table 2. Contact Resistivity on Graphene with Edge Contact

metal (thickness) ρC (Ω·μm) ref (year) graphene type
Cr/Pd/Au (1/15/60 nm) 150, *25% reduction of ref 20 34 (2013) mechanically exfoliated
Cu ∼150, *32% reduction 35 (2013) epitaxial on SiC
Pd ∼457, *22% reduction 35 (2013) epitaxial on SiC
Ni 89−11 36 (2014) mechanically exfoliated
Pd/Au (20/30 nm) 2.2 × 10−9 Ω·cm2 37 (2014) CVD
Ti 7.7 × 10−8 Ω·cm2 38 (2015) CVD, multilayer

Figure 6. Energy band diagrams for sensing NO2 gas with (A) graphene, (B) SnO2 surface, and (C) graphene−SnO2 interface. Schematic
mechanisms of (D) SnO2 NPs, (E) less graphene interacting with SnO2 NPs, and (F) more graphene interacting with SnO2 NPs.55

graphene edge, leading to larger electron orbital overlap, which Following this demonstration of effective 1D contact and
is also consistent with Xia’s strategy of reducing λm, the effective previous analysis on “current crowding”, much work has been
graphene−metal coupling length.32 done to improve the contact to graphene based on the increase
4899 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

of the edge perimeter in the contact area, often followed by electrical conductivity and surface functions. Ohmic character-
postmetal deposition annealing to aid the covalent bonding of istics have generally been observed.50
carbon and metal atoms at the graphene crystal edges. The Chung et al. fabricated a flexible H2 sensor using graphene
results are shown in Table 2. with Pd NPs.53 The optimized 3 nm thick Pd NP film shows a
Applications. Photodetection. As mentioned in the 33% gas response exposed to 1000 ppm of H2, and the
Contact Resistance section, the contact metal dopes the detection limit is 20 ppm of H2 at room temperature. They
graphene and therefore shifts its Fermi level locally, due to suggested that the sensor response time could be improved by
the difference in the metal/graphene work function as well as optimizing sample preparation, for instance, the control over
the limited DOS near the graphene Dirac point. In a graphene the size/density of Pd NPs with patterned graphene in
FET, the Fermi level EF in the graphene under the metal nanometer width. Park et al. used a periodic 200 nm wide
contact is determined by the charge transfer from the metal and graphene nanoribbon array with 1 μm pitch to protect the
not affected by the gate. The Fermi level in the channel aligns graphene from photoresist contamination on its surface by
with this EF; however, its doping level is controlled by the gate. implementing a 10 nm thick Cr interlayer.54 The fabricated Pd
An energy difference in the Dirac point is thus established NP/graphene nanoribbon H2 gas-sensing device exhibits 90%
between the graphene under the contact and in the channel, response within 60 s at 1000 ppm and 80% recovery within 90 s
leading to band bending and therefore a built-in electric field at in ambient N2. Detailed mechanisms were proposed by
the contact/channel interface. Tammanoon et al. on the basis of the formation of Ohmic
Utilizing this built-in field to separate the photogenerated M−S contacts and accessible interfaces of graphene−SnO2
carriers, Xia et al. demonstrated the pioneering work elucidating NPs.55 The band diagram and physical models are as depicted
the influence of metal contacts on photocurrent generation in in Figure 6. The synergistic metal NPs in contact with graphene
backgated graphene FETs.39 Ghosh et al.’s work on reduced modulates the overall electronic properties, leading to a
graphene oxide for position-dependent photodetection also significant enhancement of selectivity and sensitivity in gas-
supports this physical picture.40 This device concept further led sensing characteristics. In addition to various metal NPs,
to the high-frequency study by the same group.41 Urich et al. graphene with different formats, for instance, pristine graphene,
investigated the ultimate high-frequency response of the graphene oxide, and reduce graphene oxide (RGO) has also
graphene photodetectors in the same layout, extracting the been investigated.49,50,52 Chu et al. obtained high H2 sensitivity
intrinsic response time via a nonlinear photocurrent with a resistance decrease of more than 80% in response to 1%
autocorrelation signal and gave a bandwidth of ∼262 GHz, H2 at room temperature.56 The sensor showed good selectivity
corresponding to a photodetection response time of 2.1 ps, toward H2 over CH4, NH3, O2, and N2O. Vedala et al. found
which is mainly limited by the lifetime of the photogenerated that the presence of holes in the basal plane of the RGO
carriers.42 To investigate directly the temporal response at the created edges with O2 functionalities that significantly enhanced
metal/graphene interface, Prechtel et al. introduced the pump− the response of Pt NP/holey RGO to H2.49
probe technique in a suspended chemical vapor deposition An increase in device conductance for increasing H2
(CVD) graphene device and measured a photocurrent with a concentrations was attributed to the interfacial interactions of
full width at half-maximum (fwhm) of ∼4 ps originating from H2 with quinone groups located at the edges of the Pt NP/
the built-in electric field.43 Echtermeyer and co-workers further holey RGO system. In Lu et al.’s work on the detection of
elucidated the photoelectric and photothermoelectric contri- NO2/NH3 using RGO FETs, the authors pointed out that in
butions to the photocurrent based on angular-dependent addition to the charge transfer between the RGO and the
measurements using linearly polarized light, observing more adsorbed gas molecules, the variation in the Schottky barrier
prominent photoelectric effects at longer wavelengths,44 which between the RGO and the contact metal, due to the gas
is also consistent with the report by Tielrooij et al.45 The absorption, could also contribute to the overall gas-sensing
limiting factors of the photodetection responsivity (typically response.57
less than 1 mA/W) in the photocurrents generated at the two
electrodes are of opposite sign, and the total current is nearly METAL CONTACT TO 2D SEMICONDUCTORS
diminished. Mueller et al. circumvented this issue by using Based on the studies done on graphene, the metal contact to
asymmetric metal contacts at the source and the drain.46 other 2D materials, especially 2D semiconductors such as
Gas Sensing. A number of graphene-based devices for TMDs (MoS2, MoSe2, WS2, etc.) and BP, has seen a surge of
detecting various gas species have been demonstrated,47−49 as research activities in recent years. Unlike graphene, these 2D
graphene possesses high surface-to-volume ratio in conjunction semiconductors have tunable band gaps covering both the
with high conductivity, leading to the sensitive molecule narrow (∼10−1 eV, BP) and wide (1−2 eV, TMD) regimes,
detection on its surface.50 The electrons are accepted or providing great potential for various device applications. Metal
donated from gas molecules to increase or decrease the channel forms Schottky contact to 2D semiconductors, which will be
conductance of the devices. Even though the graphene-based discussed in this section. MoS2 represents one of the most
devices show high sensitivity, the difficulty of determining the frequently studied 2D semiconductors. Therefore, we will focus
type of the gas species and their concentration prevents the our systematic discussions on MoS 2 along with brief
implementation of graphene as gas sensors. introduction on more recent works of BP/phosphorene.
Therefore, metal or metal oxide nanoparticles (NPs) have As the electron mean free path is much shorter in TMDs, the
been introduced into the device structure recently.51,52 The TLM can be applied with a modified TLM model. Three
interaction between metal NPs and graphene shifts the Fermi factors should be considered when applying the TLM model.
level of graphene in order to achieve equilibrium with metal First, the contact resistance is a combination of the metal−
NPs. These NPs in the form of a thin film contacted with the semiconductor interface resistivity and the semiconductor sheet
graphene are expected to improve the selectivity of the gas resistivity. However, the metal electrodes can easily dope and
sensors because different metals on graphene result in different tune the atomic thin channels of 2D TMD semiconductors
4900 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

Figure 7. Extraction of the SBH at the Ni/MoS2 junction. (A) Source−drain current as a function of the backgate voltage under fixed VDS, at
different temperatures. (B) Effective barrier height φB as a function of the backgate voltage. (C) Arrhenius-type plot of data in (A). (D) True
SBH φSB extracted for different work function metals.23

underneath through the applied biases, thus making the contact kΩ·μm to MΩ·μm. In the backgated FET configuration, Liu et
resistivity complicated. For this reason, the resistivity of al. used the conventional TLM to characterize the contact
semiconducting TMDs under the metal contact (ρcontact) and resistivity ρC and transfer length LT of monolayer CVD MoS2
the resistivity of the semiconducting channel (ρchannel) might be contacted with Ti/Au.58 The transfer length was calculated as
very different. Second, because of current crowding effects, the ρC /R sh , where Rsh is the sheet resistance of monolayer MoS2.
transfer length (LT) and physical metal−semiconductor contact At 0 V backgate, LT ∼ 1.26 μm, and at higher Vbg, LT ∼ 0.63
length (Lc) should be characterized case by case when applying μm. The transistor performance is controlled by two Schottky
effective contact length to the TLM model. Third, the TLM barriers at the source and the drain, respectively, and the device
approach cannot accurately model metal contact to multi-2D is more accurately described as a Schottky barrier transistor, as
layers if the edge contacts dominate the contact resistance. illustrated in Figure 8. The Fermi level pinning between MoS2
Most of metal/TMD 2D materials are Schottky contacts, which
lead to the measured two-point resistance from TLM
depending on the voltage bias and polarization.
Initial Contact with MoS2. Das et al. performed the
thorough experimental study of contacts to mechanically
exfoliated few-layer MoS2,23 comparing the Schottky barrier
height (SBH) and the FET performance using different contact
metals: the low work function of Sc (φM = 3.5 eV, 30 nm Sc/20
nm Ni) and Ti (4.3 eV, 50 nm) and the high work function of
Ni (5.0 eV, 50 nm) and Pt (5.9 eV, 50 nm). Figure 7 illustrates
the details of extracting the SBH φSB at the Ni/MoS2 interface
from the FET characteristics. The temperature-dependent IDS
at each fixed gate voltage VGS gives the Arrhenius plot (Figure Figure 8. Band diagrams for a Ti/Au contacted MoS2 FET with a
7B), from which the effective barrier height φB at certain VGS is backgate, showing the Schottky barriers at the source and the drain
extracted using the high-temperature data. The linear regime of under different Vbg and VDS.58
φB as a function of VGS indicates that the thermionic emission
model applies well to the current (Figure 7C). The true SBH and the electrode is mainly caused by the gap states from the
φSB shows linear dependence on the metal work function φM, intralayer S−Mo bonding and the presence of a metal−MoS2
as shown in Figure 7D, with slope of 0.1, indicating strong alloy with a different work function.59 It is very difficult to
Fermi level pinning at the metal/MoS2 interface. eliminate this effect because it requires an ideal Schottky
To quantify the metal/2D semiconductor contact resistance, contact between metal and a perfect 2D semiconductor crystal.
the resistive network model is very much in parallel to the work Improving the signal crystal quality of the 2D semiconductor is
done for graphene, although the graphene/metal contact is an important but challenging choice to weaken the Fermi level
Ohmic while the 2D semiconductor/metal contact is Schottky pinning. Decoupling the metal from the semiconductor surface
and shows much higher contact resistivity ρC in the range of can also decrease the amount of Fermi level pinning. An
4901 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

Figure 9. (A) Schematic illustration of a graphene/WS2/graphene light-harvesting device. (B) Polarization-resolved scattering spectrum of a
Au nanorod/SiO2/Si. (C) Scattering spectrum of WS2/SiO2/Si. (D−F) Emission polarization-resolved scattering spectra of nanorods with
different lengths on the WS2/SiO2/Si. (G) Calculated near-field intensity map at 612 nm wavelength. (H−J) Calculated emission polarization-
resolved scattering spectra of nanorods with different lengths on the WS2/SiO2/Si.82,83

intervening oxide60,61 or graphene layer62 between the metal scanning tunneling microscopy (STM), and scanning tunneling
and MoS2 can significantly weaken the Fermi level pinning and spectroscopy (STS).65 They found that the natural exfoliated
also reduce the contact resistance and SBH. MoS2 has high intrinsic defect concentration, which imposes
In addition to the modeling and macroscopic measurements questions about the metal’s influence on the MoS2 electronic
of the contact resistance, material characterizations and properties reported because, in those studies, MoS2 is often
localized investigations have also been carried out to probe assumed to be pristine and a perfect crystal.
the microscopic origins of the contact properties observed. In summary, the research on contact to 2D semiconductors
Gong et al. studied the wetting and growth of nanometer thick is still in an early stage with much improvement of ρC required
evaporated Pd, Au, and Ag films on monolayer CVD MoS2 for decent device performance. MoOx interlayer66 or NO2
using atomic force microscope (AFM) to image the metal film surface doping67 was shown to improve the contacts to MoS2
morphology and Raman spectra to monitor the MoS 2 and WSe2, respectively. Reference 68 gives much information
monolayer quality.63 Pd wets MoS2 uniformly, while Au and on the approaches that the research community has taken to
Ag cluster on MoS2. The metal clusters introduce local stress search for better contacts in 2D FETs.
and strain at the metal/MoS2 interface,64 detrimental to the Contacting BP. BP, the bulk form of phosphorene, has
current injection and carrier transport. McDonnell et al. attracted much attention because it has a narrow band gap in
addressed the contact inhomogeneity issue in more detail the range of 0.3−1.0 eV, tunable by the layer number and
using localized measurements, including I−V characteristics, X- bridging the band gap between graphene (zero band gap) and
ray photoelectron spectroscopy (XPS) core-level spectra, TMDs (usually above 1 eV). Furthermore, it has high carrier
4902 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

mobilities and demonstrates ambipolar transport, which opens measurements supported by ab initio DFT calculations illustrate
up possibilities for complementary 2D semiconductor that the d-orbital of the contact metal plays a key role in
electronics. The electrical contacts to BP have been studied forming low contact resistance with monolayer WSe2. The In-
by researchers who have accumulated much knowledge in and Ag-based contacts show the smallest contact resistance and
graphene and MoS2 devices. Using TLM, Du et al. investigated the highest drive current.
Ni/Au (30/50 nm) and Pd/Au (30/70 nm) contacts on 18.7 It is known that light harvesting can be enhanced by the
nm exfoliated BP flakes.69 For the high work function Pd (5.4 presence of metal NPs on 2D materials, inducing surface
eV), ρC = 3.15 × 103 Ω·μm and LT = 0.72 μm; for Ni with φM plasmon resonance.81 As depicted in Figure 9A, a plasmonic Au
= 5.0 eV, ρC = 1.75 × 103 Ω·μm and LT = 1.18 μm at −40 V NP traps the incident light through localized surface plasmon
backgate. Because of the Schottky barrier formed between the resonance and enhances exciton generation in its vicinity.82 By
contact and the BP, the device is, in fact, a Schottky barrier coupling the 2D WS2 to Au nanorods, Kern et al. significantly
FET, similar to the case of MoS2. Das et al.’s work70 provides increased the light−matter interaction due to the plasmon
information on BP FET performance complementary to that of resonance of the Au nanorod, which enhances the optical near-
Du et al. using another high work function contact metal, Ti, fields within the WS2 monolayer, as illustrated in Figure 9B−
with ρCTi = 4.85 × 103 Ω·μm. They also found that Fermi level J.83 The photoluminescence intensity was increased by more
pinning occurs at the metal/BP interface. More recently, than 1 order of magnitude. Butun et al. also achieved the
Haratipour et al. reported ρC = 1.14 × 103 Ω·μm on a ∼10 nm enhancement of light emission from large-area monolayer
BP flake contacted with Ti/Au (5/100 nm), encapsulated by MoS2 using plasmonic silver nanodisc arrays, where enhanced
HfO2.71 photoluminescence up to 12-fold was obtained, as displayed in
Parallel studies have also been conducted on BP. Anugrah et Figure 10A−D.84
al. deposited ferromagnetic (FM) contacts on few-layer Gas Sensing. Two-dimensional materials are of interest for
phosphorene and measured φSB of 110 and 200 meV for Py gas sensors due to their high surface area, enhanced catalytic
(φM = 4.8 eV) and Co (φM = 5.0 eV), respectively.72 Similarly, properties, active sites, reactivity, low power consumption, high
TiO2/Co contacts have also been applied to exfoliated few- process compatibility, and inherent flexibility.85 MoS2-,86
layer BP and showed a reduction of the SBH from ∼206 meV WS2-,87 and SnS2-88based materials have been fabricated for
for direct Co contact to <50 meV with TiO2 tunneling layer, gas-sensing devices. Similar to graphene, the synergistic metal
tunable by the backgate. NPs in contact with 2D hybrid layers could also be employed to
Applications. Photodetection. Tsai et al. demonstrated a modulate the characteristics of gas sensors. As referred to as an
Schottky metal−semiconductor−metal (MSM) photodetector electronic sensitization effect,89 metal NPs are used to increase
using few-layer MoS2 and interdigitated Au electrodes,73 the overall active sensing surface area and the adsorption sites
achieving a responsivity of 0.57 A/W at 532 nm wavelength of gas molecules. The metals with high work function increase
with ∼70/110 μs rise/fall time. The device could endure a the hole concentration of the 2D material device. The Pd NP
temperature of 200 °C, suitable for operation in harsh film contacted with MoS2 has been found to enhance the gas
environments. The authors further improved the photodetector response to NH3.48 Similar work was also reported in which
performance in a similar device by using trilayered MoS2 and metal oxide SnO2 NPs induce a strong p-type doping effect for
shrunken contact spacing,74 demonstrating responsivity of MoS2.47 However, low work function Al depletes the hole
∼1.04 A/W, rise/fall time of 40/50 μs, and high radiation carriers of the device, thus leading to the increase of the
tolerance. Also in parallel with the work done on sensitivity for detecting NO2.
graphene,43−45,75 Zhang et al. investigated the photovoltaic After Das et al. systematically investigated the ideal metal
and the photothermoelectric contributions to the photocurrent contact for the evaluation of intrinsic transport properties of
generated in a MoS2 Schottky barrier FET using scanning MoS2-based FETs,23 Fang et al. found that degenerate vapor
photocurrent microscopy.76 Yu et al. demonstrated efficient doping of the contact regions of WSe2 FET devices could be
gate-tunable photocurrent generation in a vertical junction of effective in minimizing the Schottky barrier and lowering the
graphene−MoS2−metal (Ti), utilizing the large junction area contact resistance.67,90 This was achieved by exposing the
and the graphene as a work-function-tunable contact.16 Cao contact regions to NO2 and K vapors, as displayed in Figure 11.
and co-workers reported a GaSe MSM photodetector with high By being exposed to the gas vapor, the barrier widths at the
responsivity of 5000 A/W and response time of 10−20 ms.77 metal/2D materials’ interfaces can be significantly reduced,
Zhang et al. fabricated phototransistors based on large-area leading to a more efficient carrier tunneling and lower
WSe2 monolayers contacted with the metals of different work resistance contacts. Liu et al. reported high-performance NO2
functions.78 It has been found that the low Schottky contact (down to 20 ppb) and NH3 (down to 1 ppm) sensing using
exhibits a very high photogain of ∼105 and a specific detectivity CVD monolayer MoS2 Schottky barrier FET,91 elucidating the
of ∼1014 Jones, but the response time is ∼5 s. In contrast, the contribution of the Schottky barrier modulation upon analytical
high Schottky contact shows a fast response time of ∼23 ms, molecular adsorption in addition to the charge transfer
but the photogain and specific detectivity decrease by several mechanism in the channel. Although a BP device with similar
orders of magnitude. As the low work function of Ti is expected structure was also reported for NO2 sensing down to 5 ppb,92
to decrease the contact resistance,23 the interdigitated electro- the authors deemed the effect of Schottky barrier modulation
des were employed to substantially increase the metal−MoS2 induced by NO2 exposure minimal. Nonetheless, metals can be
contact region,79 leading to the enhancement in photocarrier employed as a catalytic layer, which is generally deposited onto
generation. A low contact resistance was achieved between the surface of the 2D materials in the form of a thin film to
MoS2 and the Ti/Au electrodes with a responsivity of 2200 A/ enhance the decomposition of gas molecules. Alsaif et al.
W under 532 nm illumination at an external bias of 1 V. The explained the H2 gas reaction using the “hydrogen spillover
WSe2-based devices fabricated by Liu et al. exhibit improved effect”.93 The Au/Pd metal layer on the 2D MoO3 induces the
metal (In, Ag, Al, and Ti) contacts to WSe2.80 Device dissociation of H2 molecules into electrons and H+ ions. The
4903 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

GRAPHENE CONTACT TO CONVENTIONAL

SEMICONDUCTORS
As a transparent semimetal, graphene has enormous potential
in Schottky-junction-based optoelectronic applications when
put in contact with conventional semiconductors, replacing
normal metals. Considerable efforts have been devoted to
understanding the fundamental properties of the graphene−3D
semiconductor Schottky junction. In this section, we start with
the basic 3D Schottky junction model, reviewing the
experimental methods for extracting the Schottky junction
parameters in graphene−conventional semiconductor junctions
and looking at models proposed for this 2D metal−3D
semiconductor system.
Schottky Junction: Barrier Height and Ideality Factor.
Tongay et al. did the systematic study on graphite−conven-
tional semiconductor contacts, providing foresight to the
graphene−semiconductor junctions.98 In this study, highly
oriented pyrolytic graphite (HOPG) contacts were made to Si,
GaAs, and SiC in “soft landing” forms: (1) spring-loaded bulk
HOPG, where HOPG is gently pressed onto the substrate; (2)
vdW adherence of cleaved HOPG flakes; and (3) HOPG
“paint”, which consists of graphite powder/flakes. All of these
forms resulted in similar junction behavior. The junctions were
characterized using the classical Schottky junction I−V and C−
V characteristics. Under forward bias V, the current through the
Schottky barrier is
⎡ ⎛ eV ⎞ ⎤
I = IS(T )⎢exp⎜ ⎟ − 1⎥
⎢⎣ ⎝ ηkBT ⎠ ⎥⎦ (9)

⎛ eΦ ⎞
IS(T ) = AA*T 2 exp⎜ − B0 ⎟
⎝ kBT ⎠ (10)
where IS(T) is the saturation current under zero bias, A is the
effective junction area, A* is the Richardson constant of the
semiconductor, and φB0 is the zero-bias SBH. η is the ideality
factor, which should be 1 if the Schottky junction current is
purely from thermionic emission. Deviation of η from 1 could
be due to bias-dependent SBH, generation recombination of
the carriers, thermally assisted tunneling current, inhomoge-
neity of the junction, and image force lowering of the SBH. To
exclude the extrinsic effects at the metal−semiconductor
Figure 10. (A) Schematic illustration of the silver nanodisc array on
interface, the flat band zero-electric-field SBH φBF can be
MoS2. (B) SEM image and (C) corresponding integrated photo- calculated as
luminescence map of the nanodisc arrays. (D) Photoluminescence ΦBF = ηΦB0 − (η − 1)(kBT /e)ln(NC/ND) (11)
spectra from the arrays.84
where NC is the effective DOS in the conduction band and ND
is the doping density. In addition to the forward I−V curve, φB
could also be extracted from the C−V characteristics of the
H+ ions are transferred and reacted with O atoms in the first reverse-biased Schottky junction. The junction capacitance is
MoO3 monolayer, forming OH2 and HxMoO3. The rest of the eεsND
H+ ions then go underneath the MoO3 layers. CJ =
2(Vbi − VR ) (12)
Solar Cell. Shanmugam et al. demonstrated Schottky
junction solar cells based on the stacked structure of ITO− where εs is the dielectric constant of the semiconductor, ND is
CVD MoS2−Au, operating in the spectral range of 350−950 the doping density, Vbi is the built-in potential, and VR is the
nm,94 with a power conversion efficiency of 0.7% for ∼110 nm reverse-bias voltage. The barrier height is then φC−V = Vbi + (EC
thick MoS2 membrane and 1.8% for ∼220 nm. Although there − EF), where EC is the conduction band bottom and EF is the
Fermi level. φC−V is often higher than φB0 from the I−V curve
are a few other reports on utilizing 2D semiconductor devices and more closely related to φBF.
for light harvesting,95−97 they attribute most of the photovoltaic Figure 12 illustrates the extraction of the SBH φB0 and the
response to the 2D p−n junction instead of the metal−2D ideal factor η from the forward I−V curve of a graphene−Si
semiconductor Schottky junction. Schottky junction.99 From eq 9, we have
4904 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

Figure 11. (A) Schematic illustration of NO2 exposure inducing p-doped source/drain contacts of a WSe2 FET. (B) Transfer characteristics of
a device before and after NO2 exposure. (C,D) Extracted effective hole mobility and output characteristics from the device in (B). (E)
Schematic illustration of K exposure inducing n-doped source/drain contacts of a WSe2 FET. (F) Transfer characteristics of a device before
exposure and after 1, 20, 40, 70, and 120 min exposure. (G) Transfer characteristics of the device in (F) after 120 min exposure with extracted
effective electron mobility displayed in the inset. (H) Output characteristics of the device in (F).67,90

eV 1 ⎛ I (T ) ⎞ ⎛ eΦ ⎞ 1
ln I ∼ + ln IS(T ) ln⎜ S 2 ⎟ = ln(AA*) − ⎜ B0 ⎟
kBη T (13) ⎝ T ⎠ ⎝ kB ⎠ T (14)
The semilogarithmic plot of ln I versus 1/T yields a linear When ln(IS(T)/T2) versus 1/T is plotted, φB0 is extracted from
regime under moderate forward bias, where the series resistance the linear fit slope; the intercept of the fitted line with the
of the Schottky diode has not emerged and the current can be vertical coordinate also gives information about the junction
described by eq 9. From the slope of the linear fit eV/(kBη), the area and the Richardson constant.
ideality factor η of the Schottky diode can be extracted. The Figure 13 demonstrates the C−V extraction of the SBH
intercept of the fitted line at zero bias voltage yields the φC−V. According to eq 12, when 1/C2 versus VR is plotted, the
saturation current IS(T) (eq 10). At a given temperature, the linear fit’s slope gives the dopant density ND, and the intercept
zero-bias SBH φB0 can be readily calculated, assuming that we with the horizontal axis gives the built-in potential Vbi, from
have reliable values of the effective junction area A and the which the SBH can be calculated as φC−V = Vbi + (EC − EF). In
Richardson constant A*. However, the accurate value of AA* is addition to these routine methods of characterizing the
not always available, and temperature-dependent measurement Schottky junction properties, Yim et al. proposed a technique
gives an alternative of obtaining φB0. From eq 10, we have for extracting the junction resistance and capacitance using
4905 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

Figure 14. Extraction of the SBH using impedance spectroscopy.

(A) Device layout and its equivalent circuit. (B) Measured
impedance spectrum. (C) Fitting parameters (C and R of parts 1,
2, and 3) extracted from (B). (D) C−V characteristics of part 2 for
extracting the SBH.100
Figure 12. Extraction of the SBH and ideality factor η of the
graphene/Si Schottky junction. (A) Device structure. (B) Linear fit easily be identified. From the extracted capacitance value, the
of the forward I−V curve according to the thermionic emission SBH was extracted using eq 15 (Figure 14D). This approach
model. (C) Temperature dependence of the saturation current.98,99 has already been used in a recent work for analyzing the
graphene/silicon junction-based vapor sensing (chloroform,
phenol, and methanol).101
Characterizing the Graphene/Conventional Semicon-
ductor Junctions. As discussed above, the pioneering work of
Tongay et al. adapted the classical Schottky junction model for
the graphite/semiconductor contacts and suggested the
possibilities of exploring and characterizing graphene/semi-
conductor systems using the well-established methods. Here we
summarize the series of parallel studies done on graphene
contacted to some most commonly used semiconductors,
namely, on Si, Ge, GaAs, GaN, SiC, and ZnO (Table 3).
Silicon. Following Tongay’s work, Chen et al. carried out the
detailed study on the exfoliated bilayer graphene/silicon
Schottky junction,102 setting out the fundamentals for the
future integration of graphene with silicon. Bilayer graphene on
both n-type and p-type silicon substrates exhibited rectifying I−
Figure 13. C−V extraction of the SBH for Schottky junctions V behavior, from which a SBH of 0.4 eV was extracted,
formed between graphite and Si, GaAs, and SiC.98
consistent with ref 98. In this early work, however, the ideality
factor η of the junctions was high, in the range of 4.89−7.69 for
impedance spectroscopy,100 which they suggested could n-type Si and 29.67−33.50 for p-type Si. The photocurrent of
differentiate the parasitic capacitance/resistance in the equiv- the junctions was also characterized. To connect with our
alent circuit of the device under testing. Here, the graphene-on- previous discussion of the contact resistance at the 2D/3D
silicon Schottky device was divided into three parts, each having interface, this work gives a Au/graphene contact resistivity of
its own resistance and capacitance (Figure 14A). From the 212 × 10−6 Ω·cm2 and ρC at the graphene/Si interface is 73.6 ×
equivalent circuit, the total impedance of the device can be 10−6 Ω·cm2, suggesting a bonding energy for graphene and Si
fitted to the experimental impedance spectrum (Figure 14B), higher than that with Au. After the initial graphite study,98
and the fitting parameters give the individual R and C of the Tongay et al. continued with a more systematic work on CVD
parts 1, 2, and 3 (Figure 14C). From the dependence of R and graphene−semiconductor interfaces,110 on n-type Si, GaAs, 4H-
C on the bias voltage, the intrinsic junction component could SiC, and GaN. Again, they used both I−V and C−V
4906 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

Table 3. Graphene/Conventional Semiconductor Schottky Junction Properties

3D semiconductor ΦSB (eV) η ref (year) graphene type
n-Si 0.4 4.89−4.76 102 (2011) mechanically exfoliated
p-Si 0.4 ∼30 102 (2011) mechanically exfoliated
GaN 0.74 2.9 103 (2011) CVD
GaN 0.7 2.4 103 (2011) CVD, *after anneal
AlGaN 0.4 104 (2014) CVD
SiC 0.08 105 (2012) epitaxial
SiC 0.58 4.5 106 (2014) epitaxial
SiC 1.3 107 (2014) CVD transferred on C−SiC
SiC 1.16 107 (2014) CVD transferred on Si−SiC
SiC 0.85 108 (2009) epitaxial, local I−V
SiC 0.36 108 (2009) exfoliated then deposited, local I−V
ZnO 0.65 1.15 109 (2012) CVD
IGZO 0.64 (off), 0.04 (on) 1.27 19 (2014) CVD

characteristics to extract the Schottky junction parameters. The Alternatively, Shiu et al. used scanning photoelectron
barrier height was rather different than the values reported microscopy and high-resolution photoemission spectra to
before with exfoliated graphene, but the Schottky−Mott model obtain the Schottky barrier properties at the graphene/GaN
holds well for all the systems measured. In this work, the interface.115 They demonstrated the visibility and the SBH of
authors pointed out an essential difference between graphene single-layer graphene on an n-type GaN/AlN/Si3N4/Si(111)
and a conventional metal is that the Fermi level shifts in substrate and found that the SBH is decreased with decreasing
graphene during the charge transfer across the graphene/ numbers of layers of graphene. Therefore, monolayer graphene
semiconductor interface because of its low DOS. Moreover, a on GaN is most desirable for decreasing the barrier between the
strong bias voltage also shifts the graphene Fermi level, which transparent electrode and GaN and improving the LED
leads to significant changes in the diode current. Because of this performance.
tunability of the graphene Fermi level, the graphene/semi- Silicon Carbide. Tadjer et al. investigated the Schottky
conductor junction properties could be adjusted by electric junction formed between epitaxial graphene (EG) on SiC and
doping of the graphene layer. Yang et al. demonstrated this the substrate,105 using its I−V and C−V characteristics over a
concept in their work,111 as “graphene barrister” with a gate- wide temperature range (15−320 K). Using TLM, they found
controlled Schottky barrier at the graphene−silicon interface the EG sheet resistance of 1408 Ω/sq and a contact resistivity
for achieving a current on/off ratio of 105. A similar idea was of 3−7 × 10−6 Ω·cm2. The forward-bias barrier height was
realized in ref 112 by controlling the graphene/silicon contact extracted to be 0.08 eV from the Richardson plot. Such a low
behavior (from Schottky to Ohmic) using a polymer electrolyte barrier height was attributed to the image force decrease and
gate. The graphene doping tunes the graphene−silicon work maximized tunneling probability to the bottom of the SiC
conduction band (field emission). From the temperature-
function difference, leading to tunable photocurrent in the
dependent measurements, the Arrhenius behavior of the barrier
Schottky diode, as well. The model and experimental works on
height only persisted down to 140 K, and the ideality factor was
charge transfer at the graphene/semiconductor interface will be
almost temperature-independent, confirming again that the
discussed in detail in another section.
field emission was the dominant current component. The
Germanium. Excellent mobility was observed in graphene
authors also determined the barrier height from the C−V
transferred onto the Ge substrate with GeOx surface layers.113
characteristics to be 0.55−0.75 eV down to 80 K, which
Baek et al. used graphene as a tunneling barrier between Ge and increased sharply to 1.09 eV at 15 K. The reverse-bias current
Co. The Ge/Co Schottky barrier could be tuned by the number increases significantly when the temperature decreases, while
of layers of the inserted graphene. Graphene here has potential the forward-bias current does not show as strong a trend. The
as the tunneling barrier for spin injection from Co to Ge.114 forward current is speculated to be dominated by defect-
Gallium Nitride. As a wide band gap (3.4 eV) semi- assisted tunneling. Dharmaraj et al. grew Bernal-stacked bilayer
conductor, GaN has been widely used in bright light-emitting EG on 4H-SiC (0001) using low-energy e-beam irradiation,
diodes (LEDs). Its special properties lead to other applications which breaks the bond in SiC and causes Si sublimation,106 and
in optoelectronics and high-power and high-frequency devices. observed the Schottky barrier formation and the rectification at
Tongay et al. investigated the graphene/GaN junction proper- the EG−SiC junction, with φB0 ∼0.58 eV and η ∼4.5.
ties at elevated temperatures to confirm the possibility of Zinc Oxide. Lee et al. fabricated Schottky junction sputtering
integrating graphene into GaN-based devices.103 They found ZnO on a prepatterned single-layer graphene sheet.109 They
decent rectifying behavior in the as-fabricated junction, with measured φB0 of 0.65 eV and an ideality factor η of 1.15. Using
φB0 of 0.74 eV and η of 2.9. The rectification was preserved a backgate, the SBH can be modulated, which leads to
until 550 K and diminished above 650 K. After a 900 K anneal, improvement in the photoresponse of the device. Liu et al.
the Schottky diode was improved with φB0 of 0.70 eV and η of fabricated gate-tunable graphene on indium gallium zinc oxide
∼2.4, possibly due to the reduction of impurities and residue at (IGZO) junction in vertical thin film transistors for potential
the interface. Graphene shows excellent thermal stability in this applications in large-scale flexible electronics.19 The Schottky
study, and the usual problem of interdiffusion of atoms at the junction had an ideality factor of 1.27 and barrier height of 640
metal/semiconductor interface in conventional Schottky meV in the off-state and 40 meV in the on-state with nearly
junctions is avoided. linear Ohmic behavior.
4907 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

Organic Crystals. The lack of knowledge of graphene

contacting ultrathin organic thin films also prevents the
development of organic electronics applications. Therefore, it
is desirable to understand the intrinsic charge transport in the
2D limit between 2D materials such as graphene and organic
crystals.116 Generally, the charge transport occurs in the first
few molecular layers near the dielectric interface assumed in
organic FETs. Zhang et al. fabricated highly ordered single-
crystalline mono- to tetralayer pentacene crystals on hexagonal
boron nitride via vdW stacking.117 They found that the charge
transport is dominated by hopping in the first conductive layer
and then transforms to band-like in subsequent layers due to
strong modulation of the molecular packing by interfacial vdW
interactions. Moreover, the structural modulation becomes
negligible beyond the second conductive layer, leading to a
mobility saturation thickness of ∼3 nm. He et al. also grew few-
layer dioctylbenzothienobenzothiophene (C(8)-BTBT) molec-
ular crystals on a graphene/boron nitride substrate via vdW
epitaxy with controlled thickness down to the monolayer.118
The crystalline layers are atomically smooth and decoupled
from the substrate due to weak vdW interactions, affording a
pristine interface. A carrier mobility of 10 cm2 V−1 s−1 and a
saturation voltage of ∼1 V have been achieved.
Junction Inhomogeneity. In most cases, the transferred
graphene interacts with the substrate via weak vdW forces, and
there could easily be wrinkles, folds, or defects in the graphene
sheet, especially in CVD graphene. Even EG exhibits
inhomogeneities. It has been discussed in previous sections
that the ideality factor η of a graphene/semiconductor usually Figure 15. Inhomogeneity characterized by CAFM in graphene/SiC
exceeds 1, suggesting inhomogeneity of the junction in addition Schottky junctions. Sample topography and localized I−V
to other possibilities such as tunneling current contribution. In characteristics of (A) few-layer DG/SiC and (B) EG/SiC junctions.
this part, we will look into the experiments done at the (C,D) Histograms and band diagrams for the sample in (A) and
nanoscale for localized I−V characteristics, exploring spatially (B), respectively.108
resolved Schottky junction properties.
Sonde et al. investigated the current distribution in
graphene/SiC junctions using scanning current spectrosco- AlGaN/GaN is also an important heterostructure for high-
power electronics. Following up on the previous work of
py,108 implementing Pt-coated conductive atomic force
graphene/SiC junction inhomogeneities,108 Giannazzo and co-
microscopy (CAFM), a technique developed by Giannazzo et
workers investigated the graphene contact to AlGaN/GaN
al. for characterizing laterally inhomogeneous Schottky
using the same methodology and found that the microstructure
junctions in wide band gap semiconductors.119 In this work,
of the AlGaN layer plays a critical role in the junction nature of
both EG and deposited/transferred graphene (DG) form the contact, which could be either Schottky or Ohmic
Schottky contact with the single-crystalline SiC substrate. As depending on the defects of the AlGaN layer.104,120 For a
shown in Figure 15, the local I−V characteristics show perfect AlGaN layer contacted by the Au-coated conductive
distribution on both the SiC substrate and the graphene area, AFM tip, φB0 ∼ 0.9 eV and φB0 ∼ 0.6 eV for an AlGaN layer
confirming the rectification yet inhomogeneous property of the with V defects, due to preferential current paths at the defect
junction. However, on the EG/SiC sample, at certain locations, locations. With graphene transferred on the AlGaN/GaN
the I−V curve is linear, exhibiting Ohmic-like behavior. The heterostructure, φB0 for the perfect AlGaN sample was reduced
histograms of the SBH give the same φB0 of the Pt/SiC to 0.4 eV, and for the V defect sample, the contact became
junction (∼1.6 ± 0.2 eV), compared to φB0 of DG/SiC ∼ 0.85 Ohmic-like. Interestingly, the SBH exhibits much lower spread
± 0.06 eV and φB0 of EG/SiC ∼ 0.36 ± 0.1 eV. It is interesting with the graphene compared with the bare CAFM tip/AlGaN/
to note that the SiC substrate has such a high level of GaN contact. This effect was attributed to the current
inhomogeneity when put into Schottky contact with the Pt- spreading/averaging effect of the graphene over the AlGaN
coated AFM tip. On the other hand, the DG demonstrates surface.
relatively high homogeneity in the contact, and the average Zhong et al. studied the electrical properties of the CVD
SBH (∼0.85 eV) is in reasonable agreement with the Mott− graphene transferred onto both n- and p-GaN using conductive
Schottky relation for graphene (WGr ∼ 4.5 eV) and SiC (χSiC ∼ AFM and found that the wrinkles on graphene form Ohmic
3.7 eV). The EG on SiC, however, experiences a Fermi level contact with the substrate while the flat regions are in Schottky
pinning ∼0.49 eV above the Dirac point, due to the positively contact with GaN.121 Using density functional theory, they
charged states at the interface between the Si face and C-rich calculated the Fermi level at the standing and folded wrinkles of
reconstructed buffer layer during the epitaxial growth. From the graphene, with zigzag or armchair directions. The local EF shifts
I−V characteristics, the energy distribution of these interface toward n- or p-type GaN for the two directions, respectively,
states is inhomogeneous and varies rapidly at the nanoscale. acting as high-conductive channels.
4908 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

Alternatively, Shivaraman et al. analyzed the inhomogeneity

of the Schottky junction formed between EG and SiC using
macroscopic characterizing methods, that is, the I−V and C−V
measurements.122 They based their analysis on the model
proposed originally by Güttler123 and Tung,124 which suggests
that φB0 extracted from the I−V curve is the effective barrier
height and φ C−V is the mean barrier height of the
inhomogeneous junction. They satisfy the following relation
qσ 2
ΦB0 = ΦC − V −
2kBT (15)
where σ is the standard deviation of the Gaussian distribution
and is assumed to be temperature-independent. This equation
provides a means of extracting the spread of the barrier height
distribution by plotting the difference between φB0 and φC−V
versus 1/T. The standard deviation was found to be 137 ± 11
meV, in fair agreement with that in ref 108. Parui et al.
characterized the Gaussian distribution of the SBH in
graphene/n-Si diodes and found σ also in the range of ∼100
meV.99
Tomer et al. combine both the localized morphology/
electrical measurements and macroscopic, temperature-depend-
ent I−V characteristics to study the intrinsic junction
Figure 16. Inhomogeneities in a CVD graphene/SiC junction. (A)
inhomogeneity of the transferred CVD graphene/SiC sys- STM image of CVD graphene on SiC; inset is the profile along the
tem.107 Nonideal behaviors such as increasing φB0 and dashed line. (B) STM image of the graphene ripples; inset shows
decreasing η with increasing temperature are related to the the corresponding dI/dV spectrum. (C) STM image of a ripple. (D)
spatial inhomogeneities in graphene, examined by STM and dI/dV spectra taken from the corresponding spots shown in (C);
STS. STM/STS measurements over graphene ripples give a normal probability distribution of the Dirac energy is shown in the
Fermi level shift of 270 meV with fwhm of 70 meV, with inset.107
respect to the Dirac point. The SBH was measured to be 1.3 ±
0.18 eV and 1.16 ± 0.16 eV for graphene on C-terminated SiC
and Si-terminated SiC, respectively. The spatial inhomogene-
ities are mostly due to ripples, ridges, and SiC steps, revealed by
the STM, as shown in Figure 16.
Toward Low-Dimensional Models: Charge Transfer
and Landauer Approach. As mentioned before, Tongay et al.
proposed the model for charge transfer at the graphene/
semiconductor interface, which is a significant difference
between the role of graphene in a Schottky junction and that
of a conventional metal, due to the low DOS near the Dirac
point in graphene.110 In contrast to a conventional Schottky
diode, the reverse-bias current is not saturating but increasing
continuously with the increased VR (Figure 17). In the reverse-
bias regime, the electron charge per unit area on the graphene is
Q = en induced = C J(Vbi + VR ) (16)
where CJ is given by eq 12. When the residual doping carrier
density n0 in graphene is taken into account, the Fermi level
shift with respect to the Dirac point can be calculated as
⎛ εsND(Vbi + VR ) ⎞
E F = −h ̅ |vF| π ⎜⎜n0 − ⎟⎟
⎝ 2e ⎠ (17)
Therefore, upon a reverse bias being applied, the SBH changes
by the amount of
eΔΦSBH(VR ) = −ΔE F(VR ) = hv̅ F⎡⎣ π(n0 − n induced) − πn0 ⎤⎦ Figure 17. J−V characteristics at the graphene/(A) n-Si, (B) n-
GaAs, (C) n-4H-SiC, and (D) n-GaN interfaces.110
1 n 1 πεsND(Vbi + VR )
≈− hv̅ F πn0 induced = − hv̅ F
2 n0 2 2en0 (18) ⎛ e Φ0 + eΔΦ (V ) ⎞
IS(T ) = AA*T 2 exp⎜ − SBH SBH
⎟
Accordingly, the saturation current in a Schottky junction, ⎝ kBT ⎠ (19)
given before in eq 10, becomes
4909 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

Figure 18. Schottky junction properties of a graphene/Si diode. (A) I−V characteristics and the fit using classical thermionic emission model;
A* is the effective Richardson constant in A·cm−2·K−2. (B) SBH as a function of the reverse-bias voltage. (C) I−V characteristics and the fit
using the Landauer transport model.126

An et al. adapted this charge transfer model, as well, and junction current given by eqs 9 and 10, the Landauer formalism
calculated the bias-dependent Fermi level shift in a graphene/Si gives
Schottky junction.125 A systematic theoretical and experimental e +∞
discussion about the charge transfer at the graphene/semi- J=
τ
∫−∞ T (E)D(E)(fg − fSi )dE
(20)
conductor interface could also be found in ref 121, with the
experiment carried out in a graphene/GaN system. Cavallo et where τ is the time scale for carrier injection from the contact,
al. demonstrated an interesting result on enhanced carrier T(E) is the transmission probability over the Schottky barrier
mobility in graphene deposited on a GeOx/Ge substrate, due to φB, D(E) is the graphene DOS given by 2/[π(hvF)2]|E| = D0|E|,
the charge transfer between the surface states in GeOx and the where vF is the Fermi velocity, and f is the Fermi−Dirac
graphene.113 distribution. Eq 20 yields
Finally, we will discuss the 2D nature of graphene as the ⎡ ⎛ eV ⎞ ⎤
“metal” part in a Schottky junction and its influence on the J = JS ⎢exp⎜ ⎟ − 1⎥
derivation of the junction current expression, compared with ⎣⎢ ⎝ ηkBT ⎠ ⎥⎦
the classical Richardson form given by eqs 9 and 10. There is ⎡ eD ⎛Φ ⎞⎤ ⎛ −Φ ⎞
one recent discussion on the Landauer transport formalism in JS = ⎢ 0 (kBT )2 ⎜ B + 1⎟⎥exp⎜ ⎟
an “ideal” graphene−silicon junction.126 Three kinds of samples ⎢⎣ τ ⎝ kBT ⎠⎥⎦ ⎝ kBT ⎠ (21)
were studied in this work: (1) CVD-grown graphene on Si;
prior to the CVD growth, the surface of the Cu foil was treated Comparing eq 21 with eq 10, it is therefore the graphene/
by ammonium persulfate solution to etch away ∼700 nm of Cu, semiconductor coupling time scale τ accounting for the low
values of A* obtained in the graphene Schottky junctions. τ is
minimizing the metallic impurities in graphene. This kind of
4.62 × 10−11 s, as obtained from the fit. Figure 18B illustrates
sample gave the most ideal Schottky junction behavior with η ∼ the reverse-bias-dependent barrier height φB observed in the
1.08, regularly CVD-grown graphene on Si with η of 1.63, and “ideal” sets of graphene/Si samples, which is also in agreement
conventional metal/silicon Schottky junction Cr/Si with η of with the charge transfer phenomenon discussed previously in
1.3. The J−V characteristics are shown in Figure 18A. The this section. Using the Landauer formalism again
curious point is that while the classical Schottky junction model +∞
gives the Richardson constant for a certain semiconductor (Si e ∫−∞ D(E)f (Ef )dE = CSi|V |
(22)
in this study), the A* extracted from the J−V characteristics
shows large variation in the three sets of devices. This the authors were able to fit the “ideal” junction I−V behavior
discrepancy was explained by the Landauer transport formal- reasonably well, as shown in Figure 18C, accounting for the
ism, as proposed by the authors. Instead of the Schottky nonsaturating reverse-bias current.
4910 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

Table 4. Graphene/3D Semiconductor Schottky Junction Photodetectors

material responsivity rise/fall time wavelength (nm) ref
Si 2.8 mA/W 1550 128
Si 29 mA/W ∼100 μs 850/940 129
Si 435 mA/W a few milliseconds 850 130
Si 0.11 A/W 633 133
Si 300 V/W 0.5 s 550 133
Si ∼18−128 mA/W 0.26/0.54 ms 445 134
porous Si 0.2 A/W 3 μs 400 135
Ge 51.8 mA/W 23 μs 1400 127
GaAs 3.73 mA/W 72/122 μs 850 131
ZnO NW array 0.55 mV/W 0.1 s 1064 132
ZnO NW array 0.2 mV/W 808 132
ZnO NW array 0.32 mV/W 532 132
ZnO nanorod array 113 A/W 0.7/3.6 ms 365 136
Ge QD Gr/ZnO 9.7 A/W 1400 137

Applications. Photodetectors. Graphene/3D semiconduc- responsivity 113 A/W and the rise/fall time of 0.7/3.6 ms. Kim
tor Schottky junction has been broadly explored for photo- et al. reported near-UV detection using graphene/porous Si
detector applications. The device performances are summarized junctions,135 with a responsivity of 0.2 A/W and ∼3 μs
in Table 4. response time at 400 nm illumination.
Infrared Detection. Zeng et al. reported a monolayer Gas Sensors. Kim et al. reported Schottky diodes made
graphene/Ge Schottky junction photodetector with a peak from monolayer graphene on p- and n-Si substrates, exhibiting
sensitivity at 1400 nm wavelength, with a responsivity of 51.8 stable sensing ability to selected gas (NH3) and aromatic
mA/W and 23 μs response time.127 The Stake group showed molecules (anisole, benzene, chlorobenzene, and nitroben-
2.8 mA/W responsivity at 1.55 μm wavelength using an zene), making them potentially useful for environmental and
exfoliated graphene/Si Schottky junction.128 Lv et al. reported biocompatible sensors.138 The sensing mechanism was
29 mA/W, with ∼100 μs response time at 850/940 nm attributed to the charge transfer from adsorbents to graphene,
illumination with graphene/Si.129 An et al. presented weak which directly affects the work function and conductivity of the
signal detectors using monolayer graphene/Si junctions with a graphene as well as the SBH. Uddin et al. demonstrated a H2
photovoltage responsivity of ∼107 V/W, a photocurrent sensor with down to sub-ppm sensitivity using a Pd- and Pt-
responsivity up to 435 mA/W, and a microsecond response functionalized graphene/Si heterostructure.139 Under reverse
time.130 These tunable and scalable devices are well-suited for bias, the junction current varies exponentially when the SBH
broad-band (400−900 nm) photodetection, with responsivity changes due to the molecule adsorption, enabling low-power
peaking at 850 nm, ideal for low-power integrated optoelec- operation and tunable sensitivity. Singh et al. demonstrated
tronic circuit. Luo et al. used a GaAs nanocone array/ NH3 and NO2 detection using the same device structure,
monolayer graphene Schottky junction to detect near-infrared confirming the SBH change using C−V measurements.140 Kim
light, with a peak sensitivity of 3.73 mA/W at 850 nm
et al. fabricated Si nanowire arrays vertically aligned by metal-
illumination.131 The rise/fall time is 72/122 μs. Liu et al.
assisted chemical etching of the Si wafer. With graphene on top,
investigated the broad-band photoresponse of a ZnO nanowire
the ends of the nanowires were prevented from being bundled
array/reduced graphene oxide film photodetector under 1064,
and formed uniform Schottky junction with the graphene.141
808, and 532 nm illumination and achieved photovoltaic
responsivity of 0.55, 0.20, and 0.33 mV/W, respectively,132 with The device showed 37 and 1280% change in resistance and 3.5/
the photoresponse time of ∼0.1 s. Wang’s group reported a 0.15 and 12/0.15 s response/recovery time for O2 and H2
high-performance Ge quantum dot decorated graphene/ZnO detection, respectively. Zhu et al. evaluated reduced graphene
infrared photodetector.137 The device demonstrated ∼9.7 A/W oxide/Si Schottky diodes for gas sensing, including CH4, O2,
responsivity under 1400 nm illumination with a fast rise/fall CO, NO2, NO, and SO2,142 and discussed the influence of the
time of ∼40/90 μs. oxygen functional groups on the surface of the reduced
Visible Light. The MSM photodetector demonstrated by An graphene oxide.
et al. works at 633 nm wavelength with a responsivity of 0.11 Solar Cells. The efficiencies of the graphene/3D semi-
A/W.125 Zhang et al. introduced the direct growth of conductor solar cells are summarized in Figure 19. Li et al.
nanocrystalline graphene/graphite on Si/SiO2 to form a reported graphene/n-Si Schottky junction solar cells with
metal-free Schottky junction,133 which achieved a 300 V/W efficiencies up to 1.5%.143 The same group later demonstrated
photovoltage responsivity with ∼0.5 s response time at 550 nm an enhanced efficiency of 2.86% using the Schottky junction
wavelength. Zhu et al. demonstrated that reduced graphene formed between graphene and a Si nanowire array with faster
oxide/Si Schottky photodiodes showed responsivity in the response time.144 Xie et al. also reported 2.15% efficiency using
range of 18.12−128.72 mA/W and rise/fall time of 0.26/0.54 a similar device structure.145 Feng et al. demonstrated up to
ms under 445 nm light.134 2.9% efficiency in graphene/silicon pillar array solar cells and
Ultraviolet. Nie et al. presented the Schottky junction UV enhanced it up to 4.35% after HNO3 treatment.146 Poly(3,4-
detector fabricated by coating a ZnO nanorod array with ethylenedioxythiophene):polystyrenesulfonic acid (PE-
graphene.136 The UV light is effectively trapped in the nanorod DOT:PSS) coating on graphene also improves the efficiency
array, enhancing the absorption of the light, achieving a slightly.
4911 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

Lee et al. grew MoS2 directly on SiC using CVD and

characterized the p−n junction formed between the 2D MoS2
and 3D SiC.155 Figure 20 demonstrates the I−V and C−V

Figure 19. Efficiency of the graphene/3D semiconductor Schottky

junction solar cells. The red dots are graphene/Si devices; the
green square is graphene/Zn3P2; the blue triangle is graphene/
CH3NH3I pasted perovskite. The numbers refer to the references.

Miao et al. improved the efficiency significantly to 8.6% using

a graphene/n-Si Schottky junction with the graphene doped by
bis(trifluoromethanesulfonyl)amide, which shifts the graphene Figure 20. Properties of a p-MoS2/n-SiC junction. (A) Temper-
chemical potential, and increased the carrier density and the ature-dependent I−V characteristics; inset shows the device
built-in potential at the junction.147 Cui et al. reported structure. (B) C−V characteristics; inset is the linear 1/C2 plot
graphene/Si Schottky junction solar cells with initial efficiencies for extracting the built-in voltage. (C) Band diagram.155
in the range of 1.94−2.66% and demonstrated an enhancement
to 5.95% by doping treatment of SOCl2.148 The influence of characteristics. The I−V curve does not show strong temper-
other volatile oxidants such as HNO3, HCl, and H2O2 was also ature dependence and follows the multistep recombination
investigated. Xie et al. carried out a study on the modification of tunneling current expression, giving an ideality factor η of 2.1.
the graphene/Si interface to tune the performance of the Chuang et al. demonstrated nearly ideal electrical properties of
Schottky junction solar cell.149 Methylated Si and the insertion the InAs/WSe2 junction with an ideality factor of 1.1.156
of a thin P3HT organic layer at the junction interface Yim et al. investigated photodiodes consisting of CVD MoS2
suppressed the surface carrier recombination effectively and (n-type) and p-type silicon (Figure 21A).157 They varied the
led to a high power conversion efficiency of 10.56%. Li et al. thickness of the MoS2 layer in the range of ∼4−16 nm. All of
demonstrated an efficiency of 9.2% by nitrate ion doping of a the devices showed current rectification and broad spectrum
graphene/n-Si solar cell.150 Ho et al. reported the selective photoresponse. The photocurrent exhibited strong dependence
filling of the topological cracks in graphene using Au NPs in a on the MoS2 thickness (Figure 21B). The large photocurrent at
graphene/Si Schottky junction solar cell151 and achieved an higher MoS2 thickness is probably due to the enhanced light
efficiency as high as 12.3%. Song et al. investigated the influence absorption volume in thicker MoS2 films. Esmaeili-Rad et al.
of the native oxide at the graphene/Si junction interface and reported the integration of 60 nm mechanically exfoliated MoS2
selected the optimal oxide thickness for efficiency enhance- onto a thin film amorphous silicon (a-Si, 100 nm) MSM
ment.152 When the graphene was doped with AuCl3 and coated structure (Figure 21C), which enhances the performance of the
with a TiO2 antireflective layer, the efficiency was improved to original a-Si photodetector.158 Amorphous Si has a low mobility
15.6%. (∼10 cm2 V−1 s−1), and the integration of MoS2 makes the
Vazquez-Mena et al. presented a graphene/Zn3P2 solar cell device 10× faster with the photocurrent rise/fall time of ∼0.3
with gate-tunable junction barrier153 and showed an efficiency ms and responsivity of 210 mA/W at 550 nm wavelength. Li et
of 1.9% with an open-circuit voltage of 0.53 V. Yan et al. al. also investigated the photodiodes formed between exfoliated
reported 11.5% efficiency in the device formed using multilayer monolayer MoS2 and both n- and p-type silicon159 and achieved
graphene and CH3NH3I pasted onto perovskite.154 a high photoresponse of 7.2 A/W in the MoS 2/n-Si
heterostructure. MoS2/p-Si demonstrated a weaker built-in
TWO-DIMENSIONAL SEMICONDUCTOR CONTACT TO voltage. Lopez-Sanchez et al. demonstrated both LEDs and
CONVENTIONAL SEMICONDUCTORS: THE p−n solar cells with the n-type exfoliated monolayer MoS2/p-Si
JUNCTION heterojunction.161 The LED shows a low threshold power
The integration of 2D semiconductors onto 3D semiconductor density of 3.2 W/cm2, significantly lower than previously
substrates opens up possibilities of device design for current reported 15 kW/cm2 for other MoS2 electroluminescent
rectification, photodetection, and energy harvesting. The devices. The solar cell has an external quantum efficiency of
layered semiconductors such as TMDs (e.g., MoS2) typically 4.4%. Tsai et al. fabricated solar cells using CVD monolayer
do not have out-of-plane bonds at the surface, therefore MoS2 transferred to p-Si.162 The built-in electric field at the
diminishing lattice mismatch and stress/strain-related problems MoS2/p-Si interface separates the photogenerated carriers and
at the heterojunction interface. produces a photocurrent or open-circuit current. In this work,
4912 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

Figure 21. (A) Schematic diagram and photograph of the MoS2/Si photodiode. (B) MoS2 thickness dependence of the photocurrent density
for devices shown in (A). (C) Schematic illustration of the Si/MoS2 MSM photodetector. (D) Schematic of the MoS2/Si photodetector,
showing the vertically standing layered MoS2.157,158,160

the MoS2/p-Si device shows a power conversion efficiency of regarding the nature of charge transport in the 2D organics
5.23%, reported to be the highest among all monolayer TMD have not been addressed.
solar cells. Hao et al. deposited bulk-like MoS2 film on the p-Si For any of the 2D materials discussed, making the electrical
substrate using dc magnetron sputtering to form MoS2/Si p−n contact is the prerequisite for any kind of research or
junctions.163 Current rectification and photovoltaic character- application, especially for high-performance electronic or
istics were also observed in these devices. Wang et al. optoelectronic devices. From this review, we see that the
demonstrated the sputtering deposition of MoS2 films with a contact to graphene has improved from ρC ∼ 10−4 to 10−5 Ω·
vertically standing layered structure on p-Si,160 as shown in cm2 around 2010 to the current state-of-art of ∼10−8 to 10−9 Ω·
Figure 21D. The junction showed an on/off ratio of 5000 and cm2 in most recent studies, by introducing well-controlled
ideality factor of 1.83. As a photodiode, the heterostructure has graphene edges in the contact area and using metals which form
a high detectivity of ∼1013 Jones and short response time of ∼3 carbides with graphene. In first section of this review,
μs, due to the special vertically standing layered MoS2 which mechanically exfoliated and CVD graphene devices are
offers strong light absorption and fast photogenerated carrier discussed in detail for high-performance applications. On the
separation and transport. other hand, liquid exfoliated graphene or solution-processable
layered materials are more important in large-scale production
PERSPECTIVE of thin film transistors aiming at lower-performance elec-
Two-dimensional electronics has become an extremely active tronics.5,166 Though there are some encouraging results on ink-
field over the past 10 years, with intriguing expectations for printed electronics,167−171 the contact to the conductive ink has
real-life applications in the foreseeable future and eventually not been sufficiently studied, which remains a blank to be filled
bringing about the ultimate flexible electronics. In this roadmap, in. TMDs and BP lag behind graphene by about 4 orders of
integrating the 2D materials with existing 3D CMOS magnitude in terms of the contact resistivity, and the
technology is the low-hanging fruit, and indeed, there have microscopic mechanisms of contact formation between metal
already been devices commercialized, such as graphene touch and 2D semiconductors are still not well understood. Graphene
screen, graphene-coated LEDs, etc. However, in these initial with etched edges could serve as the buffer layer between the
products, the outstanding electrical properties of graphene are 2D semiconductor and the metal for contact improvement.172
implemented only to a limited extent. The TMD devices are Alternatively, the 1T metallic phase in TMDs is explored for
still far away from applications because of their relatively poor Ohmic contact with the metal electrodes, while the TMD
performance and the challenges in large-scale reproducible channel is maintained at the regular 2H phase.173−176 FETs
production. BP and X-enes164,165 are expected to offer relatively with 1T phase/metal contacts are reported to show superior
high mobility and a tunable band gap but suffer from ambient mobilities and contact resistivity as low as 200−300 Ω·μm for
instability, and their investigations are still at an early stage. MoS2.175 TMDs and BP have large spin−orbit coupling and are
Two-dimensional molecular crystals held by vdW forces, which attractive candidates for spin manipulation in hybrid
could effectively eliminate interlayer screening, leading to ferromagnetic systems. The very young field of using
directly probing disorder effects and interfaces on charge ferromagnetic metal contact for spin-polarized current injection
transport, have barely been studied. The understanding of the has only seen a few published works merely characterizing the
contact with these 2D molecular crystals would make the Schottky junction barrier height, tunable by inserting a tunnel
carrier injection and modulation more efficient to improve the barrier, yet not concerning the spins. Much more work is
performance of organic FETs. However, the fundamental issues needed for injecting spins into these 2D semiconductors.
4913 DOI: 10.1021/acsnano.6b01842
ACS Nano 2016, 10, 4895−4919
ACS Nano Review

Contacting graphene or 2D semiconductors to 3D semi- REFERENCES

conductors provides alternative routes for making diodes. Most (1) Ferrari, A. C.; Bonaccorso, F.; Fal’ko, V.; Novoselov, K. S.;
of metal/2D materials are Schottky contacts, which lead to the Roche, S.; Boggild, P.; Borini, S.; Koppens, F. H. L.; Palermo, V.;
measured two-point resistance from TLM depending on the Pugno, N.; Garrido, J. A.; Sordan, R.; Bianco, A.; Ballerini, L.; Prato,
voltage bias and polarization. Most of the works applied to M.; Lidorikis, E.; Kivioja, J.; Marinelli, C.; Ryhanen, T.; Morpurgo, A.;
TLM roughly estimate contact resistance of 2D materials. Few et al. Science and Technology Roadmap for Graphene, Related Two-
works thoroughly studied this limitation. Very similar to the Dimensional Crystals, and Hybrid Systems. Nanoscale 2015, 7, 4598−
contact resistance problem, the fundamental issue in these 2D/ 4810.
3D systems is the transport of the carriers between 2D and 3D (2) Fiori, G.; Bonaccorso, F.; Iannaccone, G.; Palacios, T.; Neumaier,
systems. Although the physical picture is completely different D.; Seabaugh, A.; Banerjee, S. K.; Colombo, L. Electronics Based on
from a 3D/3D interface, the research community has been Two-Dimensional Materials. Nat. Nanotechnol. 2014, 9, 768−779.
using the conventional models for years. This approach is (3) Koppens, F. H. L.; Mueller, T.; Avouris, P.; Ferrari, A. C.; Vitiello,
M. S.; Polini, M. Photodetectors Based on Graphene, Other Two-
apparently problematic, as many of the parameters would have
Dimensional Materials and Hybrid Systems. Nat. Nanotechnol. 2014, 9,
questionable physical meaning, such as the Richardson constant 780−793.
in a Schottky diode. An understanding of the 2D/3D (4) Butler, S. Z.; Hollen, S. M.; Cao, L.; Cui, Y.; Gupta, J. A.;
heterojunction transport is necessary and in urgent need for Gutiérrez, H. R.; Heinz, T. F.; Hong, S. S.; Huang, J.; Ismach, A. F.;
either the fundamental device physics or the improvement of Johnston-Halperin, E.; Kuno, M.; Plashnitsa, V. V.; Robinson, R. D.;
the device performance. Ruoff, R. S.; Salahuddin, S.; Shan, J.; Shi, L.; Spencer, M. G.; Terrones,
M.; et al. Progress, Challenges, and Opportunities in Two-Dimen-
AUTHOR INFORMATION sional Materials Beyond Graphene. ACS Nano 2013, 7, 2898−2926.
Corresponding Author (5) Franklin, A. D. Nanomaterials in Transistors: From High-
Performance to Thin-Film Applications. Science 2015, 349, aab2750.
*E-mail: [email protected]. (6) Xia, F.; Wang, H.; Xiao, D.; Dubey, M.; Ramasubramaniam, A.
Author Contributions Two-Dimensional Material Nanophotonics. Nat. Photonics 2014, 8,
⊥
Y.X. and C.C. contributed equally to this work. 899−907.
Notes (7) Kim, K.; Choi, J.-Y.; Kim, T.; Cho, S.-H.; Chung, H.-J. A Role for
The authors declare no competing financial interest. Graphene in Silicon-Based Semiconductor Devices. Nature 2011, 479,
338−344.
(8) Novoselov, K. S.; Fal’ko, V. I.; Colombo, L.; Gellert, P. R.;
ACKNOWLEDGMENTS Schwab, M. G.; Kim, K. A Roadmap for Graphene. Nature 2012, 490,
The authors thank Prof. Kostya Novoselov, Prof. Wencai Ren, 192−200.
Prof. Manish Chhowalla, Prof. Zhongze Gu, Prof. Liwei Liu, (9) Liao, L.; Duan, X. Graphene for Radio Frequency Electronics.
Prof. Xinran Wang, and Dr. Yuan Liu for helpful discussion and Mater. Today 2012, 15, 328−338.
comments. This work is supported by the National Science (10) Weiss, N. O.; Zhou, H.; Liao, L.; Liu, Y.; Jiang, S.; Huang, Y.;
Foundation (DMR1508144), NSFC (Grant Nos. 61274123 Duan, X. Graphene: An Emerging Electronic Material. Adv. Mater.
and 61474099), micro/nano-fabrication platform of ZJU 2012, 24, 5782−5825.
(11) Wang, Q. H.; Kalantar-Zadeh, K.; Kis, A.; Coleman, J. N.;
University, and the Fundamental Research Funds for the
Strano, M. S. Electronics and Optoelectronics of Two-Dimensional
Central Universities (2016XZZX001-05). This work is also Transition Metal Dichalcogenides. Nat. Nanotechnol. 2012, 7, 699−
supported by ZJU Cyber Scholarship and Cyrus Tang Center 712.
for Sensor Materials and Applications, the Open Research Fund (12) Jariwala, D.; Sangwan, V. K.; Lauhon, L. J.; Marks, T. J.; Hersam,
of State Key Laboratory of Bioelectronics, Southeast University, M. C. Emerging Device Applications for Semiconducting Two-
the Open Research Fund of State Key Laboratory of Dimensional Transition Metal Dichalcogenides. ACS Nano 2014, 8,
Nanodevices and Applications at Chinese Academy of Sciences 1102−1120.
(No. 14ZS01), and Visiting-by-Fellowship of Churchill College (13) Geim, A. K.; Grigorieva, I. V. Van Der Waals Heterostructures.
at the University of Cambridge. Nature 2013, 499, 419−425.
(14) Wang, H.; Liu, F.; Fu, W.; Fang, Z.; Zhou, W.; Liu, Z. Two-
VOCABULARY Dimensional Heterostructures: Fabrication, Characterization, and
Application. Nanoscale 2014, 6, 12250−12272.
contact resistance, resistance due to the contacting of the (15) Yu, W. J.; Li, Z.; Zhou, H.; Chen, Y.; Wang, Y.; Huang, Y.;
materials interfaces; an inherent property which is independent Duan, X. Vertically Stacked Multi-Heterostructures of Layered
of the measuring approach; transfer length method, the Materials for Logic Transistors and Complementary Inverters. Nat.
contact resistance measuring method, in which the resistors of Mater. 2013, 12, 246−252.
several different lengths are constructed; with measured the (16) Yu, W. J.; Liu, Y.; Zhou, H.; Yin, A.; Li, Z.; Huang, Y.; Duan, X.
total resistances, the contact resistance can be extracted by Highly Efficient Gate-Tunable Photocurrent Generation in Vertical
extrapolating back to the resistor length at zero; Schottky Heterostructures of Layered Materials. Nat. Nanotechnol. 2013, 8,
junction, the boundary between a metal and semiconductor 952−958.
exhibiting rectifying characteristics; Schottky barrier height, (17) Akinwande, D.; Petrone, N.; Hone, J. Two-Dimensional Flexible
the potential energy barrier for electrons formed at the Nanoelectronics. Nat. Commun. 2014, 5, 5678.
(18) Chae, S. H.; Yu, W. J.; Bae, J. J.; Duong, D. L.; Perello, D.;
Schottky junction; ideality factor, a parameter equals between Jeong, H. Y.; Ta, Q. H.; Ly, T. H.; Vu, Q. A.; Yun, M.; Duan, X.; Lee,
1 and 2, representing the ratio between diffusion and Y. H. Transferred Wrinkled Al2O3 for Highly Stretchable and
recombination current, indicating the number of defects in Transparent Graphene−Carbon Nanotube Transistors. Nat. Mater.
the barrier region; Ohmic contact, a nonrectifying electrical 2013, 12, 403−409.
junction that exhibits a linear current−voltage curve as with (19) Liu, Y.; Zhou, H.; Cheng, R.; Yu, W.; Huang, Y.; Duan, X.
Ohm’s law; p−n heterojunctions, the boundary between p- Highly Flexible Electronics from Scalable Vertical Thin Film
type and n-type semiconductor materials Transistors. Nano Lett. 2014, 14, 1413−1418.

4914 DOI: 10.1021/acsnano.6b01842

ACS Nano 2016, 10, 4895−4919
ACS Nano Review

(20) Cheng, R.; Jiang, S.; Chen, Y.; Liu, Y.; Weiss, N.; Cheng, H.-C.; (41) Xia, F.; Mueller, T.; Lin, Y.-m.; Valdes-Garcia, A.; Avouris, P.
Wu, H.; Huang, Y.; Duan, X. Few-Layer Molybdenum Disulfide Ultrafast Graphene Photodetector. Nat. Nanotechnol. 2009, 4, 839−
Transistors and Circuits for High-Speed Flexible Electronics. Nat. 843.
Commun. 2014, 5, 5143. (42) Urich, A.; Unterrainer, K.; Mueller, T. Intrinsic Response Time
(21) Tomanek, D. Interfacing Graphene and Related 2D Materials of Graphene Photodetectors. Nano Lett. 2011, 11, 2804−2808.
with the 3D World. J. Phys.: Condens. Matter 2015, 27, 133203. (43) Prechtel, L.; Song, L.; Schuh, D.; Ajayan, P.; Wegscheider, W.;
(22) Allain, A.; Kang, J.; Banerjee, K.; Kis, A. Electrical Contacts to Holleitner, A. W. Time-Resolved Ultrafast Photocurrents and
Two-Dimensional Semiconductors. Nat. Mater. 2015, 14, 1195−1205. Terahertz Generation in Freely Suspended Graphene. Nat. Commun.
(23) Das, S.; Chen, H.-Y.; Penumatcha, A. V.; Appenzeller, J. High 2012, 3, 646.
Performance Multilayer MoS2 Transistors with Scandium Contacts. (44) Echtermeyer, T. J.; Nene, P. S.; Trushin, M.; Gorbachev, R. V.;
Nano Lett. 2013, 13, 100−105. Eiden, A. L.; Milana, S.; Sun, Z.; Schliemann, J.; Lidorikis, E.;
(24) Radisavljevic, B.; Radenovic, A.; Brivio, J.; Giacometti, V.; Kis, A. Novoselov, K. S.; Ferrari, A. C. Photothermoelectric and Photoelectric
Single-Layer MoS2 Transistors. Nat. Nanotechnol. 2011, 6, 147−150. Contributions to Light Detection in Metal−Graphene−Metal Photo-
(25) Bai, J.; Liao, L.; Zhou, H.; Cheng, R.; Liu, L.; Huang, Y.; Duan, detectors. Nano Lett. 2014, 14, 3733−3742.
X. Top-Gated Chemical Vapor Deposition Grown Graphene (45) Tielrooij, K. J.; Massicotte, M.; Piatkowski, L.; Woessner, A.;
Transistors with Current Saturation. Nano Lett. 2011, 11, 2555−2559. Ma, Q.; Jarillo-Herrero, P.; Hulst, N. F. v.; Koppens, F. H. L. Hot-
(26) Novoselov, K. S.; Geim, A. K.; Morozov, S. V.; Jiang, D.; Zhang, Carrier Photocurrent Effects at Graphene−Metal Interfaces. J. Phys.:
Y.; Dubonos, S. V.; Grigorieva, I. V.; Firsov, A. A. Electric Field Effect Condens. Matter 2015, 27, 164207.
in Atomically Thin Carbon Films. Science 2004, 306, 666−669. (46) Mueller, T.; Xia, F.; Avouris, P. Graphene Photodetectors for
(27) Venugopal, A.; Colombo, L.; Vogel, E. M. Contact Resistance in High-Speed Optical Communications. Nat. Photonics 2010, 4, 297−
Few and Multilayer Graphene Devices. Appl. Phys. Lett. 2010, 96, 301.
013512. (47) Cho, B.; Yoon, J.; Lim, S.; Kim, A.; Choi, S.-Y.; Kim, D.-H.; Lee,
(28) Nagashio, K.; Nishimura, T.; Kita, K.; Toriumi, A. Contact K.; Lee, B.; Ko, H.; Hahm, M. Metal Decoration Effects on the Gas-
Resistivity and Current Flow Path at Metal/Graphene Contact. Appl. Sensing Properties of 2D Hybrid-Structures on Flexible Substrates.
Phys. Lett. 2010, 97, 143514. Sensors 2015, 15, 24903−24913.
(29) Robinson, J. A.; LaBella, M.; Zhu, M.; Hollander, M.; Kasarda, (48) Cho, B.; Yoon, J.; Hahm, M. G.; Kim, D.-H.; Kim, A. R.; Kahng,
R.; Hughes, Z.; Trumbull, K.; Cavalero, R.; Snyder, D. Contacting Y. H.; Park, S.-W.; Lee, Y.-J.; Park, S.-G.; Kwon, J.-D.; Kim, C. S.;
Graphene. Appl. Phys. Lett. 2011, 98, 053103. Song, M.; Jeong, Y.; Nam, K.-S.; Ko, H. C. Graphene-Based Gas
(30) Moon, J. S.; Antcliffe, M.; Seo, H. C.; Curtis, D.; Lin, S.; Sensor: Metal Decoration Effect and Application to a Flexible Device.
Schmitz, A.; Milosavljevic, I.; Kiselev, A. A.; Ross, R. S.; Gaskill, D. K.; J. Mater. Chem. C 2014, 2, 5280−5285.
Campbell, P. M.; Fitch, R. C.; Lee, K.-M.; Asbeck, P. Ultra-Low (49) Vedala, H.; Sorescu, D. C.; Kotchey, G. P.; Star, A. Chemical
Sensitivity of Graphene Edges Decorated with Metal Nanoparticles.
Resistance Ohmic Contacts in Graphene Field Effect Transistors. Appl.
Nano Lett. 2011, 11, 2342−2347.
Phys. Lett. 2012, 100, 203512.
(50) Varghese, S. S.; Lonkar, S.; Singh, K. K.; Swaminathan, S.;
(31) Giovannetti, G.; Khomyakov, P. A.; Brocks, G.; Karpan, V. M.;
Abdala, A. Recent Advances in Graphene Based Gas Sensors. Sens.
van den Brink, J.; Kelly, P. J. Doping Graphene with Metal Contacts.
Actuators, B 2015, 218, 160−183.
Phys. Rev. Lett. 2008, 101, 026803.
(51) Phan, D.-T.; Chung, G.-S. Effects of Pd Nanocube Size of Pd
(32) Xia, F.; Perebeinos, V.; Lin, Y.-m.; Wu, Y.; Avouris, P. The
Nanocube-Graphene Hybrid on Hydrogen Sensing Properties. Sens.
Origins and Limits of Metal-Graphene Junction Resistance. Nat.
Actuators, B 2014, 204, 437−444.
Nanotechnol. 2011, 6, 179−184. (52) Alfano, B.; Polichetti, T.; Mauriello, M.; Miglietta, M. L.;
(33) Sui, Y.; Appenzeller, J. Screening and Interlayer Coupling in
Ricciardella, F.; Massera, E.; Di Francia, G. Modulating the Sensing
Multilayer Graphene Field-Effect Transistors. Nano Lett. 2009, 9, Properties of Graphene through an Eco-Friendly Metal-Decoration
2973−2977. Process. Sens. Actuators, B 2016, 222, 1032−1042.
(34) Wang, L.; Meric, I.; Huang, P. Y.; Gao, Q.; Gao, Y.; Tran, H.; (53) Chung, M. G.; Kim, D.-H.; Seo, D. K.; Kim, T.; Im, H. U.; Lee,
Taniguchi, T.; Watanabe, K.; Campos, L. M.; Muller, D. A.; Guo, J.; H. M.; Yoo, J.-B.; Hong, S.-H.; Kang, T. J.; Kim, Y. H. Flexible
Kim, P.; Hone, J.; Shepard, K. L.; Dean, C. R. One-Dimensional Hydrogen Sensors Using Graphene with Palladium Nanoparticle
Electrical Contact to a Two-Dimensional Material. Science 2013, 342, Decoration. Sens. Actuators, B 2012, 169, 387−392.
614−617. (54) Pak, Y.; Kim, S.-M.; Jeong, H.; Kang, C. G.; Park, J. S.; Song, H.;
(35) Smith, J. T.; Franklin, A. D.; Farmer, D. B.; Dimitrakopoulos, C. Lee, R.; Myoung, N.; Lee, B. H.; Seo, S.; Kim, J. T.; Jung, G.-Y.
D. Reducing Contact Resistance in Graphene Devices through Palladium-Decorated Hydrogen-Gas Sensors Using Periodically
Contact Area Patterning. ACS Nano 2013, 7, 3661−3667. Aligned Graphene Nanoribbons. ACS Appl. Mater. Interfaces 2014, 6,
(36) Leong, W. S.; Gong, H.; Thong, J. T. L. Low-Contact-Resistance 13293−13298.
Graphene Devices with Nickel-Etched-Graphene Contacts. ACS Nano (55) Tammanoon, N.; Wisitsoraat, A.; Sriprachuabwong, C.;
2014, 8, 994−1001. Phokharatkul, D.; Tuantranont, A.; Phanichphant, S.; Liewhiran, C.
(37) Min Song, S.; Yong Kim, T.; Jae Sul, O.; Cheol Shin, W.; Jin Ultrasensitive NO2 Sensor Based on Ohmic Metal−Semiconductor
Cho, B. Improvement of Graphene−Metal Contact Resistance by Interfaces of Electrolytically Exfoliated Graphene/Flame-Spray-Made
Introducing Edge Contacts at Graphene under Metal. Appl. Phys. Lett. SnO2 Nanoparticles Composite Operating at Low Temperatures. ACS
2014, 104, 183506. Appl. Mater. Interfaces 2015, 7, 24338−24352.
(38) Ito, K.; Ogata, T.; Sakai, T.; Awano, Y. Ultralow Contact (56) Chu, B. H.; Nicolosi, J.; Lo, C. F.; Strupinski, W.; Pearton, S. J.;
Resistivity in Annealed Titanium Edge Contacts for Multilayered Ren, F. Effect of Coated Platinum Thickness on Hydrogen Detection
Graphene. Appl. Phys. Express 2015, 8, 025101. Sensitivity of Graphene-Based Sensors. Electrochem. Solid-State Lett.
(39) Xia, F.; Mueller, T.; Golizadeh-Mojarad, R.; Freitag, M.; Lin, Y.- 2011, 14, K43−K45.
m.; Tsang, J.; Perebeinos, V.; Avouris, P. Photocurrent Imaging and (57) Lu, G.; Ocola, E. L.; Chen, J. Reduced Graphene Oxide for
Efficient Photon Detection in a Graphene Transistor. Nano Lett. 2009, Room-Temperature Gas Sensors. Nanotechnology 2009, 20, 445502.
9, 1039−1044. (58) Liu, H.; Si, M.; Deng, Y.; Neal, A. T.; Du, Y.; Najmaei, S.;
(40) Ghosh, S.; Sarker, B. K.; Chunder, A.; Zhai, L.; Khondaker, S. I. Ajayan, P. M.; Lou, J.; Ye, P. D. Switching Mechanism in Single-Layer
Position Dependent Photodetector from Large Area Reduced Molybdenum Disulfide Transistors: An Insight into Current Flow
Graphene Oxide Thin Films. Appl. Phys. Lett. 2010, 96, 163109. across Schottky Barriers. ACS Nano 2014, 8, 1031−1038.

4915 DOI: 10.1021/acsnano.6b01842

ACS Nano 2016, 10, 4895−4919
ACS Nano Review

(59) Kang, J.; Liu, W.; Sarkar, D.; Jena, D.; Banerjee, K. (79) Zhang, W.; Huang, J.-K.; Chen, C.-H.; Chang, Y.-H.; Cheng, Y.-
Computational Study of Metal Contacts to Monolayer Transition- J.; Li, L.-J. High-Gain Phototransistors Based on a CVD MoS2
Metal Dichalcogenide Semiconductors. Phys. Rev. X 2014, 4, 031005. Monolayer. Adv. Mater. 2013, 25, 3456−3461.
(60) Chen, J.-R.; Odenthal, P. M.; Swartz, A. G.; Floyd, G. C.; Wen, (80) Liu, W.; Kang, J.; Sarkar, D.; Khatami, Y.; Jena, D.; Banerjee, K.
H.; Luo, K. Y.; Kawakami, R. K. Control of Schottky Barriers in Single Role of Metal Contacts in Designing High-Performance Monolayer N-
Layer MoS2 Transistors with Ferromagnetic Contacts. Nano Lett. Type WSe2 Field Effect Transistors. Nano Lett. 2013, 13, 1983−1990.
2013, 13, 3106−3110. (81) Britnell, L.; Ribeiro, R. M.; Eckmann, A.; Jalil, R.; Belle, B. D.;
(61) Dankert, A.; Langouche, L.; Kamalakar, M. V.; Dash, S. P. High- Mishchenko, A.; Kim, Y. J.; Gorbachev, R. V.; Georgiou, T.; Morozov,
Performance Molybdenum Disulfide Field-Effect Transistors with Spin S. V.; Grigorenko, A. N.; Geim, A. K.; Casiraghi, C.; Neto, A. H. C.;
Tunnel Contacts. ACS Nano 2014, 8, 476−482. Novoselov, K. S. Strong Light-Matter Interactions in Heterostructures
(62) Leong, W. S.; Luo, X.; Li, Y.; Khoo, K. H.; Quek, S. Y.; Thong, J. of Atomically Thin Films. Science 2013, 340, 1311−1314.
T. L. Low Resistance Metal Contacts to MoS2 Devices with Nickel- (82) Eda, G.; Maier, S. A. Two-Dimensional Crystals: Managing
Etched-Graphene Electrodes. ACS Nano 2015, 9, 869−877. Light for Optoelectronics. ACS Nano 2013, 7, 5660−5665.
(63) Gong, C.; Huang, C.; Miller, J.; Cheng, L.; Hao, Y.; Cobden, D.; (83) Kern, J.; Trügler, A.; Niehues, I.; Ewering, J.; Schmidt, R.;
Kim, J.; Ruoff, R. S.; Wallace, R. M.; Cho, K.; Xu, X.; Chabal, Y. J. Schneider, R.; Najmaei, S.; George, A.; Zhang, J.; Lou, J.; Hohenester,
Metal Contacts on Physical Vapor Deposited Monolayer MoS2. ACS U.; Michaelis de Vasconcellos, S.; Bratschitsch, R. Nanoantenna-
Nano 2013, 7, 11350−11357. Enhanced Light−Matter Interaction in Atomically Thin WS2. ACS
(64) Sun, Y.; Liu, K.; Hong, X.; Chen, M.; Kim, J.; Shi, S.; Wu, J.; Photonics 2015, 2, 1260−1265.
Zettl, A.; Wang, F. Probing Local Strain at MX2−Metal Boundaries (84) Butun, S.; Tongay, S.; Aydin, K. Enhanced Light Emission from
with Surface Plasmon-Enhanced Raman Scattering. Nano Lett. 2014, Large-Area Monolayer MoS2 Using Plasmonic Nanodisc Arrays. Nano
14, 5329−5334. Lett. 2015, 15, 2700−2704.
(65) McDonnell, S.; Addou, R.; Buie, C.; Wallace, R. M.; Hinkle, C. (85) Kannan, P. K.; Late, D. J.; Morgan, H.; Rout, C. S. Recent
L. Defect-Dominated Doping and Contact Resistance in MoS2. ACS Developments in 2D Layered Inorganic Nanomaterials for Sensing.
Nano 2014, 8, 2880−2888. Nanoscale 2015, 7, 13293−13312.
(66) Chuang, S.; Battaglia, C.; Azcatl, A.; McDonnell, S.; Kang, J. S.; (86) Late, D. J.; Doneux, T.; Bougouma, M. Single-Layer MoSe2
Yin, X.; Tosun, M.; Kapadia, R.; Fang, H.; Wallace, R. M.; Javey, A. Based NH3 Gas Sensor. Appl. Phys. Lett. 2014, 105, 233103.
MoS2 P-Type Transistors and Diodes Enabled by High Work (87) Mayorga-Martinez, C. C.; Ambrosi, A.; Eng, A. Y. S.; Sofer, Z.;
Function MoOx Contacts. Nano Lett. 2014, 14, 1337−1342. Pumera, M. Metallic 1T-WS2 for Selective Impedimetric Vapor
(67) Fang, H.; Chuang, S.; Chang, T. C.; Takei, K.; Takahashi, T.; Sensing. Adv. Funct. Mater. 2015, 25, 5611−5616.
Javey, A. High-Performance Single Layered WSe2 P-FETs with (88) Ou, J. Z.; Ge, W.; Carey, B.; Daeneke, T.; Rotbart, A.; Shan, W.;
Chemically Doped Contacts. Nano Lett. 2012, 12, 3788−3792. Wang, Y.; Fu, Z.; Chrimes, A. F.; Wlodarski, W.; Russo, S. P.; Li, Y. X.;
(68) Du, Y.; Yang, L.; Liu, H.; Ye, P. D. Contact Research Strategy Kalantar-zadeh, K. Physisorption-Based Charge Transfer in Two-
for Emerging Molybdenum Disulfide and Other Two-Dimensional Dimensional SnS2 for Selective and Reversible NO2 Gas Sensing. ACS
Field-Effect Transistors. APL Mater. 2014, 2, 092510. Nano 2015, 9, 10313−10323.
(69) Du, Y.; Liu, H.; Deng, Y.; Ye, P. D. Device Perspective for Black (89) Franke, M. E.; Koplin, T. J.; Simon, U. Metal and Metal Oxide
Phosphorus Field-Effect Transistors: Contact Resistance, Ambipolar Nanoparticles in Chemiresistors: Does the Nanoscale Matter? Small
Behavior, and Scaling. ACS Nano 2014, 8, 10035−10042. 2006, 2, 36−50.
(70) Das, S.; Demarteau, M.; Roelofs, A. Ambipolar Phosphorene (90) Fang, H.; Tosun, M.; Seol, G.; Chang, T. C.; Takei, K.; Guo, J.;
Field Effect Transistor. ACS Nano 2014, 8, 11730−11738. Javey, A. Degenerate N-Doping of Few-Layer Transition Metal
(71) Haratipour, N.; Robbins, M. C.; Koester, S. J. Black Phosphorus Dichalcogenides by Potassium. Nano Lett. 2013, 13, 1991−1995.
p-MOSFETs with 7-nm HfO2 Gate Dielectric and Low Contact (91) Liu, B.; Chen, L.; Liu, G.; Abbas, A. N.; Fathi, M.; Zhou, C.
Resistance. IEEE Electron Device Lett. 2015, 36, 411−413. High-Performance Chemical Sensing Using Schottky-Contacted
(72) Anugrah, Y.; Robbins, M. C.; Crowell, P. A.; Koester, S. J. Chemical Vapor Deposition Grown Monolayer MoS2 Transistors.
Determination of the Schottky Barrier Height of Ferromagnetic ACS Nano 2014, 8, 5304−5314.
Contacts to Few-Layer Phosphorene. Appl. Phys. Lett. 2015, 106, (92) Abbas, A. N.; Liu, B.; Chen, L.; Ma, Y.; Cong, S.; Aroonyadet,
103108. N.; Köpf, M.; Nilges, T.; Zhou, C. Black Phosphorus Gas Sensors. ACS
(73) Tsai, D.-S.; Liu, K.-K.; Lien, D.-H.; Tsai, M.-L.; Kang, C.-F.; Lin, Nano 2015, 9, 5618−5624.
C.-A.; Li, L.-J.; He, J.-H. Few-Layer MoS2 with High Broadband (93) Alsaif, M. M. Y. A.; Balendhran, S.; Field, M. R.; Latham, K.;
Photogain and Fast Optical Switching for Use in Harsh Environments. Wlodarski, W.; Ou, J. Z.; Kalantar-zadeh, K. Two Dimensional α-
ACS Nano 2013, 7, 3905−3911. MoO3 Nanoflakes Obtained Using Solvent-Assisted Grinding and
(74) Tsai, D. S.; Lien, D. H.; Tsai, M. L.; Su, S. H.; Chen, K. M.; Ke, Sonication Method: Application for H2 Gas Sensing. Sens. Actuators, B
J. J.; Yu, Y. C.; Li, L. J.; He, J. H. Trilayered MoS2 Metal 2014, 192, 196−204.
Semiconductor Metal Photodetectors: Photogain and Radiation (94) Shanmugam, M.; Durcan, C. A.; Yu, B. Layered Semiconductor
Resistance. IEEE J. Sel. Top. Quantum Electron. 2014, 20, 30−35. Molybdenum Disulfide Nanomembrane Based Schottky-Barrier Solar
(75) Cai, X.; Sushkov, A. B.; Suess, R. J.; Jadidi, M. M.; Jenkins, G. S.; Cells. Nanoscale 2012, 4, 7399−7405.
Nyakiti, L. O.; Myers-Ward, R. L.; Li, S.; Yan, J.; Gaskill, D. K.; (95) Wi, S.; Kim, H.; Chen, M.; Nam, H.; Guo, L. J.; Meyhofer, E.;
Murphy, T. E.; Drew, H. D.; Fuhrer, M. S. Sensitive Room- Liang, X. Enhancement of Photovoltaic Response in Multilayer MoS2
Temperature Terahertz Detection Via the Photothermoelectric Effect Induced by Plasma Doping. ACS Nano 2014, 8, 5270−5281.
in Graphene. Nat. Nanotechnol. 2014, 9, 814−819. (96) Buscema, M.; Groenendijk, D. J.; Steele, G. A.; van der Zant, H.
(76) Zhang, Y.; Li, H.; Wang, L.; Wang, H.; Xie, X.; Zhang, S.-L.; Liu, S. J.; Castellanos-Gomez, A. Photovoltaic Effect in Few-Layer Black
R.; Qiu, Z.-J. Photothermoelectric and Photovoltaic Effects Both Phosphorus PN Junctions Defined by Local Electrostatic Gating. Nat.
Present in MoS2. Sci. Rep. 2015, 5, 7938. Commun. 2014, 5, 4651.
(77) Cao, Y.; Cai, K.; Hu, P.; Zhao, L.; Yan, T.; Luo, W.; Zhang, X.; (97) Baugher, B. W. H.; Churchill, H. O. H.; Yang, Y.; Jarillo-
Wu, X.; Wang, K.; Zheng, H. Strong Enhancement of Photo- Herrero, P. Optoelectronic Devices Based on Electrically Tunable P-N
responsivity with Shrinking the Electrodes Spacing in Few Layer Gase Diodes in a Monolayer Dichalcogenide. Nat. Nanotechnol. 2014, 9,
Photodetectors. Sci. Rep. 2015, 5, 8130. 262−267.
(78) Zhang, W.; Chiu, M.-H.; Chen, C.-H.; Chen, W.; Li, L.-J.; Wee, (98) Tongay, S.; Schumann, T.; Hebard, A. F. Graphite Based
A. T. S. Role of Metal Contacts in High-Performance Phototransistors Schottky Diodes Formed on Si, GaAs, and 4H-SiC Substrates. Appl.
Based on WSe2 Monolayers. ACS Nano 2014, 8, 8653−8661. Phys. Lett. 2009, 95, 222103.

4916 DOI: 10.1021/acsnano.6b01842

ACS Nano 2016, 10, 4895−4919
ACS Nano Review

(99) Parui, S.; Ruiter, R.; Zomer, P. J.; Wojtaszek, M.; van Wees, B. J.; Relationship of Highly Ordered Organic Semiconductors at the Two-
Banerjee, T. Temperature Dependent Transport Characteristics of Dimensional Limit. Phys. Rev. Lett. 2016, 116, 016602.
Graphene/N-Si Diodes. J. Appl. Phys. 2014, 116, 244505. (118) He, D.; Zhang, Y.; Wu, Q.; Xu, R.; Nan, H.; Liu, J.; Yao, J.;
(100) Yim, C.; McEvoy, N.; Duesberg, G. S. Characterization of Wang, Z.; Yuan, S.; Li, Y.; Shi, Y.; Wang, J.; Ni, Z.; He, L.; Miao, F.;
Graphene-Silicon Schottky Barrier Diodes Using Impedance Spec- Song, F.; Xu, H.; Watanabe, K.; Taniguchi, T.; Xu, J.-B.; et al. Two-
troscopy. Appl. Phys. Lett. 2013, 103, 193106. Dimensional Quasi-Freestanding Molecular Crystals for High-
(101) Fattah, A.; Khatami, S.; Mayorga-Martinez, C. C.; Medina- Performance Organic Field-Effect Transistors. Nat. Commun. 2014,
Sánchez, M.; Baptista-Pires, L.; Merkoçi, A. Graphene/Silicon 5, 5162.
Heterojunction Schottky Diode for Vapors Sensing Using Impedance (119) Giannazzo, F.; Roccaforte, F.; Iucolano, F.; Raineri, V.; Ruffino,
Spectroscopy. Small 2014, 10, 4193−4199. F.; Grimaldi, M. G. Nanoscale Current Transport through Schottky
(102) Chen, C.-C.; Aykol, M.; Chang, C.-C.; Levi, A. F. J.; Cronin, S. Contacts on Wide Bandgap Semiconductors. J. Vac. Sci. Technol. B
B. Graphene-Silicon Schottky Diodes. Nano Lett. 2011, 11, 1863− 2009, 27, 789−794.
1867. (120) Fisichella, G.; Greco, G.; Roccaforte, F.; Giannazzo, F. From
(103) Tongay, S.; Lemaitre, M.; Schumann, T.; Berke, K.; Appleton, Schottky to Ohmic Graphene Contacts to AlGaN/GaN Hetero-
B. R.; Gila, B.; Hebard, A. F. Graphene/GaN Schottky Diodes: structures: Role of the AlGaN Layer Microstructure. Appl. Phys. Lett.
Stability at Elevated Temperatures. Appl. Phys. Lett. 2011, 99, 102102. 2014, 105, 063117.
(104) Fisichella, G.; Greco, G.; Roccaforte, F.; Giannazzo, F. Current (121) Zhong, H.; Xu, K.; Liu, Z.; Xu, G.; Shi, L.; Fan, Y.; Wang, J.;
Transport in Graphene/AlGaN/GaN Vertical Heterostructures Ren, G.; Yang, H. Charge Transport Mechanisms of Graphene/
Probed at Nanoscale. Nanoscale 2014, 6, 8671−8680. Semiconductor Schottky Barriers: A Theoretical and Experimental
(105) Tadjer, M. J.; Anderson, T. J.; Hobart, K. D.; Nyakiti, L. O.; Study. J. Appl. Phys. 2014, 115, 013701.
Wheeler, V. D.; Myers-Ward, R. L.; Gaskill, D. K.; Eddy, C. R.; Kub, F. (122) Shivaraman, S.; Herman, L. H.; Rana, F.; Park, J.; Spencer, M.
J.; Calle, F. Vertical Conduction Mechanism of the Epitaxial G. Schottky Barrier Inhomogeneities at the Interface of Few Layer
Graphene/N-Type 4H-SiC Heterojunction at Cryogenic Temper- Epitaxial Graphene and Silicon Carbide. Appl. Phys. Lett. 2012, 100,
atures. Appl. Phys. Lett. 2012, 100, 193506. 183112.
(106) Dharmaraj, P.; Jeganathan, K.; Parthiban, S.; Kwon, J. Y.; (123) Werner, J. H.; Güttler, H. H. Barrier Inhomogeneities at
Gautam, S.; Chae, K. H.; Asokan, K. Selective Area Growth of Bernal Schottky Contacts. J. Appl. Phys. 1991, 69, 1522−1533.
Bilayer Epitaxial Graphene on 4H-SiC (0001) Substrate by Electron- (124) Tung, R. T. Electron Transport at Metal-Semiconductor
Beam Irradiation. Appl. Phys. Lett. 2014, 105, 181601. Interfaces: General Theory. Phys. Rev. B: Condens. Matter Mater. Phys.
(107) Tomer, D.; Rajput, S.; Hudy, L. J.; Li, C. H.; Li, L. Intrinsic 1992, 45, 13509−13523.
Inhomogeneity in Barrier Height at Monolayer Graphene/SiC (125) An, Y.; Behnam, A.; Pop, E.; Ural, A. Metal-Semiconductor-
Metal Photodetectors Based on Graphene/P-Type Silicon Schottky
Schottky Junction. Appl. Phys. Lett. 2014, 105, 021607.
(108) Sonde, S.; Giannazzo, F.; Raineri, V.; Yakimova, R.; Junctions. Appl. Phys. Lett. 2013, 102, 013110.
(126) Sinha, D.; Lee, J. U. Ideal Graphene/Silicon Schottky Junction
Huntzinger, J. R.; Tiberj, A.; Camassel, J. Electrical Properties of the
Diodes. Nano Lett. 2014, 14, 4660−4664.
Graphene/4H-SiC (0001) Interface Probed by Scanning Current
(127) Zeng, L.-H.; Wang, M.-Z.; Hu, H.; Nie, B.; Yu, Y.-Q.; Wu, C.-
Spectroscopy. Phys. Rev. B: Condens. Matter Mater. Phys. 2009, 80,
Y.; Wang, L.; Hu, J.-G.; Xie, C.; Liang, F.-X.; Luo, L.-B. Monolayer
241406.
Graphene/Germanium Schottky Junction as High-Performance Self-
(109) Lee, S.; Lee, Y.; Young Kim, D.; Song, E. B.; Min Kim, S. Back-
Driven Infrared Light Photodetector. ACS Appl. Mater. Interfaces 2013,
Gate Tuning of Schottky Barrier Height in Graphene/Zinc-Oxide
5, 9362−9366.
Photodiodes. Appl. Phys. Lett. 2013, 102, 242114. (128) Amirmazlaghani, M.; Raissi, F.; Habibpour, O.; Vukusic, J.;
(110) Tongay, S.; Lemaitre, M.; Miao, X.; Gila, B.; Appleton, B. R.; Stake, J. Graphene-Si Schottky IR Detector. IEEE J. Quantum Electron.
Hebard, A. F. Rectification at Graphene-Semiconductor Interfaces: 2013, 49, 589−594.
Zero-Gap Semiconductor-Based Diodes. Phys. Rev. X 2012, 2, 011002. (129) Lv, P.; Zhang, X.; Zhang, X.; Deng, W.; Jie, J. High-Sensitivity
(111) Yang, H.; Heo, J.; Park, S.; Song, H. J.; Seo, D. H.; Byun, K.-E.; and Fast-Response Graphene/Crystalline Silicon Schottky Junction-
Kim, P.; Yoo, I.; Chung, H.-J.; Kim, K. Graphene Barristor, a Triode Based Near-IR Photodetectors. IEEE Electron Device Lett. 2013, 34,
Device with a Gate-Controlled Schottky Barrier. Science 2012, 336, 1337−1339.
1140−1143. (130) An, X.; Liu, F.; Jung, Y. J.; Kar, S. Tunable Graphene−Silicon
(112) Chen, C.-C.; Chang, C.-C.; Li, Z.; Levi, A. F. J.; Cronin, S. B. Heterojunctions for Ultrasensitive Photodetection. Nano Lett. 2013,
Gate Tunable Graphene-Silicon Ohmic/Schottky Contacts. Appl. Phys. 13, 909−916.
Lett. 2012, 101, 223113. (131) Luo, L.-B.; Chen, J.-J.; Wang, M.-Z.; Hu, H.; Wu, C.-Y.; Li, Q.;
(113) Cavallo, F.; Rojas Delgado, R.; Kelly, M. M.; Sánchez Pérez, J. Wang, L.; Huang, J.-A.; Liang, F.-X. Near-Infrared Light Photovoltaic
R.; Schroeder, D. P.; Xing, H. G.; Eriksson, M. A.; Lagally, M. G. Detector Based on GaAs Nanocone Array/Monolayer Graphene
Exceptional Charge Transport Properties of Graphene on Germanium. Schottky Junction. Adv. Funct. Mater. 2014, 24, 2794−2800.
ACS Nano 2014, 8, 10237−10245. (132) Liu, H.; Sun, Q.; Xing, J.; Zheng, Z.; Zhang, Z.; Lü, Z.; Zhao,
(114) Baek, S.-h. C.; Seo, Y.-J.; Oh, J. G.; Albert Park, M. G.; Bong, J. K. Fast and Enhanced Broadband Photoresponse of a ZnO Nanowire
H.; Yoon, S. J.; Seo, M.; Park, S.-y.; Park, B.-G.; Lee, S.-H. Alleviation Array/Reduced Graphene Oxide Film Hybrid Photodetector from the
of Fermi-Level Pinning Effect at Metal/Germanium Interface by the Visible to the Near-Infrared Range. ACS Appl. Mater. Interfaces 2015,
Insertion of Graphene Layers. Appl. Phys. Lett. 2014, 105, 073508. 7, 6645−6651.
(115) Shiu, H. W.; Chang, L. Y.; Lee, K.-H.; Chen, H.-Y.; Gwo, S.; (133) Zhang, Z.; Guo, Y.; Wang, X.; Li, D.; Wang, F.; Xie, S. Direct
Chen, C.-H. Graphene as Tunable Transparent Electrode Material on Growth of Nanocrystalline Graphene/Graphite Transparent Electro-
GaN: Layer-Number-Dependent Optical and Electrical Properties. des on Si/SiO2 for Metal-Free Schottky Junction Photodetectors. Adv.
Appl. Phys. Lett. 2013, 103, 081604. Funct. Mater. 2014, 24, 835−840.
(116) Liu, X.; Luo, X.; Nan, H.; Guo, H.; Wang, P.; Zhang, L.; Zhou, (134) Zhu, M.; Li, X.; Guo, Y.; Li, X.; Sun, P.; Zang, X.; Wang, K.;
M.; Yang, Z.; Shi, Y.; Hu, W.; Ni, Z.; Qiu, T.; Yu, Z.; Xu, J.; Wang, X. Zhong, M.; Wu, D.; Zhu, H. Vertical Junction Photodetectors Based
Epitaxial Ultrathin Organic Crystals on Graphene for High-Efficiency on Reduced Graphene Oxide/Silicon Schottky Diodes. Nanoscale
Phototransistors. Adv. Mater. 2016, in press. 2014, 6, 4909−4914.
(117) Zhang, Y.; Qiao, J.; Gao, S.; Hu, F.; He, D.; Wu, B.; Yang, Z.; (135) Kim, J.; Joo, S. S.; Lee, K. W.; Kim, J. H.; Shin, D. H.; Kim, S.;
Xu, B.; Li, Y.; Shi, Y.; Ji, W.; Wang, P.; Wang, X.; Xiao, M.; Xu, H.; Xu, Choi, S.-H. Near-Ultraviolet-Sensitive Graphene/Porous Silicon
J.-B.; Wang, X. Probing Carrier Transport and Structure-Property Photodetectors. ACS Appl. Mater. Interfaces 2014, 6, 20880−20886.

4917 DOI: 10.1021/acsnano.6b01842

ACS Nano 2016, 10, 4895−4919
ACS Nano Review

(136) Nie, B.; Hu, J.-G.; Luo, L.-B.; Xie, C.; Zeng, L.-H.; Lv, P.; Li, Based Hole Conductor-Free Perovskite Solar Cells: Schottky Junction
F.-Z.; Jie, J.-S.; Feng, M.; Wu, C.-Y.; Yu, Y.-Q.; Yu, S.-H. Monolayer Enhanced Hole Extraction and Electron Blocking. Small 2015, 11,
Graphene Film on ZnO Nanorod Array for High-Performance 2269−2274.
Schottky Junction Ultraviolet Photodetectors. Small 2013, 9, 2872− (155) Lee, E. W.; Ma, L.; Nath, D. N.; Lee, C. H.; Arehart, A.; Wu,
2879. Y.; Rajan, S. Growth and Electrical Characterization of Two-
(137) Liu, X.; Ji, X.; Liu, M.; Liu, N.; Tao, Z.; Dai, Q.; Wei, L.; Li, C.; Dimensional Layered MoS2/SiC Heterojunctions. Appl. Phys. Lett.
Zhang, X.; Wang, B. High-Performance Ge Quantum Dot Decorated 2014, 105, 203504.
Graphene/Zinc-Oxide Heterostructure Infrared Photodetector. ACS (156) Chuang, S.; Kapadia, R.; Fang, H.; Chia Chang, T.; Yen, W.-C.;
Appl. Mater. Interfaces 2015, 7, 2452−2458. Chueh, Y.-L.; Javey, A. Near-Ideal Electrical Properties of InAs/WSe2
(138) Kim, H.-Y.; Lee, K.; McEvoy, N.; Yim, C.; Duesberg, G. S. Van Der Waals Heterojunction Diodes. Appl. Phys. Lett. 2013, 102,
Chemically Modulated Graphene Diodes. Nano Lett. 2013, 13, 2182− 242101.
2188. (157) Yim, C.; O’Brien, M.; McEvoy, N.; Riazimehr, S.; Schäfer-
(139) Uddin, M. A.; Singh, A. K.; Sudarshan, S. T.; Koley, G. Eberwein, H.; Bablich, A.; Pawar, R.; Iannaccone, G.; Downing, C.;
Functionalized Graphene/Silicon Chemi-Diode H2 Sensor with Fiori, G.; Lemme, M. C.; Duesberg, G. S. Heterojunction Hybrid
Tunable Sensitivity. Nanotechnology 2014, 25, 125501.
Devices from Vapor Phase Grown MoS2. Sci. Rep. 2014, 4, 5458.
(140) Singh, A.; Uddin, M. A.; Sudarshan, T.; Koley, G. Tunable
(158) Esmaeili-Rad, M. R.; Salahuddin, S. High Performance
Reverse-Biased Graphene/Silicon Heterojunction Schottky Diode
Molybdenum Disulfide Amorphous Silicon Heterojunction Photo-
Sensor. Small 2014, 10, 1555−1565.
detector. Sci. Rep. 2013, 3, 2345.
(141) Kim, J.; Oh, S. D.; Kim, J. H.; Shin, D. H.; Kim, S.; Choi, S.-H.
(159) Li, Y.; Xu, C.-Y.; Wang, J.-Y.; Zhen, L. Photodiode-Like
Graphene/Si-Nanowire Heterostructure Molecular Sensors. Sci. Rep.
2014, 4, 5384. Behavior and Excellent Photoresponse of Vertical Si/Monolayer MoS2
(142) Zhu, M.; Li, X.; Chung, S.; Zhao, L.; Li, X.; Zang, X.; Wang, K.; Heterostructures. Sci. Rep. 2014, 4, 7186.
Wei, J.; Zhong, M.; Zhou, K.; Xie, D.; Zhu, H. Photo-Induced Selective (160) Wang, L.; Jie, J.; Shao, Z.; Zhang, Q.; Zhang, X.; Wang, Y.; Sun,
Gas Detection Based on Reduced Graphene Oxide/Si Schottky Diode. Z.; Lee, S.-T. MoS2/Si Heterojunction with Vertically Standing
Carbon 2015, 84, 138−145. Layered Structure for Ultrafast, High-Detectivity, Self-Driven Visible−
(143) Li, X.; Zhu, H.; Wang, K.; Cao, A.; Wei, J.; Li, C.; Jia, Y.; Li, Z.; Near Infrared Photodetectors. Adv. Funct. Mater. 2015, 25, 2910−
Li, X.; Wu, D. Graphene-on-Silicon Schottky Junction Solar Cells. Adv. 2919.
Mater. 2010, 22, 2743−2748. (161) Lopez-Sanchez, O.; Alarcon Llado, E.; Koman, V.; Fontcuberta
(144) Fan, G.; Zhu, H.; Wang, K.; Wei, J.; Li, X.; Shu, Q.; Guo, N.; i Morral, A.; Radenovic, A.; Kis, A. Light Generation and Harvesting in
Wu, D. Graphene/Silicon Nanowire Schottky Junction for Enhanced a Van Der Waals Heterostructure. ACS Nano 2014, 8, 3042−3048.
Light Harvesting. ACS Appl. Mater. Interfaces 2011, 3, 721−725. (162) Tsai, M.-L.; Su, S.-H.; Chang, J.-K.; Tsai, D.-S.; Chen, C.-H.;
(145) Xie, C.; Lv, P.; Nie, B.; Jie, J.; Zhang, X.; Wang, Z.; Jiang, P.; Wu, C.-I.; Li, L.-J.; Chen, L.-J.; He, J.-H. Monolayer MoS 2
Hu, Z.; Luo, L.; Zhu, Z.; Wang, L.; Wu, C. Monolayer Graphene Film/ Heterojunction Solar Cells. ACS Nano 2014, 8, 8317−8322.
Silicon Nanowire Array Schottky Junction Solar Cells. Appl. Phys. Lett. (163) Hao, L.; Liu, Y.; Gao, W.; Han, Z.; Xue, Q.; Zeng, H.; Wu, Z.;
2011, 99, 133113. Zhu, J.; Zhang, W. Electrical and Photovoltaic Characteristics of
(146) Feng, T.; Xie, D.; Lin, Y.; Zhao, H.; Chen, Y.; Tian, H.; Ren, MoS2/Si P-N Junctions. J. Appl. Phys. 2015, 117, 114502.
T.; Li, X.; Li, Z.; Wang, K.; Wu, D.; Zhu, H. Efficiency Enhancement (164) Tao, L.; Cinquanta, E.; Chiappe, D.; Grazianetti, C.; Fanciulli,
of Graphene/Silicon-Pillar-Array Solar Cells by HNO3 and PEDOT- M.; Dubey, M.; Molle, A.; Akinwande, D. Silicene Field-Effect
PSS. Nanoscale 2012, 4, 2130−2133. Transistors Operating at Room Temperature. Nat. Nanotechnol. 2015,
(147) Miao, X.; Tongay, S.; Petterson, M. K.; Berke, K.; Rinzler, A. 10, 227−231.
G.; Appleton, B. R.; Hebard, A. F. High Efficiency Graphene Solar (165) Derivaz, M.; Dentel, D.; Stephan, R.; Hanf, M.-C.; Mehdaoui,
Cells by Chemical Doping. Nano Lett. 2012, 12, 2745−2750. A.; Sonnet, P.; Pirri, C. Continuous Germanene Layer on Al(111).
(148) Cui, T.; Lv, R.; Huang, Z.-H.; Chen, S.; Zhang, Z.; Gan, X.; Jia, Nano Lett. 2015, 15, 2510−2516.
Y.; Li, X.; Wang, K.; Wu, D.; Kang, F. Enhanced Efficiency of (166) Lin, Z.; He, Q.; Yin, A.; Xu, Y.; Wang, C.; Ding, M.; Cheng,
Graphene/Silicon Heterojunction Solar Cells by Molecular Doping. J. H.-C.; Papandrea, B.; Huang, Y.; Duan, X. Cosolvent Approach for
Mater. Chem. A 2013, 1, 5736−5740. Solution-Processable Electronic Thin Films. ACS Nano 2015, 9,
(149) Xie, C.; Zhang, X.; Wu, Y.; Zhang, X.; Zhang, X.; Wang, Y.; 4398−4405.
Zhang, W.; Gao, P.; Han, Y.; Jie, J. Surface Passivation and Band (167) Torrisi, F.; Hasan, T.; Wu, W.; Sun, Z.; Lombardo, A.;
Engineering: A Way toward High Efficiency Graphene-Planar Si Solar Kulmala, T. S.; Hsieh, G.-W.; Jung, S.; Bonaccorso, F.; Paul, P. J.; Chu,
Cells. J. Mater. Chem. A 2013, 1, 8567−8574.
D.; Ferrari, A. C. Inkjet-Printed Graphene Electronics. ACS Nano
(150) Li, X.; Xie, D.; Park, H.; Zhu, M.; Zeng, T. H.; Wang, K.; Wei,
2012, 6, 2992−3006.
J.; Wu, D.; Kong, J.; Zhu, H. Ion Doping of Graphene for High-
(168) Secor, E. B.; Lim, S.; Zhang, H.; Frisbie, C. D.; Francis, L. F.;
Efficiency Heterojunction Solar Cells. Nanoscale 2013, 5, 1945−1948.
(151) Ho, P.-H.; Liou, Y.-T.; Chuang, C.-H.; Lin, S.-W.; Tseng, C.-Y.; Hersam, M. C. Gravure Printing of Graphene for Large-Area Flexible
Wang, D.-Y.; Chen, C.-C.; Hung, W.-Y.; Wen, C.-Y.; Chen, C.-W. Self- Electronics. Adv. Mater. 2014, 26, 4533−4538.
Crack-Filled Graphene Films by Metallic Nanoparticles for High- (169) Hurch, S.; Nolan, H.; Hallam, T.; Berner, N. C.; McEvoy, N.;
Performance Graphene Heterojunction Solar Cells. Adv. Mater. 2015, Duesberg, G. S. Inkjet-Defined Field-Effect Transistors from Chemical
27, 1724−1729. Vapour Deposited Graphene. Carbon 2014, 71, 332−337.
(152) Song, Y.; Li, X.; Mackin, C.; Zhang, X.; Fang, W.; Palacios, T.; (170) Withers, F.; Yang, H.; Britnell, L.; Rooney, A. P.; Lewis, E.;
Zhu, H.; Kong, J. Role of Interfacial Oxide in High-Efficiency Felten, A.; Woods, C. R.; Sanchez Romaguera, V.; Georgiou, T.;
Graphene−Silicon Schottky Barrier Solar Cells. Nano Lett. 2015, 15, Eckmann, A.; Kim, Y. J.; Yeates, S. G.; Haigh, S. J.; Geim, A. K.;
2104−2110. Novoselov, K. S.; Casiraghi, C. Heterostructures Produced from
(153) Vazquez-Mena, O.; Bosco, J. P.; Ergen, O.; Rasool, H. I.; Nanosheet-Based Inks. Nano Lett. 2014, 14, 3987−3992.
Fathalizadeh, A.; Tosun, M.; Crommie, M.; Javey, A.; Atwater, H. A.; (171) Torrisi, F.; Coleman, J. N. Electrifying Inks with 2D Materials.
Zettl, A. Performance Enhancement of a Graphene-Zinc Phosphide Nat. Nanotechnol. 2014, 9, 738−739.
Solar Cell Using the Electric Field-Effect. Nano Lett. 2014, 14, 4280− (172) Liu, Y.; Wu, H.; Cheng, H.-C.; Yang, S.; Zhu, E.; He, Q.; Ding,
4285. M.; Li, D.; Guo, J.; Weiss, N. O.; Huang, Y.; Duan, X. Toward Barrier
(154) Yan, K.; Wei, Z.; Li, J.; Chen, H.; Yi, Y.; Zheng, X.; Long, X.; Free Contact to Molybdenum Disulfide Using Graphene Electrodes.
Wang, Z.; Wang, J.; Xu, J.; Yang, S. High-Performance Graphene- Nano Lett. 2015, 15, 3030−3034.

4918 DOI: 10.1021/acsnano.6b01842

ACS Nano 2016, 10, 4895−4919
ACS Nano Review

(173) Voiry, D.; Mohite, A.; Chhowalla, M. Phase Engineering of

Transition Metal Dichalcogenides. Chem. Soc. Rev. 2015, 44, 2702−
2712.
(174) Cho, S.; Kim, S.; Kim, J. H.; Zhao, J.; Seok, J.; Keum, D. H.;
Baik, J.; Choe, D.-H.; Chang, K. J.; Suenaga, K.; Kim, S. W.; Lee, Y. H.;
Yang, H. Phase Patterning for Ohmic Homojunction Contact in
MoTe2. Science 2015, 349, 625−628.
(175) Kappera, R.; Voiry, D.; Yalcin, S. E.; Branch, B.; Gupta, G.;
Mohite, A. D.; Chhowalla, M. Phase-Engineered Low-Resistance
Contacts for Ultrathin MoS2 Transistors. Nat. Mater. 2014, 13, 1128−
1134.
(176) Ma, Y.; Liu, B.; Zhang, A.; Chen, L.; Fathi, M.; Shen, C.;
Abbas, A. N.; Ge, M.; Mecklenburg, M.; Zhou, C. Reversible
Semiconducting-to-Metallic Phase Transition in Chemical Vapor
Deposition Grown Monolayer WSe2 and Applications for Devices.
ACS Nano 2015, 9, 7383−7391.

4919 DOI: 10.1021/acsnano.6b01842

ACS Nano 2016, 10, 4895−4919