Layered semiconductor molybdenum disulde nanomembrane based

Schottky-barrier solar cells

Mariyappan Shanmugam, Chris A. Durcan and Bin Yu
*
Received 21st August 2012, Accepted 27th September 2012
DOI: 10.1039/c2nr32394j
We demonstrate Schottky-barrier solar cells employing a stack of layer-structured semiconductor
molybdenumdisulde (MoS
2
) nanomembranes, synthesized by the chemical-vapor-deposition method,
as the critical photoactive layer. An MoS
2
nanomembrane forms a Schottky-barrier with a metal
contact by the layer-transfer process onto an indium tin oxide (ITO) coated glass substrate. Two
vibrational modes in MoS
2
nanomembranes, E
1
2g
(in-plane) and A
1g
(perpendicular-to-plane), were
veried by Raman spectroscopy. With a simple stacked structure of ITOMoS
2
Au, the fabricated
solar cell demonstrates a photo-conversion efciency of 0.7%for 110 nmMoS
2
and 1.8%for 220 nm
MoS
2
. The improvement is attributed to a substantial increase in photonic absorption. The MoS
2
nanomembrane exhibits efcient photo-absorption in the spectral region of 350950 nm, as conrmed
by the external quantumefciency. Asizable increase in MoS
2
thickness results in only minor change in
MottSchottky behavior, indicating that defect density is insensitive to nanomembrane thickness
attributed to the dangling-bond-free layered structure.
Introduction
Schottky-barrier solar cells have the appealing advantage of low-
cost manufacture attributable to their simple and versatile
fabrication process and thin-lm-like structure (reduced active
material consumption as compared with pn or pin junction
bulk semiconductor solar cells). Schottky-barrier solar cells rely
mostly on a single semiconductor material (either p-or n-type
doped) in association with a metal, leading to energy barrier
formation and carrier dissociation at the metalsemiconductor
(MS) interface. Various MS material systems have been used
in Schottky-barrier solar cells, including Alamorphous carbon
nitride,
1
Au, Al, Mg and CaPbSe quantum dots,
2
AlSnS,
3
and
AuGaAs.
4
Recently, carbon derivatives have attracted attention
due to their promising optical/electrical properties. Schottky
junctions made by carbon nanotubeCdSe,
5
grapheneCdS
nanowire,
6
graphene, carbon nanotubeSi,
7,8
and grapheneSi
nanowire
911
have been reported. In general, MS interfacial
behavior and carrier mobility/photon absorption in the semi-
conductor material are key factors that impact the overall
performance of Schottky-barrier solar cells. While a group of
semiconductors such as boron doped diamond,
12
GaAs,
13,14
InP,
15
SiGe,
16
GaP,
17
GaN,
18
and SiC
19
can be used, the design of
an MS system with appropriate material work function differ-
ence is an essential factor to determine the Schottky-barrier
height and efciency of carrier separation.
The development of Schottky-barrier solar cells has been
decelerated over the past decade due to several fundamental
material challenges, including interfacial oxide layer formation
and Fermi-level pinning caused by interfacial states (which are
inevitable at MS interfaces).
2024
While poor interface properties play a substantial role in the
impeded improvement of energy-conversion efciency, searching
for a new class of semiconductor materials free of interfacial
states (hence without Fermi level pinning) would be of strategic
signicance to the ultimate realization of highly efcient
Schottky-barrier solar cells.
Transition-metal dichalcogenide semiconductors exhibit two
dimensional (2D) layered lattice structures with strong in-plane
covalent bonds and weak interlayer van der Waals interaction.
Molybdenum disulphide (MoS
2
), a layered semiconductor, has
recently attracted attention in the elds of nanoelectronics,
2530
energy storage
3133
and photovoltaics.
34
Nanosheets of monolayer
or multilayer MoS
2
can be obtained via mechanical or chemical
exfoliation,
35,36
similar tothe way graphene is produced. Recently,
direct assembly of MoS
2
nanosheets on SiO
2
by the chemical-
vapor-deposition (CVD) method has been demonstrated.
37
MoS
2
exhibits direct energy bandgap (1.85 eV) in monolayer and
indirect bandgap (1.3 eV) in multilayer/bulk form, respec-
tively.
38
Onelectronic properties, it has beenrecently reportedthat
MoS
2
shows carrier mobility values of 200 cm
2
V
1
s
1
in
monolayer and 517 cm
2
V
1
s
1
in a few layers, suggesting
promising conducting capacity.
39,40
Most importantly, the unique
structure of layered semiconductors provides inherent advantages
of forming a smooth surface free of any dangling chemical bonds,
potentially enabling the design of highly efcient Schottky-barrier
College of Nanoscale Science and Engineering, State University of New
York, Albany, NY-12203, USA. E-mail: byu@albany.edu; Fax: +1 518
956-7492; Tel: +1 518 956-7492
This journal is The Royal Society of Chemistry 2012 Nanoscale, 2012, 4, 73997405 | 7399
Dynamic Article Links
C
<
Nanoscale
Cite this: Nanoscale, 2012, 4, 7399
www.rsc.org/nanoscale PAPER
P
u
b
l
i
s
h
e
d

o
n

0
3

O
c
t
o
b
e
r

2
0
1
2
.

D
o
w
n
l
o
a
d
e
d

b
y

U
n
i
v
e
r
s
i
t
y

o
f

Y
o
r
k

o
n

0
6
/
0
2
/
2
0
1
4

1
4
:
1
7
:
5
0
.

View Article Online / Journal Homepage / Table of Contents for this issue
solar cells. Various metal contacts, including Ag, Al, Au, Co and
Fe, have been investigated for the formation of Schottky-barriers
with MoS
2
bulk crystals. No Fermi level pinning was observed
even for a high density of material defect (10
14
cm
3
).
41
In this paper we demonstrate solar cells employing a layer-
structured semiconductor MoS
2
nanomembranemetal (Au)
Schottky barrier. The CVD-assembled MoS
2
nanomembrane
also functions as the key photo-active layer. The solar cells are
made by a simple stack structure of ITOMoS
2
Au. Basic
material properties of MoS
2
, major solar cell device performance
metrics, and the impact of MoS
2
nanomembrane thickness are
discussed. The MoS
2
Au Schottky-barrier interface is charac-
terized by capacitance vs. voltage (CV) measurements.
Experimental section
Large-area (up to several square centimeters) semiconductor
MoS
2
nanomembranes were synthesized by the CVD approach.
Fig. 1(a) is the schematic of the CVD growth furnace system
showing Ar as a carrier gas and a sulphur (S) source owing
towards a molybdenum (Mo)-coated SiSiO
2
substrate (where
vaporsolid reaction occurs to assemble the MoS
2
nano-
membrane). CVD growth of MoS
2
silicon coated with 100 nm of
SiO
2
was used as the substrate for the pre-deposition of Mo. A
thin Mo metal layer was deposited with an approximate thick-
ness of 50 nm by the electron-beam evaporation method. The
CVD growth process started with owing Ar of 200 sccm in the
quartz-tube chamber with a pressure level of 1 Torr. The
furnace was subsequently heated up to 750

Cand stabilized over

one hour. Pure S powder, placed upstream in the chamber, was
heated to just above its melting temperature (115

C). The S
vapor owed over the Mo thin lm pre-deposited on the SiO
2
Si
substrate for approximately 30 minutes. Vaporsolid reaction
took place at an elevated temperature (750

C), leading to the
growth of an MoS
2
nanomembrane driven by increase in
enthalpy. While the MoMo bond is relatively weak (metallic in
Fig. 1 (a) Schematic of the CVDgrowth setup showing the reaction of sulphur vapor with Mo lmpre-deposited on SiSiO
2
substrate for the growth of
an MoS
2
nanomembrane. (b) Major process steps involved in the synthesis of the MoS
2
nanomembrane, layer transfer, and subsequent fabrication of a
Schottky-barrier solar cell. The KOH wet etching process was used to detach the MoS
2
nanomembrane from SiO
2
Si substrate. (c and d) Partially
oating MoS
2
nanomembrane in KOH solution. (e) Free-oating MoS
2
nanomembrane in KOH to be transferred onto an ITO-coated glass substrate
for solar cell fabrication.
7400 | Nanoscale, 2012, 4, 73997405 This journal is The Royal Society of Chemistry 2012
P
u
b
l
i
s
h
e
d

o
n

0
3

O
c
t
o
b
e
r

2
0
1
2
.

D
o
w
n
l
o
a
d
e
d

b
y

U
n
i
v
e
r
s
i
t
y

o
f

Y
o
r
k

o
n

0
6
/
0
2
/
2
0
1
4

1
4
:
1
7
:
5
0
.

View Article Online
nature), the MoS bond is a strong ionic one and therefore more
thermodynamically stable. As the temperature increases, Mo
atoms gain sufcient thermal energy (exceeding the activation
energy) and make covalent bonds with S atoms. After the
completion of CVD growth, the sample was annealed at 1000

C
for one hour to improve the crystalline quality and the unifor-
mity of the resultant nanomembrane. The major steps in the
MoS
2
layer transfer sequence are depicted in Fig. 1(b) starting
from the SiSiO
2
substrate, Mo metal deposition, MoS
2
growth,
and transfer to the target substrate in which the solar cells were
fabricated.
The CVD-assembled MoS
2
nanomembrane was transferred
from the growth substrate (SiO
2
Si) to an indium tin oxide (ITO
serving as transparent electrode) coated device substrate,
involving a wet chemistry etching step of the SiO
2
layer in the
growth substrate to yield a freestanding MoS
2
nanomembrane in
potassium hydroxide (KOH) solution. Fig. 1(c)(e) show the
optical images of the process to separate the MoS
2
nano-
membrane from SiO
2
Si substrate. The MoS
2
nanomembrane
was transferred onto the transparent conducting oxide (ITO)
substrate to fabricate the Schottky-barrier solar cell. KOH
solution with a concentration of 15 M was prepared in which the
MoS
2
SiSiO
2
substrate was immersed to etch away the SiO
2
. A
slow and smooth KOH etching of SiO
2
released the MoS
2
nanomembrane partially from the Si substrate. Within approx-
imately 10 minutes a completely free-oating MoS
2
nano-
membrane in KOH solution was obtained. The thickness of the
MoS
2
nanomembrane is determined by that of the Mo lm
deposited by electron beam evaporation (it is hence scalable).
The surface morphology and thickness of the as-grown MoS
2
nanomembrane was characterized by Atomic Force Microscopy
(AFM, dimension 3100). Raman spectroscopy (Horiba Scien-
tic) with a laser of 532 nm was performed on the samples to
conrm the growth of MoS
2
via identied vibrational modes.
After transferring the MoS
2
nanomembrane onto the ITO-
coated glass substrate, deposition of 50 nm of Au contact was
made by electron-beam evaporation to form a Schottky-barrier
(with MoS
2
). The JV characteristics of the solar cell were
measured under both dark and standard AM 1.5 illumination
conditions using an Agilent B1500A semiconductor parameter
analyzer. An Xe arc lamp was used to simulate the solar light.
The EQE characteristics of the solar cells were measured in the
wavelength range of 350950 nm using a Newport mono-
chromator (equipped with the same Xe arc lamp that was used to
measure the JV characteristics). The CV measurement was
performed on the solar cell using a B1520A-FG Multi Frequency
Capacitance Measurement Unit Module attached to an Agilent
B1500A using a frequency of 1 kHz and AC perturbation of
10 mV at room temperature.
Results and discussion
Fig. 2(a) is the micrograph showing the surface morphology of
an MoS
2
nanomembrane after being transferred onto the ITO
substrate, as obtained from the 2D mapping measurement of an
Atomic Force Microscope (AFM). The AFM line-scan was
carried out, from point A to point B as shown in Fig. 2(a), to
measure the physical thickness of the MoS
2
nanomembrane. The
step height between MoS
2
and the ITO substrate, as seen in
Fig. 2(b), conrmed the thickness of the MoS
2
nanomembrane to
be about 55 nm, an average value from the measured data. We
believe that the as-grown MoS
2
nanomembrane on the SiSiO
2
substrate in our CVD process is highly uniform in thickness, as
reported by Y. Zhan et al. in a similar growth experiment.
37
The
variation in AFM line-scan data (approximately 5 nm) could
arise from deformation of the MoS
2
nanomembrane as a result
of the layer transfer process. We expect that micro-level wrinkles
might have been formed on the MoS
2
nanomembrane caused by
various types of mechanical interactions during the wet chem-
istry etching and transferring process.
Material characterization using Raman spectroscopy was
conducted on two transferred MoS
2
nanomembrane samples
with different physical thicknesses. While the as-assembled MoS
2
exhibits an average thickness of 55 nm, as shown in Fig. 2(b),
stacked MoS
2
nanomembranes were prepared via multiple (two
and four, respectively, in our experiment) transfers.
The samples were subsequently used for Raman spectroscopic
investigation. The expected thicknesses of the two samples are
110 nm (two transfers) and 220 nm (four transfers), respec-
tively. Prior research results show that there is a signicant
difference in the signature modes in Raman spectra for
Fig. 2 (a) 2D mapping micrograph of AFM surface scan obtained on an
as-assembled MoS
2
nanomembrane after transferring onto an ITO
substrate. The line-scan measurement was performed from point A to B
to evaluate the physical thickness of MoS
2
. (b) Measured step prole of
the MoS
2
nanomembrane on ITO substrate.
This journal is The Royal Society of Chemistry 2012 Nanoscale, 2012, 4, 73997405 | 7401
P
u
b
l
i
s
h
e
d

o
n

0
3

O
c
t
o
b
e
r

2
0
1
2
.

D
o
w
n
l
o
a
d
e
d

b
y

U
n
i
v
e
r
s
i
t
y

o
f

Y
o
r
k

o
n

0
6
/
0
2
/
2
0
1
4

1
4
:
1
7
:
5
0
.

View Article Online
monolayer, bilayer and few-layer MoS
2
.
37
In our experiment, a
much thicker stack of MoS
2
was employed to enhance the photo-
absorption in the Schottky-barrier solar cells. Fig. 3(a) shows the
Raman spectra of the MoS
2
nanomembrane samples with
thicknesses of 110 nm and 220 nm, respectively. Two signif-
icant Raman active modes were identied at the wavenumbers of
383 cm
1
and 408 cm
1
which correspond to E
1
2g
and A
1g
modes
respectively.
Fig. 3(b) displays schematically both E
1
2g
and A
1g
modes. In
the E
1
2g
mode both Mo and S atoms vibrate along the in-plane
direction (yet opposite to each other), whereas in the A
1g
mode
the S atoms vibrate in the perpendicular-to-plane direction. The
Raman signature peaks observed in our experiment conrmboth
in-plane and perpendicular-to-plane Raman active modes. As we
used relatively thick MoS
2
(>100 nm), there are no blue- and red-
shifts for the E
1
2g
and A
1g
modes, different from that observed in
monolayer/bilayer MoS
2
nanomembranes.
37
The signicant
variation in the peak intensity of Raman spectra, as observed in
the two examined samples, was due to the substantial difference
in physical thickness (100 nm).
Fig. 4(a) shows the schematic cross-sectional view of the MoS
2
nanomembraneAu Schottky-barrier solar cell. The MoS
2
nano-
membrane was transferred onto an ITO substrate (serving as the
windowlayer through which the solar cell is illuminated). Fig. 4(b)
shows the energy band diagramof the fabricated Schottky-barrier
solar cell structure with a stack of ITOMoS
2
Au.
In general, if the metal work function is higher than that of the
n-type semiconductor, then an ideal Schottky-barrier could be
formed. In our experiment, Au (work function: 5.1 eV) forms a
barrier on the MoS
2
semiconductor (work function: 4.6 eV).
The ITO used in this study is a heavily n-type doped transparent
conductive oxide with a work function of 4.5 eV. It forms an
almost perfect ohmic contact with the n-type doped MoS
2
nanomembrane. As shown in Fig. 4(b), the Fermi level of ITO is
above the conduction band due to its heavily doped nature.
Although layer-dependent optical properties (e.g., change of
bandgap with number of layers) has been demonstrated in ultra-
thin MoS
2
(from a monolayer to six layers) as reported by K. F.
Mak et al.,
39
the bulk properties dominate in thicker stacks of
MoS
2
. In our experiment, MoS
2
nanomembrane stacks
Fig. 3 (a) Raman spectra obtained on two samples of MoS
2
nano-
membranes with thicknesses of 110 nm and 220 nm, respectively. The
samples were prepared by multiple transfers/stacking of CVD-grown
MoS
2
. (b) Schematic showing the two Raman active modes corre-
sponding to the signature peaks observed in MoS
2
.
Fig. 4 (a) Schematic cross-sectional view of the demonstrated Schottky
barrier solar cell structure showing a stack of MoS
2
nanomembrane on
an ITO substrate with Au contact. (b) Energy band diagram of the solar
cell with formation of a Schottky-barrier between the MoS
2
nano-
membrane and Au metal contact.
7402 | Nanoscale, 2012, 4, 73997405 This journal is The Royal Society of Chemistry 2012
P
u
b
l
i
s
h
e
d

o
n

0
3

O
c
t
o
b
e
r

2
0
1
2
.

D
o
w
n
l
o
a
d
e
d

b
y

U
n
i
v
e
r
s
i
t
y

o
f

Y
o
r
k

o
n

0
6
/
0
2
/
2
0
1
4

1
4
:
1
7
:
5
0
.

View Article Online
(110 nm and 220 nm) were used for the fabrication of solar
cells. Therefore layer dependency of the optical bandgap
becomes negligible. We consider the MoS
2
stack with an optical
bandgap of 1.3 eV which effectively absorbs photons in the
wavelength region of 350950 nm. The Schottky-barrier solar
cells are illuminated through the ITO window layer which does
not absorb any photons in the visible light spectral region, as its
bandgap is 3.8 eV and more than 95% optically transparent.
Incident photons generate electronhole pairs in the MoS
2
nanomembrane and electrons are subsequently excited to the
conduction band (E
C
) of MoS
2
. As the work function difference
between Au and MoS
2
results in a Schottky-barrier, the photo-
excited electrons move from the E
C
of MoS
2
towards the Fermi
level of ITO(where they can be collected to the load). Alarger work
function difference between Au and n-type MoS
2
would generate a
higher electric eld in the depletion region of MoS
2
. The built-in
electric eld facilitates the dissociation of the photo-excited elec-
tronhole pairs, transporting separated carriers towards ITO and
Au contacts. The processes of photo-excitation in MoS
2
and elec-
tron transport from MoS
2
to ITO are demonstrated in Fig. 4(b).
Fig. 5(a) and (b) show the dark and illuminated JV charac-
teristics of the fabricated Schottky-barrier solar cells with MoS
2
thicknesses of 110 nm and 220 nm, respectively. The output
power as a function of voltage in each solar cell is shown in the
insets of Fig. 5(a) and (b), respectively. For the Schottky-barrier
solar cell with MoS
2
thickness of 110 nm key metrics J
SC
(short-
circuit current density), V
OC
(open-circuit voltage), P
MAX
(maximum power), FF (ll factor) and h (photo-conversion
efciency) are 2.52 mA cm
2
, 590 mV, 1.22 10
4
W, 0.48 and
0.7%, respectively. For the solar cell with MoS
2
thickness of
220 nm, J
SC
, V
OC
, P
MAX
, FF and h are 5.37 mAcm
2
, 597 mV,
2.99 10
4
W, 0.55 and 1.8% respectively. It is observed that an
increase of MoS
2
thickness helps to signicantly enhance J
SC
of
the solar cell (by 113%) due to more efcient photo-absorption in
the thicker MoS
2
stack, and hence creates a noteworthy change
in overall device performance. It should be noted that there is no
sizable change in V
OC
between solar cells with different thick-
nesses of MoS
2
stacks (merely 1% change). It is evident that an
increase in thickness of the MoS
2
stack would not cause a notable
change in the band structure (hence no change in the Fermi level)
due to the absence of layer-dependent properties.
Fig. 5(c) shows the schematic to compare carrier recombina-
tion in conventional and layered semiconductors. Process 1
represents optical excitation leading to electronhole pair
generation. In conventional semiconductors, the amount of
interface states (recombination centers) in the bandgap is huge
due to unsaturated chemical dangling bonds which capture
electrons from the conduction band (Process 2) and holes from
the valence band (Process 3), leading to recombination. In
layered semiconductors, the amount of interface states is signif-
icantly less due to the self-saturated surface bonds which do not
actively participate in the recombination process. Thus, we
attribute the enhancement in solar cell performance to the
increase in MoS
2
thickness which improved photon absorption.
We expect that as a layered material, defect density in MoS
2
does
not increase drastically with increasing thickness compared to a
conventional semiconductor.
Fig. 6 shows External Quantum Efciency (EQE) spectra of
the two Schottky-barrier solar cells with different MoS
2
thicknesses in the wavelength range of 350950 nm. The
maximum EQE values of 41% and 52% were measured for solar
cells with 110 nm and 220 nm of MoS
2
stack, respectively, at
wavelength of 600 nm. The increase in MoS
2
thickness plays a
vital role in photo-absorption. Our prior work on MoS
2
-based
Fig. 5 Measured dark and illuminated JV characteristics of the
Schottky-barrier solar cells with a stack of MoS
2
nanomembrane having
thicknesses of (a) 110 nm and (b) 220 nm. Inset of each gure shows
the power output as a function of voltage for the fabricated solar cell. (c)
Recombination process via electronhole capture by interfacial trap
states present in conventional and layered semiconductors, showing
signicantly fewer traps in the latter. This leads to reduced probability of
carrier recombination in a layered semiconductor.
This journal is The Royal Society of Chemistry 2012 Nanoscale, 2012, 4, 73997405 | 7403
P
u
b
l
i
s
h
e
d

o
n

0
3

O
c
t
o
b
e
r

2
0
1
2
.

D
o
w
n
l
o
a
d
e
d

b
y

U
n
i
v
e
r
s
i
t
y

o
f

Y
o
r
k

o
n

0
6
/
0
2
/
2
0
1
4

1
4
:
1
7
:
5
0
.

View Article Online
solar cells
34
employing a combination of monolayer/multi-layer
stacks of MoS
2
on TiO
2
(serving as electron acceptor) has shown
photon absorption of MoS
2
in the wavelength region of 350
750 nm. The measured EQE spectra in the extended wavelength
region, as compared with the previously reported values (up to
950 nm),
34
suggests that an MoS
2
nanomembrane with an
optical bandgap of 1.3 eV could effectively absorb photons in
both visible and near infrared spectral regions due to the absence
of the layer-dependent optical properties of MoS
2
as governed by
its physical thickness.
Fig. 7(a) shows the MottSchottky (C
2
V) characteristics of
the Schottky-barrier solar cells. The C
2
V characteristics were
used to estimate the Schottky-barrier heights associated with
each of the two fabricated solar cells. Fig. 7(b) shows schemati-
cally the energy-level alignment at the MoS
2
Au interface under
zero, forward and reverse voltage bias conditions. Since the
MoS
2
samples are n-type doped, Fermi levels in MoS
2
and Au
shift upwards and downwards, respectively, with reference to the
equilibrium Fermi level at zero voltage bias, yielding
the maximum barrier height. In the case of reverse voltage bias,
the situation becomes opposite. We found that both solar cells
(with different MoS
2
thickness) resulted in nearly the same
Schottky-barrier height (1.0 eV), as measured from the linear
extrapolation in the reverse-biasing region (from 1 V to 4 V). It
can be understood that a change of the MoS
2
should not make a
difference in the Schottky-barrier height. While we expect that
MoS
2
thickness (down to a few layers) may change the Schottky-
barrier height, the layer dependence becomes insignicant in the
nanomembranes used in our experiment. The observed linear
behavior of the MottSchottky plots suggests that the MoS
2
nanomembranes possess a relatively low density of surface
defects.
J. M. Luther et al. reported a Schottky-barrier solar cell using
PbSe quantum dots.
2
The result showed signicant variation in
the voltage-dependent C
2
for two PbSe samples (thickness
difference > 300 nm).
2
In our study, MoS
2
Au solar cells (with
active layer thickness difference of 100 nm) exhibit no signicant
change in the C
2
values with reverse bias. This could be
attributed to the layered MoS
2
nanomembrane (defect-free or
with negligibly low density of defects) as compared with
conventional semiconductors in which unsaturated dangling
bonds make the MS interface highly defective. In layered
semiconductors the crystal lattice planes are weakly bound by
van der Waals interactions and hence the surface exhibits self-
saturated electronic bonds which are no longer active for surface
reaction with the environment. The nature of a self-saturated
surface in layered MoS
2
nanomembrane eliminates Fermi level
pinning and interfacial oxide layer formation (both could impede
the transport signicantly). The voltage-bias-dependent C
2
values are almost identical in both device samples (with different
thicknesses) due to negligible change in the trap states at MoS
2

Au interface where carriers are stored and released, affecting the

junction capacitance. This is a key appealing property of layered
semiconductors for solar cell applications as compared with
bulk/thin lm semiconductors which possess a particular type of
electronic bonding along and across the crystal planes (leading to
defective bulk and surface upon cleaving).
We project that the performance of the MoS
2
nanomembrane-
based Schottky-barrier solar cell could be further improved.
Layered semiconductors are typically very sensitive to growth
conditions such as temperature and pressure. Temperature uc-
tuation across the growth furnace plays a signicant role in
inuencing the solidvapor reaction between pre-deposited Mo
and S vapor, and eventually the electronic quality of the MoS
2
nanomembrane. In addition, defects and impurities introduced
Fig. 6 Measured EQE spectra of the Schottky-barrier solar cells with a
stack of MoS
2
nanomembranes having thicknesses of 110 nm and
220 nm, respectively, in the wavelength region of 350950 nm. This
conrms an optical bandgap of approximately 1.3 eV for the prepared
MoS
2
nanomembranes.
Fig. 7 (a) The measured MottSchottky (C
2
V) plot of the Schottky-
barrier solar cells with a stack of MoS
2
nanomembranes having thick-
nesses of 110 nm and 220 nm, respectively, showing a linear C
2
dependency (shown in the circled region) on the applied reverse bias, V
R
.
(b) Schematic showing MoS
2
Au interfacial energy level change at zero
(equilibrium), forward, and reverse bias conditions and the effect on
Schottky barrier height.
7404 | Nanoscale, 2012, 4, 73997405 This journal is The Royal Society of Chemistry 2012
P
u
b
l
i
s
h
e
d

o
n

0
3

O
c
t
o
b
e
r

2
0
1
2
.

D
o
w
n
l
o
a
d
e
d

b
y

U
n
i
v
e
r
s
i
t
y

o
f

Y
o
r
k

o
n

0
6
/
0
2
/
2
0
1
4

1
4
:
1
7
:
5
0
.

View Article Online
during the CVD growth and deformation formed during layer
transfer would all contribute to reduced carrier mobility and
degraded carrier transport properties. We expect that further
optimization of the growth process, use of a cleaner fabrication
environment, and direct assembly of MoS
2
nanomembranes on
an ITO substrate would produce less defective materials, leading
to improved solar cell performance.
Conclusions
Schottky-barrier solar cells have been demonstrated using layer-
assembled, CVD-grown MoS
2
semiconductor nanomembranes
as the critical photo-active layer. While relatively thick MoS
2
nanomembranes are used in this experiment, the material is
scalable via a controlled CVD growth process and large-area
solar cell fabrication is potentially feasible. The dependency of
solar cell performance on MoS
2
thickness was evaluated. Further
improvement in solar cell performance is expected through
optimizing the growth process, eliminating structural deforma-
tion (induced during fabrication), and minimizing the amount of
surface defects and impurities. The reported results would
promote continued efforts towards developing highly efcient
Schottky-barrier solar cells taking advantage of the unique
interfacial properties of layered semiconductor nanostructures.
References
1 Z. B. Zhou, R. Q. Cui, Q. J. Pang, G. M. Hadi, Z. M. Ding and
W. Y. Li, Sol. Energy Mater. Sol. Cells, 2002, 70, 487493.
2 J. M. Luther, M. Law, M. C. Beard, Q. Song, M. O. Reese,
R. J. Ellingson and A. J. Nozik, Nano Lett., 2008, 8, 34883492.
3 B. Ghosh, M. Das, P. Banerjee and S. Das, Solid State Sci., 2009, 11,
461466.
4 M. Soylu and F. Yakuphanoglu, Thin Solid Films, 2011, 519, 1950
1954.
5 L. Zhang, Y. Jia, S. Wang, Z. Li., C. Ji, J. Wei, H. Zhu, K. Wang,
D. Wu, E. Shi, Y. Fang and A. Cao, Nano Lett., 2010, 10, 35833589.
6 Y. Ye, Y. Dai, L. Dai, Z. Shi, N. Liu, F. Wang, L. Fu, R. Peng,
X. Wen, Z. Chen, Z. Liu and G. Qin, ACS Appl. Mater. Interfaces,
2010, 2, 34063410.
7 C. C. Chen, M. Aykol, C. C. Chang, A. F. J. Levi and S. B. Cronin,
Nano Lett., 2011, 11, 18631867.
8 J. Wei, Y. Jia, Q. Shu, Z. Gu, K. Wang, D. Zhuang, G. Zhang,
Z. Wang, J. Luo, A. Cao and D. Wu, Nano Lett., 2007, 7, 23172321.
9 C. Xie, P. Lv, B. Nie, J. Jie, X. Zhang, Z. Wang, P. Jiang, Z. Hu,
L. Luo, Z. Zhu, L. Wang and C. Wu, Appl. Phys. Lett., 2011, 99,
133113.
10 G. Fan, H. Zhu, K. Wang, J. Wei, X. Li, Q. Shu, N. Guo and D. Wu,
ACS Appl. Mater. Interfaces, 2011, 3, 721725.
11 C. Xie, J. Jie, B. Nie, T. Yan, Q. Li, P. Lv, F. Li, M. Wang, C. Wu,
L. Wang and L. Luo, Appl. Phys. Lett., 2012, 100, 193103.
12 G. H. Glover, Solid State Electrochem., 1973, 16, 973978.
13 F. Dubecky and B. Olejnkova, J. Appl. Phys., 1991, 69, 1769.
14 C. Ghezzit, E. Gombiaz and R. Moscaf, Semicond. Sci. Technol.,
1991, 6, 3133.
15 A. Singh, P. Cova and R. A. Masut, J. Appl. Phys., 1994, 76, 2336.
16 F. Lu, D. Gong, J. Wang, Q. Wang, H. Sun and X. Wang, Phys. Rev.
B: Condens. Matter Mater. Phys., 1996, 53, 46234629.
17 D. Vanmaekelbergh, A. Koster and F. I. Marn, Adv. Mater., 1997, 9,
575578.
18 Y. Fukushima, K. Ogisu, M. Kuzuhara and K. Shiojima, Phys. Status
Solidi C, 2009, 6, 856859.
19 S. Tongay, T. Schumann and A. F. Hebard, Appl. Phys. Lett., 2009,
95, 222103.
20 F. Leonard and J. Tersoff, Phys. Rev. Lett., 2000, 84, 46934696.
21 A. Dimoulas, P. Tsipas, A. Sotiropoulos and E. K. Evangelou, Appl.
Phys. Lett., 2006, 89, 252110.
22 T. Nishimura, K. Kita and A. Toriumi, Appl. Phys. Lett., 2007, 91,
123123.
23 Y. Zhou, W. Han, Y. Wang, F. Xiu, J. Zou, R. K. Kawakami and
K. L. Wang, Appl. Phys. Lett., 2010, 96, 102103.
24 S. J. Oh, D. K. Kimand C. R. Kagan, ACS Nano, 2012, 6, 43284334.
25 Y. Yoon, K. Ganapathi and S. Salahuddin, Nano Lett., 2011, 11,
37683773.
26 Z. Yin, H. Li, H. Li, L. Jiang, Y. Shi, Y. Sun, G. Lu, Q. Zhang,
X. Chen and H. Zhang, ACS Nano, 2012, 6, 7480.
27 H. S. Lee, S. W. Min, Y. G. Chang, M. K. Park, T. Nam, H. Kim,
J. H. Kim, S. Ryu and S. Im, Nano Lett., 2012, 12, 36953700.
28 D. J. Late, B. Liu, H. S. S. R. Matte, V. P. Dravid and C. N. R. Rao,
ACS Nano, 2012, 6, 56355641.
29 Y. Zhang, J. Ye, Y. Matsuhashi and Y. Iwasa, Nano Lett., 2012, 12,
11361140.
30 S. Ghatak, A. N. Pal and A. Ghosh, ACS Nano, 2011, 5, 77077712.
31 K. Chang and W. Chen, Chem. Commun., 2011, 47, 42524254.
32 K. Chang, W. Chen, L. Ma, H. Li, H. Li, F. Huang, Z. Xu, Q. Zhang
and J. Y. Lee, J. Mater. Chem., 2011, 21, 62516257.
33 K. Chang and W. Chen, J. Mater. Chem., 2011, 21, 1717517184.
34 M. Shanmugam, T. Bansal, C. Durcan and B. Yu, Appl. Phys. Lett.,
2012, 100, 153901.
35 K. S. Novoselov, D. Jiang, F. Schedin, T. J. Booth, V. V. Khotkevich,
S. V. Morozov and A. K. Geim, Proc. Natl. Acad. Sci. U. S. A., 2005,
102, 10451.
36 J. N. Coleman, M. Lotya, A. ONeill, S. D. Bergin, P. J. King,
U. K.Young, A. Gaucher, S. De, R. J. Smith and I. V. Shvets,
et al., Science, 2011, 331, 568571.
37 Y. Zhan, Z. Liu, S. Najmaei, P. M. Ajayan and J. Lou, Small, 2012, 8,
966971.
38 H. Liu and P. D. Ye, IEEE Electron Device Lett., 2012, 33, 546
548.
39 K. F. Mak, C. Lee, J. Hone, J. Shan and T. F. Heinz, Phys. Rev. Lett.,
2010, 105, 136805.
40 B. Radisavljevic, A. Radenovic, J. Brivio, V. Giacometti and A. Kis,
Nat. Nanotechnol., 2011, 6, 147150.
41 J. R. Lince, D. J. Carre and P. D. Fleischauer, Phys. Rev. B: Condens.
Matter Mater. Phys., 1987, 36, 16471656.
This journal is The Royal Society of Chemistry 2012 Nanoscale, 2012, 4, 73997405 | 7405
P
u
b
l
i
s
h
e
d

o
n

0
3

O
c
t
o
b
e
r

2
0
1
2
.

D
o
w
n
l
o
a
d
e
d

b
y

U
n
i
v
e
r
s
i
t
y

o
f

Y
o
r
k

o
n

0
6
/
0
2
/
2
0
1
4

1
4
:
1
7
:
5
0
.

View Article Online