“This document is the Accepted Manuscript version of a Published Work that appeared in final form in Journal of Materials
Chemistry A, copyright © The Royal Society of Chemistry 2019 after peer review and technical editing by the publisher. To
access the final edited and published work see:

https://pubs.rsc.org/en/content/articlelanding/2019/ta/c9ta08818k#!divAbstract

Multi-layered photocathodes based on Cu2ZnSnSe4 absorber and
MoS2 catalyst for the Hydrogen Evolution Reaction
Received 00th January 20xx,
Accepted 00th January 20xx
DOI: 10.1039/x0xx00000x

Sergi Grau,a Sergio Giraldo,b Edgardo Saucedo,b Joan Ramón Morante,b Antoni Llobet*a,c and
Carolina Gimbert-Suriñach*a
A multilayered p-type photoabsorber based on kesterite Cu2ZnSnSe4 has been functionalized with inexpensive and easily
scalable amorphous MoS2 electrocatalyst by cathodic photoelectrodeposition. The resulting photocathodes with a final
composition of Mo/MoSe2/Cu2ZnSnSe4/CdS/ZnO/ITO/MoS2 (where ITO = indium tin oxide) feature one of the highest
hydrogen evolution catalytic current density reported to date at a potential E = 0 V vs RHE (-18 mA/cm2 in pH 2) under 1 sun
illumination. A comparison of the performance of the photoelectrode with an analogous dark electrode with composition
ITO/MoS2 allows us to estimate a photovoltage of 430 mV, which is very close to the working voltage of an equivalent
photovoltaic cell with open circuit voltage VOC = 464 mV. The photoelectrodes show sustained activity at -10 mA/cm2 above
the thermodynamic potential of the H+/H2 redox couple (E = +86 mV) for 1 hour without any loss of activity. Prolonged
reaction times up to 4 hours show no degradation of the photoabsorber multilayered architecture but a decrease in the
thickness of the the MoS2 electrocatalytic film due to the mechanical stress caused by the hydrogen bubbles forming on the
surface of the electrode. The MoS2 layer is key not only to catalyze the catalytic reaction but also to protect the
photoabsorber from the corrosive electrolyte solution. The adhesion of the MoS2 electrocatalytic film on the ITO top layer
is critical for the stability during hydrogen evolution catalysis.

Introduction
Since 2000 the worldwide energy consumption has been
increasing with an average rate of 2.3 % with contaminating
fossil fuels being one the main primary source.1,2 Thus, finding
sources of energy, that are renewable, environmentally friendly
and that can supply our energy demand is imperative.
Artificial photosynthesis is a general term used to describe the
light induced reactions that promote the oxidation of water to
dioxygen and use the generated reducing equivalents to
transform aqueous protons to hydrogen gas or reduce CO2 to
useful carbon based fuels.3,4 One common strategy for the
hydrogen production in the context of artificial photosynthesis
is the use of a photoelectrochemical cell (PEC), where one or
two photoelectrodes harvest the light and generate electronhole pairs that are used to perform the oxygen evolution
reaction (OER, photoanode) and hydrogen evolution reaction
(HER, photocathode). An example of a simple PEC based on a
photocatode and a dark anode is given in Figure 1.5-8
Thermodynamically, the water splitting reaction requires 1.23 V
(WS, eq. 1) but in practice, an additional overpotential
associated with the kinetic barriers of each half-reaction is
needed (HER in eq. 2 and OER in eq. 3). To minimize the

of Chemical Research of Catalonia (ICIQ), The Barcelona Institute of
Science and Technology, Avinguda Països Catalans 16, 43007 Tarragona, Spain. Email: cgimbert@iciq.cat; allobet@iciq.cat
b. Catalonia Institute for Energy Research (IREC), Jardins de les Dones de Negre 1,
08930 Sant Adrià de Besòs, Spain.
c. Departament de Química, Universitat Autònoma de Barcelona, 08193 Cerdanyola
del Vallès, Barcelona, Spain.
Electronic Supplementary Information (ESI) available: optical characterization of the
photoelectrodes,
additional
SEM
images,
electrochemical
and
photoelectrochemical experiments, J-V curve and EQE of equivalent PV device as
well
as
additional
details
about
experimental
procedures.
See
DOI: 10.1039/x0xx00000x

overpotential required for the HER and OER it’s necessary to use
selective and efficient catalysts for each reaction.
Platinum is the best catalyst for HER known to date, able to
reduce protons to molecular hydrogen with almost no
overpotential, but is scarce and expensive.9,10 A cheaper and
remarkably well-performing material is MoS2, which was not
investigated as a HER catalyst due to its low activity as bulk
material until 2005, when Nørskov and co-workers
demonstrated that MoS2 nanoparticles are in fact very active.12
Further experimental and theoretical studies revealed that the
morphology and size of the nanoparticles are critical for the
catalytic performance and that only the edge sites and sulfur
vacancies are the active sites on the crystallographic
edges.12,13,14 Since then, numerous studies with different
synthetic approaches to obtain MoS2 active for HER with
different crystalline nanostructures as well as amorphous MoS2
thin films have been published.15-20 On the other hand, only a

a. Institute

Figure 1. Top) water splitting photoelectrochemical cell configuration used in
this work based on a photocathode, a dark anode and an applied bias (Eapp). EF,n*
= fermi level of electrons; EF,p* = fermi level of holes. Bottom) water splitting
reaction (WS, eq. 1), hydrogen evolution reaction (HER, eq. 2) and oxygen
evolution reaction (OER, eq. 3).

Please do not adjust margins

Please do not adjust margins
ARTICLE

Journal Name

few examples can be found in the literature where MoS2 HER
electrocatalyst is combined with semiconductors to build
photocathodes.21-24
During the last decades, quaternary salts of the family of p-type
kesterites with stoichiometry Cu2ZnSn(S/Se)4 have been widely
used as photocathodes for the HER. They have been explored
for their high performance as photoabsorbers in photovoltaic
devices with efficiencies up to 12.7 % and VOC up to 514 mV.25
An additional attractive aspect of these materials is that they
are made of non-critical raw materials and low-toxicity
elements. The main family of kesterites that have been used in
photoelectrochemistry for solar fuel production are those made
of sulfur Cu2ZnSnS4, commonly abbreviated as CZTS. A
pioneering work in this regard was published by Domen and
coworkers in 2010. They used a CZTS photocathode with an
additional n-type CdS layer to create a p-n junction, a TiO2 top
layer to protect the absorbers from the electrolyte and platinum
nanoparticles on the surface to catalyse the HER.26 In 2015, a
modified electrode with an additional In2S3 inter-layer
(Mo/CZTS/CdS/In2S3/Pt) and superior performance was
reported by the same group.27 Two years later, Luo, Zou and
coworkers used a similar photocathode configuration
improving the performance further with the addition of Ge.28
More recently, selenium based thin-film kesterites have been
tested as photocathodes containing a protecting top layer of
TiO2 and Pt catalyst on the surface.29
Despite all these promising advances, there is still a lack of an
efficient and stable Cu2ZnSn(S/Se)4 photocathode that is made
of inexpensive catalysts. In addition, the stability of the
reported examples is most of the time an overlooked parameter
that is critical for a possible market exploitation. In this work,
CZTSe based p-type photoabsorbers are combined with MoS2
electrocatalyst layers to build a powerful photocathode for the
HER. The two key components have been selected for i) their
high performance as light absorbers (CZTSe) and catalytic
activity (MoS2) and ii) for the optimum band alignment of the
semiconductor25,30,31 with the overpotential of the
electrocatalyst,12,14,18-20,22 that should favour charge separation
and fast catalysis leading to high performance.

Experimental

ZnO/ITO window layers were sputtered using ZnO (99.99%) and
In2O3-SnO2 (ITO) (99.99%) targets from PhotonExport.
UV-Vis measurements were carried out on a Lambda 1050
PerkinElmer spectrophotometer equipped with a PMT, InGaAs
and PbS detectors system, double beam optics, double
monochromator and D2 and W light sources. Diffuse
reflectance measurements were carried out in the same
equipment using 150mm Integrating Sphere with PbS and PMT
detectors. Scanning electron microscopy (SEM) images were
obtained with a ZEISS Series Auriga micro- scope using 5 kV
accelerating voltage. External quantum efficiency (EQE)
measurements were performed using a Bentham PVE300
system calibrated with Si and Ge photodiodes.
Commercially available glass coated ITO glasses (SPI supplies®)
were cut in approximately 1,25 cm x 1,25 cm pieces and cleaned
with 2% HellmanexTM detergent, ethanol, acetone and dried
under air flow. The electrical connection between copper wire
and the electrode was made with solder (USS-9210 MBR
electronics) using cerasolzer alloy (#CS186 MBR electronics).
The electrical connection as well as the edges of the electrode
were covered with non-transparent epoxy resin in order to
isolate it from the solution. The electrode’s area was calculated
using J-Image (free software).
Linear sweep voltammetry (LSV) and Chronopotenciometry (CP)
experiments were measured in a one compartment three
electrode configuration cell with controlled illuminated area of
0.5 cm2 using a CHI660D potentiostat. Platinum Disk/Mesh
counter electrodes and reference electrodes (standard calomel
electrode (SCE) or AgCl/Ag) were purchased from IJ-Cambria
Ldt. Buffer solutions from pH 2 to pH 12 were prepared with
H3PO4, NaH2PO4, Na2HPO4 and Na2SO4 from Sigma-Aldrich.
Solutions at pH 13 were prepared with 0.1 M NaOH from SigmaAldrich. Photoelectrochemical experiments were ran under
simulated 1 sun irradiation (100 mW/cm2), using ABET 150LS
Xenon lamps as a source of light provided with a AM1.5 filter.
The irradiation intensity was calibrated with a Si photodiode.
A selective hydrogen Clark electrode (model H2-NP, Unisense)
was used as a detector to quantify the amount of molecular
hydrogen evolved from the (photo)electrodes during the
photoelectrochemical experiment in a standard twocompartment cell separated with a frit.

Materials and methods

Preparation of the multi-layered photoabsorber

(NH4)2[MoS4] and Indium beads were supplied from SigmaAldrich. Mo/Cu/Sn/Cu/Zn stacks were sputtered using
elemental targets of Mo (99.95%), Cu (99.99%), Zn (99.99%) and
Sn (99.99%) supplied by PhotonExport. Ge nanolayers were
thermally evaporated from Ge chips (99.999%) supplied by
Sigma-Aldrich. The reactive annealings were performed using
Se powder (99.999%, 200 mesh) and Sn powder (99.995%, 100
mesh) from Alfa Aesar. For the preparation of the CdS,
Cd(NO3)2, SC(NH2)2, sodium citrate and ammonia from Alfa
Aesar. The chemical etchings of the absorbers were performed
using a (NH4)2S (22% (w/w)) aqueous solution prepared from
commercial (NH4)2S (44% (w/w)) supplied by Alfa-Aesar.

The CZTSe films were prepared by a sequential process onto
Mo-coated soda-lime glass substrates. For this purpose, Zn-rich
and Cu-poor Cu/Sn/Cu/Zn stacks (Cu/(Zn+Sn) = 0.75 and Zn/Sn =
1.10 determined with calibrated X-ray fluorescence
(Fischerscope XVD)) were deposited using DC magnetron
sputtering (Alliance AC450). Additionally, very thin Ge layers
were thermally evaporated prior to the stack deposition (10 nm
Ge on the Mo) and on top of the metallic stack (5 nm Ge), as
reported elsewhere.32 Then, the whole precursor stack was
subsequently annealed in a Se + Sn atmosphere (100 mg of Se
and 5 mg of Sn), using graphite boxes (69 cm3 in volume) in a
conventional tubular furnace (Hobersal). The selenization was
performed in a two-step process: the first one at 400 °C (heating

2 | J. Name., 2012, 00, 1-3

This journal is © The Royal Society of Chemistry 20xx

Please do not adjust margins

Please do not adjust margins
Journal Name

ARTICLE

rate of 20 °C/min) for 30 minutes and 1.5 mbar of argon
pressure, followed by a second step at 550 °C (heating rate of
20 °C/min) for 15 minutes and 1 bar of Ar pressure, with a
natural cooling down to room temperature. To complete the
devices, a CdS buffer layer (50 nm) was grown by chemical bath
deposition, preceded by a chemical etching in (NH4)2S solution
in order to remove possible secondary phases and to passivate
the absorbers surface. Just after CdS growth, the multilayered
photoabsorber were completed by DC-pulsed sputtering
deposition of ZnO (50 nm) and In2O3-SnO2 (ITO, 200 nm) as
transparent conductive window layer (Alliance CT100).
(Photo)electrodeposition of the MoS2 electrocatalyst
An electrochemical bath containing 30mM of (NH4)2[MoS4] and
0.1M Na2SO4 as supporting electrolyte in MilliQ water was used
for both electrodeposition on commercially available ITO
glasses and photoelectrodeposition on the CZTSe absorbers.
The solution was kept under argon bubbling during the whole
experiment to prevent the formation of molybdenum oxides
and used as prepared. In a standard experiment, the
potentiostat was set to galvanostatic mode at j = -7.5·10-5 A/cm2
to cathodically electroplate the amorphous MoS2. The thickness
of the catalyst layer is controlled by the time and charge passed
through the system during the electrodeposition. In a typical
experiment, a loading corresponding to a charge of 0.05 C/cm2
was performed, which corresponds to a 200 nm of thickness
approximately (confirmed by SEM). After the electrodeposition,
the electrodes were cleaned with MilliQ water and dried under
air flow.
An analogous methodology was used for the CZTSe electrodes
functionalization. Before making the electrical contact with
indium as soldering material, the four edges of the
photoelectrode were scratched, removing the material to
expose the molybdenum back contact (approximately 0.5-1 mm
of wideness). The photoelectrochemical deposition was done
under irradiation with a light intensity calibrated to 2 suns due
to the high concentration of the dark red coloured (NH4)2[MoS4]
solution.
Photo-electrochemical experiments

In a custom-made one-compartment, three-electrode
configuration cell made of Teflon with a 0.5 cm2 window, the
working CZTSe electrode, platinum counter electrode and
reference electrode were placed close together inside the
buffered solution. Prior to the electrochemical experiment, the
solution is degassed by bubbling argon for 20 min. LSV, CP
and/or bulk electrolysis experiments were performed under
continuous light irradiation calibrated to 1 sun (100 mW/cm2).
For detection of hydrogen gas, a two-compartment cell made of
glass and separated by a frit was used. The working CZTSe
electrode was placed together with the AgCl/Ag reference
electrode in one compartment, while the platinum mesh was
placed in the second compartment facing as close to the frit as
possible. After degassing the solution with argon flow, the CP
experiment is started and the evolved hydrogen gas monitored
with a Clark sensor in the headspace of the cathodic
compartment.

Results en discussion
Preparation of CZTSe absorber layer

The photoabsorber consists of a carefully designed multilayered
configuration
that
can
be
formulated
as
Mo/MoSe2/Cu2ZnSnSe4/CdS/ZnO/ITO (where ITO is indium tin
oxide). The n-type CdS interlayer has been proved to be crucial
for device performance by creating a very efficient p-n junction
with the p-type CZTS(Se) type of absorbers.26(REF) On the other
hand, ZnO fills the defects of the CdS layer that is prepared by
chemical bath deposition as described below. It also gives a
uniform support for the deposition of the ITO protecting layer.
Detailed description of the preparation of the samples is given
in the supporting information.32,33,34 Briefly, it consists of
selenization of Cu/Zn/Sn stacks, that have been previously
deposited by sputtering on molybdenum-coated soda-lime
glass (also prepared by sputtering). The resulting
Mo/MoSe2/Cu2ZnSnSe4 samples are dipped in a chemical bath
for n-type CdS deposition, followed by the sputtering deposition
of ZnO and ITO in order to complete the electrode
configuration. Finally, the device is subjected to a soft postdeposition annealing at 250 °C for 15 minutes on a hotplate in
air atmosphere to improve cathode performance. The full
photoabsorber Mo/MoSe2/Cu2ZnSnSe4/CdS/ZnO/ITO will be
regarded as CZTSe hereafter. Powder x-ray diffraction,
resonance Raman spectroscopy and transmission electron
microscopy analysis of the prepared samples confirmed the
crystallinity and purity of the Cu2ZnSnSe4 photoabsorber layer
with no secondary phases (Figures SX and SY in the supporting
information). Composition, morphology and grain size were
also as previously reported.29,33 Optical band gap of the four
absorbing layers CZTSe/CdS/ZnO/ITO were measured using UVVIS diffuse reflectance spectroscopy (DRS) and external
quantum efficiency (EQE) experiments obtaining the expected
values according to the literature (Figure S1 and Table S1 in the
SI).35,36 In Figure S1, a representative Tauc plot obtained from
DRS spectrum shows seven transitions corresponding to CZTSe
(pure, Eg =0.94 eV), CZTSSe (Eg =1.35 eV), two for CdS (Eg =2.12.0 eV), two for ZnO/ZnS (Eg =3.0-3.4 eV)37 and one for ITO (Eg =
2.9 eV). Due to the post deposition annealing, some S2- migrate
from CdS to ZnO and CZTSe layers, doping the absorber, as well
as Cu+ ions doping CdS layer, phenomena that have proven to
be beneficial for the device efficiency.38 Thus, CdS and ZnO
feature double peaks (green dashed lines in Figure S1), very
likely due the shallow doping during the post deposition
annealing, thus corresponding one to the pure material and the
other to the doped material
Optimization of the MoS2 electrocatalyst layer

Molybdenum sulfide MoSx (x = 2 or 3) can be prepared by
electrodeposition from the precursor (NH4)2[MoS4] into
conductive substrates in three different ways: i) cathodic
deposition, ii) anodic deposition and iii) cyclic deposition.15-17
The last two approaches were ruled out due to the required

This journal is © The Royal Society of Chemistry 20xx

J. Name., 2013, 00, 1-3 | 3

Please do not adjust margins

Please do not adjust margins
ARTICLE

Journal Name

xidation potential, which could lead to the chemical
decomposition of the CZTSe absorber during the
photoelectrochemical deposition. Thus, in this work, we
focused on the cathodic deposition approach, which we
performed from the precursor (NH4)2[MoS4] (30 mM) on ITO in
galvanostatic mode (j = -7.5·10-5 A/cm2). The process was done
under weak argon bubbling to avoid the formation of
molybdenum oxides. The loading of the molybdenum sulfide
deposited on the surface of ITO was controlled by the total
amount of charge passed through the electrochemical system.
The performance of the resulting MoS2-coated ITO electrodes
(ITO-MoS2 hereafter) towards HER in pH 2 phosphate buffer
solution with different loadings of the electrocatalyst is
summarized in Table 1 (see also Figure S2-S3 in the SI). The
overpotential required to reach -1 mA/cm2 decreases with
increasing amount of MoS2 while the current density at E = -0.3V
vs RHE increases with higher amount of catalyst. The MoS2 films
are not transparent and therefore are expected to influence the
light absorbing process once deposited on the CZTSe
photoabsorber (Figure S4). Thus, although the best performing
electrode is that with a charge loading of 0.1 C/cm2, it is also the
electrode with a lower degree of transmittance (10-55 % in the
range of 350-800nm). Therefore, the ITO-MoS2 electrode with
Q = 0.05 C/cm2 with E = -0.175 V at j =-1 mA/cm2 and j = -9.46
mA/cm2 at E = -0.3 V was selected as a good compromise
between high activity and high transparency (Figure 2).
The ITO-MoS2 electrodes show a reduction process at potentials
between -0.025 and -0.150 V vs RHE, before the hydrogen
evolution reaction, attributed to the reduction of the MoS3 to
MoS2 during the first Linear Sweep Voltammetry (LSV) scan.15,17
This reduction wave was not observed again in subsequent
scans of the LSV experiments on the same sample (Figure 2,
compare red and blue traces). The activity of the ITO-MoS2
electrodes is highly dependent on the pH of the electrolyte
solution, showing drastic decrease of the cathodic current
associated to the HER in pH values higher than 2 (Figure S5).
ITO-MoS2 electrodes tested in pH higher than 10 were unstable
and the catalyst peeled off very quickly from the electrode
surface. This result is in sharp contrast to other reported
amorphous MoS2 based electrodes for HER in the literature
which are very active in extreme pH values (0 and 14).39 We
attribute these results to the lower adhesion of our MoS2 to the
ITO substrate.

Figure 2. LSV of ITO-MoS2 with a Q = 0.05 C/cm2 loading at pH 2. Counter
electrode: Pt mesh. Reference electrode: AgCl/Ag. Scan rate: 5 mV/s.
Experiments recorded under stirring conditions. iR drop corrected at 85%.

was ran for 1 hour, showing no significant loss of activity (Figure
3). The averaged overpotential required to sustain -10 mA/cm2
for the whole experiment was -328 mV vs RHE, in good
agreement with the LSV results in Figure 2 and indicating that
the scan rate used to run the LSV was close to stationary
conditions. In order to check the Faradaic efficiency (FE), the
experiment
was
repeated
in
a
two-compartment
electrochemical cell, monitoring the amount of hydrogen
evolved from the cathode by using a Clark electrode sensor,
obtaining 100% efficiency (Figure 3, inset). The activity of the
electrode after this test was also checked by LSV without
remarkable losses of current densities (Figure 2, compare blue
and green traces). Long-term CP tests up to 10h were ran to
assess the stability of the ITO-MoS2 electrodes. The catalytic
activity started to decrease linearly after 1 hour of electrolysis
until the 5th hour. At this time the required potential to run j = 10 mA/cm2 is E = -0.6 V as opposed to the E = -0.33 V required
during the first hour (Figure S6). After 5h, the HER activity of the

Table 1. Overpotential (at j = -1 mA/cm2) and current density (at η = 300 mV)
values of the ITO-MoS2 samples with different loadings of HER catalyst.

Charge (C/cm2)

E for j = -1 mA/cm2

j at Evs RHE = -0.3 V

0.01
0.02
0.05
0.1

-0.205 V
-0.195 V
-0.175 V
-0.165 V

-7.37 mA/cm2
-8.45 mA/cm2
-9.46 mA/cm2
-9.96 mA/cm2

LSV experiments in Figure 2 show that ITO-MoS2 electrodes with
a 0.05 C/cm2 loading feature an overpotential of -175 mV and 330 mV vs RHE to reach -1 mA/cm2 and -10 mA/cm2 respectively
at pH 2. A chronopotentiometry (CP) experiment at -10 mA/cm2

Figure 3. CP of ITO-MoS 2 with a Q = 0.05 C/cm2 loading in pH 2 at -10mA/cm2
in a two-compartment electrochemical cell. Counter electrode: Pt mesh.
Reference electrode: AgCl/Ag. iR drop corrected at 85%. Inset: Amount of
molecular hydrogen evolved during CP experiments (red line) and the
calculated from the charged passed (black dotted line).

4 | J. Name., 2012, 00, 1-3

This journal is © The Royal Society of Chemistry 20xx

Please do not adjust margins

Please do not adjust margins
Journal Name

ARTICLE

electrode decreases drastically requiring even lower potentials
to sustain the selected current density.
In order to get insights into the deactivation processes during
electrolysis, the morphology of the electrodes before and after
catalysis were analyzed by Scanning Electronic Microscopy
(SEM). Top view SEM images confirm homogeneous and
continuous MoS2 films covering completely the ITO surface
after deposition of the electrocatalyst (Figure 4A and Figure S7).
Only small cracks in the amorphous MoS2 layer can be observed,
that are likely produced by the mechanical stress due to the
expansion of hydrogen bubbles formed during the
electrodeposition process. After 1h of electrolysis, the SEM
images show small regions where the initial cracks start to get
wider, starting to expose the ITO underlayer to the electrolyte
(Figure 4B). The cracks become even larger after 10h of
experiment when the ITO is mostly exposed and only little
amount of the MoS2 layer is left accounting for the loss in
activity at long electrolysis time (Figure 4C). Interestingly, the
shape of some of the cracks suggest a peeling due to mechanical
stress from a gas bubble rather than a chemical process (Figure
S8). It is important to notice that the images of the damaged
regions shown in Figure 4B and 4C were chosen from the most
affected zones and they are not representative of the whole
surface, where the ITO-MoS2 mainly looks intact (Figure S9).

Figure 4. Top view SEM images of ITO-MoS 2 electrodes as prepared (A), after
1h of CP (B) and after 10h of CP (C).

MoS2 modified CZTSe and HER photoelectrochemical performance

The CZTSe absorbers were functionalized with MoS2 by cathodic
photoelectrodeposition onto the ITO top layer in galvanostatic
mode (j = -7.5·10-5 A/cm2) under 2 sun illumination due to the
strong absorption of light by the electrochemical bath
containing 30 mM of the red precursor (NH4)2[MoS4]. This
methodology resulted to be a very useful way to get consistent
rate of growth of the catalyst film despite the fact that the
actual applied potential to make the HER catalyst layer growth
is different due to the slightly distinct photovoltage generated
by different samples of CZTSe. This is exemplified in Figure S10,
which shows the E-t curves of MoS2 deposition from
representative ITO and CZTSe (photo)electrodes. The catalyst
deposition using this method is very reproducible, giving
identical E-t curves for ITO samples and only slight differences
in E-t curves for the CZTSe samples.
The photoelectrochemical activity towards HER of the CZTSeMoS2 electrodes was tested in pH 2 phosphate buffer solution
under 1 sun illumination. Analogous to the ITO-MoS2
electrodes, the CZTSe-MoS2 photocathodes also showed a
reduction process before the HER catalysis assigned to the MoS3
to MoS2 transformation at potentials from 0.4 to 0.25 V vs RHE
during the first LSV scan (Figure S11). As depicted in the red line
of Figure 5, CZTSe covered with MoS2 features an onset

potential and overpotential of +254 mV and +100 mV to achieve
-1 mA/cm2 and -10 mA/cm2 respectively (see also Figure S12).
No dark-current was observed in the range of potentials used as
indicated with the red dashed line in Figure 5. Due to the high
overpotential required to produce molecular hydrogen for raw
CZTSe photocathode no photocurrent was detected below 0 V
vs RHE without the MoS2.
By comparing the LSV of the ITO-MoS2 and the CZTSe-MoS2
electrodes we can observe the onset potential for HER is
anodically shifted by approximately +430 mV (Figure 5). This
value is close to the Open Circuit Potential (Voc = 464 mV)
obtained from the J-V curve of an equivalent photovoltaic
device, indicating minimum losses in photovoltage (Vph) by the
presence of the MoS2 layer, and suggesting we are at the limit
of the capabilities of the photoabsober (Figure S13). In addition,
the similarity of the catalytic slopes of the two electrodes
indicates that the same species (MoS2) is responsible for the
chemical reaction (compare blue and red lines in Figure 5).
Remarkably, the high current densities obtained for our CZTSeMoS2 electrode are amongst the best performing
photocathodes for the hydrogen evolution reaction described
in the literature (Table S2). While the sulfur containing kesterite
based photoelectrodes (CZTS) give higher photovoltages due to
their larger band gap, their current densities are significantly
lower than those of the CZTSe described in this work (below -10
mA/cm2 for the former as opposed to the -18 mA/cm2 for the
latter at E = 0 V vs RHE.26-28,40,41 Another important advantage
of our photoelectrodes is the use of a more readily available
MoS2 catalyst as opposed to the Pt based CZTS(Se) examples
reported to date.29 In this regard, previous works using similar
electrocatalytic MoS2 layers but different semiconductor
materials such as Cu2O show lower performance in terms of
current densities and stability.39 Only those photocathodes that
use p-Si modified with MoS2 give comparable results.21-24
CP experiments at -10 mA/cm2 were ran in a two-compartment
cell under 1 sun illumination, obtaining an averaged potential of
+86 mV vs RHE for one hour with FE of 100% and no significant
loss in activity (Figure 6 and S14). A LSV experiment of the

Figure 5. LSV of ITO-MoS2 and CZTSe-MoS2 with the same loading of catalyst.
Scan rate: 5 mV/s. LSV recorded under stirring conditions. iR drop corrected at
85%.

This journal is © The Royal Society of Chemistry 20xx

J. Name., 2013, 00, 1-3 | 5

Please do not adjust margins

Please do not adjust margins
ARTICLE

Journal Name

photocathode after CP shows a slight decrease in activity
requiring an overpotential of +233 mV and +66 mV to get -1
mA/cm2 and -10 mA/cm2 respectively (Figure S12, compare red
and blue lines). Longer CP experiments show that our
photocathode is able to produce molecular hydrogen at -10
mA/cm2 for 2h without any significant loss in activity (i.e. 1.2
μL(H2)/cm2·s) and able to perform HER below the
thermodynamic potential for 3.5 hours (Figure S16). After the
first two hours, the potential required to sustain -10 mA/cm2
increases linearly, increasing the potential by ca. 44 mV/h, a
similar trend that was observed for the ITO-MoS2 electrodes.
Interestingly, the MoS2 layer seems to last longer when it is on
the surface of the photoelectrode (2h) than when it is on the
ITO electrode (1h). This result might be attributed to the distinct
nature of the ITO surface when it is deposited on top of the
CZTSe/CdS/ZnO photoabsorber compared to when it is
deposited directly onto the glass, resulting in better adhesion to
the surface in the former case.
Due to the low stability of the ITO layer directly immersed in pH
2,42-44 it was not possible to measure more than two scans with
raw CZTSe because of the fast degradation of the
photoabsorber. Thus, the amorphous MoS2 layer not only
reduces the overpotential for HER, but also protects the ITO and
the absorber layers from the electrolyte, preventing its
corrosion and dissolution. This is also supported by aging
experiments consisting of immersing the CZTSe-MoS2
photoelectrodes in pH 2 for 5 hours, showing no change in
activity (Figure S17). Therefore, the loss of activity is attributed
to the depletion of electroactive material during catalysis most
likely due to mechanical stress from the evolution of H2 bubbles,
as suggested also for the parent ITO-MoS2 electrodes.
SEM images of the CZTSe-MoS2 electrodes show the
homogeneity and compactness of the top layer before and after
catalytic tests, despite the small cracks already mentioned for
the analogous ITO-MoS2 (Figure 7, A and D). While the top-view
SEM images did not show any change in the morphology, crosssections of the photocathodes as prepared and after 1 hour of
CP revealed that the MoS2 layer became thinner (Figure 7, D and
E) from 150-250 nm to 80-130 nm. The electrocatalyst layer
became even thinner with the extended CP, reaching 60-90nm

Figure 6. Left) CP of CZTSe-MoS2 in pH 2 at -10 mA/cm2 in a two-compartment
photoelectrochemical cell. Dashed black line indicates the averaged potential
during the CP experiment (E = +86 mV). Right) Schematic drawing of the CZTSe
photoelectrode where a-MoS2 is amorphous molybdenum disulfide.

Figure 7. Top view and cross section of SEM images of CZTSe-MoS2 as prepared
(A and D), after 1h of CP (B and E) and after 5h (C and F). In cross-section images,
MoS 2 layer was coloured with blue.

of thickness after 5 hours. No peeled regions were found during
the surface analysis as opposed to what was observed for the
ITO-MoS2 electrodes suggesting that the mechanical stability in
this case is better. On the other hand, the other layers
(Mo/MoSe2/CZTSe/ZnO/ITO) look intact, highlighting once
more the stabilizing role of the MoS2 electrocatalytic layer.

Conclusions
A simple and easily scalable method based on galvanostatic
(photo)deposition of MoS2 from (NH4)2[MoS4] has been
developed giving excellent and reproducible results on the
surface of conductive supports (ITO) or on the surface of a
multilayered absorber based on p-type kesterite CZTSe
(Mo/MoSe2/Cu2ZnSnSe4/CdS/ZnO/ITO).
The resulting CZTSe-MoS2 photoelectrodes show impressive
HER catalytic performance at pH 2 with current densities up to
-18 mA/cm2 at E = 0 V vs RHE and sustained catalysis at +86 mV
vs RHE at j = 10 mA·cm-2. A comparison with its analogous dark
electrode ITO-MoS2 allows us to calculate a photovoltage of 430
mV. This value is in good agreement with the VOC = 464 mV of a
photovoltaic device made of the same absorber, indicating that
our system takes advantage of the full photoabsorption
capabilities of the CZTSe material. In addition, the stability of
the CZTSe-MoS2 photocathodes has been studied in detail,
finding that it is able to perform HER for two hours without any
significant loss of activity. Moreover, it is able to keep the
hydrogen production for more than 3.5 hours below the
thermodynamic potential (E = 0 V vs RHE). Longer reaction times
induce loss of the electroactive layer caused by the expansion
of microbubbles on the surface of the photocathode as
demonstrated by SEM images. A similar phenomenon has been
observed for the analogous dark ITO-MoS2 electrodes. In sharp
contrast, the photoactive CZTSe layer remains intact,
highlighting the key role of the MoS2 film in isolating it from the
electrolyte. All these results allow us to conclude that,
amorphous MoS2 obtained by (photo)electrodeposition is a
very good catalyst for HER, inexpensive and easy to prepare that
not only reduces the overpotential for HER but also protects the
sensitive photoactive underlayers of the photocathode from
the corrosive catalytic media. According to these findings,
future endeavors should be focused on improving the adhesion
between the electrocatalyst MoS2 layer and the photoabsorber.

6 | J. Name., 2012, 00, 1-3

This journal is © The Royal Society of Chemistry 20xx

Please do not adjust margins

Please do not adjust margins
Journal Name

ARTICLE

Conflicts of interest
“There are no conflicts to declare”.

Acknowledgements
Support from MINECO, FEDER, AGAUR and H2020 Programme
are gratefully acknowledged through grants CTQ2016-80058-R,
CTQ2015-73028-EXP, SEV 2013-0319, ENE2016-82025-REDT,
CTQ2016-81923-REDC,
2017-SGR-1631,
2017-SGR-862,
ENE2016-80788-C5-1-R, ENE2017-87671-C3-1-R, BES-2014068533, and H2020-NMBP-03-2016-720907.

Notes and references
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

BP Statistical Review of World Energy. Statistical Review of
World Energy June 2018; BP p.l.c., 1 St James’s Square,
London, SW1Y 4PD, UK, 2018.
IEA. Global Energy & CO2 Statuts Report; 2018.
J. Barber, Chem. Soc. Rev. 2009, 38, 185.
N. S. Lewis, Science 2016, 351, aad1920.
M. G. Walter, E. L. Warren, J. R. Mckone, S. W. Boettcher, Q.
Mi, E. A. Santori, N. S. Lewis, Chem. Rev. 2010, 110, 6446.
M. S. Prévot, K. Sivula, J. Phys. Chem. C 2013, 117, 17879.
S. D. Tilley, Adv. Energy Mater. 2019, 9, 1802877.
K. Sivula, R. van de Krol, Nature Rev. Mater. 2016, 1, 15010.
M. Saurat, S. Bringezu, J. Ind. Ecol. 2008, 12, 754A.
E. Alonso, E.; F. R. Field, R. E. Kirchain, Environ. Sci. Technol.
2012, 46, 12986.
I. Chorkendorff, I. K. P. JørgensenJ. H. Nielsen, B. Hinnemann,
B. P. G. Moses, J. K. Nørskov, S. Horch, J. Bonde, J. Am. Chem.
Soc. 2005, 127, 5308.
G. Li, D. Zhang, Q. Qiao, Y. Yu, D. Peterson, A. Zafar, R. Kumar,
S. Curtarolo, F. Hunte, S. Shannon, Y. Zhu, W. Yang, L. Cao, J.
Am. Chem. Soc. 2016, 138, 16632.
Z. He, W. Que, Appl. Mater. Today 2016, 3, 23.
I. Chorkendorff, J. Bonde, K. P. Jorgensen, S. Horch, T. F.
Jaramillo, J. H. Nielsen, Science 2007, 317, 100.
H. Vrubel, X. Hu, ACS Catal. 2013, 3, 2002.
D. Merki, S. Fierro, H. Vrubel, X. Hu, Chem. Sci. 2011, 2, 1262.
C. G. Morales-Guio, X. Hu, Acc. Chem. Res. 2014, 47, 2671.
A. Albu-Yaron, C. Lévy-Clément, A. Katty, S. Bastide, R. Tenne,
Thin Solid Films 2000, 361, 223.
E. A. Ponomarev, G. Hodes, Thin Solid Films 1996, 280, 186.
Y. Huang, R. J. Nielsen, W. A. Goddard, M. P. Soriaga, J. Am.
Chem. Soc. 2015, 137, 6692.
L. A. King, T. R. Hellstern, J. Park, R. Sinclair, T. F. Jaramillo, ACS
Appl. Mater. Interfaces 2017, 9, 36792.
M. J. Shearer, S. Jin, Q. Ding, D. Liang, M. Cabán-Acevedo, A.
S. Daniel, C. R. English, F. Meng, R. J. Hamers, J. Am. Chem.
Soc. 2014, 136, 8504.
C. G. Morales-Guio, K. Thorwarth, B. Niesen, L. Liardet, J.
Patscheider, C. Ballif, X. Hu, J. Am. Chem. Soc. 2015, 137, 7035.
K. C. Kwon, S. Choi, K. Hong, C. W. Moon, Y.-S. Shim, D. H. Kim,
T. K. W. Sohn, J.-M. Jeon, C.-H. Lee, K. T. Nam, S. Han, S. Y. Kim,
H. W. Jang, Energy Environ. Sci. 2016, 9, 2240.
Dhawale, D. S.; Ali, A.; Lokhande, A. Sustain. Energy Fuels
2019, 3, 1365.
D. Yokoyama, T. Minegishi, K. Jimbo, T. Hisatomi, G. Ma, M.
Katayama, J. Kubota, H. Katagiri, K. Domen, Appl. Phys.
Express 2010, 3, 2.
F. Jiang, Gunawan, T. Harada, Y. Kuang, T. Minegishi, K.
Domen, S. Ikeda, J. Am. Chem. Soc. 2015, 137, 13691.

28 X. Wen, W. Luo, Z. Guan, W. Huang, Z. Zou, Nano Energy 2017,
41, 18.
29 C. Ros, T. Andreu, S. Giraldo, V. Izquierdo-Roca, E. Saucedo, J.
R. Morante, ACS Appl. Mater. Interfaces 2018, 10, 13425.
30 S. Chen, L. W. Wang, Chem. Mater. 2012, 24, 3659.
31 R. Haight, A. Barkhouse, O. Gunawan, B. Shin, M. Copel, M.
Hopstaken, D. B. Mitzi, Appl. Phys. Lett. 2011, 98 (25), 253502.
32 S. Giraldo, E. Saucedo, M. Neuschitzer, F. Oliva, M. Placidi, X.
Alcobé, V. Izquierdo-Roca, S. Kim, H. Tampo, H. Shibata, A.
Pérez-Rodríguez, P. Pistor, Energy Environ. Sci. 2018, 11, 582.
33 M. Neuschitzer, E.-R. Moises, M. Guc, J. Prieto, S. Giraldo, I.
Forbes, P.-R. Alejandro, E. Saucedo, J. Mater. Chem. 2018, 6,
11759.
34 S. Giraldo, M. Neuschitzer, T. Thersleff, S. López-Marino, Y.
Sánchez, H. Xie, M. Colina, M. Placidi, P. Pistor, V. IzquierdoRoca, V.; K. Leifer, A. Pérez-Rodríguez, E. Saucedo, Adv. Energy
Mater. 2015, 5, 1501070.
35 K. Kaur, N. Kumar, M. Kumar, J. Mater. Chem. A 2017, 5, 3069.
36 S. Li, S.; C. Ghinea, T. J. M. Bayer, M. Motzko, R. Schafranek, A.
Klein, J. Phys. Condens. Matter 2011, 23, 33.
37 A. Torabi, V. N. Staroverov, J. Phys. Chem. Lett. 2015, 6, 2075.
38 M. Neuschitzer, Y. Sanchez, T. Olar, T. Thersleff, S. LópezMarino, F. Oliva, M. Espindola-Rodríguez, H.; Xie, M. Placidi,
V. Izquierdo-Roca, I. Lauermann, K. Leifer, A. Pérez-Rodríguez,
E. Saucedo, Chem. Mater. 2015, 27, 5279.
39 C. G. Morales-Guio, L. Liardet, M. T. Mayer, S. D. Tilley, M.
Grätzel, X. Hu, Angew. Chem. Int. Ed. 2015, 54, 664.
40 K. Wang, D. Huang, L. Yu, H. Gu, S. Ikeda, F. Jiang, J. Colloid
Interface Sci. 2019, 536, 9.
41 L. Rovelli, S. D. Tilley, K. Sivula, ACS App. Mater. Interfaces
2013, 5, 8018.
42 M. Senthilkumar, J. Mathiyarasu, J. Joseph, K. L. N. Phani, V.
Yegnaraman, Mater. Chem. Phys. 2008, 108, 403.
43 J. D. Benck, B. A. Pinaud, Y. Gorlin, T. F. Jaramillo, PLoS One
2014, 9, 10.
44 S. Geiger, O. Kasian, A. M. Mingers, K. J. J. Mayrhofer, S.
Cherevko, Sci. Rep. 2017, 7, 1.

This journal is © The Royal Society of Chemistry 20xx

J. Name., 2013, 00, 1-3 | 7

Please do not adjust margins

“This document is the Accepted Manuscript version of a Published Work that appeared in final form in Journal of Materials
Chemistry A, copyright © The Royal Society of Chemistry 2019 after peer review and technical editing by the publisher. To
access the final edited and published work see:

https://pubs.rsc.org/en/content/articlelanding/2019/ta/c9ta08818k#!divAbstract
Table of Contents

Key words
Hydrogen evolution, photocathode, kesterite, molybdenum disulphide, photoelectrochemical cell

Please do not adjust margins