Homogeneous electrochemical deposition of in on a Cu-covered Mo substrate for fabrication of efficient solar cells with a CuInS2 photoabsorber

Article abstract of DOI:10.1016/j.electacta.2012.06.103

Electrochemical deposition of indium (In) on a copper-covered molybdenum-coated glass substrate from several acidic InCl₃ solutions was studied for fabrication of CuInS₂-based solar cells. When In was deposited using a simple acidic InCl₃ solution at -0.80 V (vs. Ag/AgCl), island-shaped growth was observed, whereas a homogeneous In film was obtained from InCl₃ solution containing citric acid and sodium citrate at -0.98 V (vs. Ag/AgCl). Electrochemical and structural analyses revealed that the citric acid additive had a function for smoothing the surface of the In deposit. The mixing with sodium citrate induced appreciable inhibition of H₂ evolution during the In deposition, leading to high current efficiency of >90%. The CuInS₂ film derived from the homogeneous In had a uniform thickness with a smooth surface, while the CuInS₂ film obtained from the island-shaped In deposit showed a large variation in thickness with recessed areas. The CuInS₂ film derived from the homogeneous In was showed better photoelectrochemical response than that of the film fabricated from the island-shaped In. As expected from these differences, the solar cell with an Al:ZnO/CdS/CuInS₂/Mo structure derived from the homogeneous In film showed the best conversion efficiency of 7.8% with relatively high reproducibility.

Full text of DOI:10.1016/j.electacta.2012.06.103

Electrochimica Acta 79 (2012) 189–196
Contents lists available at SciVerse ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Homogeneous electrochemical deposition of in on a Cu-covered Mo substrate for
fabrication of efficient solar cells with a CuInS2 photoabsorber
Sun Min Lee, Shigeru Ikeda∗, Yasunari Otsuka, Wilman Septina, Takashi Harada, Michio Matsumura
Research Center for Solar Energy Chemistry, Osaka University, 1-3 Machikaneyama, Toyonaka, Osaka 560-8531, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Electrochemical deposition of indium (In) on a copper-covered molybdenum-coated glass substrate
from several acidic InCl3 solutions was studied for fabrication of CuInS2-based solar cells. When In
was deposited using a simple acidic InCl3 solution at −0.80 V (vs. Ag/AgCl), island-shaped growth was
observed, whereas a homogeneous In film was obtained from InCl3 solution containing citric acid and
sodium citrate at −0.98 V (vs. Ag/AgCl). Electrochemical and structural analyses revealed that the citric
acid additive had a function for smoothing the surface of the In deposit. The mixing with sodium citrate
induced appreciable inhibition of H2 evolution during the In deposition, leading to high current efficiency
of >90%. The CuInS2 film derived from the homogeneous In had a uniform thickness with a smooth surface,
while the CuInS2 film obtained from the island-shaped In deposit showed a large variation in thickness
with recessed areas. The CuInS2 film derived from the homogeneous In was showed better photoelec-
trochemical response than that of the film fabricated from the island-shaped In. As expected from these
differences, the solar cell with an Al:ZnO/CdS/CuInS2/Mo structure derived from the homogeneous In
film showed the best conversion efficiency of 7.8% with relatively high reproducibility.
Received 12 March 2012
Received in revised form 28 June 2012
Accepted 28 June 2012
Available online 6 July 2012
Keywords:
Electrochemical deposition
Homogeneous In film
Organic additives
H2 evolution
Solar cell properties
© 2012 Elsevier Ltd. All rights reserved.
1. Introduction
Among the above-mentioned non-vacuum methods, electro-
chemical deposition of the CuInS2 film has been shown to be
Copper indium disulfide (CuInS2) is a ternary semiconductor
compound that has been studied as an absorber for thin film solar
cells, due to its optimal band gap value for sunlight radiation
(1.5 eV) [1], sufficiently large absorption coefficient of 3 × 105 cm−1
desirable because of its advantages, including low equipment cost,
negligible waste of chemicals with utilization efficiencies close
to 100%, possible formation of a compact film required for solar
cell application, scalability, and manufacturability of a large-area
polycrystalline film. For CuInS2 film deposition, a sequential route
consisting of electrochemical deposition of a Cu/In bilayer followed
by sulfurization has been applied as the most usual process [19–27].
In this process, the latter sulfurization step is performed by using
H2S gas or sulfur vapor in much the same way as the above-
mentioned sputtering technique. Very recently, a solar cell based on
electrochemically fabricated CuInS2 films has achieved conversion
efficiency close to that of a cell fabricated by the vacuum process
(11%) [27], indicating that the electrochemical deposition method
has sufficient potential for future practical applications. For the
further improvement of the conversion efficiency of the solar cell
based on the electrochemically prepared CuInS2 film, there is room
for optimizing structures of metal staked precursors.
One of the critical problems of electrochemical deposition of a
Cu/In film is the difficulty in obtaining a homogeneous In layer,
i.e., the In layer tends to form an island-shaped morphology using
typical electrochemical deposition solutions based on both chlo-
ride and sulfate solutions [28,29]. The In island formation would
be detrimental because of inductions of thickness and composi-
tion variations of resulting CuInS2 films. Kinetic and mechanistic
studies have shown that uncontrollable nucleation and growth of
below 1000 nm [2], and efficient carrier mobility (2–20 cm2 v−1 s−1
)
[3,4]. The ideal substrate solar cell consisting of stacked layers of
the CuInS2 film, a thin buffer, and transparent conductive oxide
(TCO) on a Mo-coated glass (Mo/glass) substrate showed highest
efficiency of 11.4% [5,6]. The CuInS2 absorber in a solar cell is typi-
cally prepared by sequential deposition of Cu and In metallic layers
(i.e., a Cu/In bilayer) on the Mo/glass substrate by vacuum methods
such as sputtering and evaporation [7–9], followed by sulfurization
[5–11]. However, the use of these vacuum methods leads to high
equipment cost and significant losses of raw materials [12]. In order
to lower electricity-generating costs of solar cells to a level compet-
itive with commercial electricity generation, replacements of these
vacuum methods with facile non-vacuum methods are promising.
Hence, several non-vacuum methods, such as spraying [13–16] and
electrochemical deposition [17–27], have been reported in the lit-
erature.
∗
Corresponding author. Fax: +81 06 6850 6699.
E-mail address: sikeda@chem.es.osaka-u.ac.jp (S. Ikeda).
0013-4686/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.electacta.2012.06.103

190
S.M. Lee et al. / Electrochimica Acta 79 (2012) 189–196
In metal in such simple deposition solutions induced the forma-
tion of In islands [28,29]. These facts have motivated us to find an
electrochemical deposition condition to form a flat and homoge-
neous In layer on the Cu-deposited Mo/glass substrate. In order to
achieve this objective, we focused on evaluation of the effects of
organic additives on surface morphology of the thus-obtained In
layer. In this study, we employed citric acid and sodium citrate as
ideal organic additives to fabricate Cu/In bilayers in homogeneous
In distributions. Effects of the different Cu/In film morphologies
on structural, photoelectrochemical, and photovoltaic properties
of the CuInS2 films obtained from these Cu/In bilayers were also
studied.
2. Experimental
A Mo/glass substrate with sheet resistance of ca. 20 ꢀ/square
was used as a cathode for the electrochemical deposition of Cu/In
bilayers. Before the deposition, the Mo/glass substrate was pre-
cleaned by sonication in acetone, followed by immersion in 25 vol%
ammonia solution for 5 min to remove the molybdenum oxide
layer (MoOx) on the surface [30]. Cu/In bilayers were deposited
successively on the Mo/glass substrate by using a common three-
electrode cell with a Pt foil as a counter electrode and an Ag/AgCl
reference electrode under potentiostatic control using a Hokuto
Denko HSV-100 potentiostat–galvanostat. Temperature of all the
deposition solutions was kept at 24 1 ◦C by immersion in a
thermostated water bath. Deposition of the Cu layer was per-
formed at −0.4 V (vs. Ag/AgCl) in an aqueous solution containing
10 mmol dm−3 CuSO4 and 10 mmol dm−3 citric acid (pH 2.3) [26].
The amount of deposited Cu metal was controlled by the total
electric charge at 0.83 C cm−2 using a Hokuto Denko HF-201A
coulomb/ampere hour meter. For deposition of the In layer onto
the top of the Cu-deposited Mo/glass (Cu/Mo/glass), three kinds
of acidic solutions with pH adjusted to 2.2 by a concentrated HCl
solution were used in this study: one was a 12 mmol dm−3 InCl3
solution (In-bath(A)) [26], another was a 30 mmol dm−3 InCl3 solu-
tion containing 10 mmol dm−3 citric acid (In-bath(B)), and the other
consisted of 30 mmol dm−3 InCl3, 10 mmol dm−3 citric acid, and
35 mmol dm−3 sodium citrate (In-bath(C)) [31]. Applied potentials
for the In deposition were adjusted to −0.80 V (vs. Ag/AgCl) for In-
bath(A) and −0.98 V (vs. Ag/AgCl) for In-bath(B) and In-bath(C). The
total charge during the In deposition was fixed at 0.96 C cm−2, irre-
spective of the kind of In solution used. Amounts of deposited In
metals thus-formed were quantified by measuring the changes in
weights of the Cu/Mo/glass substrates before and after In deposi-
tion. As separate experiments, contributions of H2 evolution during
the In deposition were studied using aqueous HCl solutions (pH 2.2)
containing 10 mmol dm−3 citric acid and containing 10 mmol dm−3
citric acid and 35 mmol dm−3 sodium citrate (see below).
Fig. 1. A top view SEM image of an electrochemically deposited Cu film. Inset shows
the corresponding cross-sectional image.
patterns were measured using a Rigaku MiniFlex X-ray diffrac-
tometer (Cu K␣, Ni filter).
Photoelectrochemical properties of CuInS2 films were measured
in an aqueous solution containing 0.1 mmol dm−3 Eu(NO3)3 as an
electron acceptor at pH 4. A Pyrex electrolytic cell having a flat
window was used. The photocurrent response of the film was mea-
sured under potentiostatic control using a three-electrode system
with a Pt foil counter electrode and an Ag/AgCl reference electrode.
Transient photocurrents at potentials ranging from 0.2 V to −0.45 V
(vs. Ag/AgCl) were measured using the lock-in technique under
chopped illumination of 600 nm monochromatic light. Photocur-
rent external quantum efficiency (EQE) spectra at the potential of
−0.4 V (vs. Ag/AgCl) were also measured using the lock-in technique
under chopped illumination of monochromatic light (400–900 nm).
The number of incident photons was determined by an OPHIR Orion
Laser power meter equipped with a photodiode. All the photoelec-
trochemical measurements were performed under N2 purging.
For evaluation of solar cell properties of the CuInS2 films, the
films were processed to form an Al:ZnO/CdS/CIS/Mo/glass struc-
ture. On the CuInS2 films, a CdS buffer layer was deposited by
chemical bath deposition (CBD) [26], and an Al:ZnO window layer
was then deposited on the top of the CdS layer by radio frequency
(RF) magnetron sputtering [26,32]. Current density–voltage (J–V)
characteristics under simulated AM1.5 irradiation (100 mW cm−2
)
through the Al:ZnO window layer and those under dark of thus-
obtained solar cells were measured with a Bunkoh–Keiki CEP–015
photovoltaic measurement system.
Sulfurization of as-prepared Cu/In bilayer films to form CuInS2
films was performed by using a three-step temperature profile as
reported previously [26], with slight modifications. First, the Cu/In
bilayer films were heated to 110 ◦C in Ar, and then heating at this
temperature was continued for 60 min. In the second step, the film
treated at 110 ◦C was heated to 520 ◦C in Ar using a heating ramp
of 25 ◦C min−1. Finally, sulfurization of the film was performed by
introduction of H2S (5% in Ar) at 520 ◦C for 10 min, and then the
film was allowed to cool down to ambient temperature in Ar or H2S
(5% in Ar). Since all of the Cu/In bilayer films had relatively Cu-rich
compositions, the CuInS2 films thus-formed include a CuxS impu-
rity component: these films were treated with a 10% KCN solution
to remove the CuxS component [5].
3. Results and discussion
3.1. Effects of additives for In deposition
Fig. 1 shows a typical SEM image of the Cu film on Mo/glass.
The top view image shows that the surface of the film has a rugged
morphology composed of small lumps of several hundred nanome-
ters in size. The corresponding cross-sectional image of the Cu film
shown in inset of Fig. 1 shows the formation of a condensed film
at the bottom layer with a thickness of ca. 250 nm, while the upper
part has a bumpy structure, as expected from the surface mor-
phology. Since there is no appreciable crack or pinhole in the Cu
film, we employed this film as the substrate (i.e., the Cu/Mo/glass
substrate) for successive electrochemical deposition of In. Results
Surface and cross-sectional morphologies of the Cu, Cu/In, and
CuInS2 films prepared were examined using a Hitachi S-5000
FEG scanning electron microscope (SEM). X-ray diffraction (XRD)

S.M. Lee et al. / Electrochimica Acta 79 (2012) 189–196
191
Fig. 3 shows SEM images of In films formed on the Cu/Mo/glass
substrate from In-bath(A), In-bath(B), and In-bath(C) solutions. The
SEM image of the In film obtained from In-bath(A) shows inhomo-
geneous distributions of In deposits having many crevices between
their grains (Fig. 2a). As can be seen in Fig. 3d, the corresponding
cross-sectional SEM image shows a very rough structure: thick-
nesses of the film depend strongly on the area from ca. 200 nm to
ca. 800 nm, as expected from the surface morphology. On the other
hand, when the In-bath(B) solution containing an organic additive
of citric acid was used instead of In-bath(A), the morphology of
the surface of the In film thus-obtained was smooth (Fig. 3b). The
cross-sectional image of the film in Fig. 3e also indicates the occur-
rence of uniform growth to make a condensed In film with large
grains of over 1 ␮m in breath and 550–600 nm in thickness. A uni-
form In film with similar morphology was also obtained by using
In-bath(C), which contains citric acid and sodium citrate, as shown
in Fig. 3c and f. These results suggest that the citric acid additive
has a function to smooth the surface of the In film.
Fig. 2. Linear sweep voltammograms in (a) In-bath(A), (b) In-bath(B), and (c) In-
bath(C) obtained by using the Cu deposited Mo/glass substrate.
Fig. 4 shows XRD patterns of Cu/In bilayers on Mo/glass obtained
from In-bath(A) and In-bath(C) solutions. In addition to the intense
reflection derived from the Mo film, both samples showed several
diffraction peaks assignable to In, Cu, and a CuIn alloy. Significantly
weak reflections derived from Cu metal on both samples suggest a
less crystalline nature of the Cu layer. It is known that metallic In
easily diffuses into metallic Cu and forms intermetallic compounds
[25]. Thus, appreciable amounts of the CuIn alloy were formed on
both samples even at ambient temperature. Based on the fact that
reflections derived from the CuIn alloy appeared predominantly
in the XRD patterns of In-deposited samples after a short electro-
chemical deposition period (less than 0.3 C cm−2, data not shown),
CuIn alloy formation is preferable at the initial stage of electro-
chemical deposition of In.
Previous studies on electrochemical deposition of In in acidic
solutions without any organic additives have suggested that
growth of the In layer on Cu/Mo/glass follows a two-step behav-
ior: formation of a continuously smooth thin film followed by
three-dimensional (3D) growth to form an island-shaped surface
morphology [28,29]. The former smooth film formation results
from the rapid interdiffusion between Cu and In, leading to layer-
by-layer growth of the CuIn alloy without nucleation, and the latter
of further studies on morphological controls of the Cu film were
reported previously [33].
Fig. 2 shows linear sweep voltammetry (LSV) plots obtained
with the Cu deposited Mo/glass substrate immersed in In-bath(A),
In-bath(B), and In-bath(C) solutions. The potential sweep started at
−0.1 V vs. Ag/AgCl and was performed in the cathodic direction with
a scan rate of 10 mV s−1. In all LSV plots, remarkable rises in cathodic
currents at potentials more negative than −0.65 V (vs. Ag/AgCl)
were observed. Since the redox potential of the In3+/In couple is
−0.54 V (vs. Ag/AgCl), these rises are likely to be due to the reduc-
tion of In3+ ions. It is clear that almost the same current onset for
the In3+ reduction was observed irrespective of presences of citric
acid and sodium citrate in the solution, suggesting no appreciable
formation of citrate complexes in the present system. Based on the
fact that appreciable cathodic currents seem to appear at applied
potentials more negative than −0.70 V (vs. Ag/AgCl), we fabricated
In films at applied potentials more negative than this potential (i.e.,
−0.80 V (vs. Ag/AgCl) for In-bath(A) and −0.98 V (vs. Ag/AgCl) for In-
bath(B) and In-bath(C)). It should be noted that appreciable bobbles
were also evolved by applying such negative potentials, indicating
significant contribution of water reduction to the observed cathodic
currents (see below).
Fig. 3. SEM images of electrochemically deposited In films from (a, d) In-bath(A), (b, e) In-bath(B), and (c, f) In-bath(C): (a–c) top views and (d–f) cross-sectional views.

192
S.M. Lee et al. / Electrochimica Acta 79 (2012) 189–196
deposition from the In-bath(C) solution at −0.98 V (vs. Ag/AgCl)
showed a high current efficiency of more than 90%. Since applied
potentials of these In depositions are much more negative than the
equilibrium potential of the H2/H+ couple (−0.35 V (vs. Ag/AgCl) at
pH 2.2), those current efficiencies should strongly depend on the
degree of contribution of H2 evolution.
In order to evaluate roles of organic additives for the cur-
rent efficiency, current density measurements of H2 evolution
from acidic chloride solutions (pH 2.2) with or without citric acid
and/or sodium citrate were performed. Since the substrate sur-
face should change from Cu to In through the intermetallic CuIn
alloy (see above), we employed Cu/Mo/glass, CuIn/Cu/Mo/glass,
and In/CuIn/Cu/Mo/glass substrates for the H2 evolution experi-
ment. CuIn/Cu/Mo/glass and In/CuIn/Cu/Mo/glass substrates were
obtained by deposition of In for a short period (ca. 0.3 C cm−2
)
and a long period (ca. 0.9 C cm−2) using In-bath(A), In-bath(B),
and In-bath(C) solutions. Table 1 summarizes current densities
on these substrates in three different solutions, i.e., aqueous HCl,
that containing 10 mmol dm−3 citric acid, and that containing
10 mmol dm−3 citric acid and 35 mmol dm−3 sodium citrate. Irre-
spective of solution compositions, relatively high current densities
were observed on the Cu/Mo/glass substrate; it is clear that addition
of both citric acid and sodium citrate induces significant enhance-
ment of H2 evolution on this substrate. On the other hand, a
suppressive effect of citric acid addition on H2 evolution appeared
when we used CuIn/Cu/Mo/glass and In/CuIn/Cu/Mo/glass sub-
strates, i.e., moderate current densities comparable to those in the
simple HCl solution at −0.80 V (vs. Ag/AgCl) were observed in the
citric acid solution even at relatively high applied potential (−0.98 V
(vs. Ag/AgCl)). Furthermore, H2 evolution currents on these sub-
strates were significantly suppressed in the presence of both citric
acid and sodium citrate: current densities in the solution at −0.98 V
(vs. Ag/AgCl) were only about half of those in the simple HCl solu-
tion at −0.80 V (vs. Ag/AgCl). As discussed above, the surface of the
Cu/Mo/glass substrate should be covered with the CuIn alloy layer
in a short period during In deposition. Thus, the surface composi-
tion of the substrate during In deposition should be the CuIn alloy
or the In metal in almost all the deposition durations. The high
current efficiency of In deposition achieved in the In-bath(C) solu-
tion as mentioned above is, therefore, explained by the efficient
suppression behavior.
For further evaluation of the efficient suppression of H2 evo-
lution in the HCl solution containing both citric acid and sodium
citrate, effects of inorganic sodium salts, e.g., NaCl, KCl, Na2SO4,
and Na2CO3, in the acidic citric acid solution, instead of sodium
citrate, on the H2 evolution properties were examined by using
CuIn/Cu/Mo/glass substrates. As a result, addition of these salts was
found to be ineffective for suppression of H2 evolution. Moreover,
an increase in the concentration of citric acid had a detrimental
effect. Based on the fact that all of the carboxyl protons are not
fully dissociated in the present pH range, the difference between
sodium citrate and citric acid could not be rationally understood. At
present, the efficient suppression of H2 evolution over CuIn and In
Fig. 4. XRD patterns of electrochemically deposited Cu/In bilayers from (a) In-
bath(A) and (b) In-bath(C).
typical 3D growth is likely to be followed by instantaneous nucle-
ation limited by diffusion. The above-described XRD results suggest
that growth of In layers in the present study followed the same
mechanism, regardless of the presence of organic additives. Hence,
the citric acid in acidic solution should play a role in suppression
of island-shaped growth of In onto the preformed Cu/In alloy layer
at the second stage during the In deposition. One probable reason
for the function of citric acid is reduction of the overpotential of
In deposition induced by adsorption onto the Cu/In alloy surface
to reduce its surface energy. This function should enhance nucle-
ation probability, resulting in the formation of a smooth surface.
In addition, the enhancement of nucleation can be explained by
the difference in the applied potential, i.e., the In deposition in
In-bath(B) and In-bath(C) solutions was performed at a relatively
negative potential (−0.98 V (vs. Ag/AgCl)) compared to that in the
In-bath(A) solution (−0.80 V (vs. Ag/AgCl)). However, the fact that
a similar rough surface morphology was obtained on the Cu/In
bilayer from the In-bath(A) solution even at −0.98 V (vs. Ag/AgCl)
indicates an intrinsic effect of citric acid on the homogeneous In
deposition. It should also be noted that the application of such a
highly negative potential for the In-bath(A) solution often leads to
partial peeling of the Cu/In bilayer probably due to the evolution
of considerable amounts of hydrogen (H2) (see below). Hence, the
use of relatively less potential of −0.80 V (vs. Ag/AgCl) is a suitable
condition for In deposition from the In-bath(A) solution.
Current efficiencies of the In deposition from different solutions
were calculated by measuring weights of In deposits. When the
In deposition was performed using In-bath(A) solution at −0.80 V
(vs. Ag/AgCl), the efficiency only reached 60%. Similar current effi-
ciency was also obtained when the In-bath(B) solution was used
with applied potential of −0.98 V (vs. Ag/AgCl), even though a
smooth In film without crevices was formed. On the other hand, In
Table 1
Current densities for H2 evolution on several substrates under various conditions.
substrates
Current density (mA cm−2
)
Nonea (−0.80 V)d
Nonea (−0.98 V)d
CAb (−0.98 V)d
CAb + SCc (−0.98 V)d
Cu/Mo/glass
CuIn/Cu/Mo/glass
In/CuIn/Cu/Mo/glass
1.36
0.45
0.13
1.74
0.64
0.17
3.20
0.40
0.10
6.02
0.24
0.07
a
Aqueous HCl solution (pH 2.2) without any additives.
b
c
Aqueous HCl solution containing 10 mmol dm−3 citric acid (pH 2.2).
Aqueous HCl solution containing 10 mmol dm−3 citric acid and 35 mmol dm−3 sodium citrate (pH 2.2).
Applied potentials referred to the Ag/AgCl electrode.
d

S.M. Lee et al. / Electrochimica Acta 79 (2012) 189–196
193
Fig. 5. XRD patterns of CuInS2 films derived from electrochemically deposited Cu/In
bilayers in (a) In-bath(A) and (b) In-bath(C) solutions.
surfaces in the present acidic mixture of HCl, citric acid, and sodium
citrate is likely to be due to a concerted mechanism of components
in the solution.
3.2. Structural characterizations of sulfurized CuInS2 films
The Cu/In bilayers with different In morphologies, which were
obtained from In-bath(A) and In-bath(C) solutions, were converted
into CuInS2 films (labeled CuInS2(A) and CuInS2(C), respectively)
by sulfurization in H2S (5% in Ar) at 520 ◦C followed by chemical
etching of a CuxS component. As shown in Fig. 5, XRD analyses
of thus-obtained CuInS2(A) and CuInS2(C) films indicate typi-
cal diffraction peaks of (1 1 2), (0 0 4)/(2 0 0), (2 0 4)/(2 2 0), and
(1 1 6)/(3 1 2) reflections that coincide with CuInS2 crystal with
a chalcopyrite structure (JCPDS 27-0159) [6] and no appreciable
detection of any secondary phases, except for the Mo substrate.
The fact that the Cu/In/S composition ratios of these films were
confirmed to be almost stoichiometric (i.e., 24/25/51) also indi-
cates formation of films having a single CuInS2 phase. On both films,
moreover, the intensity of the major peak at 2ꢁ = 27.8◦, correspond-
ing to the (1 1 2) plane, was much stronger than those of others
derived from (2 0 0) and (2 0 4)/(2 2 0) planes, indicating preferable
(1 1 2) orientation. In addition, the CuInS2(C) film showed more
intense reflections than those of the CuInS2(A) film (e.g., the inten-
sity of (1 1 2) reflection of CuInS2(C) was ca. 1.2 times larger than
that of CuInS2(A)), suggesting that the former film achieves a rel-
atively large grain growth in comparison with that of the latter
one.
Fig. 6 shows SEM images of CuInS2(A) and CuInS2(C). Top view
SEM images of both films indicate that they consist of angular-
shaped microcrystallites, as shown in Fig. 6a and b. It is clear that the
CuInS2(A) film has appreciable recessed areas (Fig. 6a), while a rela-
tively flat surface composed of compact agglomerates was observed
in the SEM image of the CuInS2(C) film (Fig. 6b). The significant dif-
ferences in surface roughness between these films should be due
to morphological differences of Cu/In bilayers before sulfurization.
As expected from these surface morphologies, the CuInS2(A) film
showed a relatively large variation of thicknesses compared to that
of the CuInS2(C) film. Moreover, the size of each CuInS2 grain in the
CuInS2(C) film is larger than that in the CuInS2(A) film, which is in
good agreement with the above-described XRD results. Another
point worth noting is that CuInS2(C) appears to have relatively
poor adherence compared to that of CuInS2(A), though both of the
films have negligible voids between CuInS2 crystallites. The flat
surface and well-grown crystallite without voids of the CuInS2(C)
Fig. 6. (a, b) Top view and (c, d) cross-sectional view SEM images of CuInS2 films
derived from electrochemically deposited Cu/In bilayers in (a, c) In-bath(A) and (b,
d) In-bath(C).

194
S.M. Lee et al. / Electrochimica Acta 79 (2012) 189–196
Fig. 7. Photocurrent–potential curves of (a) CuInS2(A) and (b) CuInS2(C) films in a
0.1 mol dm−3 Eu3+ aqueous solution at pH 4.
film would be beneficial for solar cell application, but the poor
adherence would be detrimental because of the increment in series
resistance: details of these effects are discussed below.
3.3. Photoelectrochemical properties
Fig. 7 shows photocurrent–potential curves of CuInS2 films
taken in a solution containing Eu3+ as an electron acceptor. Owing
to p-type semiconductive characters, cathodic photocurrents were
observed and their magnitude increased with negative shifts of the
potential. Although the onset potential of photocurrents observed
on both films was essentially the same (ca. −0.05 V (vs. Ag/AgCl)),
the CuInS2(C) film showed relatively large photocurrents, suggest-
ing higher film quality than that of the CuInS2(A) film. As shown
in Fig. 8, corresponding EQE spectra of photocurrent on both film
measured at −0.4 V (vs. Ag/AgCl) also indicated better photoelectro-
chemical response of the CuInS2(C) film than that of the CuInS2(A)
film. It is noted that both films showed almost the same onset wave-
length, i.e., band gap energies (Egs) of these films determined from
intercepts of the photon energy (hꢂ) axis of (EQE × hꢂ)2 − hꢂ plots
[34] (not shown) were equivalent (ca. 1.5 eV). The value is also in
good agreement with the reported Eg value of CuInS2 films [1].
The EQE value is correlated with the width of the space charge
layer (W) as follows [35]:
Fig. 9. J–V characteristics of Al:ZnO/CdS/CuInS2/Mo/glass cells. These cells were
made from CuInS2 films obtained from (a) In-bath(A) and (b) In-bath(C). CP denotes
the crossing point of illuminated and dark J–V curves.
where ˛ denotes absorption coefficient. This equation indicates
possible estimation of W from a slope of −ln(1 − EQE) vs. ˛ plot.
Using ˛ values at wavelengths close to photoabsorption onset [36]
W values of CuInS2(A) and CuInS2(C) films at an applied potential (E)
of −0.4 V (vs. Ag/AgCl) were estimated to be 0.10 ␮m and 0.17 ␮m,
respectively. The accepter density (Na) is correlated with W by
ꢀ
ꢁ
2(EFB − E)ε0εr
1
2
W =
,
(2)
qNa
where, EFB, ε0, εr, and q are flatband potential, dielectric constant
of the vacuum, the relative dielectric constant of CuInS2 film, and
electric charge, respectively. Since the photocurrent onset of a p-
type semiconductor is likely to occur at potentials of ca. 0.2–0.3 V
more negative than EFB [35,37–39] the EFB of CuInS2 in the present
electrolyte solution would lie at ca. 0.15–0.25 V (vs. Ag/AgCl) (see
Fig. 7). Moreover, assuming that εr of CuInS2 is [10,35,40] Nas of
CuInS2(A) and CuInS2(C) films were calculated to be 6.8–7.2 × 1016
and 2.1–2.5 × 1016, respectively. The Nas of chalcopylite films for
efficient solar cells were in the order of 1016 cm−3 [41–43]. Hence,
Na values of present CuInS2 films would be preferable to apply the
use of solar cell as discussed below.
EQE = 1 − exp(−˛W),
(1)
3.4. Solar cell properties
Solar
cells
with
a
device
structure
of
Al:ZnO/CdS/CuInS2/Mo/glass (without an antireflection coat-
ing) were prepared by depositing CdS and Al:ZnO layers on
CuInS2(A) and CuInS2(C). The light and dark J–V characteristics and
cell parameters obtained from these J–V curves are shown in Fig. 9
and Table 2, respectively. The solar cell derived from CuInS2(C)
exhibits good performance with conversion efficiency (Á) of 7.8%
(Fig. 9b) compared to that of the cell based on CuInS2(A) (Á = 4.4%,
Fig. 9a). It should be noted that the Á values of the former cells were
relatively constant (more than 7%), whereas those of the latter
cells were significantly deviated in the range of about 4–7% despite
Fig. 8. EQE spectra of cathodic photocurrents obtained on (a) CuInS2(A) and (b)
CuInS2(C) films. The measurement was carried out at −0.4 V (vs. Ag/AgCl) in a
0.1 mol dm−3 Eu(NO3)3 aqueous solution at pH 4.

S.M. Lee et al. / Electrochimica Acta 79 (2012) 189–196
195
Table 2
Solar cell parameters obtained from illuminated J-V curves shown in Fig. 8.
a
b
e
f
Photo-absorber
JSC (mA cm−2
)
VOC (V)
FFc
Ád (%)
Rs (ꢀ cm2)
Rsh (kꢀ cm2)
CuInS2(A)
CuInS2(C)
16.0
21.3
0.57
0.67
0.49
0.55
4.4
7.8
10.2
8.0
2.48
3.69
a
Short-circuit current density.
Open circuit voltage.
Fill factor.
Conversion efficiency.
Series resistance.
Shunt resistance.
b
c
d
e
f
applying the same procedure. The insufficient reproducibility of
the cell based on the CuInS2(A) film is due to the poor homogeneity
of CuInS2, e.g., large thickness variations and inhomogeneous
composition distributions, leading to reduced cell parameters.
In accordance with the above photoelectrochemical measure-
ments, the CuInS2(C) film should have more efficient carrier
utilization properties than that of the CuInS2(A) film (see
Figs. 7 and 8). As discussed above, this would be due to better
structural properties and electric properties of the CuInS2(C) film
(flat surface, well-grown crystallite, and relatively small Na) than
those of the CuInS2(A) film. These resulted in relatively small series
resistance of the cell derived from CuInS2(C), thus leading rela-
tively large short-circuit current density (JSC) and fill factor (FF).
Another difference in cell parameters between the CuInS2(A)-based
cell and the CuInS2(C)-based cell summarized in Table 2 is open cir-
cuit voltages (VOCs). As has been reported in the literature [42–44]
the buffer–absorber (CdS–CuInS2) interface is considered to be a
“cliff-type” band alignment, which differs from the “notch-type”
alignment of the CdS–Cu(In,Ga)Se2 interface. Since Nas of present
CuInS2 films are of the same order as those reported in studies on
the band alignment calculations (e.g., the Na value of 5 × 1016 was
given in the literature [43]), Fermi levels of present CuInS2 films
should be comparable to that in the literature. Hence, the reported
“cliff-type” band alignment can be applied to present materials. In
such a band alignment, it is known that interface recombination is
dominant in the present cell [43,44]. As discussed above, the rough
surface morphology of the CuInS2(A) film results in enlargement
of interface areas between the CdS buffer layer and the CuInS2
absorber; thus, this should limit VOC as well as FF of the cell. In
addition, VOC of an illuminated solar cell depends on the ratio of
the active area (Aa) and the total area (At) of the junction interface,
i.e., as the Aa/At ratio increases, VOC tends to decrease [45,46]. Since
microscopic roughness of the junction interface increases At but
not Aa, high VOC can be obtained by a cell having a smooth junc-
tion interface. Hence, the CuInS2(C) film having a relatively smooth
CdS–CuInS2 interface achieved larger VOC than that obtained for the
cell based on the CuInS2(A) film.
of a smooth and homogeneous In film on the surface of a Cu-covered
Mo/glass substrate were examined. Homogeneous electrochemical
deposition of In was found to be achieved by efficient enhancement
of the In nucleation induced by citric acid, whereas the addition of
sodium citrate was effective for reduction of H2 evolution, a side
reaction. Photoelectrochemical analyses of CuInS2 films obtained
by sulfurization of Cu/In bilayers revealed that relatively high qual-
ity film was obtained from the homogeneous In film compared
to the film derived from the inhomogeneous In island film. As
expected the solar cell with a Al:ZnO/CdS/CuInS2/Mo structure
derived from the homogeneous In was shown to be more efficient
than the cell derived from the inhomogeneous In. Hence, we have
proved the importance of structural controls of the In precursor film
to obtain an efficient CuInS2-based solar cell. Since the basic idea
should be applicable to other thin-film compound solar cells based
on Cu(In,Ga)Se2 (CIGS) and Cu2ZnSnS4 (CZTS) absorbers, studies
are now in progress to find optimal conditions of electrochemical
depositions in order to obtain smooth and homogeneous metallic
precursor stacks for solar cell applications.
Acknowledgements
This work was partially supported by The KAITEKI Institute, Inc.
This work was also carried out as part of a program supported by
a research grant from City of Osaka. The authors thank Dr. Ahmed
Ennaoui (Helmholtz–Zentrum–Berlin, Germany) for his stimulat-
ing suggestions and discussion.
References
[1] M.A. Green, K. Emery, D.L. King, S. Igari, W. Warta, Progress in Photovoltaics:
Research and Applications 10 (2002) 355–360.
[2] F.C. Akkari, R. Brini, M. Kanzari, B. Rezig, Journal of Materials Science 40 (2005)
5751–5755.
[3] L.L. Kazmerski, M.S. Ayyagari, G.A. Sanborn, Journal of Applied Physics 46 (1976)
4865–4869.
[4] Y.B. He, T. Krämer, I. Österreicher, A. Polity, B.K. Meyer, M. Hardt, Semiconductor
Science and Technology 20 (2005) 685–692.
[5] J. Klaer, J. Bruns, R. Henninger, K. Siemer, R. Klenk, K. Ellmer, D. Bräunig, Semi-
conductor Science and Technology 13 (1998) 1456–1458.
[6] K. Siemer, J. Klaer, I. Luck, J. Bruns, R. Klenk, D. Bräunig, Solar Energy Materials
and Solar Cells 67 (2001) 159–166.
[7] S. Bandyopadhyaya, S. Chaudhuri, A.K. Pal, Solar Energy Materials and Solar
Cells 60 (2000) 323–339.
[8] C.V. Klopmann, J. Djordjevic, E. Rudigier, R. Scheer, Journal of Crystal Growth
289 (2006) 121–133.
[9] J. Klaer, R. Klenk, H.-W. Schock, Thin Solid Films 515 (2007) 5929–5933.
[10] M. Krunks, O. Bijakina, T. Varema, V. Mikli, E. Mellikov, Thin Solid Films 338
(1999) 125–130.
Further inspection of the solar cell parameters shown in Table 2
indicates that a limiting factor of the present best cell derived from
the CuInS2(C) film is its small FF, as a result of high series resis-
tance (Rs) as compared to several literature devices. As discussed
above, the CuInS2(C) film showed relatively poor adhesion; high
Rs is speculated to be mainly due to high interfacial resistance
between the CuInS2(C) film and the Mo substrate. The observa-
tion of cross-over behavior at a lower current than that observed
on the cell derived from CuInS2(A) (Fig. 9) supports this hypothesis.
In order to obtain further increase in solar cell performance of the
present device, therefore, optimization of sulfurization parameters
to improve adhesion is required.
[11] R. Klenk, J. Klaer, R. Scheer, M.C. Lux-Steiner, I. Luck, N. Meyer, U. Rühle, Thin
Solid Films 480–481 (2005) 509–514.
[12] C.J. Hibberd, E. Chassaing, W. Liu, D.B. Mitzi, D. Lincot, A.N. Tiwari, Progress in
Photovoltaics: Research and Applications 18 (2010) 432–452.
[13] I. Oja, M. Nanu, A. Katerski, M. Krunks, A. Mere, J. Raudoja, A. Goossens, Thin
Solid Films 480–481 (2005) 82–86.
[14] M. Krunks, O. Kijatkina, A. Mere, T. Varema, I. Oja, V. Mikli, Solar Energy Mate-
rials and Solar Cells 87 (2005) 207–214.
[15] A. Katerski, A. Mere, V. Kazlauskiene, J. Miskinis, A. Saar, L. Matisen, A. Kikas,
M. Krunks, Thin Solid Films 516 (2008) 7110–7115.
[16] C. Camus, N.A. Allsop, S.E. Gledhill, W. Bohne, J. Röhrich, I. Lauermann, M.C.
Lux-Steiner, C.H. Fischer, Thin Solid Films 516 (2008) 7026–7030.
4. Conclusion
In this study, the effects of additions of certain amounts of citric
acid and sodium citrate to an acidic InCl3 solution on the formation

196
S.M. Lee et al. / Electrochimica Acta 79 (2012) 189–196
[17] S. Nakamura, A. Yamamoto, Solar Energy Materials and Solar Cells 49 (1997)
415–421.
[18] B. Asenjo, A.M. Chaparro, M.T. Gutiérrez, J. Herrero, Thin Solid Films 511–512
(2006) 117–120.
[19] S. Nakamura, A. Yamamoto, Solar Energy Materials and Solar Cells 75 (2003)
81–86.
[20] C.D. Lokhande, G. Hodes, Solar Cells 21 (1987) 215–224.
[21] J. Herrero, J. Ortega, Solar Energy Materials and Solar Cells 20 (1990) 53–65.
[22] A. Antony, A.S. Asha, R. Yoosuf, R. Manoj, M.K. Jayaraj, Solar Energy Materials
and Solar Cells 81 (2004) 407–417.
[31] B.M. Basol, S. Aksu, J. Wang, U.S. Patent 20090283415 (2009).
[32] S. Canegallo, V. Demeneopoulos, L.P. Bicelli, G. Serravalle, Journal of Alloys and
Compounds 228 (1995) 23–30.
[33] S.M. Lee, S. Ikeda, T. Harada, M. Matsumura, Journal of Non-Crystalline Solids,
http://dx.doi.org/10.1016/j.jnoncrysol.2011.12.043.
[34] S. Ikeda, T. Nakamura, S.M. Lee, T. Yagi, T. Harada, T. Minegishi, M. Matsumura,
ChemSusChem 4 (2011) 262–268.
[35] J.J. Scragg, P.J. Dale, L.M. Peter, G. Zoppi, I. Fobes, Physica Status Solidi (b) 245
(2008) 1772–1778.
[36] These ˛ values were determined by measuring photoabsorption of the 100-
nm-thick CIS film deposited on a fluorine-doped tin oxide (FTO)-coated glass
substrate using a spray technique.
[23] R.P. Wijesundera, W. Siripala, Solar Energy Materials and Solar Cells 81 (2004)
147–154.
[24] D. Lincot, J.F. Guillemoles, S. Taunier, D. Guimard, J. Sicx-Kurdi, A. Chaumont,
O. Roussel, O. Ramdani, C. Hubert, J.P. Fauvarque, N. Bodereau, L. Parissi, P.
Panheleux, P. Fanouillere, N. Naghavi, P.P. Grand, M. Benfarah, P. Mogensen, O.
Kerrec, Solar Energy 77 (2004) 725–737.
[25] Y. Onuma, K. Takeuchi, S. Ichikawa, Y. Suzuki, R. Fukasawa, D. Matono, K. Naka-
mura, M. Nakazawa, K. Takei, Solar Energy 80 (2006) 132–138.
[26] S.M. Lee, S. Ikeda, T. Yagi, T. Harada, A. Ennaoui, M. Matsumura, Physical Chem-
istry Chemical Physics 13 (2011) 6662–6669.
[37] P.L. Dale, Peter in Photoelectrochemical Materials and Energy Conversion Pro-
cesses, Wiley–VCH, Weinheim, 2010.
[38] L.M. Peter, Chemical Reviews 90 (1990) 753–769.
[39] J. Li, R. Peat, L.M. Peter, Journal of Electroanalytical Chemistry 165 (1984) 41–59.
[40] S. Ikeda, R. Kamai, T. Yagi, M. Matsumura, Journal of the Electrochemical Society
157 (2010) B99–B103.
[41] T. Minemoto, T. Matsui, H. Takakura, Y. Hamakawa, T. Negami, Y. Hashimoto,
T. Uenoyama, M. Kitagawa, Solar Energy Materials and Solar Cells 67 (2001)
83–88.
[42] S.S. Schmidt, D. Abou-Ras, T. Unold, T. Eisenbarth, H. Wilhelm, Journal of Applied
Physics 110 (2011) 064515.
[27] C. Broussillou, M. Andrieux, M. Herbst-Ghysel, M. Jeandin, J.S. Jaime-Ferre,
S. Bodnar, E. Morin, Solar Energy Materials and Solar Cells 95 (2011)
S13–S17.
[28] R.C. Valderrama, M. Miranda-Hernández, P.J. Sebastian, A.L. Ocampo, Elec-
trochimica Acta 53 (2008) 3714–3721.
[29] Q. Huang, K. Reuter, S. Amhed, L. Deligianni, L.T. Romankiw, S. Jaime, P.-P. Grand,
V. Charrierb, Journal of the Electrochemical Society 158 (2011) D57–D61.
[30] P.P. Prosini, M.L. Addonizio, A. Antonaia, Thin Solid Films 298 (1997) 191–196.
[43] R. Klenk, Thin solid films 387 (2001) 135–140.
[44] M. Turcu, U. Rau, Thin Solid Films 432 (2003) 158–162.
[45] B.J. Stanbery, Critical Reviews in Solid State and Materials Sciences 27 (2002)
45.
[46] H.J. Hovel, Solar Cells, Academic Press, New York, 1975.