Effect of potassium hydrogen phthalate (C8H5KO4) on the one-step electrodeposition of single-phase CuInS2 thin films from acidic solution

Article abstract of DOI:10.1016/j.electacta.2010.02.077

The effect of potassium hydrogen phthalate (C₈H₅KO₄) as a special additive on the one-step electrodeposition of single-phase CuInS₂ thin films from acidic solution (pH 2.5) was investigated in detail. The XRD, SEM and UV-vis-NIR characterization confirms that the addition of an adequate concentration of C₈H₅KO₄ (23 mM) to the electrolytic bath containing 12.5 mM Cu²⁺, 10 mM In³⁺, 40 mM S₂O₃^2- and 100 mM LiCl can contribute greatly to the controllable growth of pure chalcopyrite CuInS₂ films with uniform surfaces and an ideal band gap of approximately 1.54 eV. Complexation studies of C₈H₅KO₄ with Cu²⁺ and In³⁺ in electrolytic solutions indicated that C₈H₅KO₄ can complex Cu²⁺ more strongly than In³⁺ and move the electrode potentials of Cu²⁺ and In³⁺ near each other as determined by polarization analysis. Furthermore, the potentiodynamic polarization and electrochemical impedance spectroscopy (EIS) analysis performed in a series of solution systems revealed a three-step reaction mechanism for CuInS₂ deposition and considerable adsorption of C₈H₅O₄^- and Cu(C₈H₅O₄)⁺ to the cathode surface. This deposition shows that the synergetic effects of complexation and adsorption originated from the additive on the Cu²⁺ electro-reduction, thus promoting the co-deposition of copper, indium and sulfur in the form of single-phase CuInS₂.

Full text of DOI:10.1016/j.electacta.2010.02.077

Electrochimica Acta 55 (2010) 4428–4435
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Effect of potassium hydrogen phthalate (C8H5KO4) on the one-step
electrodeposition of single-phase CuInS2 thin films from acidic solution
Xinhua Xua,b, Feng Wanga,b,∗, Jingjun Liua,b, Jing Jia,b
a State Key Laboratory of Chemical Resource Engineering, Beijing University of Chemical Technology, Beijing 100029, PR China
b Institute of Carbon Fibers and Composites, Beijing University of Chemical Technology, Beijing 100029, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
The effect of potassium hydrogen phthalate (C8H5KO4) as a special additive on the one-step electrode-
position of single-phase CuInS2 thin films from acidic solution (pH 2.5) was investigated in detail. The
XRD, SEM and UV–vis–NIR characterization confirms that the addition of an adequate concentration
Received 25 November 2009
Received in revised form 22 February 2010
Accepted 22 February 2010
Available online 1 March 2010
of C8H5KO4 (23 mM) to the electrolytic bath containing 12.5 mM Cu2+, 10 mM In3+, 40 mM S2O3 and
2−
100 mM LiCl can contribute greatly to the controllable growth of pure chalcopyrite CuInS2 films with
uniform surfaces and an ideal band gap of approximately 1.54 eV. Complexation studies of C8H5KO4 with
Cu2+ and In3+ in electrolytic solutions indicated that C8H5KO4 can complex Cu2+ more strongly than In3+
and move the electrode potentials of Cu2+ and In3+ near each other as determined by polarization anal-
ysis. Furthermore, the potentiodynamic polarization and electrochemical impedance spectroscopy (EIS)
analysis performed in a series of solution systems revealed a three-step reaction mechanism for CuInS2
Keywords:
Copper indium disulfide
One-step electrodeposition
C8H5KO4 additive
Reaction mechanism
deposition and considerable adsorption of C8H5O4 and Cu(C8H5O4)+ to the cathode surface. This depo-
−
sition shows that the synergetic effects of complexation and adsorption originated from the additive on
the Cu2+ electro-reduction, thus promoting the co-deposition of copper, indium and sulfur in the form of
single-phase CuInS2.
© 2010 Elsevier Ltd. All rights reserved.
1. Introduction
These advantages make CuInS2 an interesting semiconductor for
photovoltaic applications. Intense efforts by many research groups
CuInS2 is a ternary chalcopyrite semiconductor that is known
to have great potential as an absorption layer material for devel-
oping low-cost, less-toxic and highly-efficient thin film solar cells
[1]. CuInS2 has outstanding physical properties including a direct
optical band gap around 1.5 eV that matches well with the solar
spectrum [2]. Furthermore, the material has a large absorption
coefficient of about 105 cm−1 that permits CuInS2-based solar cells
with a theoretical efficiency up to 26% [3] with a CuInS2 thick-
ness of only 1–2 ␮m [4]. This compound also has good radiation
stability and environmental compatibility that should keep the
CuInS2 photovoltaic devices working well in various conditions [5].
Furthermore, CuInS2 can exist in nonstoichiometric composition
within the detection limit of 10% or more [6], thus its n/p conduc-
tion type can be controlled easily by adjusting the molar ratio of
the compositional elements, i.e., p type as the S or Cu is excess and
n type as the In is excess [7,8].
have yielded the fundamental understanding that the structural,
electrical and optical properties of CuInS2 largely depend on its
preparation methods [9–11]. The technique of co-evaporation [12]
or sulfurization of a metallic precursor [13] can produce high-
quality films from which CuInS2 solar cells have achieved the
highest conversion efficiency of nearly 12%. However, these meth-
ods are not suitable for large-scale fabrication due to the expensive
cost and toxic post-annealing required in a S or H2S atmosphere.
Therefore, a great deal of effort is currently directed toward finding
a large-scale, high-quality, low-cost and eco-friendly technology
for fabricating CuInS2 thin films.
One-step electrodeposition is a potential method for achiev-
ing this goal [14]. Several attempts have been made to employ
the one-step electrodeposition process in which the CuInS2 thin
film is potentiostatically grown by co-deposition of Cu, In and S
simultaneously from aqueous solution containing Cu2+, In3+ and
2−
2−
SO3
or S2O3
species. Asenjo et al. [15] reported that the co-
electrodeposition of CuInS2 films in a solution of CuSO4, In2(SO4)3
and Na2SO3 shows not only a CuInS2 phase predominant for the
films but also has secondary phases of Cu2O/Cu2S, CuO/CuS and
In2S3. In contrast, Romain et al. [16] electrodeposited CuInS2 films
from a solution containing 10 mM Cu(O2CCH3)2, 10 mM InCl3 and
∗
Corresponding author at: Institute of Carbon Fibers and Composites, Beijing Uni-
versity of Chemical Technology, Beijing 100029, PR China. Tel.: +86 10 64411301;
fax: +86 10 64411301.
E-mail address: wangf@mail.buct.edu.cn (F. Wang).
0013-4686/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2010.02.077

X. Xu et al. / Electrochimica Acta 55 (2010) 4428–4435
4429
400 mM Na2S2O3 with pH 1.5 and obtained pure CuInS2 films but
with rough and porous surfaces.
surface. The deposition was then performed at a temperature of
30 ◦C and a potential of −1.0 V for 30 min. Finally, the deposited
films were annealed in an Ar atmosphere at 350 ◦C for 60 min to
increase the crystallinity. All potentials mentioned in this paper are
reported relative to SCE. Analytical grade reagents and ultra pure
water were used in this study.
The crystallographic phase structures of the deposited films
were characterized by an X-ray diffractometer (XRD, MAC Sci-
ence MX Labo2) with Cu K␣ radiation (ꢀ = 1.541874 Å) at 35 kV and
25 mA and a scanning rate of 0.04◦/s in the 2ꢁ range of 10–70◦.
The surface morphology was observed by a scanning electron
microscope (SEM, Hitachi S-4800). The optical reflectance spectra
were recorded in the wavelength range of 190–1100 nm using an
ultraviolet–visible–near infrared spectrophotometer (UV–vis–NIR,
Shimadzu UV-3150), and the band gaps (Eg) of the deposited films
were calculated according to the formula of Eg = 1240/ꢀ (eV), where
ꢀ is the wavelength (nm) with respect to the strong peak of the
corresponding differential reflectance spectra.
Nevertheless, the purity and stoichiometry of CuInS2 thin films
prepared by one-step electrodeposition are still rather difficult
to control [6,15,17] due to the much more positive potentials
of copper and sulfur than that of indium. This difference causes
the preferential deposition of either copper, sulfur or CuxS [15]
and may affect the photoelectricity properties of the resulting
CuInS2 thin films. Moreover, the electrodeposition potentials for
the three elements are not close to each other, even when the
pH or electrolyte concentration, which may introduce significant
concentration polarization, are adjusted [18]. To the best of our
knowledge, the mechanism and kinetics of CuInS2 electrodeposi-
tion have not been reported in detail. A great deal of study will be
required to improve the growth of high-quality CuInS2 by a one-
step electrodeposition process and develop a better understanding
of the co-deposition behavior of Cu, In and S.
In this work we added an organic additive, potassium hydro-
gen phthalate (C8H5KO4), to the electrolytic bath containing CuCl2,
InCl3, Na2S2O3 and LiCl for the preparation of single-phase CuInS2
thin films. The special additive C8H5KO4 was used in hopes of nar-
rowing the potential gap of Cu, In, S by metal ion complexation
while buffering the solution pH to modify the electrodeposition
potential of sulfur [19]. For the co-deposition of Cu, In and S, we
adjusted the pH and electrolyte concentration so that the electrode
potential of the individual deposits became closer to each other. The
effect of C8H5KO4 concentration on the structure, morphology and
optical properties of deposited films was investigated. Additionally,
the effect of C8H5KO4 on the electrodeposition mechanisms and
kinetic behavior of CuInS2 thin films on Ni foil was investigated in
detail for the first time.
2.2. Complexation examination
The complexation properties of C8H5KO4 with Cu2+ and In3+
were examined by ultraviolet light (UV) absorption spectra
recorded on the Shimadzu UV-3150 UV–vis–NIR spectrophotome-
ter. Spectra were obtained from solutions of C8H5KO4, CuCl2 or
InCl3 individually and from solutions containing C8H5KO4 together
with CuCl2 or InCl3. To assure the accuracy and comparability of
measurements, the solution concentration of C8H5KO4 as low as
possible, about 0.1 mM, and the concentration of CuCl2 or InCl3
was varied from 0 to 10 mM. The pH value of these solutions was
adjusted to 2.5 with diluted HCl acid.
2.3. Electrochemical measurements
2. Experimental
To investigate the effect of C8H5KO4 on the reduction of Cu2+ and
In3+ while simultaneously investigating the formation mechanisms
and deposition kinetics of CuInS2 thin films, potentiodynamic
polarization and electrochemical impedance spectroscopy (EIS)
measurements were performed in a series of solutions. Three-
electrode cells, as mentioned above, and an electrochemical
workstation consisting of an EG&G PAR Model 273 Potentio-
stat/Galvanostat, a Solartron 1250 Frequency Response analyzer
and an EG&G PAR Model 5210 Lock-in amplifier were used. The
potentiodynamic polarization curves were scanned from 0 to
−1.5 V with a constant scan rate of 10 mV s−1. The EIS were mea-
sured in a frequency range of 100 kHz to 1 mHz with an alternated
signal of 10 mV amplitude at a potential of −1.0 V. All solutions used
were adjusted to a pH value of 2.5 by HCl and deaerated by sparging
with high-purity nitrogen for 30 min before the experiments.
2.1. Films preparation and characterization
The CuInS2 thin films were electrodeposited potentiostatically
in a three-electrode cell. Ni foil with an exposure surface of 2.0 cm2
wasusedas theworking electrode, a2.5 cm × 2.5 cmplatinum sheet
was used as the counter electrode and a standard calomel electrode
(SCE) was used as the reference electrode. The electrolytic bath con-
tained 12.5 mM CuCl2, 10 mM InCl3, 40 mM Na2S2O3, 100 mM LiCl
and 0–23 mM C8H5KO4, with a pH value of 2.5 adjusted by diluted
HCl. Prior to the deposition the Ni foil electrode was electropolished
in a 72–75 wt% H2SO4 solution until achieving a mirror-like smooth
3. Results and discussion
3.1. Deposited films preparation and characterization
Fig. 1 shows the XRD patterns of deposited films with different
C8H5KO4 concentrations in the electrolytic baths. The thin films
prepared at [C8H5KO4] = 0 and 15 mM are composed of both CuInS2
(JCPDS Card File, 27-0159) and Cu2S phases (JCPDS Card File, 33-
0490). In contrast, the thin film prepared at [C8H5KO4] = 23 mM
shows an XRD patterns characteristic of a single CuInS2 phase
except for two significant diffraction peaks from the Ni substrate.
The intense peak at 27.8◦ and two weak peaks at 46.2◦ and 55.0◦
are indexed to the (1 1 2), (2 0 4) and (3 1 2) reflections of chalcopy-
rite structured CuInS2, respectively (JCPDS Card File, 27-0159). The
typical SEM image shown in Fig. 2 reveals that the pure CuInS2 thin
film has a relatively uniform surface that consists of densely packed
Fig. 1. XRD patterns of thin films electrodeposited from electrolytic baths with dif-
ferent C8H5KO4 concentrations. [C8H5KO4] = (a) 0, (b) 15 and (c) 23 mM. All of the
thin films were annealed at 350 ◦C for 60 min in Ar atmosphere.

4430
X. Xu et al. / Electrochimica Acta 55 (2010) 4428–4435
Fig. 2. SEM micrograph of the CuInS2 thin films electrodeposited from electrolytic
bath with [C8H5KO4] = 23 mM.
spherical clusters of smaller nanoparticles. The differential spectra
(dR/dꢀ vs. ꢀ) of the reflectance vs. wavelength plot (Fig. 3) revealed
an optical band gap (Eg) [20] near 1.54 eV, which matches well with
the solar spectrum.
3.2. Complexation studies between C8H5KO4 additive and Cu2+
In3+ cations
,
The above results indicate that the one-step electrodeposition
combined with 23 mM C8H5KO4 is an excellent method for the con-
trollable growth of high-quality CuInS2 thin films. The presence of
C8H5KO4 leads to a change in the Cu–In–S co-electrodeposition
behavior. Different aspects of the mechanism of action of the
C8H5KO4 additive were investigated in detail for the first time.
As is well known, the complexing effect plays an important role
in alloy co-electrodeposition processes [21]. The UV absorption
spectra of individual CuCl2, InCl3, C8H5KO4 solution and C8H5KO4
solutions mixed with various concentrations of CuCl2 or InCl3, pre-
sented in Fig. 4, were compared to identify complex formation
between C8H5KO4 and Cu2+ or In3+ cations.
Fig. 4A clearly shows that the UV spectrum for CuCl2 has
an absorption at ꢀ = 235.69 nm, whereas the UV spectrum for
C8H5KO4 exhibits an intense absorbance at 239.96 nm and a weak
absorbance at 280.73 nm. These peaks likely correspond to the
characteristic E2 and B absorption bands of aromatic compounds,
respectively [22]. The E2 band located at higher energy originates
Fig. 4. UV spectra of various solution systems (pH 2.5). A: (a) 5 mM CuCl2; (b)
0.1 mM C8H5KO4 + 0–1 mM CuCl2; (c) 0.1 mM C8H5KO4 + 5 mM CuCl2; (d) 0.1 mM
C8H5KO4 + 10 mM CuCl2. B: (e) 5 mM InCl3; (f) 0.1 mM C8H5KO4; (g) 0.1 mM
C8H5KO4 + 5 mM InCl3; (h) 0.1 mM C8H5KO4 + 10 mM InCl3.
from the ␲ → ␲* transition of conjugated vinylene system in the
benzene ring. The B band at lower energy originates from the
overlapping of ␲ → ␲* transition and benzene ring structure vibra-
tion. Adding a small quantity of Cu2+ (lower than 1 mM) into the
C8H5KO4 solution did not result in a change in the absorbance and
all spectra maintained the same pattern as in Fig. 4A(b). Increas-
ing the added concentration of Cu2+ from 1 to 10 mM, however,
intensifies the E2 band of C8H5KO4 and results in a red shift from
239.96 to 244.51 nm. The B band exhibits a gradual distortion in
shape until it disappears, indicating the formation of the Cu2+ com-
plex in C8H5KO4 solution. As is well known, C8H5KO4 (structure
presented in Fig. 5(a)) can dissociate to phthalate mono- and di-
−
anions (i.e., C8H5O4− and C8H4O42−) in solution [23]. The C8H5O4
species is the dominant anion due to the much smaller primary
dissociation constant of H2C8H4O4 relative to the secondary dis-
sociation constant (pK1 = 2.96, pK2 = 5.41) [24]. In the structure of
C8H5O4−, the most active carboxyl oxygen atom can coordinate
metal cations to various degrees through its lone pair electrons
[25]. Therefore, it is deduced that th−e most probable complexation
happens between Cu2+ and C8H5O4 in the stoichiometric ratio of
1/1, as shown in Fig. 5(b) with a chemical formula of Cu(C8H5O4)+.
The formation constant of Cu(C8H5O4)+ is log K = 7.10 according to
previous literatures [24,25].
The UV spectrum of InCl3 in Fig. 4B, however, does not show an
absorption, which implies that the electrons bound in the ground
state level of indium are especially stable because the UV spectrum
reveals the electron transition properties between energy levels
[26]. Furthermore, the addition of 0–10 mM In3+ to the C8H5KO4
solution has no effect on the E2 and B bands, but rather leads to a
slight blue shift of B band, suggesting that a relative weak complex-
Fig. 3. Differential spectra (dR/dꢀ vs. ꢀ) of the reflectance vs. wavelength plot
(shown in the inset) for the CuInS2 thin films electrodeposited from electrolytic
bath with [C8H5KO4] = 23 mM.
ation between C8H5O4 and In3+ in 1:1 stoichiometry as shown in
−

X. Xu et al. / Electrochimica Acta 55 (2010) 4428–4435
4431
Fig. 5. Structures of (a) potassium hydrogen phthalate (C8H5KO4); (b) Cu2+–phthalic acid monodentate complex (Cu(C8H5O4)+); (c) In3+–phthalic acid monodentate complex
(In(C8H5O4)2+).
peak IICu is possibly ascribed to the further reduction of Cu+ to Cu
by reaction (2) [27]. After the addition of 15 or 23 mM C8H5KO4
to the above Cu2+ + LiCl solution system, the reduction peak ICu
remains nearly unchanged at −0.13 V, whereas the reduction peak
IICu shifts to more negative potential values of −0.95 or −1.1 V,
respectively. Simultaneously, a significant increase of cathodic cur-
rent was observed, showing that C8H5KO4 has little effect on the
reduction of Cu2+ to Cu+ but largely inhibits the reduction of Cu+ to
Cu. As confirmed above, Cu2+ exists mainly in the Cu(C8H5O4)+ form
in C8H5KO4-containing solution, so it is deduced that the reduc-
tion peaks ICu and IICu may correspond to reactions (3) and (4),
respectively.
Fig. 5(c). The weaker complexing interaction of C8H5KO4 with In3+
compared with Cu2+ is probably due to the more stable electronic
state of indium than that of copper.
3.3. Effect of C8H5KO4 additive on the cathodic processes of Cu2+
and In3+
The distinct complexing ability of C8H5KO4 with Cu2+ and In3+
may affect the cathodic reaction processes of the two metal cations
differently, which were investigated by the cathodic polarization
comparison of Cu2+ or In3+ in LiCl supporting electrolyte both with-
out and with various concentrations of C8H5KO4, as shown in Fig. 6.
The polarization curve of Cu2+ in LiCl solution without C8H5KO4 has
two reduction peaks, ICu and IICu, at the potential around −0.13 and
−0.87 V, respectively, as shown in Fig. 6(A). The peak ICu is likely
attributed to the reduction of Cu2+ to Cu+ via reaction (1), and the
Cu2+ + 2Cl− + e− = CuCl2
(1)
(2)
(3)
(4)
−
CuCl2− + e− = Cu + 2Cl−
Cu(C8H5O4)+ + e− = Cu(C8H5O4)
Cu(C8H5O4) + e− = Cu + C8H5O4
−
The intermediate Cu(C8H5O4) cuprous species obtained from
reaction (3) is likely more stable than CuCl2−generated from reac-
tion (1) because of the stronger binding of Cu(I) to the oxygen atoms
in C8H5O4−, thus hindering reaction (4). Furthermore, the charged
groups in solution like Cu(C8H5O4)+ and C8H5O4− can be adsorbed
on the cathodic surface (to be confirmed by EIS analysis in the later
section of this paper), which leads to a reduction of the effective
electrode area for Cu deposition. The two reasons give an accept-
able explanation of why the peak IICu related to the reduction of
Cu(I) to Cu(0) shifts toward more negative potential in the presence
of added C8H5KO4.
The polarization curve of In3+ in the solution without C8H5KO4
shows two reduction peaks IIn and IIIn centered at the potential of
−0.83 and −1.1 V (Fig. 6(B)) that may correspond to reactions (5)
and (6), respectively. Reaction (5) is the rapid reduction of In(III)
to In(I) and reaction (6) is the rate-controlling In(I) → In(0) [28].
Adding 15 or 23 mM C8H5KO4 to the InCl3 + LiCl system, the polar-
ization curves exhibit nearly no change in shape or position of
the two reduction peaks and shows only an increase in cathodic
current, implying that C8H5KO4 has little effect on the electrode
potentials of indium.
In3+ + 2e− = In+
In+ + 1e− = In
(5)
(6)
The aforementioned analyses suggest that the addition of
C8H5KO4 can bring the reduction potentials of Cu(II) → Cu(0) and
In(III) → In(0) close to each other, due to the stronger complexa-
tion of copper ions over indium ion. As the C8H5KO4 concentration
is increased to 23 mM, both the reduction potentials of copper and
indium fall in the range of −1.2 to −0.95 V and is centered around
−1.1 V. In the practical experiments, however, it was noticed that
potentials more negative than −1.0 V usually result in the peeling of
films from substrates. Therefore a C8H5KO4 concentration of 23 mM
Fig. 6. Cathodic polarization curves for (A) Cu2+ and (B) In3+ deposition on Ni sub-
strates in absence and presence of different concentration of C8H5KO4. Electrolyte
(a), (b), (c) containing 12.5 mM Cu2+ + 100 mM LiCl + 0–23 mM C8H5KO4; Electrolyte
(d), (e), (f) containing 10 mM In3+ + 100 mM LiCl + 0–23 mM C8H5KO4; pH 2.5; scan
rate = 10 mV s−1
.

4432
X. Xu et al. / Electrochimica Acta 55 (2010) 4428–4435
Fig. 7. Cathodic polarization curves of Ni electrode in various solution systems: (a) 12.5 mM Cu2+ + 100 mM LiCl; (b) 40 mM S2O32− + 100 mM LiCl; (c) 12.5 mM Cu2+ + 40 mM
S2O32− + 100 mM LiCl; (d) 12.5 mM Cu2+ + 40 mM S2O32− + 100 mM LiCl + 23 mM C8H5KO4; (e) 10 mM In3+ + 100 mM LiCl; (f) 40 mM S2O32− + 100 mM LiCl; (g) 10 mM
In3+ + 40 mM S2O32− + 100 mM LiCl; (h) 10 mM In3+ + 40 mM S2O32− + 100 mM LiCl + 23 mM C8H5KO4. Experimental conditions: pH 2.5; scan rate = 10 mV s−1
.
and applied potential of −1.0 V were used as the optimum condi-
tions for the one-step electrodeposition of CuInS2 thin films and
give a theoretical explanation for the XRD results shown in Fig. 1.
nificant current slope detected in the wide potential range from 0
2−
to −0.7 V suggests that the reduction of S2O3
starts before 0 V.
The approximate flat appearing between −0.4 and −0.6 V denotes
2−
the maximum reduction rate of S2O3
to S, and the small peak
occurred at more negative potential than −0.6 V may correspond
to the further reduction of S to S2− [19,29]. Based on the above
results the two successive current waves illustrated from 0 to
−0.6 V in the polarization curve scanned from Cu2+ + S2O32− + LiCl
3.4. CuInS2 electrodeposition mechanism and kinetics
3.4.1. Polarization studies
To investigate the CuInS2 electrodeposition mechanism and
kinetics, it is first necessary to examine the effect of C8H5KO4 on
the deposition behaviors of some secondary phases like Cu, In, S,
CuxS and InxS by using polarization measurements. These impuri-
ties likely emerge during the one-step electrodeposition process of
CuInS2 films due to the wide differences in the electrode potentials
for copper, indium and sulfur [15]. Fig. 7(a)–(h) shows the cathodic
system, shown in Fig. 7(c), are attributed to the reduction of
2−
Cu(II) to Cu(I) and S2O3
to S. The third reduction wave cen-
tered at −0.8 V is ascribed to the formation of S2− and CuxS species
such as CuS and Cu2S. By adding 23 mM C8H5KO4 to the mixture
of Cu2+ + S2O32− + LiCl, a new reduction peak at potential around
−1.1 V is observed in the polarization curve shown in Fig. 7(d). The
unique and reasonable explanation is that the deposition poten-
tial of CuxS shifts towards the more negative potential under the
influence of C8H5KO4 additive.
polarization curves of the individual species Cu2+, In3+, S2O32− and
the mixture Cu2+ + S2O32−, In3+ + S2O3
trolyte without and with 23 mM C8H5KO4.
in LiCl supporting elec-
2−
In Fig. 7(e), the polarization curve of In3+ in LiCl solution exhibits
two reduction peaks in the potential ranges of −0.6 to −1.0 V and
−1.0 to −1.2 V, from the transform of In(III) to In(I) and In(I) to
In Fig. 7(a) the two observed reduction peaks at potentials of
−0.13 and −0.87 V correspond to the reduction of Cu(II) to Cu(I)
and Cu(I) to Cu(0), respectively [27]. In Fig. 7(b) and (f), the sig-

X. Xu et al. / Electrochimica Acta 55 (2010) 4428–4435
4433
In(0), respectively [28]. As S2O3 is mixed into the In3+ solution,
the polarization current shown in Fig. 7(g) displays a sharp decrease
in the range of 0 to −0.4 V and sustaining stabilization from −0.4
2−
2−
to −0.6 V, which agrees well with the reduction process of S2O3
to S as illustrated in Fig. 7(b) and (f). Whereas the flat region in the
more negative potentials of −0.6 to −0.9 V is inferred to correspond
to the reduction of S to S2− and the conversion of In(III) to In(I).
Thus the significant peak appeared at around −1.1 V is assigned
to the formation of InxS such as InS and In2S3. After the addition
of 23 mM C8H5KO4 to In3+ + S2O32− + LiCl system, the polarization
process shown in Fig. 7(h) changes little, the deposition of InxS still
occurs at the potential around −1.1 V.
From the discussion above, it is known that C8H5KO4 has lit-
tle effect on the reduction of Cu(II) → Cu(I), S2O32− → S → S2− and
In(III) → In(I). The reduction rates of Cu(II) → Cu(I) and S2O32− → S
achieve to the maximum values in the potential ranges of 0 to
−0.4 V and −0.4 to −0.6 V, respectively, whereas the maximum
reduction of S → S2− and In(III) → In(I) take place at potentials more
negative than −0.6 V, which is consistent with the values deduced
from Nernst equations. For the binary phases, however, the depo-
sition potential of CuxS (−0.8 V) is more positive than that of InxS
(−1.1 V) in the absence of C8H5KO4, which may cause the pref-
erential formation of CuxS over InxS. In contrast, the addition of
23 mM C8H5KO4 brings the deposition potential of CuxS closer to
that of InxS (−1.1 V), which likely contributes to the preparation of
pure CuInS2 films in one-step electrodeposition.
Fig. 9. Electrochemical impedance spectra of Ni electrode in the electrolytic
baths containing 12.5 mM Cu2+ + 10 mM In3+ + 40 mM S2O32− + 100 mM LiCl with-
out C8H5KO4 (a) and with 23 mM C8H5KO4 (b). Experimental conditions: pH 2.5;
scan rate = 10 mV s−1. E = −1.0 V vs. SCE.
and the reduction peaks at −1.0 V in the both polarization curves
are deduced to correspond to the deposition of CuInS2. Therefore,
the three-step deposition mechanism of CuInS2 in the electrolytic
bath free of C8H5KO4 can be illustrated by reactions (1), (7) and (8).
After the addition of 23 mM C8H5KO4, however, the CuInS2 depo-
sition mechanism can be depicted by the reactions (3), (7) and (9).
Accordingly, the possibility of forming other compounds like Cu, In,
S and InxS can be ruled out.
On the basis of the above investigation of the reduction of indi-
vidual Cu2+, In3+, S2O3
and the formation of binary CuxS and
2−
InxS impurities, the deposition of ternary CuInS2 in the mixture
solution of Cu2+, In3+, S2O32−, and LiCl both without and with
23 mM C8H5KO4 was also examined by cathodic polarization mea-
surements (Fig. 8). The polarization curves scanned in both of the
solution systems exhibit three significant current waves in the
potential ranges of 0 to −0.4 V, −0.4 to −0.8 V and −0.8 to −1.2 V,
implying a three-step reaction mechanism. According to the anal-
yses above, the first wave is mainly attributed to the reduction
of Cu(II) → Cu(I) and the second is attributed to the transforma-
tion of S2O32− → S → S2− and In(III) → In(I). Both of these processes
change little before and after the addition of C8H5KO4. Neverthe-
less, the third wave corresponding to the deposition of sulfide
exhibits some differences in the shape and apparent reduction cur-
rent without and with the influence of C8H5KO4. Referring to the
XRD results and the aforementioned forming conditions of binary
impurities, the small current inflexion appeared between −0.8 and
−0.95 V in the absence of C8H5KO4 and disappeared in the pres-
ence of C8H5KO4. This peak is assigned to the formation of Cu2S
ꢀ
S2O2− + 6H+ + 4e− = 2S + 3H2O
3
S + 2e− = S2−
(7)
In3+ + 2e− = In+
ꢁ
3CuCl−2 + In+ + 3S + 4e− = Cu2S + CuInS2 + 6Cl−
(8)
CuCl−2 + In3+ + 2S2− = CuInS2 + 2Cl−
ꢁ
Cu(C8H5O4) + In+ + 2S + 2e− = CuInS2 + C8H5O−
4
(9)
Cu(C8H5O4) + In3+ + 2S2− = CuInS2 + C8H5O−
4
3.4.2. Electrochemical impedance spectroscopy analysis
Fig. 9 compares the EIS plots for the Ni cathode in the electrolytic
baths containing Cu2+, In3+, S2O32−, and LiCl both without and with
23 mM C8H5KO4. The EIS plot measured in the absence of C8H5KO4
exhibits two small distorted capacitive loops at high frequencies
followed by a slightly larger loop at intermediate frequencies and an
oblique line (similar to 45◦) at low frequencies as shown in Fig. 9(a).
The capacitive loops correspond to the charge transfer across the
double layer [21]. The oblique line, a typical feature of Warburg
impedance [30], associated with a Nernst diffusion of reactant ions
from the electrode surface to bulk solution. By combining these
results with the polarization measurements above, an equivalent
circuit connected a Rs in series with three (RC) and a W is proposed
to simulate the EIS data, as depicted in Fig. 10(a). Rs is the solu-
tion resistance, W is the Warburg impedance, R1, R2 and R3 are the
charge-transfer resistance corresponding to reactions (1), (7) and
(8), respectively. The Cdl-1 paralleled with R1 and Cdl-2 paralleled
with R2 represent the double layer capacitance related with the
reactions (1) and (7), whereas the Cad paralleled with R3 reveals
formation of an adsorption film on the cathode surface during the
reaction (8). This adsorption is tho−ught to derive from a chloride-
containing intermediate like CuCl2 produced by reaction (1) [27].
The adsorption effect of this process is not strong enough to result
in the observation of an inductive loop in the EIS plot.
Fig. 8. Cathodic polarization curves of Ni electrode in the electrolytic bath
containing 12.5 mM Cu2+ + 10 mM In3+ + 40 mM S2O32− + 100 mM LiCl without
C8H5KO4 (a) and with 23 mM C8H5KO4 (b). Experimental conditions: pH 2.5; scan
According to the assumption established through the polariza-
tion tests, the first two reactions (1) and (7) proceed independently
rate = 10 mV s−1
.

4434
X. Xu et al. / Electrochimica Acta 55 (2010) 4428–4435
Accordingly, an equivalent circuit like that shown in Fig. 10(b)
is used to simulate the EIS data measured in the pres-
ence of C8H5KO4. (R1Cdl-1), (R2Cdl-2) and (R3Cad) represent the
resistance–capacitance circuits associated with the charge-transfer
reactions (3), (7) and (9), respectively. As discussed above, Cdl-1 and
Cdl-2 also denote the common double layer capacitance, whereas
Cad describes the formation of a transitional adsorption layer at
the cathode that is derived from the intermediate Cu(C8H5O4)
produced by the reaction (3). In the equivalent circuit (R1Cdl-1
)
is connected in parallel with (R2Cdl-2) and both are connected in
series with (R3Cad). This connectivity is based on the assumption
that the initial two reactions (3) and (7) proceed independently
but are dependent on reaction (9). Furthermore, a parallel arrange-
ment (RLL), where RL and L represent the resistance and inductance
related to the adsorption/desorption process of Cu(C8H5O4)+ and
C8H5O4−, is connected in parallel with the three (RC) circuits
because the adsorption/desorption phenomenon runs throughout
the entire CuInS2 deposition process in the C8H5KO4-containing
electrolytic bath. Additionally, the solution resistance Rs is con-
nected in series in the equivalent circuit depending on the general
rule. Notice that the circuit model shown in Fig. 10(b) gives a rea-
sonable illustration to the EIS data presented in Fig. 9(b).
4. Conclusions
The effect of C8H5KO4 as a special additive on the one-step elec-
trodeposition of CuInS2 thin films from acidic solution containing
Fig. 10. Equivalent circuits utilized to simulate the electrochemical impedance
spectra of Ni electrode in the electrolytic baths containing 12.5 mM Cu2+ + 10 mM
In3+ + 40 mM S2O32− + 100 mM LiCl without C8H5KO4 (a) and with 23 mM C8H5KO4
(b).
the precursor ions Cu2+, In3+ and S2O3 has been investigated in
2−
detail for the first time. The results show that the addition of ade-
quate concentration of C8H5KO4 (23 mM) to the electrolytic bath
can contribute significantly to the controllable growth of pure chal-
copyrite CuInS2 films with uniform surfaces and an ideal band gap
of 1.54 eV. The studies on the mechanism of action of C8H5KO4
additive demonstrate that (1) Cu(C8H5O4)+ and In(C8H5O4)2+ com-
plexes are formed from the binding of Cu2+, In3+ ions and –COO−
during the entire process of CuInS2 deposition from C8H5KO4-free
solution. Therefore, the connection of (R1Cdl-1) in the equivalent cir-
cuit is in parallel with (R2Cdl-2). Both circuits, however, are designed
to connect in series with the (R3Cad) circuit because the initial two
reactions (1) and (7) provide the reactants for reaction (8). As the
reactants are consumed, the diffusion, rather than charge trans-
fer, becomes the major controlling factor to the CuInS2 deposition,
thus the Warburg impedance is connected in series with (R3Cad).
Additionally, the solution resistance Rs, which is often a significant
factor in the impedance of an electrochemical cell, is also connected
in series in the circuit model as in many other reports. The circuit
model given in Fig. 10(a) can simulate the EIS plot presented in
Fig. 9(a), thus illustrating that the CuInS2 electrodeposition in the
absence of C8H5KO4 is controlled by a combination of kinetic and
diffusion processes. These processes, however, are not beneficial to
form a single-phase CuInS2 film.
Upon the addition of C8H5KO4 the EIS plot displays a distinct
dynamic response as shown in Fig. 9(b). The much larger capac-
itive loops present in the high and intermediate frequency range
may be associated with the electrode reactions (3), (7) and (9). The
inductive loop appeared at low frequencies due to relaxations of the
electrode area by the formation and growth of monolayers formed
on the facets of crystallites [21] and suggests that some adsorbed
species significantly participate in the CuInS2 electrodeposition
process. The adsorbed species are deduced to be Cu(C8H5O4)+ and
C8H5O4−. The former Cu(C8H5O4)+ adsorption originates from the
electrostatic attraction between heterogeneous charges, since the
Ni working−electrode is used as the cathode, whereas the lat-
ter C8H5O4 adsorption probably derives from the interaction of
metallic Ni with the carboxylate group of C8H5O4− [31]. The strong
adsorption on the Ni cathode surfaces limits the effective elec-
trode areas for Cu deposition and consequently inhibits the rate
of the Cu2+ reduction process, which in turn promotes the co-
deposition of copper, indium and sulfur in the form of single-phase
CuInS2.
−
group in C8H5O4−; (2) the complexing capability of C8H5O4 is
strong for Cu2+ but weak for In3+ and can bring the electrode poten-
tials of Cu2+ and In3+ close to each other; and (3)−the considerable
adsorption of charged Cu(C8H5O4)+ and C8H5O4 on the cathode
surfaces can make the effective electrode area for Cu deposition
limited, consequently inhibiting the kinetics of the Cu2+ reduction
process. In conclusion, the synergetic effects of complexation and
adsorption interfere with the Cu2+ reduction and promote the co-
deposition of copper, indium and sulfur in the form of chalcopyrite
CuInS2. Accordingly, a three-step reaction mechanism was pro-
posed to illustrate the CuInS2 deposition process in the presence
of C8H5KO4 additive. Step 1: the reduction of Cu(II) → Cu(I); Step
2: the reduction of S2O32− → S → S2− and In(III) → In(I); and Step 3:
the formation of pure CuInS2.
Acknowledgments
The authors acknowledge the financial support from the pro-
gram of Introducing Talents of Discipline to Universities (B08003).
References
[1] S.S. Lee, K.W. Seo, J.P. Park, S.K. Kim, I.I.-W. Shim, Inorg. Chem. 46 (2007) 1013.
[2] R. Garuthara, R. Wijesundara, W. Siripala, Sol. Energy Mater. Sol. Cells 79 (2003)
331.
[3] H.W. Schock, Sol. Energy Mater. Sol. Cells 34 (1994) 19.
[4] R.N. Bhattacharya, W. Batchelor, J.E. Granata, F. Hasoon, H. Wiesner, K.
Ramanathan, J. Keane, R.N. Noufi, Sol. Energy Mater. Sol. Cells 55 (1998) 83.
[5] S.J. Peng, J. Liang, L. Zhang, Y.H. Shi, J. Chen, J. Cryst. Growth 305 (2007) 99.
[6] T. Yukawa, K. Kuwabara, K. Koumoto, Thin Solid Films 286 (1996) 153.
[7] S.P. Grindle, C.W. Smith, S.D. Mittleman, Appl. Phys. Lett. 35 (1979) 24.
[8] G. Hodes, T. Engelhard, C.R. Herrington, L.L. Kazmetski, D. Cahen, Prog. Cryst.
Growth Charact. 10 (1984) 241.

X. Xu et al. / Electrochimica Acta 55 (2010) 4428–4435
4435
[9] H.Y. Ueng, H.L. Hwang, J. Phys. Chem. Solids 50 (1989) 1297.
[10] K. Topper, J. Bruns, R. Scheer, M. Weber, A. Weidinger, D. Brauning, Appl. Phys.
Lett. 71 (1997) 482.
[20] S.K. Panda, S. Gorai, S. Chaudhuri, Mater. Sci. Eng. B 129 (2006) 265.
[21] Q.B. Zhang, Y.X. Hua, Y.T. Wang, H.J. Lu, X.Y. Zhang, Hydrometallurgy 98 (2009)
294.
[11] H.J. Lewerenz, N. Dietz, J. Appl. Phys. 73 (1993) 4975.
[12] D. Braunger, D. Hariskos, T. Walter, H.W. Schock, Sol. Energy Mater. Sol. Cells
40 (1996) 97.
[13] J. Klaer, J. Bruns, R. Hrnninger, K. Slemer, R. Klenk, K. Ellmer, D. Brauning,
Semicond. Sci. Technol. 13 (1998) 1456.
[14] S. Nakamura, A. Yamamoto, Sol. Energy Mater. Sol. Cells 49 (1997) 416.
[15] B. Asenjo, A.M. Chaparro, M.T. Gutierrez, J. Herrero, Thin Solid Films 511–512
(2006) 117.
[22] Y.K. Ke, H.R. Dong, Analytical Chemistry Handbook – Spectroscopy Analysis,
Chemical Industry Press, Beijing, 1998 (Chapter 13).
[23] S.E. Cabaniss, P.A. Maurice, G. Madey, Appl. Geochem. 22 (2007) 1646.
[24] S.W. Zhang, F.L. Tang, T. Zhang, Modern Chemistry Reagent Handbook – Chem-
ical Analysis Reagent, Chemical Industry Press, Beijing, 1987, p. 410.
[25] M.A. Ali, D.A. Dzombak, Geochim. Cosmochim. Acta 60 (1996) 302.
[26] I. Baraldi, G. Ginocchietti, U. Mazzucato, A. Spalletti, Chem. Phys. 337 (2007)
168.
[16] C. Romain, B. Florence, B. Marc, P. Marcel, K. Philippe, C. R. Chimie 11 (2008)
1016.
[17] A.M. Martinez, A.M. Fernandez, L.G. Arriaga, U. Cano, Mater. Chem. Phys. 95
(2006) 271.
[27] M.A. Pasquale, L.M. Gassa, A.J. Arvia, Electrochim. Acta 53 (2008) 5891.
[28] M. Mohamedi, J. Bouteillon, J.C. Poignet, Electrochim. Acta 41–9 (1996) 1495.
[29] N. Sato, M. Ichimura, E. Arai, Y. Yamazaki, Sol. Energy Mater. Sol. Cells 85 (2005)
159.
[18] Y.L. Lai, F.Y. Liu, Z. Zhang, J. Liu, Y. Li, S.S. Kuang, J. Li, Y.X. Liu, Electrochim. Acta
54 (2009) 3004.
[19] B. Subramanian, C. Sanjeeviraja, M. Jayachandran, Mater. Chem. Phys. 71 (2001)
41.
[30] J. Zhao, N. Li, S. Li, S. Gao, G.F. Cui, Electrochem. Commun. 9 (2007) 2262.
[31] S.G. Baca, I.G. Filippova, P. Franz, C. Ambrus, M. Gdaniec, H.S. Evans, Y.A.
Simonov, O.A. Gherco, T. Bejan, N. Gerbeleu, S. Decurtins, Inorg. Chim. Acta
358 (2005) 1763.