Fabrication of a CuInS₂ photoelectrode using a single-step electrodeposition with controlled calcination atmosphere

Abstract

n-type CuInS₂ thin films were fabricated by a one-step electrodeposition of Cu/In/S precursors on a transparent fluorine-doped tin oxide (FTO) substrate followed by calcination in a mixture of N₂ and H₂ at 500 °C without the need of toxic H₂S gas for sulfurization. The analyses of structural and optical properties revealed that tetragonal chalcopyrite CuInS₂ thin films were obtained when heat treated in a mixture of 8% H₂ and 92% N₂; whereas calcination in pure N₂ resulted in the formation of CuInS₂ with a considerable amount of In₂O₃ and Cu_xS impurities. The synthesized n-CuInS₂ thin film has an optical bandgap of 1.3 eV with the conduction band at ca. −1.2 V vs. Ag/AgCl at pH 6.5. The heat treatment in a mixture of H₂ and N₂ provided a reducing atmosphere to suppress the oxide formation of the electrodeposited Cu/In/S elements. It also increased the donor density of CuInS₂ sixfold by increasing the density of the sulfur vacancies (known to be electron donors) at the grain surface. As a result of the improved purity and increased donor density, the hydrogen treated CuInS₂ thin film yielded a photocurrent density of ∼8 mA cm⁻² at 0.5 V vs. Ag/AgCl in 0.25 M Na₂S and 0.35 M Na₂SO₃ under visible light illumination, which is a fourfold enhancement as compared to that obtained from the CuInS₂ film calcined in pure N₂.

---

1. Introduction

The ternary CuInS₂ semiconductor is one of the promising chalcopyrite compounds with the I–III–VI₂ configuration. It has a narrow direct band gap of ∼1.3 to 1.5 eV which covers most of the visible component in the solar spectrum.¹ Combined with a large absorption coefficient of more than 10⁵ cm⁻¹, CuInS₂ is an ideal candidate for a photoabsorber in solar cells.² Although its selenide-analogues, such as CuInSe₂, CuGaSe₂ and Cu(In,Ga)Se₂, yield a higher energy conversion efficiency, the higher toxicity of Se drives the research community towards replacing Se with the less hazardous CuInS₂.³ The recent development of CuInS₂ has extended its applications from solid state photovoltaics to liquid-junction photoelectrochemical (PEC) cells for hydrogen generation.^4,5 The tunability of the p-type and n-type attributes of CuInS₂ by manipulating the copper vacancies or the indium interstitials⁶ further encourages the investigation of both types of CuInS₂, with the ultimate goal of tandem-type photoelectrochemical cells that split water with no externally applied bias. Although p–n junction photoelectrodes have been mainly created based on p-type CuInS₂,^7,8 the high quality n-type CuInS₂ has also manifested its importance in the construction of an effective p–n junction for solar energy conversion applications: Tell and Thiel reported the formation of the a p–n junction electrode using both n-type and p-type CuInS₂ to yield a promising quantum efficiency;⁹ Ito et al. introduced a thin interlayer of n-type CuInS₂ between p-type CuInS₂ and n-type CdS for thin film solar cells;¹⁰ Arici and coworkers formed flat-interface donor–acceptor hybrid solar cells using n-type CuInS₂ and p-type PEDOT:PSS.¹¹ Therefore, further efforts in synthesizing CuInS₂, either in p-type or n-type behavior, will clearly yield a valuable insight on its photo-chemical/physical properties in relation to the performance in sunlight conversion applications.

Fabrication of chalcopyrite thin films traditionally involves a high-vacuum evaporation/deposition technique.¹² To reduce the high preparation cost associated with the high vacuum, a two-step process has been developed where the metallic components (Cu, In, Ga, Zn or Sn) are initially deposited on a conducting substrate at milder conditions followed by a sulfurization step at a high temperature.¹³ A variety of techniques has been used for metallic deposition such as sputtering,¹⁴ spray pyrolysis,¹⁵ molecular beam deposition¹⁶ and electrodeposition.² In the subsequent sulfurization process at a high temperature, H₂S or elemental sulfur are the most used precursors despite their toxicity in the gas phase. Regardless of the deposition technique of the metallic precursors and the source of sulfurization, an annealing process is necessary to crystallize the deposited components into their chalcopyrite structure. The morphology (formation of pinholes) as well as the film stability (adherence) can be optimized by altering the annealing conditions.¹⁷

Considering the advantages of using the non-vacuum and two-step film synthesis, electrodeposition is an attractive option for depositing the metallic components of CuInS₂ since it can proceed at low/room temperature, requires a simple equipment setup, and generates no waste emission. As a greener fabrication process is increasingly required, efforts have been made in excluding H₂S throughout the synthesis of chalcopyrite semiconductor thin films.^18–20 This motivated us to investigate the simultaneous deposition of sulfides during the metal deposition in this work eliminating the use of H₂S in the subsequent high temperature annealing. We have recently reported the formation of a CuInS₂–TiO₂ nanotube composite using a square wave pulse electrodeposition technique that enables the penetration of soluble CuInS₂ (CuCl₂, InCl₃ and Na₂S₂O₃) precursors into the capillary force resistant pore channels of the TiO₂ nanotubes for deposition.²¹ Despite the successful synthesis of TiO₂ nanotubes completely wrapped with CuInS₂ nanoparticles, the formation of In₂O₃ impurities was observed. The oxidation of the deposited metallic In induced a stoichiometric imbalance of metallic Cu : In ratio and, therefore, promoted the formation of oxides of Cu or Cu_xS. Thus, the suppression of the formation of In₂O₃ during the synthesis is important. In this study, we attempted to fabricate CuInS₂ thin film using a single-step electrodeposition method followed by crystallization of the deposited precursors in a reducing atmosphere. Specifically, the effects of using a mixture of N₂ and H₂ on the suppression of impurities as related to the electronic properties and the PEC performance is discussed.

2. Experimental

Preparation of a CuInS₂ thin film on the FTO substrate

CuInS₂ on an FTO substrate film was prepared by a single-step electrodeposition method using 60 mL of an aqueous solution containing 30 mM CuCl₂ (99.9%, Ajax Finechem), 30 mM of InCl₃ (98%, Sigma-Aldrich) and 300 mM of Na₂S₂O₃ (98%, Sigma-Aldrich). The FTO substrates were pre-cleaned with Milli-Q water, ethanol and acetone under mild sonication, and followed by drying in vacuo before use. By changing the volume ratio of the three precursor solutions the Cu : In : S molar ratios were adjusted prior to electrodeposition. A three-electrode system was used for electrodeposition, with the FTO substrate as the working electrode and Ag/AgCl and platinum foil as the reference and the counter electrodes, respectively. In a cyclic voltammetry (CV) measurement, the applied voltage is scanned forward and backward in the presence of metallic ions. Reduction of copper (Cu) and indium (In) ions was observed to initiate in the range of −0.1 V and −0.7 V. The maximum reduction reactions occurred at −0.6 V and −1.2 V, respectively, for Cu and In. To achieve simultaneous electrodeposition of Cu and In, the voltage was fixed at −1 V as both metallic ions can be reduced effectively. The voltage The electrodeposition duration was set at 15 min. The-obtained CuInS₂ film was dried and annealed at 500 °C for 1 h either in pure N₂ or in a gas mixture of 92% N₂ and 8% H₂ to facilitate crystallization. Before the heating, N₂ was purged through the tube furnace for 2 h to completely remove the air. Subsequently, the heating was performed with the ramping rate of 5 °C min⁻¹ from room temperature to 500 °C. The gas flow rate was maintained at 50 mL min⁻¹ throughout the process.

Characterization of the obtained CuInS₂ films

The crystallographic phase structures of the deposited films were characterized using an X-ray diffractometer (X'pert Pro MRD, Philips) with Cu K_α radiation at 45 kV and 40 mA, a step size of 0.013° and a scan step time at 97.92 s in the 2θ range of 25° to 60°. The morphological features and the film thickness were determined using a scanning electron microscope (SEM, S900 Hitachi). The UV-Vis diffuse reflectance spectra were obtained using a UV/Vis/NIR spectrophotometer (Perkin Elmer LAMBDA 1050) with GaP (1200–900 nm) and Si (900–250 nm) detectors.

The electrochemical impedance experiments were carried out in a three-electrode electrochemical system with a 0.5 M K₂SO₄ electrolyte (pH 6.3–6.5). The counter and the reference electrodes were platinum wire and Ag/AgCl, respectively. The Mott–Schottky plot of the CuInS₂ film was measured using a potentiostat (Autolab Model PGSTAT 12) with a frequency response analyzer (Autolab FRA2 modules). The frequencies of 10 Hz, 100 Hz and 1000 Hz were used for the measurements. Additionally, the applied potential was scanned from −1 V to 0.1 V (vs. Ag/AgCl) with a scan rate 50 mV s⁻¹.

Photoelectrochemical (PEC) characterizations

The photoelectrochemical properties of the CuInS₂ films were examined under potentiostatic conditions in a three-electrode system using a CuInS₂/FTO as the working electrode, an Ag/AgCl as the reference electrode and a platinum wire as the counter electrode in an aqueous solution containing 0.25 M of Na₂S and 0.35 M of Na₂SO₃ at pH 12. An electrolytic cell made out of Teflon with a flat quartz window was used. The photocurrent responses were measured under visible light illumination using a 300 W Xenon lamp with a cut off filter (λ ≥ 435 nm). The illuminated area of the working electrode was fixed at 0.196 cm².

3. Results and discussion

Upon determining the electrodeposition voltage using cyclic voltammetry, the precursor solution containing the pre-calculated molar ratio of Cu : In : S was electrodeposited on FTO glass for 15 min. A homogeneous layer of dark-brown color film was observed on the FTO surface immediately following the application of −1 V bias, indicating the deposition of a sulfur-containing Cu and In. Fig. 1 shows XRD patterns of the deposited films with different Cu : In : S precursor molar ratios after annealing in N₂ at 500 °C for 1 h. The formation of chalcopyrite CuInS₂ was observed in the films prepared using the precursor solutions with a slightly higher In-content (In/Cu = 0.8–2.5) and a largely sulfur-rich environment.

Fig. 1 XRD patterns of electrodeposited films with different precursor molar ratios of Cu : In : S. All films were electrodeposited for 15 min and annealed at 500 °C in N₂ for 1 h.

It is known that a partial evaporation of In component during the synthesis will result in an actual lower amount of deposited In; therefore, a higher concentration of the In-precursor was necessary to ensure the desired stoichiometry of the resulting crystalline CuInS₂.²² For a similar reason, sulfur depletion was expected during the annealing treatment, especially when it was performed in the absence of a gas phase sulfur source in the calcination environment. Hence, an excess amount of sulfur in the precursor solution was needed. Besides the crystalline CuInS₂, a significant amount of In₂O₃ was detected in all samples independent of the precursor molar ratios used. The incomplete replacement of air by N₂ purging together with the diffusion of oxygen from the FTO substrate at 500 °C may contribute to the oxidation of CuInS₂ or In in its elementary form. As the intermediate formation of the Cu_xIn_x alloy phases (such as Cu₇In₃ and Cu₁₁In₉) takes place before CuInS₂ is formed, the reaction between Cu and In as the temperature is raised is important. When the deposited In was melted at 157 °C the inter-diffusion of solid Cu and liquid In led to an alloy formation. The contact of liquid In with a possible oxygen source from the annealing atmosphere and the FTO substrate at a high temperature promoted the crystallization of In₂O₃.

Based on Fig. 1, the optimum precursor molar ratio of Cu : In : S = 1 : 0.8 : 10 was chosen for the subsequent preparation of CuInS₂ with a suitable film thickness. The high absorption coefficient of CuInS₂ implies that the film thickness of only a few μm is required to absorb the light across the visible spectrum.⁷ The electrodeposition duration is a well-established tool to control the film thickness as it systematically follows Faraday's law where the coulomb of electrodeposition (charge transported, i.e. the amount of deposited elements) is a function of the deposition time under the potentiostatic conditions. Fig. 2 shows XRD patterns of a CuInS₂ thin films deposited with increasing deposition duration at −1.0 V with the precursor molar ratio of Cu : In : S = 1 : 0.8 : 10. The diffraction peaks at 27.9°, 46.5° and 54.8° attributed to the CuInS₂ planes (112), (204)/(220) and (116)/(312) were clearly seen in the samples with the electrodeposition time of 15 min and longer. The comparison of the peak intensity for 27.9° (CuInS₂) and 26.6° (SnO₂ from the FTO substrate) confirms the increasing amount of CuInS₂ deposited onto FTO substrates with time. However, the peaks attributed to In₂O₃ and Cu_xS appeared to be overwhelming when prolonging deposition time (120 min) was employed. As In₂O₃ and Cu_xS are in general undesired impurities in a sulfide-type chalcopyrite crystal,^23,24 a 30 min electrodeposition time was picked as the optimum also supported by the suitable film thickness as revealed by the SEM analysis discussed below.

Fig. 2 XRD patterns of CuInS₂ samples with different deposition times. All films were the electrodeposited with precursor molar ratio of Cu : In : S = 1 : 0.8 : 10 and annealed at 500 °C in N₂ for 1 h.

Fig. 3 shows the SEM images of the CuInS₂ film electrodeposited for 30 min. An evenly distributed layer composed of compact aggregates of angular-shaped particles was observed (Fig. 3(a)). The surface coverage of the film was considerably uniform and the grain size was about 400 nm. The relatively rough surface of the film could facilitate light trapping through the internal reflection of the incoming illumination on the particles and thus benefit the light absorption. A cross-sectional SEM image of the corresponding thin film (Fig. 3(b)) indicates a CuInS₂ thickness of ∼1.27 μm. Given the high optical absorption coefficient of CuInS₂ at 10⁵ cm⁻¹, 1–2 μm thick layer is sufficient for absorbing most of the incident light. Furthermore, cracks at the interface between CuInS₂ particles and the FTO substrate normally found in samples synthesized using the high temperature sulfurization methods were not observed. This is highly advantageous as such defects cause complications when films are used in photoelectrochemical cells with liquid electrolytes.

Fig. 3 SEM images of CuInS₂ electrodeposited for 30 min, with the precursor molar ratio of Cu : In : S= 1 : 0.8 : 10 and annealed at 500 °C in N₂ for 1 h: (a) top view; (b) cross sectional view.

In the absence of H₂S in the annealing process, N₂ purging was ineffective in sealing the thin films from the remaining air as O₂ has a higher density than N₂. This resulted in the formation of In₂O₃. When In forms In₂O₃, the resulting Cu/In/S stoichiometry makes Cu_xS more favorable. Therefore, the suppression of the liquid In oxidation (as In was first melted at 157 °C under the annealing conditions) is important. Instead of pure N₂, a mixture of 92% N₂ and 8% H₂ was introduced in the annealing process to provide a reducing atmosphere during the crystallization. (The sample is named “H₂-annealed CuInS₂” for simple reference.) Fig. 4 shows the XRD patterns of CuInS₂ films electrodeposited for 30 min and annealed in either pure N₂ or a mixture of N₂ and H₂. It is clearly seen from the XRD patterns that the H₂-annealed CuInS₂ film demonstrated high purity since the contribution from In₂O₃ and Cu_xS was hardly detected. Thus the strategy of suppressing the oxidative reactions was found to be effective to prepare CuInS₂. As-deposited film without annealing was also included for comparison to clarify the amorphous nature of the deposited elements. There was no appreciable difference in the morphological features of the H₂-annealed CuInS₂ as compared with the N₂-annealed CuInS₂ (SEM data shown in the ESI† document).

Fig. 4 XRD patterns of CuInS₂ electrodeposited with the precursor molar ratio of Cu : In : S = 1 : 1 : 10 for 30 min followed by (a) H₂/N₂ annealing; (b) N₂ annealing; (c) no annealing; and (d) the FTO substrate.

Fig. 5(a) and (b) shows the Tauc plots of N₂- and H₂-annealed annealed CuInS₂. The direct bandgap for N₂ annealed CuInS₂ derived from the intercept of the absorbance edge and the background line was 2.25 eV, indicating the composite nature of the sample containing significant amount of In₂O₃. In contrast, Tauc plot of H₂-annealed CuInS₂ reveals a 1.3 eV direct bandgap, which is in good agreement with the XRD pattern of CuInS₂ in Fig. 4. The improvement in purity of CuInS₂ influenced the electronic properties of the films. In the study of the semiconductor–electrolyte interface, Mott–Schottky analysis is used to determine the flat band potential, carrier (doping) density and the nature (p-type or n-type) of the semiconductor. When the carrier distribution and the electric field within the space charge region are described by Boltzmann distribution and Gauss' law, the Mott–Schottky equation can be expressed as:

1/C² = (2/εε₀eN_D)[E − E_fb − (kT/e)] (1)

where C is the capacitance of the space charge region, ε is the dielectric constant, ε₀ is the permittivity of vacuum, e is the electronic charge, N_D is the donor density (electron donor concentration for an n-type semiconductor), E is the applied potential, E_fb is the flatband potential, k is the Boltzmann's constant and T is the temperature. Therefore, the Mott–Schottky plot of C⁻² versus applied voltage affords a curve with linear region, in which the slope yields the carrier density and its extrapolation with voltage intersection yields the flat-band potential.

Fig. 5 Tauc plots and Mott–Schottky plots of N₂-annealed CuInS₂ (a and c) and H₂-annealed CuInS₂ (b and d).

The Mott–Schottky plots of the N₂-annealed and H₂-annealed CuInS₂ are shown in Fig. 5(c) and (d). The positive slopes of 1/C² verified that both samples are n-type semiconductors. The E_fb value was determined from extrapolating for 1/C² to 0. The E_fb potentials for N₂-annealed and H₂-annealed CuInS₂ were estimated to be −0.62 V and −1.2 V vs. Ag/AgCl at pH 6.5. These negative values of E_fb are also indicative of the n-type behavior. For H₂-annealed CuInS₂, considering that the Fermi level of an n-type semiconductor is close to the conduction band, its conduction band edge (V_CB) and valance band edge (V_VB) were estimated to lie at ∼−1.2 V and ∼+0.1 V vs. Ag/AgCl, respectively. These band edges are in good agreement with the previously reported values.⁷ On the other hand, due to the presence of considerable amounts of In₂O₃ as indicated by the XRD and Tauc analyses, the equilibrated E_fb has shifted to a more positive value in N₂-annealed CuInS₂. Meanwhile, the donor density level (N_D) can be calculated from the slope of the Mott–Schottky plot (N_D is inversely proportional to the slope). However, since the Mott–Schottky method is based on a flat electrode model which the CuInS₂ films in this work do not fall under, the estimation of the absolute value of the donor density may result in inaccuracy. Hence, the donor densities for N₂-annealed and H₂-annealed CuInS₂ were compared qualitatively based on the slopes of the plots. The slope values as shown in Fig. 5(c) and (d) revealed a sixfold enhancement in donor density of H₂-annealed CuInS₂ over N₂-annealed CuInS₂. Improved purity of CuInS₂ is one of the plausible explanations. The activation of sulfur vacancies in the grain surface in the presence of H₂ during calcination is another contributing factor since it is known to be an electron donor for CuInS₂.²⁵ The increased donor density is beneficial for the charge transport within the bulk CuInS₂ film thus improving its overall photoconversion performance.

Fig. 6 shows the photocurrent generated from the N₂-annealed and H₂-annealed CuInS₂ films under repeated on-off visible light illumination (λ ≥ 435 nm) cycles. Both CuInS₂ films generated anodic current upon photoexcitation which is another indication of their n-type behavior. Although the photoresponses were reproducible during the illumination cycles, the promptness of the response was generally inferior to the oxide-based photoactive semiconductors (e.g. TiO₂ and ZnO).²⁶ The slow photoresponses are associated with the relatively slow electron transfer kinetics and the gradual photocorrosion of CuInS₂ (its intrinsic instability). In these n-type CuInS₂ films, the generated holes, when not consumed effectively, might be responsible for the self-oxidation of CuInS₂.^27,28 This is supported by the poorer photostability of the CuInS₂ film when Na₂SO₄ solution was used as the electrolyte in PEC experiments (data not shown). A careful selection of electrolyte (including holes scavenging or redox electrolytes) is, therefore, helpful to improve its photostability. A photocurrent density of 8 mA cm⁻² was generated in H₂-annealed CuInS₂ whereas a value of only 2 mA cm⁻² was observed in N₂-annealed CuInS₂ under the identical conditions. This suggests that the improved purity and the populated donor density in H₂-annealed CuInS₂, as proven by the XRD and Mott–Schottky analyses, are beneficial to the overall electron–hole pair generation and the collection efficiency.

Fig. 6 Photocurrent responses of N₂- and H₂-annealed CuInS₂ measured at 0.5 V vs. Ag/AgCl in aqueous electrolyte containing 0.25 M Na₂S and 0.35 M Na₂SO₃ under visible illumination (λ ≥ 435 nm).

4. Conclusions

A simultaneous electrodeposition of elementary Cu/In/S in a tailored precursor molar ratio followed by annealing in a reducing atmosphere containing 92% N₂ and 8% H₂ was found to be effective in preparing CuInS₂ without the need of utilizing H₂S. The formation of the undesirable impurities, such as In₂O₃ and Cu_xS, is completely suppressed under these conditions in contrast to samples calcined in an inert atmosphere. In addition to the improved purity, the electrochemical impedance spectroscopy measurements revealed that the electron donor density of CuInS₂ annealed in reducing atmosphere was approximately six times that of the samples calcined in an inert atmosphere due to the activation of the sulfur vacancies during H₂ calcination. As result of the improved purity and populated donor density, the current generation of the CuInS₂ thin films calcined in the mixture of N₂ and H₂ manifests a fourfold enhancement in the photochemical performance under visible light illumination as compared to the films prepared under N₂ alone.

Acknowledgements

This work has been supported by the Australian Research Council Discovery Project (DP0986398 and DP110101638). Yiming Tang also acknowledges the scholarship under the State Scholarship Fund awarded by China Scholarship Council. The authors would also like to acknowledge the UNSW Mark Wainwright Analytical Centre, and particularly thank Dr Yu Wang for his generous help in the XRD analysis.

Notes and references

1. H. J. Lewerenz, H. Goslowsky, K. D. Husemann and S. Fiechter, Nature, 1986, 321, 687.
2. R. P. Wijesundera and W. Siripala, Sol. Energy Mater. Sol. Cells, 2004, 81, 147.
3. S. T. Connor, C. M. Hsu, B. D. Weil, S. Aloni and Y. Cui, J. Am. Chem. Soc., 2009, 131, 4962.
4. S. Ikeda, M. Nonogaki, W. Septina, G. Gunawan, T. Harada and M. Matsumura, Catal. Sci. Technol., 2013, 3, 1849.
5. K. Kobayakawa, A. Teranishi, T. Tsurumaki, Y. Sato and A. Fujishima, Electrochim. Acta, 1992, 37, 465.
6. G. Massé, N. Lahlou and C. Butti, J. Phys. Chem. Solids, 1981, 42, 449.
7. S. Ikeda, T. Nakamura, S. M. Lee, T. Yagi, T. Harada, T. Minegishi and M. Matsumura, ChemSusChem, 2011, 4, 262.
8. D. Yokoyama, T. Minegishi, K. Jimbo, T. Hisatomi, G. Ma, M. Katayama, J. Kubota, H. Katagiri and K. Domen, Appl. Phys. Express, 2010, 3, 101202.
9. B. Tell and F. Thiel, J. Appl. Phys., 1979, 50, 5045.
10. K. Ito, N. Matsumoto, T. Horiuchi, K. Ichino, H. Shimoyama, T. Ohashi, Y. Hashimoto, I. Hengel, J. Beier and R. Klenk, Jpn. J. Appl. Phys., Part 1, 2000, 39, 126.
11. E. Arici, N. S. Sariciftci and D. Meissner, Adv. Funct. Mater., 2003, 13, 165.
12. S. Bandyopadhyaya, S. Chaudhuri and A. K. Pal, Sol. Energy Mater. Sol. Cells, 2000, 60, 323.
13. S. Nakamura and Y. Akio, Sol. Energy Mater. Sol. Cells, 1997, 49, 415.
14. Y. B. He, T. Kramer, I. Osterreicher, A. Polity, B. K. Meyer and M. Hardt, Semicond. Sci. Technol., 2005, 20, 685.
15. M. H. Valdes, M. Berruet, A. Goossens and M. Vazquez, Surf. Coat. Technol., 2010, 204, 3995.
16. M. Gossla, T. Hahn, H. Metzner, J. Conrad and U. Geyer, Thin Solid Films, 1995, 268, 39.
17. S. M. Lee, S. Ikeda, T. Yagi, T. Harada, A. Ennaoui and M. Matsumura, Phys. Chem. Chem. Phys., 2011, 13, 6662.
18. B. Asenjo, A. M. Chaparro, M. T. Gutierrez and J. Herrero, Thin Solid Films, 2006, 511, 117.
19. M. Martinez, A. M. Fernandez, L. G. Arriaga and U. Cano, Mater. Chem. Phys., 2006, 95, 270.
20. X. H. Xu, F. Wang, J. J. Liu, K. C. Park and M. Fujishige, Sol. Energy Mater. Sol. Cells, 2011, 95, 791.
21. J.-H. Yun, Y. H. Ng, S. Huang, G. Conibeer and R. Amal, Chem. Commun., 2011, 47, 11288.
22. M. Krunks, V. Mikli, O. Bijakina and E. Mellikov, Appl. Surf. Sci., 1999, 142, 356.
23. S. Marsillac, M. C. Zouaghi, J. C. Bernède, T. Ben Nasrallah and S. Belgacem, Sol. Energy Mater. Sol. Cells, 2003, 76, 125.
24. T. Wilhelm, B. Berenguier, M. Aggour, K. Skorupska, M. Kanis, M. Winkelnkemper, J. Klaer, C. Kelch and H. J. Lewerenz, Thin Solid Films, 2005, 480–481, 24.
25. K. Töpper, J. Bruns, R. Scheer, M. Weber, A. Weidinger and D. Bräunig, Appl. Phys. Lett., 1997, 71, 482.
26. P. Kumar, L. S. Panchakarla, S. V. Bhat, U. Maitra, K. S. Subrahmanyam and C. N. Rao, Nanotechnology, 2010, 21, 385701.
27. K. W. Cheng, Y. C. Wu and Y. T. Hu, Mater. Res. Bull., 2013, 48, 2457.
28. S. Cattarin, P. Guerriero, N. Dietz and H. J. Lewerenz, Electrochim. Acta, 1995, 40, 1041.

---

Footnote

† Electronic supplementary information (ESI) available: Surface morphology of CuInS₂ thin films on FTO substrate using 30 min deposition time calcined under (a) N₂; (b) 92% N₂ + 8% H₂ for 1 h at 500 °C, with the precursor molar ratio for electrodeposition of CuCl₂ : InCl₃ : Na₂S₂O₃ at 1 : 1 : 10. See DOI: 10.1039/c3ra45691a

---