Morphological and compositional analysis of electrodeposited indium (III) sulfide (In2S3) films

In the last few years, notable progress in understanding the growth mechanism of thin solar films deposited by numerous techniques have been made. Electrodeposition continues to be a complex deposition technique that can lead to low-quality material regions (crack) in the semiconductor material. Such cracks form porous zones on the substrate and diminish the heterojunction interface quality of a photovoltaic (PV) cell. In this paper, electrodeposition of In2S3 films was systematically and quantitatively investigated by varying the electrodeposition parameters including bath composition, current density, deposition time, and deposition temperature. Their effects upon the film growth mechanism, composition, and morphology were studied with the help of scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy (EDS), and fracture and buckling software (digital image analysis). In addition, the effect of different glass-substrates (Mo, ITO, and FTO) and annealing treatments upon the performance of the electrodeposited In2S3 film was analyzed. Furthermore, the Taguchi Method was used to determine the optimal electrodeposition parameters and study their influence upon the morphological and compositional properties of In2S3 films.


I. INTRODUCTION
In2S3 is a promising indirect bandgap (2.3 eV) semiconductor material that is applicable for optoelectronic, photovoltaic, and photoelectric devices [1]. In2S3 has been used as an alternative to CdS as a buffer layer in copper indium gallium selenide/sulfide (CIGS)-based solar cells [2]. Several reports have been published on deposition of In2S3 thin films with diverse morphologies by numerous deposition techniques. In In2S3-based CIGS solar cells, efficiencies of 15.7% have been achieved, which is slightly less than the 16% efficiency reported for CdS-based solar cells deposited by chemical bath deposition (CBD) [3].
Electrodeposition is considered to be a low-cost alternative to vacuum-based technique for applying thin films to a substrate with ful coverage and high growth yield. It provi to a substrate with full coverage and high growth yield. It provides high material utilization efficiency and precise control with proper bath chemistry [4] [5]. Electrodeposited In2S3-based CIGSe solar cells have yielded the highest efficiency of 10.2% [6]. However, electrodeposition still remains a complex technique for synthesizing thin solar films due to several deposition parameters which can affect the properties of the films. Poor adhesion, non-uniformity, flaking, and cracking of films are common problems when electrodepositing thin films onto smooth surfaces, as nucleation growth is influenced by various deposition parameters and contamination of the surface [7] [8]. In view of this, an effort has been made to study the morphological and compositional properties of electrodeposited In2S3 films.
In this paper, we present studies on electrodeposition of In2S3 films using an organic solvent, ethylene glycol. The morphological and compositional properties of In2S3 films were analyzed by varying deposition parameters including bath composition, current density, deposition time, and deposition temperature. In addition, the stoichiometric ratio between sulfur and indium was tailored to the optimal ratio by providing annealing treatment. In2S3 films were deposited onto three different substrates namely indium tin oxide (ITO), fluorine-doped tin oxide (FTO), and molybdenum (Mo)-coated glass substrates, to study their effects upon the morphology and composition with respect to the crack density and stoichiometry of the films. The statistical Taguchi Method was used to obtain optimal deposition parameters based upon the data obtained from scanning electron microscopy (SEM), digital image (from SEM) analysis (using fracture and buckling software), and energy dispersive X-ray spectroscopy (EDS). The primary goal was to avoid non-uniformity, cracks, and improper stoichiometry.
Taguchi (statistical) analysis helped us to analyze the effect of each deposition parameter upon the morphological and compositional properties of the In2S3 films.

II. EXPERIMENTAL DETAILS
The electrodeposition was performed in an electrochemical cell using Ag/AgCl reference electrode, cathode (glass substrate), and anode (graphite). Voltage and current was applied to the electrodes to deposit the material onto the substrates.
An electrolytic solution was prepared containing 150 ml of ethylene glycol (C2H6O2), 0.1 M sulfur (S), 0.1 M sodium chloride (NaCl), and 0.05 M indium chloride (InCl3). S was dissolved in (C2H6O2) by heating the solution to 150 o C. After S melted and fully dissolved, the solution was cooled down to 80 o C, and then InCl3 and NaCl was added to solution. Later, sodium thiosulfate (Na2S2O3·5H2O) was added to solution and used as an additional source of sulfur. A digital potentiostat (WaveNow) from Pine Research Instrumentation was used in a galvanostat mode to perform chronopotentiometry (supplying voltage and current). The substrates were thoroughly cleaned in soapy water and then ultrasonically treated in distilled water for 45 minutes at high temperature. In2S3 films were deposited over 3.75 cm 2 surface area of the substrate.

A. Statistical Analysis
The electrodeposited In2S3 films were statistically analyzed using Minitab 16. The Taguchi analysis was performed to optimize electrodeposition parameters in order to improve the morphology and composition of In2S3 films. The analysis helped us to analyze the effect of each deposition parameter upon the performance of the electrodeposited material. Table I lists the deposition parameters and their levels involved in the study. The values of the levels were changed to study the effects of individual deposition parameters upon the responses, with the least number of experimental runs. The crack density was chosen as a signal factor (response) in this study. Later, the stoichiometric ratio between sulfur and indium was studied as a function of annealing temperature to achieve the proper stoichiometry. An orthogonal array (design of experiments DOE) of L18 (2 1 x 3 4 ) was selected for conducting experiments at two and three levels for bath composition and the rest of the four deposition parameters. Therefore, 18 experiments were performed to complete the study. For each experiment, two samples were generated to minimize the error in the data and achieve balance in the DOE. The effect of each deposition parameter at a given level upon the response was estimated using analysis of means (ANOM and analysis of variance (ANOVA).

Fig.1. Optimal Deposition Parameters
The main effect plot for means revealed the optimal values for each deposition parameter as shown in Fig. 1. The optimal values obtained from the plot for deposition parameters are encircled. The points in the plot are the mean of the response (crack density) at three levels of each parameter, with a reference line (in red) drawn at the grand mean of the response data. The lowest average mean of mean crack density defines the optimal level (encircled) for that deposition parameter. Therefore, the optimal deposition parameters for the electrodeposition of In2S3 films were A2, B1, C1, D3, and E1. It is clear from the response table from Taguchi analysis that current density was the most significant deposition parameter while bath composition was the least significant deposition parameter to the morphology of In2S3 films as shown in Table  II.

B. Digital Image Analysis
In the past, our electrodeposited In2S3 films have shown good stoichiometries but poor morphologies [9]. Films were usually non-uniform and thicker around the edges, and sometimes flaked away into the solution when pulled out from the electrochemical cell. The films have shown cracks, which form porous zones on the substrate and could possibly result in loss of charged carriers and insufficient electrical transport within the PV cells [9] [10].
To study the crack behavior as a function of deposition parameters, digital image analysis of SEM images was performed using fracture and buckling analysis software. This software runs on a program that has been written using MATLAB Version 7.1 (R14). It investigates and detects the dark features of the image compared to the average intensity of the image. The crack density is determined by linear analysis. A line pattern (over the selected area of interest) is superimposed onto the SEM image, where this line pattern is perpendicular to cracks. The intersection between the analysis lines and cracks are determined [11]. Hence, the crack density is determined by dividing the number of intersections with the cumulative line length, as shown in Fig.2. Hence, SEM images (with same magnification scale, 5 kX) were taken for all In2S3 films to calculate crack density (number of cracks/54.9742 µm square). The mean crack density for films deposited at current density of 0.75 mA/cm 2 was 0.0685 cracks/54.972 µm square, whereas, the mean crack density for films deposited at 1.25 mA/cm 2 and 1.75 mA/cm 2 were 0.2846 cracks/54.972 µm square and 0.2055 cracks/54.972 µm square, respectively. However, the crack density for films deposited at current density of 1.25 mA/cm 2 and 1.75 mA/cm 2 were unknown for four In2S3 films. These films were assumed to have high crack densities and, therefore, flaked away from the substrate while rinsing with distilled water after electrodeposition. Similarly, the mean distance between the cracks was found to be 0.9271 µm for films deposited at current density of 0.75 mA/cm 2 , whereas, the mean distance between cracks increased to 3.725 µm and 5.95 µm for films deposited, respectively, at 1.25 mA/cm 2 and 1.75 mA/cm 2 .

C. Morphological Analysis
The cracked morphology of electrochemically deposited In2S3 thin films have been reported previously by several scientists, for instance in [12], [13], and [14]. The reason is still misunderstood and not fully known at most times. The potential causes for the cracking behavior of the electrodeposited semiconductor thin films include thickness, poor incorporation of solutes with solvents [15], deposition voltage [16], substrates, surface contamination [17], and piezoelectric effect, current density [18], and electrolytes [19]. In our study, we found that the current density and substrate, respectively, the most and second-most significant deposition parameter responsible for the cracking behavior in electrodeposited In2S3 films. Films on FTO-and ITO-coated glass substrates were non-uniform and part of the films flaked away when rinsed with distilled water, whereas the films deposited onto Mo-coated glass substrates were uniform and adherent. However, In2S3 films on all three substrates featured cracks and porous zones, as were evident from SEM micrographs (see Fig. 3). Current densities were found to have a significant impact upon the morphology of In2S3 films. The mean crack density and mean distance between cracks were significantly higher for films deposited at higher current densities (1.25 mA/cm 2 and 1.75 mA/cm 2 ) than for films deposited at lower current density (0.75 mA/cm 2 ). The mean distance between the cracks for films deposited at 1.75 mA/cm 2 was close to 6 µm.
Moreover, the morphology of the In2S3 films on ITO glass substrates was also influenced by the substrate itself and the bath composition.
The films deposited using sodium thiosulfate as an additional sulfur source showed poor adhesion. Films flaked-off of the substrate after postdeposition rinsing with distilled water. However, films (Number of cracks/54.9742 µm square) deposited using only sulfur showed good adherence, but high crack density. On the other hand, In2S3 film morphology on FTO glass substrate was far better than ITO glass substrates in terms of uniformity (coverage), appearance, and adherence. Current density was the only significant factor affecting the morphology of the films. Meanwhile, molybdenum glass substrates yielded the best-electrodeposited In2S3 films. The films showed nice uniformity and adherence regardless of the bath composition, deposition time, and deposition temperature. High current density (1.25 mA/cm 2 and 1.75 mA/cm 2 ) produced films with high crack densities, whereas, low current density (0.75 mA/cm 2 ) produced crack-free In2S3 films (see Fig. 3).  [10], we stated that deposition voltage has the highest significant impact upon the stoichiometry of In2S3 films. Therefore, in this current study, we found, as expected, the current density (a strong function of deposition voltage) to have a significant impact upon the composition of the films. In2S3 films deposited at low current density produced crack-free films, however, the composition of the films was slightly sulfur-rich as-deposited (see Table III). Subsequently, In2S3 films were annealed at 250 o C for an hour in air to evaporate the excess sulfur. After the annealing treatment, the stoichiometric ratio between sulfur and indium was nearly perfect (S/In : 3/2).

IV. SUMMARY OF WORK
In2S3 was electrodeposited onto Mo-coated glass substrates using an organic solvent. In2S3 films with planar surface morphologies and near-to-perfect stoichiometric ratios were successfully deposited. Morphological and compositional properties of In2S3 were studied by varying deposition parameters. Current density had the most significant impact upon the morphology and composition of films. In2S3 films deposited at low current density (0.75 mA/cm 2 ) produced uniform, adherent, and crack-free films. In our intensive study, we observed that the In2S3 films tend to start cracking from the edge of the substrate. The nucleation growth is higher near the edges and therefore films initially start to flake-off from the edge of the substrate. As the current density/deposition time increases, those cracks tend to propogate into the middle part of the substrate and, eventually, the film either falls into the solution while pulling-off the susbtrate or flakes-off while rinsing with distilled water. The films deposited on FTO-glass substrates also showed nice morphology at low current density, while films on ITO-glass substrates showed poor uniformity and adherence at all levels. This indicates that the different substrates can also lead to different morphologies with the same deposition parameters. The Taguchi Method helped us to optimize the deposition parameters and understand their influence upon the cracking behavior of the In2S3 films. The mean crack density and the mean distance between the cracks increased approximately by 400-and 600 times for the films grown at higher current density for all three substrates. Similarly, composition of the films was highly influenced by the current density. However, In2S3 films electrodeposited at low current density produced slightly sulfur-rich films, which were then annealed for an hour at 250 o C. After the annealing treatment, excess sulfur was removed