1. Introduction

The substitution of cadmium sulfide (CdS), commonly used as a buffer layer in copper indium gallium (di)selenide (CIGS) thin film solar cells, by another nontoxic material with a wider band gap in order to improve light transmission in the blue wavelength region has been a challenge for researchers since the last decade. To meet these requirements, indium sulfide is a good candidate owing to its stability^[1], band gap energy between 2 eV^[2] and 2.8 eV^[3], as well as its transparency and photoconductivity behavior^[4]. Conversion efficiencies up to 12.4% have already been demonstrated for CIGS solar cells with In₂S₃ buffer layer co-evaporated from In and S powder^[5].

Indium sulfide has been elaborated using various techniques like successive ionic layer adsorption and reaction (SILAR)^[6], thermal evaporation^[7], atomic layer deposition (ADL)^[8], chemical bath deposition (CBD)^[9], chemical spray deposition (CSP)^[10], spin coating^[11], electrodeposition^[12], etc. It is worth noting that the quality of films strongly depends on the growth technique. According to the literature the band gaps range between 2.0 eV to 3.7 eV depending on the deposition techniques used and assuming a direct band gap transition^[13].

Chemical spray pyrolysis is a technique being considered in research to prepare thin and thick films, ceramic coatings, and powders^[14]. It is chosen because it is economical and fast and a vacuum is not required to prepare thin film^[15]. In addition, it is suitable for industrial scale production.

In this study we have elaborated In₂S₃ thin films prepared by CSP varying the [S]/[In] ratio from 2.5 to 4.5 and fixing the starting indium concentration. Some amount of alcohol (5%) was added to aqueous solvent to reduce the surface tension. In₂S₃ thin films were sprayed on substrates heated at 250 ℃. The crystalline structure, morphology, chemical composition and optical properties of CSP-deposited In₂S₃ thin films are reported.

2. Experimental details

2.1 Film preparation

Indium sulfide (In₂S₃) thin films were elaborated onto glass substrate from aqueous solution containing indium (Ⅲ) chloride (InCl₃) (99.999%), thiourea (CS(NH₂)₂ ($\geqslant$ 99%) and alcohol (5% in volume) to reduce the surface tension. The concentration of indium chloride was fixed at 0.026 M and the [S]/[In] ratios were varied from 2.5 to 4.5 by steps of 1. The substrate temperature was kept at 250 ℃ for all ratios. The volume sprayed was 10 mL, the spray rate 1.5 mL/min, the air compressed pressure 0.7 bar and the distance between the glass substrate and nozzle was kept to 30 cm.

Glass substrates were previously washed with HNO₃ and subsequently rinsed with water, ethanol and acetone.

2.2 Film characterization

Structural properties were characterized by means of X-ray diffraction (XRD) measurement using a Rigaku Ultima Ⅳ diffractometer in the θ-2θ configuration and using Cu K$\alpha$ radiation (1.5418 Å). Structural properties were also characterized by Raman scattering measurements performed with a LabRAM HR UV spectrometer coupled to a Peltier-cooled CCD camera and using a 632.81 nm laser excitation line with a spectral resolution of 3 cm^-1. Microanalysis by energy dispersive X-ray spectroscopy (EDS) was obtained on a Jeol-JSM6300 operating at 20 keV. The surface morphology was studied by atomic force microscopy (AFM) using a Bruker Multimode 8 AFM Nanoscope V controller. Optical properties were monitored by transmittance using a Deuterium-Halogen lamp (DT-MINI-2-GS Micropark) in association with a 500 mm Yvon-Jobin HR460 spectrophotometer using a back-thinned Si-CCD detector (Hamamatsu) optimized for the UV-VIS range.

3. Results and discussion

3.1 X-ray diffraction analysis

The X-ray diffraction of as-deposited thin films elaborated at a substrate temperature of 250 ℃, and [S]/[In] = 2.5, 3.5 and 4.5, respectively, are presented in Fig. 1. The deposited films are $\beta$-indium sulfide polycrystalline with tetragonal structure. The main diffraction peak is located at 33.50$^\circ$ corresponding to (0, 0, 12) crystal plane perpendicular to the substrate plane. Other diffraction peaks located at 14.27$^\circ$, 27.63$^\circ$, 43.95$^\circ$ and 48.09$^\circ$ and attributed to (1, 0, 3), (1, 0, 9), (1, 0, 15) and (2, 2, 12) planes, respectively, are also observed. The sharpness of the peak (0, 0, 12) suggests a good crystallinity of the films. The lattice parameters "$a=b$" and c determined by XRD data were found to be $a=$ 7.5919 Å, $c=$ 32.0317 Å; $a=$ 7.5976 Å, $c=$ 32.07 Å and $a=$ 7.6084 Å, $c=$ 32.0316 Å for samples elaborated at [S]/[In] equal to 2.5, 3.5 and 4.5, respectively. These values are in good agreement with the values given in reference JCPDS#25-0390. One noted also a small peak, located at 31.656$^\circ$ and corresponding to some residual In₂O₃, in film elaborated from precursors with the lowest S content [S]/[In] = 2.5. The residual In₂O₃ phase is no longer observed when the [S]/[In] ratio exceeds 2.5. The intensity of (0, 0, 12) peak increases when the [S]/[In] ratio in the solution varied from 2.5 to 3.5 and decreases afterwards.

The variation of crystallite grain size with [S]/[In] ratio in the starting solution was investigated using the Debye-Scherrer formula from (0, 0, 12) diffraction line:

$\begin{equation} D = K \lambda /\beta \cos \theta. \end{equation}$

(1)

$\beta$ is the full width at half maximum (FWHM), $\lambda$ wavelength of X-ray whose value is 1.5418 Å (Cu K$\alpha )$, K the Scherrer constant which generally depends on the crystallite shape, close to 1 ($K =$ 0.9 was used) and θ is the Bragg angle at the center of the peak. The crystallite size D obtained from this equation corresponds to the mean minimum dimension of a coherent diffraction domain. Figure 2 shows that the grain size decreased with the increase of [S]/[In] ratio. The crystallite grain size, as measured using Debye-Scherrer method, decreases when the sulfur to indium ratio in the sprayed solution increases.

3.2 Raman spectroscopy analysis

The quality of the deposited films has been characterized by micro-Raman. Raman spectroscopy is a nondestructive and relatively fast experimental technique to determine the phase and quality of the deposited film. Raman spectroscopy analysis depends not only on the nature of the compound but also on its crystallographic form. The Raman modes are observed in the energy region between l00-550 cm^-1. Figure 3 shows the following Raman modes corresponding to $\beta$-In₂S₃ thin films: $E_{\rm g}$ = 268 cm^-1, $F_{\rm 2g}$ = 329 cm^-1 and $A_{\rm 1g}$ = 369 cm^-1, respectively. These Raman modes confirm the composition and structure of $\beta$-In₂S₃ phase and are similar to those reported in Ref. [16]. We also noted that Raman modes in sample S2 are sharper than in the other samples so confirming the XRD results.

3.3 Surface morphology

Figure 4 shows scanning electron micrographs of thin films sample of indium sulfide at 10000 magnifications with a measuring scale of 6 $\mu$m in the pictures. The surface of the samples revealed continuous films with no cracks and voids and it is clearly seen also from the SEM photographs that the films are dense, smooth, and homogeneous without pinholes and perfectly covering the entire substrates for all [S]/[In] ratios. It is worth noting that the morphology of samples was improved when the amount of sulfur in solution increased. The small grain size of our sample observed in SEM images is caused by the amount of alcohol added in the initial solution to reduce the tension of surface of water^[15].

3.4 Surface topography of sprayed In₂S₃ thin films

The surface topography of In₂S₃ thin films were also investigated by AFM. Figure 5 shows the AFM images of samples S1, S2 and S3 with different sulfur concentration over a 3 × 3 $\mu$m² scanning range. The color z axis scales is 70 nm. Table 1 gives the average grain size and RMS roughness of the surface of the samples. The images of Fig. 5 reveal samples with small grains and almost well covered with little surface "microvoid" due to back reflection event of reactant atoms hitting the surface substrate. Films are dense and practically constituted of equal-sized grains with a relatively uniform distribution. The grain size obtained from AFM statistics is about 120 $\pm$ 10 nm, which is about three times the crystallite size calculated from the Scherrer equation. When the amount of sulfur increases, the films are more homogeneous and exhibit densely-packed grains. The sample S2 has lower roughness and grain size compared to others.

3.5 EDS measurements

Table 2 shows samples with good stoichiometry for almost all samples and we noted that the measured S/In ratio in thin films is close to the theoretical stoichiometry. Even when the [S]/[In] ratio in the solution was changed the [S]/[In] ratio in thin films remained very near to stoichiometry, unlike some researchers who assert that when indium chloride was used to prepare indium sulfide, no stoichiometry was obtained^[17]. The composition of the sprayed thin film is little influenced by the sulfur-indium ratio in the starting solution. When the [S]/[In] ratio in solution increases, the [S]/[In] ratio in thin films increases and goes from 1.43 to 1.46. Then, the amount of sulfur in the initial solution improves the stoichiometry of samples. The reason is that sulfur produces an exhaustion of oxygen in the proximity of the film that avoids the formation of indium oxide phases.

3.6 Optical transmittance analysis

Figure 6 shows the optical transmittance for In₂S₃ thin films sprayed from solutions with different [S]/[In] ratios. Thin films obtained for all [S]/[In] ratios revealed good transparency in the spectral range 600-1000 nm.

The rise of the optical band gap has been attributed to the presence of oxygen^[18], the presence of excess sulfur^[19] or grain size of sample^[20]. Optical band gap of films, $E_{\rm g}$, were deduced from ($\alpha h \nu )$$^{2}$ versus $h \nu$ by extrapolating the straight line portion of the graph in the absorption regime, where $\alpha$ is the absorption coefficient and $h \nu$ the photon energy. The results are shown in Fig. 7.

The band gap of samples slightly increased from 2.57 to 2.63 eV when the [S]/[In] ratio increased from 2.5 to 4.5, which means that the optical band gap is insensitive to the starting [S]/[In] ratio. This slight variation of optical band gap can be related to the decrease of the crystallite size for increasing [S]/[In] values.

4. Conclusion

Synthesis of In₂S₃ thin film was carried out by spray pyrolysis with different [S]/[In] ratio solutions using indium chloride, thiourea and 5% of alcohol to reduce the surface tension in order to improve the adherence of films on the glass substrate. Based on the SEM analysis, we noted that films are dense, well-covered substrate with no cracks and voids for all [S]/[In] ratios. Nevertheless the AFM analysis reveals some microivoids due to the reflection of the reactant hitting the substrate. The atomic ratio S/In in films given by EDS is stoichiometric for almost all samples and increased when the sulfur to indium ratio in starting solution increased. Raman spectroscopy confirms the presence of $\beta$-In₂S₃ phase and the energy band gap slightly changes with the [S]/[In] ratio, from 2.57 eV to 2.63~eV, when the sulfur to indium ratio increased from 2.5 to 4.5, respectively. Finally we can state that in our study the sample S2 possesses the best physical and chemical characteristics for photovoltaic applications even though more experiments are underway to apply sprayed In₂S₃ thin films in PV devices.

Acknowledgments: This work was supported by the Generalitat valenciana through grant PROMETEUS 2009/2011 and the European Commission through NanoCIS project (FP7-PEOPLE-2010-IRSES ref. 269279).