Introduction
Recently, there has been considerable interest for developing new polycrystalline thin film semiconductors using various techniques. Among them, chemical bath deposition (CBD) has found out a special significance being a low temperature method as well as not highly expensive. Many chalcogenide semiconductors such as insoluble sulphides and selenides have been successfully deposited using this technique. If the precipitation is controlled through the use of suitable metal complexing agents and the amount of anions in the bath is controlled through setting up of appropriate chemical equilibria, thin film deposition can take place. The deposition of the thin film takes place through the condensation of the metal and sulphide/selenide ions over the substrate surface.
In the last years, chemical bath deposition has emerged as an excellent method for the deposition of the CdS buffer layer in efficient thin film solar cells with either chalcopirytes (CIS and CIGS) [1] or CdTe [2] absorbers. Although very good efficiencies have been reached using the CBD-CdS, its high toxicity makes difficult its public acceptance. In this way, a new topic in the R&D activities about these solar cells is to fabricate less environmentally polluting devices by replacing the CdS buffer layer by other compounds with similar or even better properties [3]. One promising candidate to achieve this goal is the CBD-prepared indium hydroxy sulphide In(OH)xSz, which has been used as buffer layer in CIS solar cells obtaining successful I-V characteristics [4, 5].
In previous works we have studied the chemistry of the indium hydroxy sulphide deposition process as well as different thin film properties [6,7,8,9]. Some of the results obtained are being presented at the QUANTSOL 99 Workshop.

Experimental
Thin films of In(OH)xSz were prepared from an acidic bath (pH=2.2-2.5) containing 0.025 M indium (III) chloride, 0-0.1 M acetic acid (AcOH) and 0.05-0.3 M thioacetamide (TA) at 70 ^oC. To produce the test samples the reaction solution was stirred in a 20 ml quartz beaker and the deposition was made on 76x26 mm2 commercial-quality glass microscope slides washed thoroughly with distilled water and dried in air. Some films were also air annealed during 10 min at 300^oC and 400^o C.
Optical film thickness was calculated from the transmission and near-normal specular reflection spectra following an iteration method [10]. Both transmission and reflection spectra were measured at room temperature by using unpolarized light at normal incidence in a spectral wavelength range from 300 to 2500 nm. X-ray photoemission spectroscopy (XPS) measurements were performed to investigate the surface composition of the as-deposited and air-annealed thin films. X-ray spectra were recorded using the grazing angle technique (GAXRD) and transmission electron diffractograms were obtained from samples deposited on nickel grids covered by a cellulose-carbon thin film using a transmission electron microscope (TEM).

Results and discussion
The formation of the In₂S₃ films is based on the slow release of In₃₊ and S_2- ions in an acidic medium and their subsequent condensation on the substrates when the ionic product exceeds the solubility product Kps (10^-73) [11]. Sulphide ions are provided by the hydrogen sulphide produced during the thioacetamide hydrolysis in dilute acid solutions [12]. Finally, the hydrogen sulphide is dissociated to give rise to the sulphide ions needed for the In₂S₃ precipitation. The chemical deposition mechanism of indium sulphide films is not clear, however, the tentative reactions can be described by the following schemes:

CH₃-CS-NH₂ + H⁺ + 2H₂O <==> CH₃COOH + H₂S + NH₄⁺

H₂S <==> HS^- + H⁺

HS^- <==> S^2- + H+

2In³⁺ + 3S^2- <==> In₂S₃

Because we are in an aqueous medium and the indium (III) ion is very acid, undesirable hydrolyzed species could appear in the solution. Therefore, acetic acid is added to the reaction mixture either to reduce the pH, which would favour the TA hydrolysis and avoid the formation of hydrolyzed species, or to complex the In³⁺ ions avoiding also these last species. As reported elsewhere [8], the molecular formula of the compound deposited as thin film is not well defined and an indium (III) hydroxy sulphide is believed to form.
Chemical bath deposition of indium sulphide thin films is a rather complicated task because the deposition conditions leading to good quality films and high reproducibility are very restricted. It has been observed that in general average growth rate increases as more AcOH is added to the solution because it avoids the hydroxide-complex formation. In contrast the average growth rate decreases as TA amount is increased for a constant AcOH concentration because a TA excess in the solution favours the homogeneous precipitation instead of the film deposition.
Our experiments have proved that the deposition conditions leading to the best quality films (good adherence at any deposition time, yellow-bright colour, homogeneity and high reproducibility) are: [InCl3] = 0.025 M, [TA] = 0.1 M and [AcOH] = 0.1 M at 70ºC which agrees with the higher film growth rate. Terminal thickness has been obtained by using the above conditions lasting the deposition time till 60 min. Figure 1 shows that the higher thickness that can be obtained is about 1800 Å being reached after about 45 min.

Figure 1. Optical film thickness as a function of deposition time for films prepared using the conditions: [InCl3] = 0.025 M, [TA] = 0.1 M and [AcOH] = 0.1 M at 70ºC.

XPS characterization has proved that films are mainly composed by indium hydroxy sulphide, indium oxide and indium sulphate having adsorbed on the surface some contaminant species from the solution and also CO₂ from the air. The non-annealed samples have more than 80% indium hydroxy sulphide in their composition and the indium oxide percentage seems to increase if no acetic acid is added and low TA concentration is used. All samples annealed have shown a high increase (more than 50%) in the oxygen amount that demonstrates an oxygen incorporation from the air. In contrast, an important sulphur amount (about 50-60% when annealing temperature is 400^oC) has escaped from films during the annealing processing either as sulphur or as SO₂ vapours. As for film composition, an increase in the indium oxide and sulphate percentages with annealing temperature is produced reaching the former 40-50% when samples are heated at 400^oC.
The S/In atomic ratio in the indium hydroxy sulphide compound is about 1 for the non-annealed samples being the highest S/In ratios obtained when deposition conditions favour the sulphide precipitation reducing the hydroxide formation. On the other hand, it can be demonstrated from the Auger InMNN spectra that some hydroxide ions must be bound to the indium (III) compensating the sulphur deficiency. When samples are heated in air atmosphere at 400^oC the S/In ratio in the indium hydroxy sulphide reaches even 0.3 showing that the sulphur under sulphide form is mainly oxidized to volatile species (S and SO₂).
Transmission spectra of these films have been measured over the wavelength range 300-2500 nm and corrected for glass substrate absorption. All samples are highly transparent in the near infrared range and the transmittance of visible light does not exceed 80% for a great number of them. At a constant thioacetamide concentration a decrease in transmission is produced as acetic acid concentration is increased. The same behaviour has been observed as TA concentration is increased remaining constant the acetic acid amount. Refractive index of the films has been calculated from transmission and reflection spectra. The values are not constant varying from 1.59 to 2 and seeming to depend on deposition conditions. The less transparent films have shown the highest refraction index values but in any case, those are rather lower than the value 2.56 reported for In2S3 thin films [13].
The absorption coefficient (a) of the films has been calculated from the transmission and specular-reflection data by means the Chopra equation [14]. All samples (as-prepared and annealed) show a linear behaviour near the absorption edge when a2/3 , (ahn)2/3 or (a/hn)2/3 is plotted versus hn indicating a direct forbidden transition for the material. Energy gap values of many samples have been represented versus the optical thicknesses in figure 2. A clear increase in the band gap is observed for decreasing film thickness being the data very well fitted to a curve

Eg = A + B/T²- C/T

This decrease in the Eg with thickness of In2S3 films can be understood on the quantum size effect observed in the thin films of semiconductors. Similar blue shifts in Eg values for the films with smaller thickness and/or grain size have been reported for many chemically deposited chalcogenide films [15, 16].
In order to reveal the structural aspects of indium hydroxy sulphide thin films, x-ray diffraction has been performed. Owing to its small thickness, it has been necessary the use of the grazing angle X-ray diffraction (GAXRD) technique to obtain a diffraction pattern from the films. Figure 3 displays the GAXRD spectrum of a thin film deposited using 0.025 M InCl3, 0.5 M TA and 0.15 M AcOH during 25 min. at 70ºC.

Figure 2. Energy gap values vs.optical thickness of samples as-deposited (¦) and air-annealed (). Fitting exponential curve is also plotted.

The XRD spectrum of the residual powder removed from the reaction solution is also plotted in this figure. Both spectra are quite similar showing three wide peaks that reveal a low cristallinity. Peak values for the film are 2Q = 29.1, 33.6 and 48.3 degrees and for the powder 2Q = 28.8, 33.5 and 48.3 degrees. These peaks fit quite well with the three main diffraction planes of either the a-In2S3 or the b-In2S3 (see table 1). Therefore indium compound chemically deposited could have one of these structures or even a mixture of them.

Figure 3. XRD spectra for the thin film and residual powder prepared using the conditions: [InCl3] = 0.025 M, [TA] = 0.5 M and [AcOH] = 0.15 M during 25 min. at 70ºC.

To better elucidate the CBD-indium hydroxy sulphide structure, electron diffraction rings have been obtained using a TEM apparatus. Table 1 shows the interplanar spacings calculated from the rings patterns obtained for a sample deposited on a cellulose-carbon-covered nickel grid using the conditions: 0.025 M InCl3, 0.5 M TA and 0.15 M AcOH during 25 min. at 70ºC. As in XRD spectra, three rings are clearly defined which correspond to the three main diffraction planes. The rest of rings are too weak to be measured (marked in asterisk) but they demonstrate that both a an b-In2S3 phases exist in the CBD-indium hydroxy sulphide.

(*) Rings observed but not measured
Table 1. Interplanar spacing calculated from the electron diffraction ring pattern of a sample prepared using the conditions: [InCl3] = 0.025 M, [TA] = 0.5 M and [AcOH] = 0.15 M during 25 min. at 70ºC. Comparison with the a an b In2S3 phases.

References

[1]	K. Zweibel, H. S. Ullal and B. von Roedem, Proceeding 25th IEEE Photovoltaic Specialist Conference, p. 159, Washington 1996.
[2]	C. Ferekides, J. Britt, Y. Ma and L. Killian, Proceeding 23th IEEE Photovoltaic Specialist Conference, p. 389, Louisville 1993.
[3]	H. W. Schock, Solar Energy Materials and Solar Cells, 34 (1994) 19.
[4]	D. Braunger, D. Hariskos, T. Walter and H. W. Schock, Solar Energy Materials and Solar Cells, 40 (1996) 97.
[5]	D. Hariskos, M. Ruckh, U. Rühle, T. Walter, H. W. Schock, J. Hedström and L. Stolt, Solar Energy Materials and Solar Cells, 41-42 (1996) 345.
[6]	R. Bayón, C. Guillén, M. A. Martínez, M. T. Gutierrez and J. Herrero, Proceeding 14th European Photovoltaic Solar Energy Conference and Exhibition, p. 2050, Barcelona 1997.
[7]	R. Bayón, C. Guillén, M. A. Martínez, M. T. Gutierrez and J. Herrero, J. Electrochem. Soc., 145 (8) (1998) 2775.
[8]	R. Bayón, C. Guillén, M. A. Martínez, A. Martínez-Chaparro, M. T. Gutiérrez and J. Herrero, Proceeding 1st World Conference and Exhibition on Photovoltaic Solar Energy conversion, Vienna 1998. In press.
[9]	R. Bayón, C. Maffiotte and J. Herrero, Thin Solid Films. Sent.
[10]	J. L. Hernandez-Rojas et al., Appl. Opt., 31 (1992) 1606.
[11]	A. I. Busev, The Analytical Chemistry of Indium, Pergamon Oxford (1962).
[12]	E. H. Swiftand and E. A. Butler, Anal. Chem., 28 (2) (1956) 146.
[13]	J. George, K. S. Joseph, B. Pradeep and T. Y. Palson, Phys. Stat. Sol. (a), 106 (1988) 123K.
[14]	K. L. Chopra and S. K. Bahl, Phys. Rev. B, 1 (1970) 2545.
[15]	S. Gober and G. Hodes, J. Phys. Chem., 98 (1995) 5338.
[16]	R. Rossetti, R. Hull, J. N. Gibson and L. E. Brus, J. Phys. Chem., 82 (1989) 552.

[BACK] [MAIN] [FURTHER]

---

emails: rocio.bayon@ciemat.es

---

Last updated March 25, 1999