Introduction

Tin sulfide SnS material has been a subject of various recent investigations due to its physical properties and photovoltaic applications [1–3]. This compound belongs to groups IV–VI of compounds formed with Sn as the cation and S as the anion. The constituent elements are nontoxic and abundant in nature leading to the development of devices that are environmentally safe and have public acceptability. Many techniques have been used to produce tin sulfide such as: spray [4, 5], electrodeposition [6], silar method [7], and chemical bath deposition (CBD) [8, 9]. In previous study [8], we have deposited high quality SnS(ZB) films on glass substrates by CBD. This technique which is simple and inexpensive method is useful for large area applications. Optical characterization of SnS and In2S3 thin films can be used as an effective diagnostic tool to assess film quality. The envelope method [10], based on the optical transmission spectrum taken at normal incidence, has been successfully applied to determine the film thickness as well as the optical constants of tin sulfide thin films. To date, envelope method has not yet used for optical characterization of tin sulfide CBD thin films.

In2S3 is a III–VI binary compound. It appears to be one of the most promising candidates for photovoltaics owing to its stability, interesting structural characteristics [11, 12], electronical [13], optical, photoelectrical [14, 15], and acoustical properties [16] as well as its photocatalytic [17] and environmental interest [18].

During the last two decades, indium sulfide and especially β-In2S3 material has been the subject of much research because of its photoconducting behavior which makes it as promising optoelectronic material [19, 20]. Such material can be classified as mid-band gap semiconductor (2.2 eV) as compared to materials such GaAs (1.4 eV, narrow band gap), or to ZnO (3.2 eV, wide band gap) [21, 22]. Nevertheless an efficient photovoltaic structure such as SnS/β-In2S3 requires good structural and optical properties of the β-In2S3 thin layers.

In2S3 exists generally in four phases [23, 24]. Recently, we have shown that β-In2S3 layers can be obtained by the CBD technique [25, 26]. It is interesting to note that Ag/SnS/β-In2S3/SnO2:F heterojunction [without aluminum (Al) inclusion in the In2S3 film] was elaborated in our laboratory but no photovoltaic effect has been observed.

In this study, to increase the optical band gap of β-In2S3 thin films, Al was introduced in the starting solution. The addition of Al as doping agent in the thin layers of β-In2S3 leads to move the intrinsic absorption of this material toward ultra-violet zone, and therefore to, widen the domain of transmitted wavelength to the absorber material SnS in the SnS/β-In2S3:Al/SnO2:F photovoltaic structure. β-In2S3:Al thin films were deposited on SnO2:F/glass with different concentration ratio \( y = \left({{\left[ {{\text{Al}}^{3 + } } \right]}}/{{\left[ {{\text{In}}^{3 + } } \right]}}\right) = 0, 20, 30 \,{\text{and}}\,40\% \). We have already noticed that in β-In2−x Al x S3 prepared by the spray pyrolysis [27], the presence of Al atoms in the material contributes to the growth and to the structure of the deposited films. Moreover it was observed from experiment that the best crystallinity of β-In2S3 was obtained on SnO2/Pyrex substrate; the contamination was also reduced.

In this study, we have also studied the effect of successive deposition runs on the refractive index “n”, the extinction coefficient “k”, and the real and imaginary parts of the dielectric constant of tin sulfide homogeneous thin layer.

In this way, tin sulfide and indium sulfide thin films have been elaborated on glass substrates and SnO2:F/glass, respectively, with the aim to establish the optimized experimental conditions leading to grown absorber SnS(ZB) and window In2S3:Al thin films with good structural and optical properties such as high SnS absorption coefficient “α” and wide In2S3:Al optical band gap.

To date, the use of CBD multilayers of SnS(ZB) and In2S3:Al as absorber and window material, respectively, to elaborate solar cell Ag/SnS/β-In2S3:Al/SnO2:F has not yet been tested. Haleem et al. [28], have developed and characterized InS x O y (OH) Z /SnS heterojunction; they showed that the photo-response and the conversion efficiency of such cell are still so low.

Experimental details

Deposition of SnS(ZB) absorber layer

Multi-layers SnS thin films have been deposited via CBD process by means of basic solution containing SnCl2·2H2O, diluted triethanolamine (TEA), NH4OH(15 M) as well as thioacetamide (TA) (0.1 M) as precursors. The temperature of the reaction mixture was kept constant using a water bath to 25 °C. The substrates were kept vertically in the hermetically closed deposition bath mounted on a heating magnetic agitator, which controls temperature of the solution. Then, the deposited thin films were taken out from the bath, washed with water, and finally put in a dried box for approximately 15 min. The Thickness “e” of one layer of SnS is in the order of 100 nm. This value of SnS thin monolayer is rather low [8]. In order to increase the thickness of SnS to be used as an absorber in solar cell we have made multilayer. The thicknesses of the films were evaluated using envelope method as 305, 340, and 455 nm for four, five and six deposition runs, respectively.

Deposition of In2S3:Al window layer

Multilayer indium sulfide thin film has been deposited by a CBD, on SnO2:F/glass substrates with different concentration ratios \( y = \left({{\left[ {{\text{Al}}^{3 + } } \right]}}/{{\left[ {{\text{In}}^{3 + } } \right]}}\right) = 0, 20, 30\;{\text{and}}\;40\%.\) Such SnO2:F films exhibits 75–80% of transparency, and surface resistance of 10–15 Ω/sq [29]. SnO2:F thin layers act as ohmic contact in SnS/In2S3:Al solar cells.

The solution of indium sulfide contains indium chloride and TA as sulfur precursor with concentration (0.1 M). In order to get Al inclusion in β-In2S3 thin films, the necessary mass of AlCl3 are directly added into the deposition solution. Bath temperature is 80 °C; the pH of the solution is controlled to 2 by adding acetic acid into the reaction mixture. The substrates are introduced vertically into the hermetical closed deposition cell. This cell is mounted on a heating magnetic agitator, which controls both temperature and homogeneity of the solution. The bath color leaves clear to yellow. β-In2S3 films nucleated onto submerged surfaces, including beaker walls in about 47 min for one deposition run. The obtained yellow films were taken out from the bath, washed with water, and finally put in a dried box for approximately 15 min [25, 26]. For two deposition runs, we emerged the first obtained thin films in the freshly prepared solution for 47 more min. In this study, we repeat the same bath up to four times with the aim to increase the film thickness.

It is noted that for a number of deposit greater than four times the β-In2S3 thin film collapses, which is caused by the gravitational forces due to the film mass. Gravitational forces affect indeed the forces of adhesion and cohesion between layer and substrate. The similar behavior was observed by Pramanik et al. [30], for orthorhombic SnS thin films grown by CBD on glass substrates during t d equal to 24 h.

Optical analysis of zinc blend tin sulfide

The optical transmission measurements of zinc blend tin sulfide were obtained by means of a Varian spectrophotometer at room temperature under normal incidence over a large spectral range (0.2–2.5 μm) [8]. Using the experimental spectra of the optical transmission of our earlier study [8], this study deals with a theoretical simulation of these spectra using the envelope method [10]. Points obtained theoretically allow us to calculate the thickness and optical constants of SnS thin films.

Figure 1 shows both transmission curves experimentally and theoretically for the crystalline SnS(ZB) multilayer formed by six deposition runs. The interference phenomena between the wave fronts generated at two interfaces (air and substrate) defines the sinusoidal behavior of the curves’ transmittance in terms of the wavelength. SnS(ZB) thin film exhibits interference fringe pattern in transmission spectrum. This revealed the smooth reflecting surface of the film and there is not much loss of scattering at the surface. Good surface state and homogeneity of the film were confirmed from the appearance of interference fringes in the transmission spectra. The refractive index “n 1” at different wavelengths was calculated in the transparency region using the envelope curve for T max(λ) and T min(λ) (Fig. 1). The expression for refractive index is given by [10]:

$$ n_{1} = \left[ {N + \left( {N^{2} - n_{0}^{2} n_{2}^{2} } \right)^{{{1 \mathord{\left/ {\vphantom {1 2}} \right. \kern-\nulldelimiterspace} 2}}}} \right]^{{{1 \mathord{\left/ {\vphantom {1 2}} \right. \kern-\nulldelimiterspace} 2}}} $$
(1)

with \( N = 2n_{0} n_{2} \left[ {\frac{{\left( {T_{\max } - T_{\min } } \right)}}{{T_{\max } T_{\min } }} } \right] + \left( {\frac{{n_{2}^{2} + n_{0}^{2} }}{2}} \right), \) n 0 is air refractive index, n 1 is the refractive index of tin sulfide thin film, and n 2 is the refractive index of glass substrate (in our case n 2 = 1.52).

Fig. 1
figure 1

Experimentally (T) and theoretically (T th) transmission of zinc blend tin sulfide elaborated by CBD on glass substrates after six depositions runs

Full size image

The thickness d of tin sulfide layer can be calculated from two maxima or minima using this equation [10]:

$$ d = \frac{{M\lambda_{1}\lambda_{2} }}{{2\left( {\lambda_{2} n_{1} (\lambda_{1}) - \lambda_{1} n_{1} (\lambda_2 })\right) }} $$
(2)

where M is the number of oscillations between two extrema (M = 1 between two consecutive maxima or minima); λ 1, n 1(λ 1) and λ 2, n 1(λ 2) are the corresponding wavelengths and refractive index.

Knowing n 1 and d values, the extinction coefficient “k 1” at different wavelengths in the transparency region can be calculated using the following formula [10]:

$$ k_{1} = \frac{\lambda }{4\pi }\alpha $$
(3)

where α is the absorption coefficient given by:

$$ \alpha = - \frac{1}{d}{ \ln }(x) $$
(4)

and

$$ x = \frac{{C_{1} \left[ {1 - \left( {\frac{{T_{ \max } }}{{T_{ \min } }}} \right)^{{{1 \mathord{\left/ {\vphantom {1 2}} \right. \kern-\nulldelimiterspace} 2}}}} \right]}}{{C_{2} \left[ {1 + \left( {\frac{{T_{ \max } }}{{T_{ \min } }}} \right)^{{{1 \mathord{\left/ {\vphantom {1 2}} \right. \kern-\nulldelimiterspace} 2}}}} \right]}}, $$

where: C 1 = (n 1 + n 0)(n 2 + n 1) and C 2 = (n 1 − n 0)(n 2 − n 1).

Using the obtained values of refractive index (n 1) and extinction (k 1) coefficient, the real and imaginary parts of dielectric constants are calculated using following expression [29]:

$$ \varepsilon_{\text{r}} = n_{1}^{2} - k_{1}^{2} $$
(5)
$$ \varepsilon_{\text{i}} = 2n_{1} k_{1} . $$
(6)

The thickness of SnS(ZB) multilayers are calculated from Eq. 2. All results are summarized in Table 1. We note that both theoretically thickness values obtained in this study and experimental calculated ones obtained in previously study [8] were already correlated.

Table 1 Theoretical values of thicknesses of SnS(ZB) multilayer thin film deduced using envelope method
Full size table

The refractive index n 1(λ) and the extinction coefficient k 1(λ) values were calculated in the transparency domain (750–2500 nm) using Eqs. 1, 3. The variations of n 1(λ)and k 1(λ) with the wavelength are shown in Fig. 2a, b. It has been found that the refractive index n 1(λ) is varying in the range of 2.3–3.2, whereas the extinction coefficient k 1(λ) is of the order of 0.05. The low value of extinction coefficient as observed of these films is a qualitative indication of surface smoothness and homogeneity of the films. Subramanian et al. [31] and Koteeswara Reddy et al. [32] are also observed a similar variation of optical parameters in electrodeposition and sprayed deposited SnS thin films. It is clearly show in Fig. 2a that the refractive index obtained in case of five deposition runs of SnS(ZB) is higher than others obtained for four and six deposition runs. This behavior may be due to the existence of an orthorhombic phase revealed by X ray diffraction reported previously [8] which can perturb the refractive index values.

Fig. 2
figure 2

The variations of refractive index “n” (a) and extinction coefficient “k” (b) of zinc blend tin sulfide elaborated by CBD on glass substrates after four (a), five (b), and six (c) depositions runs with wavelength in the 750–2500 nm domain

Full size image

On the other hand, the real (εr) and imaginary (εi) parts of the dielectric constant ε are related to the n 1 and k 1 values. εr and εi values were calculated using the Eq. 5, 6. Figure 3 shows εr and εi values dependence of photon energy. It has been found that the real part of the dielectric constant value of zinc blend tin sulfide thin film remain practically constant with increasing of the photon energy. Also the imaginary part of the dielectric constant value of SnS is found to increase with increasing of the photon energy.

Fig. 3
figure 3

The real (εr) and imaginary (εi) parts of the dielectric constant ε of SnS(ZB) thin layers grown for four (a), five (b), and six (c) depositions runs versus photon energy (hν)

Full size image

Also, the absorption coefficient α of the SnS(ZB) thin films was determined from transmittance measurements. First, α is calculated in the transparency region using the envelopes method, however, this method is not valid in the strong absorption region. In 450–750 nm domain, α is indeed calculated using the following expression [29]:

$$ T = \frac{{16n_{s} \left( {n^{2} + k^{2} } \right){ \exp }( - \alpha d)}}{{\left[ {\left( {1 + n} \right)^{2} + k^{2} } \right] \left[ {\left( {n_{s} + n} \right)^{2} + k^{2} } \right]}} \approx (1 - R)^{2} { \exp }( - \alpha d) $$
(7)
$$ \alpha \approx - \frac{1}{d}{ \ln }\left[ {\frac{T}{{\left( {1 - R} \right)^{2} }}} \right]. $$
(8)

The variation of α with the wavelength for the SnS(ZB) multilayers is shown in Fig. 4. It can be observed that zinc blend tin sulfide thin film exhibits a high absorption coefficient >1.5 × 106 cm−1 in the front absorption region 450–750 nm. SnS(ZB) multi films system are therefore suitable for absorber materials in solar cell [33]. In fact, the optical band gap is found to be equal to 1.76 eV [8].

Fig. 4
figure 4

The variation of absorption coefficient α with a wavelength λ for the SnS(ZB) thin films elaborated by CBD on glass substrates after six depositions runs

Full size image

Structural and chemical analysis of In2S3:Al (40%)

X ray spectrum of β-In2S3 (40% Al) thin layer deposited by CBD on SnO2:F/glass is shown in Fig. 5. We point out the presence of three principal orientations toward (610), (551), and (311) planes, according to JCPDS Card no. 32-0456 of the cubic structure of β-In2S3 phase with oriented preferentially of crystallites toward (610) plan perpendicular to the substrate. The film is well crystallized but it contains undesirable In6S7 and InS secondary phases corresponding to (\( \bar{4}10 \)) and (133) planes according to JCPDS Card nos. 19-0587 and no 65-1472. We also remark the existence of the tetragonal phase of Al2S3 assigned to (314) and (413), according to JCPDS Card no. 24-0014, as well as the hexagonal phase AlInS3 corresponding to (2210) (JCPDS Card no. 71-0622). Planes assigned to (220), (110), and (301) indexes correspond to SnO2:F thin layer deposited under β-In2S3:Al thin layers. In the same line, In2S3:Al (40%) thin films was also chemically analyzed by energy dispersive spectroscopy (EDS), we selected several points on the surface and the atomic percentage of Al, S, and In, values are gathered in Table 2. It is obviously shown that Al element exists which is consistent with X-ray diffraction analysis, Fig. 5.

Fig. 5
figure 5

X-ray spectrum of the β-In2S3:Al (40%) thin layer deposited by CBD on SnO2:F/glass substrate. Planes assigned to (200), (110), and (301) indexes correspond to the SnO2:F thin layer above it β-In2S3:Al thin layers was grown

Full size image
Table 2 EDS analysis of β-In2S3:Al multilayer thin film (four deposition runs) deposited on SnO2:F/glass substrate for \( y = \left({{\left[ {{\text{Al}}^{3 + } } \right]}}/{{\left[ {{\text{In}}^{3 + } } \right]}}\right) = 40{\text{\% }} \)
Full size table

The EDS results (Table 2) shows the existence of Al in the thin layers of β-In2S3 with a ratio equal to 0.1% value comparable to that found by Bhira et al. [34] who synthesized \( {\text{In}}_{2 - 2x} {\text{Al}}_{2x} {\text{S}}_{3 - 3y} \) by the spray pyrolysis technique.

Optical analysis of cubic indium sulfide

The optical reflection–transmission of cubic indium sulfide grown for different ratios \( y = \left({{\left[ {{\text{Al}}^{3 + } } \right]}}/{{\left[ {{\text{In}}^{3 + } } \right]}}\right) = 0, 20, 30,\;{\text{and}}\;40\% \) are obtained with an NkdGen spectrophotometer at room temperature over 0.2–1.1 μm spectral range. It is clearly shown in Fig. 6 that Al content influences the optical properties of β-In2S3:Al thin films. In fact, the broadening and/or shift of the short wavelength absorption edge for larger Al inclusion can be ascribed both to the good optical quality and to a more ordered structure of β-In2S3:Al thin films. Furthermore, shift of the intrinsic absorption edge toward ultraviolet region leads to a maximal transmission coefficient in visible domain which allows us to use β-In2S3:Al as good optical window layer in photovoltaic devices. Moreover, in case of 40% Al contents we notice more clearly in Fig. 5 the presence of rather weak interference fringes on both T(λ) and R(λ) spectra which confirms that for this percentage content the indium sulfide thin films have uniform thickness as well as flat and smooth surfaces.

Fig. 6
figure 6

Transmission–reflection of cubic indium sulfide multilayers elaborated by CBD with different Al content, \( y = \left({{\left[ {{\text{Al}}^{3 + } } \right]}}/{{\left[ {{\text{In}}^{3 + } } \right]}}\right) = 0, \,20, \,30,\,{\text{and }}40\% \)

Full size image

Layer thickness of the In2S3:Al multilayer is estimated by a Dektak3 instrument that is a microprocessor-based contact stylus surface profiler used for making accurate measurements on vertical features ranging in height from 100 to 655000 angstroms. This instrument acquires data by slowly moving the sample beneath the diamond-tipped stylus. Vertical movements of the stylus are sensed by a transducer, digitized, and stored in the memory for plotting and data manipulation. Figure 7 shows the average thickness of In2S3:Al as a function of Al content. To check the thickness value, several points on the surface were selected and the thickness is measured for each of them. These measures are shifted by \( \pm 20 {\text{nm}} \); we concluded that the surface homogeneity was good at the scale of the Dektak3 measurements accuracy. It is clearly shown in Fig. 7 that the film thicknesses of β-In2S3:Al multilayer decrease from 1250 to 945 nm when the Al content increase from zero to 40%.

Fig. 7
figure 7

The average thickness of In2S3:Al multilayers as a function of Al content, \( y = \left({{\left[ {{\text{Al}}^{3 + } } \right]}}/{{\left[ {{\text{In}}^{3 + } } \right]}}\right) = 0,\,20,\,30,\;\text{and}\;40{\text{\% }} \)

Full size image

In the same way, the plot of αhν versus photon energy (hν) allows us to determine the absorption edge which can be expressed versus photon energy by the following equation:

$$ \alpha h\nu = B(h\nu - E_{\text{g}} )^{n} , $$
(9)

where B is a parameter that depends on the transition probability and n is an index depending on the nature of the electronic transitions. In2S3:Al multilayer system has a direct energy band gap \( \left( {n = \frac{1}{2}} \right) \).

The extrapolation of linearly part of \( (\alpha h\nu )^{2} \) versus \( h\nu \) intercepts on the \( h\nu \) axis and give a value of the energy band gap E g, such plots are shown in Fig. 8. The straight lines imply that In2S3 (Al) samples have a direct energy band gap, which is a desirable result. In fact, at the absorption of a photon in the case of \( h > E_{\text{g}} \), there will be creation of an electron–hole. The electron goes directly to the conduction band, this transition occurs between the maximum of valence band (VB) and the minimum of conduction band (CB). In the case of a direct band gap two extrema of VB and CB are located at the same value of wave vector “k” so at the same of the photon energy. Whereas in the case of an indirect gap these two extrema are located at different energies so the transfer of electron from the valence band to the conduction band is necessarily with the intervention of a phonon.

Fig. 8
figure 8

Plots of \( (\alpha h\nu )^{2} \) versus \( h\nu \) of indium sulfide elaborated by CBD with different Al content, \( y = \left({{\left[ {{\text{Al}}^{3 + } } \right]}}/{{\left[ {{\text{In}}^{3 + } } \right]}}\right) = 0,\,20,\,30,\,\text{and}\;40{\text{\% }} \)

Full size image

The variations of the optical energy band gap E g values with Al inclusion are summarized in Table 3. This result is in agreement with the reported values [12, 35]. It is clear that the band gap E g of indium sulfide multilayer system depends on the Al concentration. In fact, Al contents cause the increase of E g, Table 3. This behavior may be due to the existence of Al2S3 and AlInS3 undesirable phases as reveled by X-ray analysis, Fig. 5. Moreover, this result is in good agreement with those reported by Bhira et al. [34]. In fact, Bhira et al. showed that the optical band gap of \( {\text{In}}_{2 - 2x} {\text{Al}}_{2x} {\text{S}}_{3 - 3y} \) thin film increases with Al concentration from 0 to 100%.

Table 3 Variation of energy band gap \( E_{\text{g}} \) of β-In2S3:Al multilayer thin film with different Al content, \( y = \left({{\left[ {{\text{Al}}^{3 + } } \right]}}/{{\left[ {{\text{In}}^{3 + } } \right]}}\right) = 0,\;20,\;30,\;\text{and}\;40{\text{\% }} \)
Full size table

The refractive index “n” and extinction coefficient “k” of indium sulfide thin films grown for different contents of Al are determined from T(λ) and R(λ) spectra using the following expressions [29]:

$$ n = \frac{{1 + [1 - \left( {\frac{1 - R}{1 + R}} \right)^{2} \left( {1 + k^{2} } \right)]^{1/2} }}{{\left( {\frac{1 - R}{1 + R}} \right)}} $$
(10)
$$ k = \frac{\lambda }{4\pi e} { \ln }\left[ {\frac{{\left( {1 - R} \right)^{2} }}{T}} \right]. $$
(11)

Figure 9 shows the plot of “n” and “k” as a function of wavelength (λ). We note that for \( y = \left({{\left[ {{\text{Al}}^{3 + } } \right]}}/{{\left[ {{\text{In}}^{3 + } } \right]}}\right) = 40\% \), “n” and “k” varied in the range of 1.61–2.5 and 0.03–0.1, respectively . Furthermore, β-In2S3:Al thin films grown for \( y = \left({{\left[ {{\text{Al}}^{3 + } } \right]}}/{{\left[ {{\text{In}}^{3 + } } \right]}}\right) = 40\% \) show high refractive index and lower extinction coefficient values in the visible domain, which indicates good optical quality. This last finding is in agreement with those reported by Trigo et al. [35].

Fig. 9
figure 9

Plot of (n) and (k) of β-In2S3:Al grown for \( y = {{\left[ {{\text{Al}}^{3 + } } \right]}}/{{\left[ {{\text{In}}^{3 + } } \right]}} = 0,\;20,\;30,\;\text{and}\;40{\text{\% }} \) as a function of wavelength λ

Full size image

Conclusions

Zinc blend tin sulfide and cubic indium sulfide have been prepared by CBD technique.

The number of deposition runs influences the physical properties of CBD deposited SnS thin layer. It is found that better thickness of SnS thin film 455 nm is obtained after six depositions run estimated by envelope method which is correlated to the experimental value given in our previously study [8]. Using the theory of Manifacier et al. in transparency region the optical properties were found for SnS material. The refractive index “n” value was found to be of the order of 2.5. SnS thin films exhibits a high absorption coefficient \( > 1.5 \times 10^{6} {\text{cm}}^{ - 1} \) in the front absorption region 450–750 nm and low value of extinction coefficient which is a qualitative indication of excellent surface smoothness and homogeneity of the films. Therefore, SnS(ZB) thin films are suitable for absorber layers in solar cell. This indicates good optical quality which permits to use this binary material as absorber layer in photovoltaic device.

Moreover, we note that Al content affects the physical properties of β-In2S3 CBD thin layer. In fact, β-In2S3:Al grown for \( y = {{\left[ {{\text{Al}}^{3 + } } \right]}}/{{\left[ {{\text{In}}^{3 + } } \right]}} = 40{\text{\% }} \) show a higher refractive index, the transmission upper to 55%, the wider band gap is close to 2.76 eV, and a lower extinction coefficient in the visible domain. This indicates good optical quality of these films which allows us to use them as window layers in photovoltaic devices.

Study is also in progress to incorporate these layers in some possibly future photovoltaic devices such as: SnS/In2S3:Al (40%)/SnO2:F, SnS:Cu/In2S3:Al (40%)/SnO2:F, and SnS:In/In2S3:Al (40%)/SnO2:F.