Glasses were used as the substrates in the experiment. Before deposition, we cleaned the glasses with deionized water, acetone, and ethanol using an ultrasonic cleaner for 15  min respectively. Sn and Ag-doped In₂S₃ thin films were deposited on the glass substrates using the thermal evaporation method (DMDE-450 deposition equipment). The Ag patch with 99.9% purity, Sn powder (99.99%), and In₂S₃ powder (99.98%) were used as the source materials. The distance between the source materials and the substrates was kept constant 11.5  cm. During evaporation, the pressure of the chamber was about 2 × 10^{− 3}  Pa. The doped films can be obtained by doping a thin layer of Ag or Sn on the In₂S₃ layer using the thermal evaporation technique. The doping concentrations of Ag and Sn were both set at 4  at.%, which was estimated by the weight of the In₂S₃ source material. The Ag/In₂S₃, Sn/In₂S₃ bilayer and undoped In₂S₃ samples were annealed in Ar atmosphere at 300  ° C for 1.5  h for the diffusion of the metals.

The structure of the prepared films was characterized by the x-ray diffraction (XRD) in the range of 10° – 80° with Cu Kα radiation source (λ = 1.5406  Å ). The optical reflectance and transmittance spectra were measured in the range from 350  nm to 1200  nm with a Cary 5000-scan UV– vis– NIR spectrometer. The optical constants such as the extinction coefficient and the refractive index were also obtained from the reflectance and transmittance spectra. Compositional analysis was carried out with x-ray photoelectron spectroscopy (XPS). The surface morphologies were investigated by atomic force microscopy (AFM) and scanning electron microscopy (SEM). An HMS-3000 Hall measurement system was used to study the electrical properties such as the carrier concentration and the resistivity of the films.

The x-ray diffraction patterns of the prepared films are shown in Fig.  1. We can see that all samples are polycrystalline and all the diffraction peaks can be attributed to the cubic In₂S₃ phase (PDF 01-084-1385). No new phases are observed either after Ag or Sn doping, which means that there is no new crystalline formed and no change in the structure of In₂S₃ due to the incorporation of Ag or Sn. This phenomenon is in accordance with that reported in the previous works.^{[11, 12]}

Fig.  1.

	Figure Option ViewDownloadNew Window
	Fig.  1. XRD patterns of the doped and the undoped samples.

The average crystallite size (D) can be estimated from the strongest diffraction peak (311) using Scherrer’ s equation^[13]

where k is a constant, θ is the Bragg angle, λ is the wavelength of the Cu Kα radiation line, and β is the full width at half maximum corresponding to the preferred orientation (311). The calculated crystallite sizes are summarized in Table  1. Since the ionic radius of In³⁺ is 0.81  Å , which is much smaller than that of Ag⁺ (1.26  Å ), a considerable increase in the crystallite size can be observed after Ag doping, as shown in Table  1. After Sn doping, there is a slightly decrease in the crystallite size, which may be due to the smaller ionic radius of Sn⁴⁺ (0.71  Å ) compared to that of In³⁺ (0.81  Å ).^[14]

Table 1.

Table 1.

Table 1. Grain sizes of doped and undoped samples. Sample In2S3 In2S3:Ag In2S3:Sn Grain size/nm 32.29 42.13 28.09

Table 1. Grain sizes of doped and undoped samples.

The XPS can evaluate the elemental compositions of the films and the bonding states for each constituent element. All the XPS spectra are calibrated by the C 1s peak (284.6  eV). The XPS spectra of the Ag-doped In₂S₃ film are shown in Fig.  2, which proves the diffusion of Ag into the In₂S₃ film after annealing. Binding energies that correspond to the elemental form of Ag⁺ (373.9  eV for Ag 3d_3/2, 367.8  eV for Ag 3d _5/2) can be observed. The binding energies of In and S (452.4  eV and 444.8  eV for In 3d_3/2 and In 3d_5/2, respectively, 161.9  eV for S 2p) indicate the formation of In₂S₃.^[15] The binding energy of oxygen (532.2  eV) corresponds to surface contaminant.^[16] For the Sn-doped In₂S₃ film, the Sn 3d photoelectron peaks are shown in Fig.  3, which verifies the diffusion of Sn into the In₂S₃ film. The peak at 486.8  eV for Sn 3d_5/2 corresponds to the Sn⁴⁺ state and the binding energy of 495.2  eV corresponds to Sn 3d_3/2.^[17]

Fig.  2.

	Figure Option ViewDownloadNew Window
	Fig.  2. XPS spectra of the Ag-doped In2S3 film.

Fig.  3.

	Figure Option ViewDownloadNew Window
	Fig.  3. Sn 3d photoelectron peaks in the XPS spectra of the Sn-doped In2S3 film.

The plane view AFM images of the doped and the undoped In₂S₃ films are shown in Fig.  4. The surface of the undoped film is composed of spherical particles uniformly distributed over all the scanned area. These spherical particles become bigger after the Ag incorporation, and the particle is of a large size exceeding 60  nm, which is much larger than the grain size (42.13  nm) calculated from the XRD result. It might be that the grain has undergone a significant growth to incorporate the nearby grains during the annealing process. The surface of the Ag-doped film is not smooth enough and the height difference can reach about 60  nm. The unsmooth surface obviously results from the large particles that are produced in the In₂S₃ layer. For the Sn-doped film, plenty of little particles are uniformly dispersed in the layer without obvious cracks.

Fig.  4.

	Figure Option ViewDownloadNew Window
	Fig.  4. AFM images (1  μ m× 1  μ m) of (a) the un-doped, (b) the Ag-doped, and (c) the Sn-doped films.

The SEM micrographs of the films are shown in Fig.  5. It is obvious that the crystallite size is increased after Ag doping and decreased after Sn incorporation, which supports the results obtained from the XRD analysis.

Fig.  5.

	Figure Option ViewDownloadNew Window
	Fig.  5. SEM images of (a) the un-doped, (b) the Ag-doped, and (c) the Sn-doped films.

The transmission spectra of In₂S₃, In₂S₃:Sn, and In₂S₃:Ag films in the wavelength range from 350  nm to 1200  nm are illustrated in Fig.  6. All the three samples show good transparency in the visible region and exhibit a sharp drop near the band edge, indicating the presence of direct optical transition in the films. From Fig.  6, we can see that all the transmission spectra exhibit interference fringes, which is an indication of the good uniformity and homogeneity of the films.

Fig.  6.

	Figure Option ViewDownloadNew Window
	Fig.  6. Optical transmittance of the samples.

From the optical reflection and transmission measurement, the optical band gap E_g can be obtained. For the direct allowed transition, the relation between E_g, absorption coefficient α , and photon energy hv can be expressed as^[18]

Here hν is the incident photon energy and A is a constant related to the effective mass. As shown in Fig.  7, the band gap can be determined by extrapolating the linear portion of the plot (α hν )² versus hν . The obtained band gaps of In₂S₃, In₂S₃:Ag, and In₂S₃:Sn films are 2.78  eV, 2.65  eV, and 2.85  eV, respectively. That is to say, Ag-doped In₂S₃ shows a decrease in band gap while Sn-doped In₂S₃ exhibits an increase in band gap compared to that of the undoped film. The higher band gap can be attributed to the decreasing size of the nanocrystalline particles induced by Sn doping.^[19] The wider band gap induced by Sn incorporation can promote the light transmission in the blue wavelength region. Therefore, the In₂S₃ film doped with Sn is more suitable for buffer layer application in solar cells than that doped with Ag.

Fig.  7.

	Figure Option ViewDownloadNew Window
	Fig.  7. The influence of Ag and Sn doping on the bandgap.

The expression of extinction coefficient k in the fundamental absorption section is given by ^[20] Figure  8 shows the variation of the extinction coefficient as a function of wavelength for the In₂S₃, In₂S₃:Ag, and In₂S₃:Sn thin films. In the strong absorption section, the extinction coefficients are high, then they decrease sharply with increasing wavelength. It can also be observed that k is slightly decreased after Sn doping, while the incorporation of Ag leads to an increase in the extinction coefficient.

Fig.  8.

	Figure Option ViewDownloadNew Window
	Fig.  8. The dependence of the extinction coefficient of the deposited films on the wavelength.

The refractive index can be obtained through the minima T_m and the maxima T_M of the envelopes of the transmission spectra.^[21] The minima T_m and the maxima T_M are obtained from Fig.  6. The refractive index n can be calculated by the following equation:

where

Here s is the refractive index of the substrate (1.5 for glass). Using Eqs.  (3) and (4), we can obtain the refractive indices of the deposited samples, as illustrated in Fig.  9. It can be seen that, in the zone of transparency, all the refractive indices vary between 1.9 and 2.37, which are a little smaller than those reported in Ref.  [22]. Besides, it can also be observed in Fig.  9 that the refractive index slightly decreases after either Ag or Sn doping.

Fig.  9.

	Figure Option ViewDownloadNew Window
	Fig.  9. The relation between the refractive index and the wavelength for the deposited films.

The electrical properties of the deposited films are summarized in Table  2. The Hall measurement proves that all these samples are of n-type conductivity. Both the Ag and Sn dopings result in a reduction of the electrical resistivity and an increase of the carrier concentration. For Ag doping, the grain boundary scattering is reduced due to the considerable increase of the crystallite size induced by the doping. Since there is a high concentration of In vacancies in the In₂S₃ thin film, the Ag ions can be easily doped into the In sites serving as donors. As a result, doping with Ag leads to the rapid increase of the carrier concentration. For Sn doping, the vacant cationic sites might be occupied by Sn atoms, or the Sn atoms might replace the In atoms progressively.^[10] As the Sn element belongs to group IV, whether it is in the substitutional position or in the interstitial position, it will act as a donor. Therefore, it will provide electrons and increase the electron concentration, thus promoting the conductivity.^[23] The In₂S₃ buffer layers with low resistivity can reduce the series resistance of the solar cell, which will increase the short circuit current of the solar cell. Therefore, the Sn-doped In₂S₃ is more suitable for buffer layer application in solar cells concerning the electrical properties.

Table 2.

Table 2.

Table 2. Electrical properties of the samples. Sample Carrier concentration/cm− 3 Resistivity/Ω · cm In2S3 2.657× 1014 2357 In2S3:Ag 1.162× 1018 2.92 In2S3:Sn 4.512× 1017 26.36

Table 2. Electrical properties of the samples.