Y. Bchiria,
N. Bouguila*a,
M. Krainia,
R. Souissibc,
C. Vázquez-Vázquezd,
M. A. López-Quintelad and
S. Alayaa
aLaboratoire de Physique des Matériaux et des Nanomatériaux appliquée à l’Environnement, Faculté des Sciences, Université de Gabès, Cité Erriadh Manara Zrig, 6072, Gabès, Tunisia. E-mail: nour.bouguila@fsg.rnu.tn
bTunis University, Ecole Nationale Supérieure d'ingénieurs de Tunis (ENSIT), Tunisia
cUniversité de Carthage, Laboratoire des Matériaux, Molécules et Applications IPEST, BP 51, La Marsa 2070, Tunis, Tunisia
dLaboratory of Magnetism and Nanotechnology (NANOMAG), Department of Physical Chemistry, Faculty of Chemistry, Universidade de Santiago de Compostela, 15782 Santiago de Compostela, Spain
First published on 10th June 2020
Indium sulfide (In2S3) thin films have been synthesized on glass substrates using the spray technique (CSP). The S:In molar ratio was varied from 1 to 4 in the starting solution. The Raman analysis confirms the formation of the β-In2S3 material and the absence of a secondary phase. The EDS analysis reveals that our layers are pure. The thin film surface is free of cracks, as observed in AFM images. Optical transmission achieved 80% in the visible and near infrared region. The refractive index (n) is affected by the changes in the S/In molar ratio. The optical parameters, single oscillator energy (E0), dispersion energy (Ed) and high frequency dielectric constant (ε∞), are calculated via the Wemple–DiDomenico model. In addition, the photoconductivity kinetics in In2S3 films for S/In = 2 were investigated and analyzed. The I–V characteristics and the photoresponse were also studied.
The efficiency of MEH-PPV/H-CuInS2 hybrid solar cells was studied by Yue et al.1 Hierachical CuInS2 was deposited with the induction of histidine by solvothermal method. The authors found that devices based on H-CuInS2 synthesized with smaller size (0.5 HD) displayed higher energy conversion efficiency of 0.59%. The chalcopyrite CuInS2 quantum dots (CuInS2-QDs) with different sizes synthesized by the solvothermal method was used in MEH-PPV–CuInS2/TiO2-NA solar cells.2 The authors obtained a peak efficiency of 1.60% for the 5.4 nm CuInS2-QDs. In addition, this efficiency is comparable to the QDSSCs based on CdS sensitized TiO2-NA.3
Recently, a series of Fe1−xCoxS2 (x = 0–0.5) solid solutions have been synthesized and served as counter electrodes in dye sensitized solar cells (DSSCs).4 The results reveal high power conversion efficiency of 8.36% than those solid solutions materials over Pt of 7.66%. Wu et al.5 studied hybrid TiO2/Sb2S3/P3HT n–i–p solar cells with Sb2S3 layers of different thickness by CBD method. They reported an optimum efficiency of 1.65% with Sb2S3 thickness of 175 nm.
However, the use of cadmium (Cd) in PV devices is undesirable from the viewpoint of environmental safety,6 serious efforts have been made to substitute the CdS by other non-toxic (Cd-free). Indium sulfide (In2S3) is one of the possible candidates to replace CdS a buffer layer. This compound is an n-type semiconductors which belongs to III–VI group. Hetro-juncture such as CIGS/β-In2S3 (ref. 7) achieved a conversion efficiencies of 15.7%. This demonstrates that comparable conversion efficiencies can be obtained respect to commonly used CdS buffer (16%).8 Osman et al.9 observed that there is no difference between both structures of CdS and In2S3, they found also that the efficiency of CIGS solar cells with both layers is comparable. It's 24.14% and 24.28% with In2S3 and CdS respectively.
In2S3 is currently researched as a result of its important properties in relation with the optoelectronic devices.10 β-In2S3 is recognized as a direct band gap semiconductor ranging from 2.0 to 2.75 eV, with a high stability, high optical transmittance (>80%) and controllable electrical properties, modified chemical composition and deposition parameters.11,12 These characteristics of In2S3 films make it a useful material for several applications. In2S3 can be used in light-emitting diode (LED) and lithium-ion batteries.13,14 Due to its absorption in the UV and visible region, In2S3 is considered as photodetector.15 Also, this material can be used for gas sensing and photocatalytic applications owing to its thermal stability, good response and reversibility, low defect density and simple synthesis process.16,17
In2S3 exists in three different crystalline phases α, β and γ which depend on the growth temperature.12,18 β-In2S3 is the most stable phase at room temperature.19,20 It crystallizes in a defect spinel structure with a high degree of tetrahedral and octahedral vacancy sites.21,22
Various techniques were commonly used for preparing In2S3 such as spray pyrolysis,23,24 ultrasonic dispersion,25 physical vapor deposition,26 chemical bath deposition,27 vacuum thermal evaporation,28 etc. Among these methods, the spray pyrolysis technique is of particular interest because it is economical and allows growing large area In2S3 thin films. Furthermore, this technique has been applied to fabricate various type of materials such as ZnO and carbon nanofibres (ECN/TiO2).29,30 Indeed, Ngo et al.30 presented a comparative study between ZnO films deposited by spray coating and ZnO prepared by spin coated. They found that the spray coated sample have higher efficiency (3.76%) than spin coated ZnO (3.11%) due to the increase in FF from 42.38% to 63.14%. So that, to choice of technique is one of important factor in solar cells application.
The physical properties of In2S3 thin films are strongly affected by the deposition parameters.25 Several authors have reported that 340 °C is a good substrate temperature for obtaining best crystallinity.31,32 Thus, various works showed that the S/In ratio has an effect on the structure, the crystallite size33–35 and the optical band gap of the In2S3 films.11,36 Zhang et al.37 have reported that the grain size increases from 32 nm to 34.1 nm with the increase of S/In molar ratio from 1 to 4. Furthermore, the high transmittance is found in the visible and near-infrared range. The energy band gap increases significantly when S/In ratio varies from 1 to 2, then decreases from 2.46 eV to 2.4 eV when S/In increases from 2 to 4. John et al.11 have reported that the use of sulfur-rich solution (S/In = 8/2 instead of S/In = 2/1) decreases the energy band gap from 2.81 to 2.64 eV. The photoluminescence (PL) properties of In2S3 at different S/In ratios have been reported by Elfarrass et al.33 They showed that all the films have two emission bands (green and red band). In addition, Bhira et al.38 have demonstrated that In2S3 thin film has high photoconductivity. Therefore, investigations on the effect of S/In ratio on the In2S3 optical properties are very important, to obtain films that are capable to ensure stable and high efficient devices.
In this work, we report the study of the effect of S/In ratio on the physical properties of In2S3 thin films.
The formation of In2S3 results from the following endothermic reaction:
2InCl3 + 3SC(NH2)2 + 6H2O → In2S3 + 3CO2 + 6NH4Cl | (1) |
The Raman analysis were conducted using a Jobin-Yvon microRaman spectrometer (T64000) with an argon laser line of 514 nm. The composition of films was done by energy dispersive spectroscopy (EDS). The surface morphology was performed by atomic force microscopy using an XE-100 instrument (Park Systems Corporation). Photoluminescence (PL) measurements were performed at room temperature using laser excitation wavelength of 655 nm. Transmittance spectra measurements have been done using a Shimadzu UV 3101 PC spectrophotometer in the wavelength range from 350 to 1900 nm. The I–V characteristics measured in dark and under illumination were performed using a (HP4140B) source/pico-ammeter. Moreover, the photoresponse of In2S3 thin films was investigated by illuminating the film with a neon light source. The resistance of the sample is measured by using a (HP4140B) Keithley Digital multimeter, interfaced to a computer for data acquisition.
Fig. 2 (a) EDS spectra of thin films at different S/In molar ratios. (b) EDX spectrum of glass substrate.45 |
S/In | S (at%) | In (at%) | S/In |
---|---|---|---|
1 | 24.60 | 25.01 | 0.98 |
1.5 | 52.18 | 42.34 | 1.23 |
2 | 41.93 | 33.51 | 1.24 |
2.5 | 46.92 | 36.76 | 1.27 |
3.5 | 41.02 | 32.08 | 1.28 |
4 | 41.02 | 32.08 | 1.28 |
Fig. 3 (a) 2D and 3D NC-AFM images of thin films for S/In molar ratios equal to 1, 2 and 4. (b) Grain size distribution histogram of thin films for S/In molar ratios equal to 1, 2 and 4. |
S/In | 1 | 1.5 | 2 | 2.5 | 3.5 | 4 |
RMS (nm) | 20 | 15 | 11 | 9 | 8 | 6 |
In order to evaluate the average grain size, a statistical count of grain size was performed on the AFM images using Image-J software. Grain size distribution histograms for the different films are shown in Fig. 3b. The average grain size of samples is found to lie in the nanometer range (44–86 nm).
Room temperature photoluminescence spectra of In2S3 at different S/In molar ratios are exhibited in Fig. 4. PL spectra reveal a peak at 1.46 eV, 1.52 eV, 2.32 eV and 2.55 eV.
The red emission at 1.45 is due to S and In vacancies in the host lattice.46 The emission at 1.52 eV and green emission (at 2.32 eV and 2.55 eV) are attributed to transitions from the excited states of the sulfur vacancy (Vs) to the In vacancy (VIn) level.46–48 Mathew et al.22 reported a green emission from In2S3 which was attributed to the donor level formed by indium interstitials. The PL emission intensity is found stranger for sample at S/In = 1 than other samples, which proved the presence of higher defect concentration.49 Thus, the emission intensity decreased indicated the reduction of defects in films. The low emission intensity is obtained for sample at S/In = 3,5, which the crystallinity is the best one. The same result was reported by Ajili et al.50
In addition, the last peak at 2.55 eV is relative to the Eg values estimated from Tauc formula in earlier work by Bouguila et al.44
n = (N + (N2 − n20n21)1/2)1/2 | (2) |
(3) |
The values of n are reported in Table 3. It can be observed that this value is minimum at S/In = 2.5, beyond this value it increases slightly. This is related to the variation in packing density of the films.54
S/In | n | B | C (μm2) | β (eV) |
---|---|---|---|---|
1 | 2.40 | 3.14 | −1.0 | 0.16 |
1.5 | 2.06 | 2.68 | −0.3 | 0.29 |
2 | 2.10 | 2.73 | −0.3 | 0.26 |
2.5 | 1.87 | 2.79 | −0.4 | 0.27 |
3.5 | 2.11 | 2.81 | −0.2 | 0.31 |
4 | 2.03 | 2.84 | −0.4 | 0.26 |
Fresnel equation55 was used to evaluate the refractive index n(λ) of the films as given by:
(4) |
Fig. 6 displays the dispersion of n for different molar ratios. In the high absorption region, the refractive index increases strongly. The sharp onset of n at the absorption edge is attributed to Van Hove singularity in the joint density state in the excitation of transition between two bands.56 Thus, it follows the normal dispersion in the medium and weak absorption regions. This latter phenomenon is ascribed to both light scattering effect and absorbance decrease. From Fig. 6, it can be noticed that n varies from 2.4 to 2.75 with S/In ratio. The higher values of n make In2S3 thin films suitable for use in optoelectronic devices. Indeed, thin films are used for optical containment in waveguides, lasers and light-emitting diodes. The good value of n depends on the active layer used.
Thus, the increase of n indicates the better crystallinity of the films. Indeed, the material with amorphous nature has a low refractive index when compared to the polycrystalline one.57
The variation of n along the Cauchy distribution is given by:58
(5) |
Table 3 represents the values of these constants, which are evaluated from the fitting of n(λ).
The dispersion energy was also calculated using the single oscillator model as described by Wemple and DiDomenico:59
(6) |
Fig. 7 represents the linear variation of (n2 − 1)−1 versus (hν)2 for these films. This permits us to determine both E0 and Ed from the slope (EdE0)−1 and (E0/Ed) from the intercept at the origin for each straight line.
The results are listed in Table 4.
S/In | 1 | 1.5 | 2 | 2.5 | 3.5 | 4 |
E0 (eV) | 3.8 | 5.6 | 5.1 | 4.9 | 5.4 | 4.8 |
Ed (eV) | 15.5 | 27.8 | 25.3 | 25.5 | 30 | 25 |
EWDDg (eV) | 1.9 | 2.8 | 2.5 | 2.4 | 2.7 | 2.4 |
ε∞ | 9.6 | 7.5 | 7.8 | 7.4 | 8 | 7.9 |
ωp (rad s−1) | 14 | 8 | 9 | 10 | 8 | 10 |
N/m* (m−3) | 6 | 1.5 | 2 | 2 | 1.5 | 2.4 |
The highest dispersion energy value is obtained at S/In = 3.5, indicating that this film has a more ordered microstructure compared to the other films. The microstructure of In2S3 thin films varied with S/In ratio. Indeed, the influence of S/In molar ratio on microstructure is related with the nucleation and grows of In2S3 crystal.60 Furthermore, sulfure deficiency or sulfur excess would involve the presence of impurities that can affect or destroy the microstructure.61
It can also be seen that the values of E0 vary from 3.8 to 5.6 eV and oscillation energy describes the expressions E0 ≈ 1.5Eg at S/In = 1 and E0 ≈ 2Eg for S/In > 1. We can thus conclude that the Wemple–DiDomenico model describes well the behavior of these films.62,63 The lower value of ratio E0 is obtained at S/In = 1, this could be attributed to the higher rate of diffusion of atoms in this film.64 Furthermore, the change in the values of E0 and Ed reveals that the molar ratio S/In is a suitable parameter for the change of refractive index and oscillator parameters. Thus, Tanaka63 proposed the first approximate value of the optical band gap EWDDg from the Wemple–DiDomenico model using the expression (). The EWDDg values are given in Table 4. The small difference between the two methods is ascribed to the calculations in two different regions. Indeed, in the case of Tauc method, the value of Eg is calculated in the absorption region of the spectrum, while in the Wemple–DiDomenico model the value of EWDDg is acquired in the transparent region of the spectrum.
The following empirical relation for Ed is established in crystals containing a single anion species:65
Ed = βNeZaNc | (7) |
(8) |
(9) |
The free carrier concentration-to-effective mass ratio can be obtained from the relation:
(10) |
Fig. 8a and b, show that the behavior of ε1 is similar to the refractive index (n). This is due to the smaller value of k compared to n. However, the behaviors of ε2 and α are the same, because ε2 depends on the k values which are correlated to the variation of absorption coefficient (α). We observed high value of ε1 at low wavelengths; this is due to the presence of space charge polarization at the grain boundaries, which generates a potential barrier. Thus, a charge accumulation at the grain boundary occurred, which leads to highest value of ε1.68 However, the real part decreases with the increase of wavelength. This trend is due to the reduction of space charge polarization effect.
From the intercepts and slopes of the ε1 plots as a function of λ2 (see Fig. 8a), the values of ε∞ and N/m* for the films are estimated and reported in Table 4.
The increase in refractive index induces a rise in the electromagnetic radiation absorption which in turn increases the frequency. In addition, the value of n becomes high when the radiation frequency corresponds to the electron characteristic frequency. The physical interpretation of the variation of plasma frequency can be ascribed to the variation of the concentration of the charged carriers in In2S3 thin films. Hence, the high value of ωp at S/In = 1 might be due to the higher concentration of the charged carriers in this film. This finding confirms the results of optical band gap value at S/In = 1. As can be seen, all N/m* values are of the order of 1049 m−3. Bouguila et al.69 reported comparable values for In2S3 thin films for different thicknesses. The higher value obtained at S/In = 1, could be explained by higher defect states in the film.70 Besides, Nicholas et al.71 reported that the low value of effective mass leads to high carrier mobility. Thus, it is clear from Table 2 that the values of N/m* are affected by the S/In molar ratio. So, we can say that ε∞ and N/m* are assigned to the internal microstructure.72
I–V characteristics measured in dark and under illumination by neon lamp with a bias voltage ranging from −10 to 10 volts are shown in Fig. 10a. The linear I–V curves from the sample confirm the ohmic nature of Au/In2S3 contacts. This ohmic contact with Au electrodes enables excellent photo-response behaviors. Furthermore, the formation of ohmic contacts between semiconductors and metals serving as electrodes is an important requirement for photodetector device. Indeed, this contact is required to inject the maximum current density across the contact.73
Fig. 10 (a) I–V characteristics of In2S3 thin film in dark and under illumination, (b) adjustment of In2S3 thin film photoresponse with two exponentials evolutions. |
The resistance evolution was measured, while the sample was illuminated by reversible switching (ON/OFF) cycles of the neon lamp. The light was ON for 15 min and OFF for 15 min for all cycles. Fig. 10b shows that the resistance of In2S3 film increases in dark until reaching its equilibrium value and decreases quickly to the baseline value in light. We also observe that repetitive cycles profile reveals the ability of the material to produce the same response, by exhibiting an acceptable reproducibility and reversibility of the response. This evolution suggests that more photons with higher energy are involved in the photon-induced charge-transfer process. This behavior is suitable given the fact that typical absorption is located for wavelength < 500 nm, as shown in Fig. 5. Likewise, the optical excitation allows the generation of electron–hole pairs. The fact that the transit time of the holes is much greater than that of the electrons; a single hole can cause the circulation of several electrons in order to ensure the electrical neutrality of the material, which enhances photocurrent response.74
The increase of photocurrent is traduced by the decrease of the resistance. In this context, we have adjusted in Fig. 10b, the resistance curves versus time with a bi-exponential equations that have two evolution components in both decaying and rising progress:74
(11) |
(12) |
The determination of the constants A1, A2, τ1 and τ2 is not obvious given the nonlinearity of the eqn (11) and (12). As a result, we divide each film response into two parts:
-the first part is dominated by the fast kinetic (τ1) and corresponds to a significant variation of the resistance.
-the second part is dominated by a slow kinetic (τ2) and results in a weak variation of the resistance.
We have adjusted the last points of the second domain (t ≫ τ1) by the function A2e−t/τ2 and we have deduced A2 and τ2 values using the least squares method. Then, the quantity A2e−t/τ2 is subtracted from the experimental points of the first domain (0 < t < τ1). The obtained results are similarly adjusted via the same technique to determine A1 and τ1. Several iterations were made in order to adjust precisely the theoretical fit with the experimental results. The fit is constructed correspond to the theoretical function. The τ1 and τ2 values are regrouped in Table 5.
Cycle number | Light | Dark | ||
---|---|---|---|---|
τ1 (s) | τ2 (s) | τ1 (s) | τ2 (s) | |
1 | 48 | 24877 | 90 | 4361 |
2 | 22 | 20362 | 77 | 4194 |
3 | 14 | 10223 | 84 | 4695 |
The results of the theoretical adjustment demonstrate good conformity with the experimental data. The evolution of the resistance under illumination and in dark follows two kinetics: the first is fast, the other is slow. For the fast kinetic, the averages values of τ1 are found around 28 s in light and 84 s in dark. We can note that time constants associated in light are relatively small compared to those associated with other in dark. This result is expected because the charge carrier generation time is shorter than the recombination time.75
This journal is © The Royal Society of Chemistry 2020 |