1 Introduction

It is known that the study of II–VI semiconductors has generated a lot of attention in the material science community over the pasts decades due to the possibility of structural and optical modifications given by the quantum confinement effect [1, 2] and the very interesting properties achieved by the metallic ion doping [3,4,5]. One of the most attractive II–VI-type semiconductors is cadmium sulfide (CdS), since it is a n-type photo-sensible semiconductor, with a direct band gap value of 2.4 eV at room temperature. Many research groups have been investigating this material due to its optical and opto-electronical properties, which allow its use in applications such as solar cells [6, 7], thin film transistors [8, 9], photo diodes [10] and light-emitting diodes (LED) [11, 12]. CdS crystals can be grown in two distinct structural phases: cubic (Zinc-blende) and hexagonal (wurtzite) [13]. In addition, CdS thin films can be produced by different techniques including: sol–gel [14], close spaced vapor transport (CSVT) [15], pulsed laser deposition (PLD) [16], RF sputtering [17], chemical vapor deposition (CVD) [18, 19], spray pyrolysis (SP) [20, 21] and chemical bath deposition (CBD) [7, 8, 22]. The latter technique was chosen in the present work due to the fact that it does not require high vacuums nor high temperatures so there is no emission of toxic gases [23]. Chemical bath deposition employs a controlled chemical reaction to deposit nanostructured thin films through a complexing agent [24]. It is worth to mention that typical chemical bath deposition of CdS thin films employs ammonia as complexing agent for the cadmium ions in the reaction solution. It is evident that the large-scale fabrication of CdS thin films, employing great amounts ammonia, could rise some serious environmental issues due to its high volatility and toxicity. Furthermore, the high volatility of ammonia could produce unwanted variations on the pH of the reaction solution in the deposition process, leading to difficulties in the reproducibility of the films. Due to these issues, there is a consistent effort in the search for alternative complexing agents to ammonia in order to make the CBD process more environmentally friendly [25,26,27,28]. The films can be grown in large areas with good reproducibility [29] making CBD a feasible technique for mass production. Since CdS is a promising material for application in solar cells, it has been doped with several materials such as gallium [30], indium [31], boron [32], manganese [33], tin [34], sodium [35], silver [18, 36] and copper [21, 37, 38]. In particular, copper doping has gained our attention since it is capable to change the properties of CdS in a relatively feasible way, increasing its resistivity in dark, but at the same time increasing its conductance in the presence of light, which means more photo-sensibility due to the increase in carrier Cu ions. In addition, it is also known that copper doping changes the CdS n-type semiconductor to a p-type semiconductor [39]. The present work focuses on the synthesis and characterization of nanostructured CdS thin films obtained by an ammonia-free chemical bath deposition process, where different levels of Cu2+ ion doping were achieved by the ion exchange method. This doping technique involves the exchange of ions between an electrolyte solution (aqueous phase) and similarly charged ions (cations in the Cu2+ doping) immobilized in an ion exchange material (solid phase) that takes place through a stoichiometric reversible reaction [40]. In this technique, the exchange is produced only between cations while leaving the anionic part of the lattice intact [41]. In contrast, in the in situ doping technique, there is a probability to produce stronger structural changes during the growth process since the dopant is present in the precursor solution [42, 43]. In addition, the ion exchange process is driven by diffusion, hence, the Cu2+ doping ions will be present mostly on the surface of the thin films [44]. Trough optical characterization, it is possible to evaluate the impact of the metallic dopant on the band gap value.

2 Experimental

The CdS and Cu-doped CdS thin films were deposited by the chemical bath deposition method in soda lime glass substrates. The fabrication method consists of a five steps process: cleaning, deposition, a second cleaning, doping and annealing treatment. All substrates were cleaned three times before deposition; to this end, the substrates were submerged in acetone, isopropyl alcohol and deionized water (one after the other), for 3 min in an ultrasonic bath. Once the cleaning was completed, the deposition process is performed. The chemical bath deposition consisted of an immersion of the substrates on a solution composed by 4 ml of cadmium nitrate 0.1 M, 5 ml of glycine 0.1 M, 2 ml buffer pH 11, 5 ml of thiourea 1 M and 44 ml of deionized water at 70 °C for 25 min. The samples were then cleaned with deionized water for 5 min in an ultrasonic bath. After the second cleaning, the produced CdS films were dipped in a solution comprising of 60 ml of cupric nitrate 0.001 M at room temperature for the ionic interchange of Cu and Cd. Different times of immersion were performed on the ion exchange solution to obtain different copper concentrations. The immersion time was measured, and the samples were labeled as MXX, being XX the time in seconds of immersion in the Cu ion exchange solution. The objective was to control the band gap value adjusting the doping time. Different samples with the same initial CdS film were dipped for different amount of time. Finally, the annealing process was performed at a temperature of 200 °C for 20 min in a nitrogen atmosphere at a pressure of 10 mmHg to improve the crystallinity of the material and to diffuse the Cu ions into the CdS structure.

The morphology and thickness of the deposited films were characterized by scanning electron microscopy (SEM), employing a JEOL JSM-7800F microscope operated at 15 kV. UV–Vis spectroscopy measurements were performed with a Perkin Elmer Lambda 19 Spectrometer, in a wavelength range from 400 to 800 nm. X-ray photoelectron spectroscopy (XPS) was performed with a Perkin Elmer PHI5100 to study the chemical composition on the different samples; the studies were done with a Mg source at 15 kV and 300 W with a pressure of 4 × 10−8 Torr. A Rigaku Ultima III X-ray diffractometer was used for the X-ray diffraction.

3 Results

Scanning electron microscopy measurements were performed on the CdS and Cu-doped CdS thin films to study their surface morphology and thickness. Figure 1 shows the micrographs of the films with different doping times. As shown in the SEM images, the undoped CdS films are composed of nanostructures with amorphous shapes of CdS with sizes around 100 nm. Upon the inclusion of the Cu-doping, it can be seen that the nanostructure’s shapes remained similar with a slight size decrease of up to 15 nm compared with the undoped material. In addition, the SEM cross-section micrograph shows that the thin film thickness is around 100 nm. This value was not significantly modified by the Cu-doping inclusion.

Fig. 1
figure 1

SEM micrographs of the deposited CdS and Cu-doped CdS films

Full size image

Figure 2 shows the optical absorption spectra of the Cu-doped CdS and undoped films. The absorbance measurements were performed in a range of wavelengths from 400 to 700 nm. It can be seen that from 600 to 700 nm, the spectra present an almost flat absorption region. The undoped CdS sample (M00) present an increase of absorption at around 500 nm. As the doping time (and hence the doping concentration) increases, the absorbance edge shifts up to 575 nm approximately. This effect was visible to the naked eye since the samples changed its color from a cadmium yellow to a reddish yellow as the doping time increased.

Fig. 2
figure 2

Absorption spectrum of the CdS and Cu-doped CdS thin films with different doping times

Full size image

Tauc’s plot method [45] was employed to obtain the optical band gap values of the thin films deposited on a glass substrate. To this end, the equation:

$$\alpha hv = {\text{B}}\left( {hv - E_{\text{g}} } \right)^{n}$$
(1)

is used, where α is the absorption coefficient, B is a constant, hv is the photon energy, Eg is the band gap and n equals 1/2 for a direct allowed transition [46]. The optical band gap is determined from the parameters obtained of a linear fit of the term (αhv)2 (calculated from the absorption spectrum), whose intersection with the x-axis (in α = 0) provides the optical band gap. This graphic interpretation starts by transforming the wavelength values to energy level using the Eq. (1), which comes from:

$$E_{\text{g}} = \frac{c \times h}{\lambda }$$
(2)

To obtain the absorption coefficient from the measured absorbance, it was necessary to assume that all the samples had a thickness of 100 nm which was corroborated from a cross-section SEM micrograph of the samples. To calculate α the following equation was used:

$$\alpha = 2.303 A/t$$
(3)

where A is the experimental absorbance and t is the thickness of the sample (100 nm). Figure 3 shows the calculations for the band gap value using the Tauc’s plot method.

Fig. 3
figure 3

Optical band gap values from the samples measured using spectroscopy

Full size image

As shown in Fig. 3, it is observed that the pure CdS film has a band gap of 2.40 eV. In addition, it was observed that as the doping concentration increases, it produces an enhancement in the band gap reaching a value up to 2.46 eV for the 18 s treated sample, while higher Cu-doping concentration seems to produce a decrease of the band gap. This last effect was previously observed in the Cu-doped CdS material and it was attributed to the effect of the band gap of the CuS, which is 2.16 eV [43].

Figure 4a shows the XPS measurements performed on three samples with different Cu-doping concentrations: (i) an undoped CdS thin film (M00), (ii) a sample with a low doping concentration (M06) and (iii) one with a high doping concentration (M60). As can be seen in Fig. 4a, the three samples exhibit the typical photoelectronic peaks of CdS thin films [47], while the copper ion signals were found only in the Cu-doped thin films as expected. The observed signals correspond to possible CuS compounds with a 2p doublet (2p1/2 and 2p3/2) with binding energies around 952 eV and 932 eV, respectively [48] and the Auger line L3M45M45 (335 eV) associated with copper when a Mg source is utilized for XPS characterization [48] was also found in the doped samples. High-resolution XPS spectra were collected from both the undoped and Cu-doped CdS thin films. Figure 4b shows the characteristic peaks for the undoped CdS 3d doublet at 411.66 eV and 404.92 eV for the Cd 3d3/2 and Cd 3d5/2 respectively, while in the case of the Cu-doped film (Fig. 4c), it can be seen that the incorporation of the copper ions through ion exchange, introduced two small signals at around and 404.5 eV and 411.3 eV, these new signals are thought to be produced by the metallic Cd ions that were replaced in the ion exchange and/or the CdO produced during the reaction. In addition, analyzing the Cu 2p doublet signal from Fig. 4d and e, it can be seen that most of the copper ions were successfully exchanged with the cadmium ions producing CuS. However, a small signal possibly composed by unreacted metallic copper and/or copper oxides are still present. In addition, from the XPS measurements, it was found that the M06 sample, presented a composition of 8% copper, 38% cadmium and 54% sulfur, while the M60 sample had a composition of 17% copper, 34% cadmium, and 49% sulfur. This information confirms that the ion exchange method employed in the present study, produce the replacement of cadmium ions by copper ions, preserving the nearly stoichiometric composition.

Fig. 4
figure 4

a X-ray photoelectron spectroscopy spectra obtained of the undoped and Cu-doped CdS films, b, c high-resolution spectra of the Cd 3d doublet for the undoped and Cu-doped CdS thin films and d–e high-resolution spectra of the Cu 2p doublet for the Cu-doped CdS thin films

Full size image

The X-ray diffractograms of the undoped CdS and Cu-doped CdS thin films are shown Fig. 5. The results exhibit the preferential cubic CdS \(\left( {1 1 1} \right)\) growth plane at the angular position \(2\theta = 26.75^\circ\) with small contributions of the cubic \(\left( {2 0 2} \right)\) and \(\left( {3 1 1} \right)\) growth planes at \(2\theta = 44.28^\circ\) and \(2\theta = 52.41^\circ\) respectively, which correspond to the zinc-blende structure of CdS indexed in the crystallography open database [49] entry COD-1011251. This crystallographic structure is typically found in CdS thin films obtained by CBD [30, 50]. As the doping concentration increases, the growth of the planes \(\left( {100} \right)\) and \(\left( {101} \right)\) of hexagonal CdS (COD-1011054) located around \(2\theta = 25.35^\circ\) and \(2\theta = 28.52^\circ\) respectively, becomes more evident, suggesting a possible coexistence of the two different crystal structures. Additionally, the grown of the \(\left( {0 1 8} \right)\) plane of the hexagonal phase of CuS (COD-9000062) located around the angular position \(2\theta = 52.51^\circ\) could help to confirm the ion exchange from Cd2+ to Cu2+ with the possible formation of a small population of CuS crystals that could explain the partial growth of the hexagonal wurtzite structure in the CdS matrix of the doped thin films. However, this last signal may be superimposed by the cubic CdS (3 1 1) signal. Scherrer Equation was employed to calculate the crystallite size of the samples before and after the doping process, however, the calculated values were found to be around 13.4 nm with almost negligible variations (within the standard deviation) upon the inclusion of the Cu-doping ions.

Fig. 5
figure 5

XRD measurements of the CdS and Cu-doped CdS films

Full size image

A pseudo-MOSFET (prototype) structure was fabricated employing both the undoped and the Cu-doped CdS thin film to investigate its possible application as a field effect transistor. The electrical behavior of the undoped and Cu-doped CdS thin films in a pseudo-MOSFET structure (as shown in Fig. 6a) was analyzed by current vs voltage measurements at room temperature under dark conditions. Figures 6b, c show the IDS–VDS output characteristics measured in the voltage range from 0 to 50 V, for several gate voltages (VGS). As shown in Fig. 6b, it can be observed that the channel current of the “undoped device” does not reach saturation even when 30 volts were applied to the gate contact. In contrast, in Fig. 6c it can be observed that the channel current of the “Cu-doped device” saturate even with 5 volts applied in the gate terminal. This observation could be considered an indication that field effect transistors can be developed by doping CdS thin films, and that they could be controlled with greater channel current precision.

Fig. 6
figure 6

a Pseudo-MOSFET prototype structure (top), and the electrical characterization of b undoped CdS device (bottom left) and c Cu-doped CdS device (bottom right)

Full size image

4 Conclusions

Cadmium sulfide thin films were fabricated by the chemical bath deposition technique. The doping process of the CdS thin films was obtained by ion exchange method at room temperature. Analysis of absorbance, energy band gap, XPS, SEM, and XRD were performed to the fabricated thin films to study the effect of the doping concentration on the properties of the material. It was found that a low doping concentration of copper produces an increase in the bandgap value of the CdS thin films, while higher concentrations produces de opposite effect since it could be promoting a greater formation of CuS. Considering the presented results and some previous studies cited here, it can be argued that the copper ions are not just in the surface of the deposited film, but actually they merged into CdS material. This effect is proposed since the thin films shown changes in their band gap values, differences in their crystallinities as well as the elemental composition for the different copper concentrations. In addition, a prototype of a pseudo-MOSFET structure was fabricated employing both the undoped and the Cu-doped CdS thin film to corroborate its possible application as a field effect transistor. However, the copper doping proposed in the present work requires a more precise control in the ion interchange method to obtain smaller and more adjustable copper concentrations. A better control in the doping process will improve the manipulation of the band gap values, producing a more suitable semiconductor for a wider variety of applications since the capability to produce a more reliable semiconductor with a controllable band gap is a priority for semiconductor industry.