The effect of copper doping on the structural, optical, and electrical properties of cadmium oxide thin films using the spray pyrolysis technique


 In this study, the pure and Cu doped CdO thin films with various doping concentrations (0 to 5 at.%) were deposited on soda-lime glass substrates, using the chemical spray pyrolysis technique. The effects of Cu doping on the structural, optical, and electrical properties of thin films were, then, investigated. The films were characterized through X-ray diffraction (XRD), field emission scanning electron microscopy equipped with an energy dispersive x-ray analyzer (FESEM-EDX), atomic force microscopy (AFM), ultraviolet-visible spectroscopy, and electrical resistance; van der Pauw techniques were also used to measure the Hall effect. X-ray diffraction studies showed that the thin films were polycrystalline only with cadmium oxide phase with the cubic face-centered crystal structure and the preferred orientations were along (111), (200), (220), (311), and (311) planes. The FE-SEM and AFM images also showed that with an increase in Cu doping levels, the grain size and surface roughness of the thin films decreased from 472 to 38 nm and from 163 to 54 nm, respectively. The expected element compositions were confirmed by EDX. The optical bandgap of the thin films ranged from 2.42 to 2.56 eV, and as Cu dopant increased, so, too, the optical bandgap. As the Cu doping concentration increased from 0 to 5 (at.%), the electrical resistivity and Seebeck coefficient were found to increase from 3.74⋅10− 4 to 8.77⋅10− 3 Ω.cm and 8.28⋅10− 6 to 2.52⋅10− 5 v/k (at 100 Co temperature difference), respectively; the carrier concentration and carrier mobility were also found to decrease from 3.24⋅1020 to 1.76⋅1020 cm− 3 and from 55.5 to 4.05 cm2/v.s, respectively. The Hall effect and thermoelectric studies revealed that the films exhibited an n-type conductivity.


Introduction
Electrical and optical properties are more significant in metal oxides such as cadmium oxide (CdO), tin oxide (SnO 2 ), indium oxide (In 2 O 3 ), titanium oxide (TiO 2 ), zinc oxide (ZnO), and nickel oxide (NiO), this issue has never escaped the attention of optoelectronic industries (Benhaliliba, et al. 2012;Dounia, et al. 2016;Ramarajan, et al. 2020;Şahin 2019). CdO is an important II-VI semiconductor with an n-type high electrical conductivity of ~10 2 -10 4 S/cm thanks to its native defects and good transparency, especially in the NIR spectral region, with a direct bandgap of 2.2-2.7 eV depending on the deviation from the ideal CdO stoichiometry (Zargar, et al. 2020). CdO has a cubic (NaCl) structure based on octahedral coordination around Cd atoms and an Fm-3m symmetry with = 4.695 ̇ (Dakhel 2014a;Dakhel and Bououdina 2014;Sahare, et al. 2017). These suitable physical properties have helped CdO thin films to have applications in photovoltaic solar cells, optoelectronic devices, smart windows, panel displays, thermally insulating glass, gas sensor devices, transparent electrodes, photodiodes, liquid crystal displays, IR detectors, and anti-reflection coatings, to name but a few (Aydin, et al. 2019;Dakhel and Bououdina 2015;Elttayef, et al. 2013). However, the optoelectronic properties of CdO thin films need to be adjusted and tailored for many optoelectronic applications. It is possible to engineer the physical properties of cadmium oxide by doping it with different elements like Ni, Fe, Y, In, Sm, Al, F, Ga, W, B, Ni, Mn, Mo, In, Au, Ag, Mn, Eu, Cr, Ti, Sn, Yb, and Cu (Desai 2018;Hymavathi, et al. 2017a;Jadhav, et al. 2017;Raaif and Mohamed 2017;Sankarasubramanian, et al. 2018). It was experimentally observed that it is possible to control the optoelectronic properties of CdO by doping it with different types of metallic ions with ionic radii smaller and/or equal to that of host lattice atoms (Dakhel 2014a;Dakhel and Bououdina 2015;Noorunnisha, et al. 2020a). Conduction properties of CdO are caused by its intrinsic natural structural point defects.
Cadmium interstitials (Mi) and oxygen vacancies (VO) are point defects responsible for n-type conduction and conduction properties. Therefore, doping metallic ions increases metal interstitials (Mi) and oxygen vacancies (VO) and can improve the conduction parameters (Çolak and Türkoğlu 2013;Dakhel 2014a). In general, Cu can be easily used as a dopant in Cd; In fact, because the ionic radius of Cu +2 with six coordinations (0.073 nm) is less than that of Cd +2 ion (0.095 nm), it could improve the conduction parameters (Dakhel and Bououdina 2014). The CdO-based thin films can be produced through all manner of techniques ranging from thermal evaporation (Alharbi and Qasrawi 2021), pulsed laser deposition , sputtering (Hymavathi, et al. 2017b), e-beam evaporation (Purohit, et al. 2017), sol-gel spin coating (Sahbeni, et al. 2020), chemical vapor deposition (CVD) (Sharma, et al. 2019), chemical bath deposition (Das and Mitra 2017), successive ionic layer adsorption and reaction (SILAR) (Sharma, et al. 2020), and solution growth (Onwuemeka and Ekpunobi 2018), to spray pyrolysis (Noorunnisha, et al. 2020b). Among all these techniques, the spray pyrolysis technique presents some advantages over other techniques in terms of affordability and suitability, which also convinced us to opt for it. In this work, we have studied the effect of copper doped CdO thin films on their physical properties.

Film preparation
Thin films were deposited on soda-lime glass substrates using a typical spray pyrolysis coating system. [ ] (at.%) was changed from 0 to 1, 2, 3, 4, and 5 (at.%) to examine the effects of the Cu content on the physical properties of the thin films. The solution was, subsequently, stirred well and heated for about 45 min at 50 o C until it became bright and suitable for spray. The substrate temperature in all samples was set to 340 °C with an accuracy of ± 5 °C using a digital temperature controller. The spray deposition rate was 10 mL/min, with a 35 cm distance between the nozzle and the substrate. The hot plate rotation speed was 30 rpm.
Before preparing the films, glass substrates were cleaned with ethanol using an ultrasonic bath for 15 min and placed on a hot plate. Eventually, the system was allowed to cool naturally to room temperature for about 80 min. The thin films prepared from the spray solutions with an atomic ratio of copper per tin equal to 0, 1, 2, 3, 4, and 5 (at.%) were named S 1 , S 2 , S 3 , S 4 , S 5 , and S 6 , respectively.

Sample characterization
The structural properties were examined through X-ray diffraction (XRD), using a D8 Advance Bruker X-ray diffractometer with Cu-Kα radiation (λ = 0.154056 nm) spectra in a 2θ range of 10°-80°. The surface morphology was investigated by field emission scanning electron microscopy (FESEM; HITACHI S-4160) and atomic force microscopy (ARA-AFM Brisk). A Unico 1800 double-beam spectrophotometer was also utilized to record the optical absorption of the thin films at the 450-1000 nm wavelength. (220), (311), and (311) planes, respectively. The peaks match the standard JCPDS data (05-0640) and confirm the structure of the pure CdO phase with a face-centered cubic crystal structure (Benhaliliba, et al. 2012;Chandiramouli and Jeyaprakash 2013;Sahare, et al. 2017). It can also be inferred from Figure 1 that the XRD peak intensity of the (111) plane is relatively higher than that of the other reflections, indicating that the CdO films have a preferential growth along the (111) direction (Dakhel 2014b;Elttayef, et al. 2013;Jadhav, et al. 2017;Sahare, et al. 2017;Sankarasubramanian, et al. 2014). Another interesting point is that Bragg's positions of the diffraction planes shift slightly due to Cu doping (Benhaliliba, et al. 2012;Sahare, et al. 2017). Otherwise, the results of XRD do not show any CdO, CuO, and metallic Cu structure in the films, suggesting that the incorporation of Cu ions in the CdO lattice does not affect its crystal structure . Typically, Cu doping in CdO thin films involves replacing the host Cd 2+ ions with Cu 2+ ions in the host CdO lattice. To ensure a successful replacement of the Cd 2+ ions with Cd 2+ ions, both elements must possess the same valance and a more or less equal ionic radius. However, the ionic radius of Cu 2+ (0.73 ) is smaller than that of Cd 2+ (0.95 ) . When Cu 2+ is introduced in the lattice site of CdO, it might be easy to diffuse in the host CdO material and easily replaces Cd 2+, causing small voids in the host lattice. This explains the slight shifts in reflections from the Cu doped CdO vis-à-vis the corresponding reflections from the pure CdO films (Benhaliliba, et al. 2012;Jadhav, et al. 2017;Sahare, et al. 2017). However, the absence of any secondary phase in the XRD patterns indicates that Cu 2+ ions are well assimilated in the CdO lattice.
The crystallite size (D s ) of the grains can be evaluated using the Scherrer's formula (Sahare, et al. 2017): where K is a constant (0.9), λ the wavelength of the X-ray, β broadening of the diffraction line measured at half of its maximum intensity (in radian), and θ Bragg's diffraction angle. Table 1 presents the variation of the measured grain size of the pure and Cu-doped CdO thin films.
The average crystallite size of the samples was determined by the Halder-Wagner method. In this method, the full width at half maximum of the physical profile can be written as follows: where, and are the full width at half maximum of the Lorentzian and Gaussian functions. that the advantage of this method is that it gives more weight to the peaks at low-and mid-angle ranges, where the overlapping of the diffracting peaks is significantly less. Now, according to the Halder-Wagner method, the relation between the crystallite size and lattice strain, is given as follows (Rebhi, et al. 2009): where ℎ * = ℎ cos ℎ and ℎ * = 2 sin . The plot of equation (2), with � ℎ * ℎ * 2 � terms along X-axis and � ℎ * ℎ * � 2 along Y-axis for each peak of the XRD pattern, is shown in Figure 2. The slope of the plotted straight line provides the average crystallite size, whereas the intercept gives the intrinsic strain of samples S 1 , S 2 , S 3 , S 4 , S 5 , and S 6 . The values of ε and D W-H are presented in detail in Table 1. As can be seen, there is an overall downward trend for the average crystallite size with an increment in Cu doping (Gupta, et al. 2011;Sahare, et al. 2017).
The dislocation density (δ), which gives direct information about the defects in the crystal structure, can be calculated using the following relation (Sahare, et al. 2017): The dislocation density is defined as the "length of dislocation lines per unit volume of the crystal", and higher values of δ indicate lower crystallinity levels for the films. The variation of the calculated δ value of the pure and Cu doped CdO thin films is shown in Figure 3. It can be seen that as the doping concentration in the CdO films increases, so, too, does the δ value (Sahare, et al. 2017).
Figure 4 represents field emission scanning electron microscopy (FE-SEM) images to provide information about the effect of copper doped on the surface morphology of the thin films. As can be seen, grains are almost uniformly distributed over the surface and they were uniformly and spherically shaped in nature without any cracks and porosity in any of the samples (Jadhav, et al. 2017). The average grain size was estimated using FE-SEM images, the results of which can be seen in Table 2. These results indicate that when the doped Cu increases, it causes a reduction in the grain size; therefore, the morphology of the samples changes considerably with Cu doped (Çolak and Türkoğlu 2013;Sahare, et al. 2017). In general, this agrees well with the XRD results. An enhanced surface uniformity and better substrate coverage were observed in the Cu-doped CdO films, which is very important for the devices (Sahare, et al. 2017).
The thickness of the thin films was estimated via the cross-sectional FE-SEM micrographs (Figure 4), as presented in The EDX technique was applied to study the elemental composition of the thin films. Figure 5 illustrates the typical EDX spectrum of the CdS thin film prepared with 0.1% (at.%) of Cu. The observed elemental peaks have confirmed the presence of Cd and Cu in a thin film.
The two-(2D) and three-dimensional (3D) surface morphology of the pure and Cu-doped CdO thin films obtained from the AFM analysis can be seen in Figure 6. The roughness and root mean square (RMS) values of the films on dimensions of a 5×5 µm 2 area are given in Table 2. The surface roughness of the CdO films decreases as the doping concentration of Cu increases (Gupta, et al. 2011;Sahare, et al. 2017;Sankarasubramanian, et al. 2018). The variation in surface roughness values indicates that the Cu doping affects the surface topography of the thin films.

Optical study
The optical properties of the prepared films were studied through transmittance and absorbance spectra. The optical absorption and transmittance spectra of the pure and Cu doped CdO thin films were recorded in the wavelength range of 450-1100 nm, which are shown in Figure 7 and Figure 8, respectively. Figure 7 shows the variation of absorption with wavelength. In general, the absorption of the thin films tends to decrease with the erasing value copper doped.
Moreover, Figure 8 shows that the optical transmittance of the CdO thin films improves with Cu doping, with all of them showing a sharp transmittance drop at the fundamental absorption band edge (λ ~500 nm); This figure also shows that, in general, it is blue-shifted for Cu doped (Benhaliliba, et al. 2012;Sahare, et al. 2017). This sharp edge corresponds to electron excitation from the valence band to the conduction band and is related to the nature and value of the optical bandgap. Hence, the CdO: Cu thin films could be used as a buffer material in solar cells (Sahare, et al. 2017).
The direct optical band gap is obtained through optical absorption measurements, plotting ( ℎ ) 2 versus photon energy ℎ ,and that Tauc's equation as follows: where is the absorption coefficient, is a constant, and is the direct optical band gap for the pure and Cu-doped CdO films, as shown in Figure 9. For all the samples, each curve of ( ℎ ) 2 versus ℎ exhibits an excellent straightline fit over a higher energy range above the absorption edge, suggesting a direct optical transition near the absorption edge. The obtained optical band gap values are listed in  (Elttayef, et al. 2013;Jadhav, et al. 2017). It is maybe due to an increase in the average crystallite size (Elttayef, et al. 2013). Besides, it might be attributed to the structural modification of the CdO films. The structural modification of the CdO: Cu thin films could be due to the Cd interstitial replacement or substitution with copper ions. This copper atom introduces some additional energy levels in the CdO bandgap close to the valance band edge. This might result in a reduction in transition energy (Jadhav, et al. 2017). The optical gap thin films with Cu dopant higher than 1% are greater than the pure CdO thin film (Benhaliliba, et al. 2012;Dakhel 2014b;Şahin 2019). These values show that the energy bandgap increases through doping which is usually attributed to the Moss-Burstein (B-M) effect (Benhaliliba, et al. 2012;Dakhel 2014a;Dakhel 2014b). Under the BM effect, optical band gap widening occurs as the Fermi level is raised into the conduction band of a degenerate nanostructured metal-oxide material (Şahin 2019). The effect of electron confinement in thin films can increase the bandgap energy by the decrease average crystallite size, and grain size (Vigil, et al. 2001). It can happen as a result of an increase in the amorphous nature, which in turn causes a decrease in the average crystallite size and grain size, according to XRD and FESEM results, respectively. That explains why, in this case, the rate of the electron-hole recombination becomes lower compared to the pure CdO thin film . Another possible reason for the broadening in the optical band gap energy via the doping process can be the impurity bands closer to the lower edge of the conduction band as a result of the Cd site in CdO, which are occupied by Cu atoms; these atoms have a strong orbital coupling with O atoms (Şahin 2019). Therefore, the valence electrons are excited to a higher energy level in the conduction band because they cannot occupy the states already occupied by the Pauli Exclusion Principle (Al-Ghamdi, et al. 2014;Manikandan, et al. 2017). It is possible, at high-level copper doping, for copper oxide in its amorphous structure to accumulate on grains and crystallite boundaries, increasing the values of transmittance and bandgap (Dakhel 2014b).

Electrical properties
The electrical properties of the pure and Cu doped CdO thin films were studied using the Hall measurement at room temperature in the van der Pauw configuration. The results are presented in Table 3. The variations of resistivity (ρ), carrier concentration (n), and mobility (µ) with the Cu dopant CdO thin films are also shown in Figure 10. The resistivity of either deposited pure films or Cu-doped films was found to range from 3.74×10 −4 to 1.01×10 −2 Ω.cm; our measured values of resistivity are in good agreement with another study (Sankarasubramanian, et al. 2018). In general, it can be noticed that the electrical resistivity of the films drastically increases as the Cu concentration in the CdO films increases (Dakhel 2014b;Sankarasubramanian, et al. 2018). The grain boundary scattering mechanism can contribute to the controlling of the mobility of the current carriers in the films; otherwise, the mobility is mainly controlled by impurity and lattice scattering mechanisms. According to the FE-SEM images, an increase in the Cu dopant can cause a reduction in the grain size. Therefore, the boundary grains increase, which in turn further scatters the free carriers away from the grain boundaries. Finally, resistance increases (Şahin 2019). It should be noted that the increasing resistivity by the Cu dopant CdO film can be attributed to the influence of scattering mechanisms like phonon and ionized impurity scattering (Sankarasubramanian, et al. 2018). In general, carrier concentration and mobility decreased, while electrical resistivity increased as a result of Cu doping. However, the variation of carrier concentration with Cu doping is small compared to the variation in other electrical parameters. This can be explained by considering that Cu dopant ions have an identical valency as that of Cd ions. Thus, Cu doping affects n by occupying Cd-ion vacancies and/or interstitial positions (Dakhel 2014b). The negative sign of the Hall coefficient strongly confirms the n-type conducting nature of the thin films. This suggests that the Cu atoms are placed into the CdO lattice sites and act as a donor by supplying free electrons when they occupy the sites of Cd 2+ (Sankarasubramanian, et al. 2018).
The thermoelectric e.m.f. of the CdO: Cu films was measured by applying a temperature gradient between the two ends of the samples. Figure 11 shows the variation of the thermoelectric emf versus the temperature difference in the thin films. The variation trend of the thermoelectric emf versus the temperature difference in the thin films was linear, which could be attributed to the increase in the carrier concentration and mobility of the charge carriers with rising temperatures (Sayas and Fadavieslam 2020). The negative slope in the thermo-EMF plots ( Figure 11) shows that the thin films are n-type semiconductors in nature. These results tie in nicely with the results of the Hall effect measurement. The thermo-EMF of all the presented data shows an increase in magnitude with respect to the temperature difference. This may be attributed to the rise in the mobility charge carrier with a rise in temperature (Anwar, et al. 2015a;Anwar, et al. 2015b). The thermoelectric emf depends on the location of the Fermi energy in the material, the type of the scattering mechanism, and the density of the charge carriers. It increases as the Fermi energy moves further into the energy gap from the bottom edge of the conduction band. It, therefore, inevitably follows that the smaller the carrier concentration, the larger the Seebeck coefficient (Fadavieslam, et al. 2016). For a better perspective, Sebeck coefficient � = ∆ � at the 100 C o temperature difference of all the samples is presented in Table   2. Decreased carrier density causes an increase in the Seebeck coefficient (Cappelli, et al. 2015).

Conclusion
In this paper, thin films of CdO: Cu with different contents of Cu% were prepared on soda-lime glass substrates, using the spray pyrolysis technique. Their structural, optical, and electrical properties were, then, studied. The XRD results show that the thin films exhibit a polycrystalline nature, and the incorporation of Cu ions in the CdO lattice does not change their crystal structure. The FE-SEM and AFM study shows that the spherically shaped grains and surface roughness decreased with an increase in the copper doping level. The optical properties of the thin films are sensitive to Cu doped. All the deposited films had a good transmittance and optical band gap for optoelectronic device applications, both of which increase with a Cu concentration in the CdO films. The electrical characteristics of the thin films show that all of them are n-type and their electrical resistivity increases as the doped Cu increases. % molar ratio in the thin films Fig. 11. Variation of the thermoelectric emf versus the temperature difference in samples S 1 , S 2 , S 3 , S 4 , S 5 , and S 6