1 Introduction

Zinc oxide semiconductor films have been very useful, particularly in the field of spintronics, because of their wide band gap (Eg = 3.31 eV), exciton band energy (60 meV) at room temperature [1] besides its significant transmittance value of 0.9 in the visible region. ZnO crystallizes within a wurtzite-type structure defined by a hexagonal lattice where Zn2+ ions occupy the tetrahedral sites formed by O2− ions. Numerous researches have already fabricated ZnO films using various techniques, including spray pyrolysis [2], reactive magnetron sputtering [3, 4], chemical vapor deposition (CVD) [5], sol–gel [6] molecular beam epitaxy [7], and pulsed laser deposition [8, 9].

Recently, the doping of semiconductors by transition metals is of great importance, because of their unusual optical properties and promising potential for optoelectronic applications [10,11,12,13]. The doping of ZnO with metal ions has been found to influence its structure and particularly its physical properties such as particle size, dislocation density, bond length, strain deformation, etc.[14, 15].

Copper (Cu) element is of particular interest, thanks to its low toxicity and large abundance [9, 16]. The incorporation of Cu within ZnO host lattice has been found to enhance its physical and chemical properties, such as photocatalysis activity, gas sensitivity, magnetic susceptibility, optical and electrical properties [14, 17].

In this work, thin films of pure and Cu doped ZnO films are deposited by ultrasonic spray pyrolysis (USP) technique onto glass substrates (30 × 12 × 1.2 mm3) at 450 °C for 30 min. The effects of the doping on structure, microstructure and optical properties are discussed.

2 Experimental part

2.1 Film preparation

In this study, the use of ultrasonic pyrolytic spray technique for the deposition of thin films of ZnO, Zn1−xCuxO [x = 2, 5, 10, 15 and 20 at%], offers various benefits, such as better homogeneity, low-power processing, easy-to-control stoichiometry, and large surface covering. Zinc acetate [Zn(CH3COO)2·2H2O)] (Fulka 99.9%) (0.01 M) and copper chloride (CuCl2·2H2O) (2–20) (Cu, at%) as precursors are dissolved in 20 ml methanol (Merck 99.5%), 30 ml ethanol (Merck 99.5%). This mixture has been chosen for its easy processing, fluidity, fast transformation, and precursor fog in steam. 50 ml deionized water (resistivity = 18.2 MΩ cm) has been added as oxidizing agent. A small amount of acetic acid is added to the aqueous solution for adjusting the pH to about 4.8 to prevent the formation of hydroxides. This solution containing various constituents is then deposited after being sprayed in fine drops, by an atomizer using a 40 kHz ultrasonic generator with the power of 130 W. This system makes it possible to transform the solution into a jet of very small droplets, i.e. fine drop of 45 μm for a frequency of 40 kHz. The remaining experimental conditions are kept fixed: the distance separating the substrate from the end of the pipe 5 cm, the solution stream rate 0.1 ml/min, surface of glass substrate (30 × 12 × 1.2 mm3), deposition temperature 450 °C, and deposition time 30 min. The temperature of the substrate has been monitored by a thermocouple and electronically registered [18].

2.2 Characterization techniques

The structure of the films is characterized by X-ray diffraction (Rigaku Ultima IV equipped with Cu–Kα radiation, 0.15418 nm). The morphology of the films was checked by scanning electron microscopy (Quanta™ 250 FEG-SEM from FEI company) equipped with energy dispersive X-ray spectroscopy (EDS) for chemical analysis. The optical properties are determined by measuring the transmittance of the films using a SpectroScan 80D spectrophotometer UV–vis in the 190–1100 nm spectral range.

3 Results and discussion

3.1 Structure and microstructure analysis

The XRD spectra of pure and copper-doped ZnO (2, 5, 10, 15 and 20 at%) are illustrated in Fig. 1, which shows that the films obtained are zinc oxide (ZnO). We find almost all the typical diffraction peaks of powdered ZnO which crystallizes in the wurtzite structure with a preferential grains orientation to the c axis which is perpendicular to the surface of the substrate. This information was obtained using reference code (01-097-0207) and was easily indexed in hexagonal wurtzite phase with P63mc space group and JCPDS card (36-1451) [19]. These diffraction peaks corresponding to (100), (002), (101), (102), (110), (103) and (112) planes of ZnO lattice are present in all spectra. No secondary phase corresponding to copper oxide (CuO) is observed within doped films (xCu ≤ 0.10). For the 15 and 20 at% doped films we noticed that a new peak emerges respectively at 38.47° and 38.60° corresponds to (111) that belongs to the CuO phase (JCPDS#05-0661) [20, 21], which are in agreement with the literature showing the presence of a wurtzite type phase for a copper doping rate lower by 15 at%, while those with a higher copper content include a tenorite type oxide phase (ZnxCu1−xO) [22]. Which means the Cu-doping concentration limit in ZnO is around xCu = 0.15 in our experiment settings.

Fig. 1
figure 1

XRD patterns of ZnO and Zn 1−xCuxO (with x = 2, 5, 10, 15 and 20 at%) films

Full size image

Compared to the undoped ZnO, the intensity and the positions of the diffraction peaks vary considerably and non-homogeneously following the increase in the doping level illustrated by the decrease in the intensity of the peaks in the doped ZnO, and we note as well that the peak (002) has been shifted to a higher value of 34.26°–34.29°. This shift shows the substitution of Zn by Cu in the hexagonal lattice prompting the decrease in the crystallinity of ZnO [23,24,25]. On the other hand, the intensity of the peak (101) increases proportionally with the doping confirming the substitution of Zn2+ by Cu2+ with the preservation of the preferred orientation in the hexagonal structure of copper-doped ZnO films perhaps because of the ionic radius difference [26,27,28,29]. Moreover, at the doping rate (20 at.%), the presence of impurity in the turn of 35.39° was noticed, which explains, firstly, that the Cu would not be more interstitial but more likely to be fused in clusters of metallic copper and, secondly, interprets the dissuasive effect on the peak (002), but it remains energetically a favorable factor [30,31,32]. Which confers that the interstitial substitution of Cu2+ influence on the concentration of interstitial Zn, oxygen and Zn vacancies [27, 33]. It is suggested that the incorporation of Cu2+ into the ZnO structure is done regularly if enters octahedral interstitial sites since tetrahedral coordination is not favorable for this species [22, 34].

For a doping rate < 15 at% we observe the presence of a single-phase Zn1-xCuxO material having the wurtzite (hexagonal) structure of ZnO with a linear variation of the lattice parameters “a” and “c”, then decreases for 15 at% and 20 at% with the presence of additional peaks corresponding to Cu, CuO or any other impurity in X-ray diffraction patterns at 2θ = 38.47° and 38.60° for 15 at% and 20 at%, respectively. This indicates that the Cu atoms were scattered in the ZnO and substituted Zn, thus increasing the lattice parameter (Table 1) [35]. This modification of lattice parameters confirming that the substitution in the tetrahedron coordinate is caused by the ionic radius of Cu2+ (ionic radius = 0.057 nm) which is smaller than the Zn2+ sites (ionic radius = 0.060 nm) [27, 36].

Table 1 X-ray diffraction data (2θ, FWHM, d-spacing, cell parameters ‘a’ and ‘c’, c/a ratio, average crystallite size (D) and micro-strain (ε) of different Zn1−xCuxO thin films)
Full size table

The real distance between two Zn atoms in the studied systems is about 0.326 nm, while the Zn atoms in ZnO are present in tetrahedral structures with Zn–O distances of 0.196 nm. For uniformly distributed Cu ions, the mean distance between Cu ions that have been substituted into Zn sites within ZnO crystal lattice can be estimated via [37] Nat = (4/3) (Z/VC) πr3, where r is the average radius of an atomic sphere and Nat is the number of atoms in the sphere. Figure 2 shows the number of zinc atoms in a sphere with a radius of r. The structural parameters for Zn0.98Cu0.02O are: a = 0.3254 nm, c = 0.5212 nm; volume of unit cell (VC) = 47.79 × 10–3 nm3; Z = 2. For zinc atoms in a wurtzite structure of Zn0.98Cu0.02O, Z/VC = 41.84 nm−3. In Zn0.98Cu0.02O film, 2% of zinc sites are occupied by Cu ions. Therefore, the 50th zinc site from the probe atom is occupied by Cu. Using the above calculation, the average distance between Cu ions is estimated to be around 0.65 nm, which means that the nearest Cu ion to a probe Cu ion is located in the next unit cell. On the other hand, In the Zn0.80Cu0.20O film, the fifth zinc site from the probe atom is occupied by Cu, the mean distance of Cu–Cu pairs is 0.30 nm. This is consistent with the XRD results.

Fig. 2
figure 2

The number of zinc atoms in a sphere with a radius of r for wurtzite ZnO

Full size image

The estimation of the micro-stress (ε) using the (Eq. (1)) [38]:

$$ \varepsilon = \frac{\beta \cos (\theta )}{4} $$
(1)

where β is the angular peak width at half maximum in radian along (002) plane, and θ is Bragg’s diffraction angle. The micro-stress of the Cu-doped sample increases at xCu = 15 at%, then reaches the minimum when xCu = 0.20 at%, the atoms trapped in the non-equilibrium position could shift to a greater positional equilibrium [27].

The parameters of the microstructure are calculated using the Scherrer formula to estimate the average crystallite size (D) (Eq. (2)) [39]:

$$ D = \frac{{0.9{\kern 1pt} \lambda }}{\beta \cos (\theta )} $$
(2)

where λ is the wavelength of the incident X-ray (Cu Kα radiation, λ = 1.5418 Å), β is the total half-height width (FWHM) of the peak XRD and θ is the maximum of the Bragg diffraction peak (in radians). The calculation (the results are reported in Table 2) shows that the crystallite size values are increasing for a doping level (2, 5 and 10 at%) then it goes down at (15 and 20 at%). This decrease can be interpreted by the segregation of Cu atoms in the grain boundaries due to the increase in surface energy density. The crystallite sizes were in the 18–40 nm range; these results are in agreement with those found in the literature [22, 27, 40].

Table 2 Composition and stoichiometry of the thin films obtained by statistical analysis of EDS spectra
Full size table

SEM images sample pure and copper-doped ZnO (2–20 at%) are shown in Fig. 3. For pure ZnO and Zn1−xCuxO films (x < 15 at%) having a homogeneous surface morphology with the particles containing the mixer of spheroid-like and rod-like particles with diameters ranging from 20 to 100 nm, confirmed the dominance of the wurtzite phase. For a Cu doping rate (15 at% and 20 at%) reveal a net change of the surface with cylindrical shape and size [22], varying between 90 and 100 nm, with an average particle size of 43 nm due to the presence of defects and deformed crystalline structures due to the substitution of Cu [27].

Fig. 3
figure 3

High–resolution SEM images of a pure ZnO film and b–f Cu–doped Zn1−xCuxO thin films. The images b–f correspond to the Cu atomic content of 2%, 5%, 10%, 15% and 20%, respectively

Full size image

The results obtained from the energy-dispersive X-ray spectroscopy are given in Fig. 4. The EDX spectra of copper-doped ZnO show the coexistence of Zn, Cu, O with the presence of foreign elements in Si, Ca, Mg synthesized deposits appearing as minor peaks belonging to the soda-lime glass substrate. The latter varies from one sample to another because of the variation in the thickness of the Cu-doped ZnO films [41]. The height of the Cu peak increases with the doping rate, which shows that Cu has been successfully doped with ZnO [18, 27].

Fig. 4
figure 4

EDX elemental composition analysis of a pure ZnO film and b–f Cu–doped Zn1−xCuxO thin films. The images b–f correspond to the Cu atomic content of 2%, 5%, 10%, 15% and 20%, respectively

Full size image

Suppose stoichiometric films of the form Zn1−xCuxO. The stoichiometric of the films is measured exclusively from the Cu and Zn signals. According to the following relation (Eq. (3)):

$$ x \, = \, r{/}\left( {1 + \, r} \right), $$
(3)

where x represents the Cu atom content and r the ratio of the Zn and Cu EDS signals (see Table 2). The incorporation of the copper in the high network is very clear, for all the doping levels, in Fig. 5, which presents the ratio of the EDS signals of Cu as a function of the calculation of the atomic content of Cu, which shows the doping efficiency with the copper.

Fig. 5
figure 5

The calculated Cu atomic content (from EDS analysis) plotted as function of the expected Cu content/doping

Full size image

3.2 Optical properties

Figure 6 shows the transmittance spectra for Cu doped ZnO thin films with different Cu concentration. The spectra of Zn1−xCuxO films are taken at room temperature to study their optical properties. The dispersion parameters of the films were evaluated using a single-effective-oscillator fit, proposed by Wemple and DiDomenico [18, 42]. The solid curves in Fig. 6 correspond to the curve fitting and the symbols represent the experimental data. The figure reveal a reasonably good fitting to the experimental data, the values of d, Eg, n at 598 nm, and n∞ extracted by fitting [18] the experimental data are listed in Table 3. The transmittance shows a decreasing tendency with increasing Cu concentration. The average value of transmittance decreases from 75 to 5% as the dopant concentration increases from 5–20%. This can be attributed to the incorporation of Cu2+ ions, which replaces the substitutional or interstitial Zn2+ ions from the host lattice. The dopant ion (Cu2+) transforms the colourless host lattice (ZnO) into brown. In the visible region when the defects increase the absorption takes maximum values and the transmittance decreases [43].

Fig. 6
figure 6

Transmission spectra of ZnO and Zn1−xCuxO (with x = 2, 5, 10, 15 and 20 at%) and transmittance of the bare glass substrate. Measured (full circles) and calculated (solid lines) transmittance spectra of films

Full size image
Table 3 Dispersion parameters of the films extracted by fitting the experimental data [18]
Full size table

Figure 7 presents the calculated refractive indices [18] of Zn1−xCuxO films. It is worthy to note that the refractive indices of the films increases with increase in Cu dopant concentration. The refractive index is sensitive to structural defects (e.g. voids, dopants, inclusions) [18]. Therefore, this change may be due to the effect of Cu incorporation [44] which increases the impurities in the host material.

Fig. 7
figure 7

Refractive index of ZnO and Zn1−xCuxO (with x = 2, 5, 10, 15 and 20 at%) films

Full size image

Figure 8 shows the energy band gap and refractive index as function of the Cu concentration. It is interesting to note that the band gap of the films decreases with increase in Cu dopant concentration. The Cu 3d orbital is much shallower that the Zn 3d orbital. When a Cu atom occupies a Zn site in ZnO, it introduces two main effects: (1) the strong d-p coupling between Cu and O moves O 2p up, which narrows the direct fundamental band gap and (2) the Cu 3d orbital creates impurity bands above the ZnO valance band. As a result, this exchange interaction give rise to a negative and a positive correction to the energy of the conduction band and valence band respectively. These facts may be related to the red shift of Eg values by Cu-doping in ZnO thin films. The transitions from the impurity bands to conduction bands are usually a mixture of direct and indirect. Thus, for heavily doped ZnO:Cu thin films, their optical band gap contains two parts—the direct fundamental band gap and the mixed impurity band gap [45]. The measured optical band gap for pure ZnO films was 3.24 eV, which is consistent with the results reported elsewhere [46]. The direct optical band gaps measured for Zn1-xCuxO films gradually decreased from 3.25 to 1.90 eV with the increase of the Cu concentration, as shown in Fig. 8. These results are in agreement with the theoretical calculations based on the functional density theory (DFT) which shows the reduction of the Cu-doped ZnO optical band gap [47]. Finally, the refractive index of Zn1−xCuxO thin films increases with decreasing energy band gap.

Fig. 8
figure 8

Energy band gap and refractive index as function of the Cu concentration

Full size image

4 Conclusion

Cu-doped ZnO thin films were deposited on glass substrates by ultrasonic spray pyrolysis. The XRD spectra of Zn1−xCuxO indicate that the films are polycrystalline in nature. The Cu doping into ZnO solid solution is found to influence considerably the film morphology, the grain size, and the stoichiometry of the oxide. The doping effectiveness is revealed by EDX analysis of the chemical composition of the films and it shows good agreement between the expected and measured Cu atomic content. The transmittance shows a decreasing tendency with increasing Cu concentration. The refractive indices increase, whereas the values of forbidden energy gap decrease with the increase in Cu dopant concentration.