1. Introduction

In the last few decades,ZnO films and devices have attracted the intention of many researchers. The ZnO film became an encouraging multifunctional material due to its large excitonic binding energy of 60 meV,an n-type wide direct band gap of 3.3 eV,and a high transparency in the visible and infrared spectra^[1]. The ZnO layers have been grown by several routes such as spray pyrolysis^[2],sputtering^[3] and sol-gel^[4]. Our films are prepared with the ultrasonic spray pyrolysis route. We selected this deposition route because it is facile,non-toxic,unobtrusive,inexpensive and has a low processing temperature method. In this work,the effect of substrate temperature and indium incorporation into the solution on the physical and surface properties of the sprayed ZnO films grown onto glass is investigated. The most important experimental techniques discussed here include the Hall-effect and the room temperature photoluminescence measurements,because they alone can provide donor and acceptor concentrations,and donor energies. To the best of our knowledge,rare works reported on the transmittance,surface morphology,Hall measurement,optical and detailed AFM parameters of pure and indium-doped ZnO thin layers fabricated by the ultrasonic spray process at several substrate temperatures.

2. Samples preparation and characterization

ZnO thin films were produced by a facile and low cost ultrasonic spray pyrolysis technique on corning glass substrates at various temperatures of 250-400 $^\circ C$. The 0.05 M of zinc acetate dehydrated (Zn(C$_{2}$H$_{3}$O$_{2})$$_{2}$$\cdot$2H$_{2}$O) was used as the starting material to spray ZnO thin films. The doping source was indium (3+) chloride (InCl$_{3})$,the ratio In/Zn was fixed at 1% in the solution. Both precursor and doping compound were dissolved in methanol at room temperature. The spray rate and the substrate to nozzle distance were kept respectively at 1 mL/min and 5 cm and the deposition time was 10 min. The glass substrate was heated within the 250-400 $^\circ C$ temperature range,which was controlled by a digital thermometer connected to the heater. The detailed technique is described in our previous work^[5]. The X-ray pattern of as-grown films was carried out with a PAN analytical 790 CACX'Pert PRO diffractometer and the optical parameters were measured using a UV-VIS-IR Shimadzu UV-3101PC double beam spectrophotometer. The atomic force microscope was the Quesant model 250 system having an (80 $\times$ 80 $\mu$m$^{2})$ head,in the wave mode in air. For the (3 $\times$ 3 $\mu$m$^{2})$ square images,the resolution was (300 $\times$ 300) pixels at a fixed scan rate of 2 Hz. The analyzed AFM parameters were determined using the WSXM system software. The room temperature photoluminescence analysis was carried out in an experimental setup consisting of a 325 nm,15 mW He-Cd laser (Kimmon type),a 0.85 m double monochromator (SPEX,model 1404),and a GaAs photon counting photomultiplier (Hamamatsu). The band explored was from 350 to 600 nm,in steps of 0.5 nm and at a speed of 0.2 s per measured point. Electrical measurements were accomplished by an MMR Hall and Van der Pauw measurement system.

3. Results and discussions

3.1 Structural properties of sprayed films

Figure 1(a) shows the X-ray diffraction spectra of as-synthesized films at different substrate temperatures in the 20$^\circ$-100$^\circ$ range of 2$\theta$ angle. The X-ray pattern shows that the films are polycrystalline in nature with a hexagonal wurtzite crystal structure. The most important planes are as follows: films have grown at 250 $^\circ C$,the main orientations,(100),(002),(101),(102),(110) and (103) are located respectively at 31.8$^\circ$,34.4$^\circ$,36.2$^\circ$,47.7$^\circ$,56.7$^\circ$ and 62.8$^\circ$ (2$\theta)$ angle set. While for the 300-350 $^\circ C$ temperature range,it is seen that all planes are formed except (100) which is very feeble. It is observed at 400 $^\circ C$ that the planes (100) and (110) are not well shaped. It is seen at 350 $^\circ C$ that the (112) orientation,peaked at 67.9$^\circ$,is well formed. It is manifested that $T_{\rm s}$ plays a drastically important role for obtaining a high-quality structure of ZnO film. The X-ray pattern confirms that the sprayed ZnO phase is grown along a (002) preferential plane,caused by the lower free surface energy of this plane. These results corroborate those reported in Reference ^[6]. We may conclude that $T_{\rm s}$ has influenced the crystalline structure and the quality of as-synthesized layers,which becomes better at $T_{\rm s}$ of 350 $^\circ C$. This obtained detail is in good agreement with that found by Benramache et al.^[7]. The change in $T_{\rm s}$ leads to a change in the kinetic of ZnO droplets (diameter of $\sim$40 $\mu$m) when they reach the heated substrate. The controlled substrate heat and the manner of how sprayed droplets are deposited and migrated on the substrate free surface could influence the crystalline structure. Indeed,this behavior is confirmed by XRD investigation. Since the rate and speed of sprayed droplets are fixed,substrate temperature exceeding 400 $^\circ C$ could expose sprayed droplets to thermal shock and even to a partial evaporation of solution,or sometimes the sprayed film has a whitish powder. This situation could disadvantage the formation of a good TCO candidate although some researchers have increased $T_{\rm s}$ up to 700 $^\circ C$^[8]. The effect of indium incorporation into the crystalline structure of ZnO at $T_{\rm s}$ of 350 $^\circ C$ is shown in Figure 1(b). The difference between Zn ($r_{\rm Zn(+2) }$ $=$ 0.074 nm) and In ($r_{\rm In(+3)}$ $=$ 0.080 nm) ionic radius modifies the crystalline structure of the host material. Thus indium ions incorporation improves the (100) plane and disadvantages the others orientations as indicated by the arrows in Figure 1(b). This change in crystalline structure is due to the migration of doping ions from the surface to the interstitials and/or vacancies sites.

3.2 Optical measurements

The transmittance of films presents a rapid increase from the UV band to the visible range with two successive oscillations,whose wavelengths are respectively $\lambda_{1}$,$\lambda_{2}$ as indicated by an arrow in Figure 2(a). In all samples,the transmission in the visible range and for longer wavelengths up to 2500 nm was included between 75% and 91%. In both layers produced at 250 and 300 $^\circ C$,the transmittance varies with the same profile in the VIS-IR band while the variation of transmittance is significantly different for the higher substrate temperatures (350 and 400 $^\circ C$) as indicated in Figure 2(a). The highest transmittance at a wavelength of 550 nm is obtained for the ZnO layers,which have grown at 400 $^\circ C$ ($>$ 84%). A similar transmittance curve was roughly observed for those prepared at 250 and 300 $^\circ C$. In visible and IR bands,the sample synthesized at 350 $^\circ C$ exhibited a lower transmittance. Thus,the transmittance is found to increase at high $T_{\rm s}$ and at 350 $^\circ C$ the doping reduces the transmittance (550 nm) a little,as listed in Table~1. The transmittance of IZO layers also presents two successive oscillations as indicated by two arrows in Figure 2(b). The transmittance of the films is increasingly dependent on $T_{\rm s}$ and ranged between 76% and 85% at the middle of the VIS band (550 nm). In-doping induces a slight change in variation of transmittance in both VIS and IR spectra as shown in Figure 2(b). The transmittance grew up rapidly from UV to the first pointed top in the visible range in the case of the doped sample. The obtained maxima of transmittance for ZnO and IZO are respectively 85% and 87%. A second broadened top of transmittance around 87% is detected in the visible band edge. Finally,the transmittance exhibits a discrepancy of about 4% between pure and In-doped ZnO in the infrared range. No significant decrease in transmittance is detected for longer wavelengths particularly above 1500 nm,so we assume that there is no interaction with the free electron plasma in the films; this effect is probably related to the increase in resistivity (≥50 $\Omega \cdot$cm). The absorption parameter is calculated using the following expression^[1],

$\begin{equation} \label{eq1} \alpha =-\frac{\ln T}{t}, \end{equation}$

(1)

where $T$ is the transmittance,and $t$ is the film's thickness. The following formula expresses the optical band gap^[1],

$\begin{equation} \label{eq2} (\alpha h\nu )^{2}=h\nu -E_{\rm g}. \end{equation}$

(2)

The optical bandgap ($E_{\rm g})$ of the film is determined by plotting ($\alpha h\nu)^{2}$ versus photon energy $h \nu$ and extrapolating the straight line portion of this plot to the $h \nu$ axis (not plotted here). The linear dependence of ($\alpha h\nu)^{2}$ to $h \nu$ signifies that pure and doped ZnO films are direct transition type semiconductors. The photon energy at the point where ($\alpha h \nu)^{2}$ is zero is $E_{\rm g}$. Consequently,the bandgap increases slightly with an increase in substrate temperature. As listed in Table 1,the obtained values of band gaps are ranged within 3.21-3.25 eV. Hence it is observed that the optical band gap shifts to the smaller values as $T_{\rm s}$ decreases. Comparable optical band gaps of 3.168,3.250 and 3.229 eV ranges have been found in literature respectively at 300,350 and 400 $^\circ C$ for the sprayed ZnO films^[7]. The films' thickness was obtained using the following formula^[9],

$\begin{equation} \label{eq3} t=\left[2\left(\dfrac{1}{\lambda _1 }-\dfrac{1}{\lambda _2 }\right)n\right]^{-1}, \end{equation}$

(3)

where $t$ is the film's thickness,$\lambda_{1}$ and $\lambda_{2}$ are the wavelengths of two consecutive maxima in the transmittance plot,and $n$ is the refractive index of ZnO found to be equal to 2^[9]. The obtained values are tabulated below (Table 1). Prasada et al. found similar values of the 265-335 nm range at (350 ＜ $T_{\rm s}$ ＜ 450 $^\circ C$)^[10] while Zahedi et al. obtained a thickness of film in the 550-2000 nm range^[6]. Table 1 summarizes the thickness of as-synthesized films at various substrate temperatures.

3.3 The (3 $\times$ 3 $\boldsymbol{\mu}$m$\boldsymbol{^{2}}$) area scanned AFM surface studies

The obtained images by AFM microscope are sketched in Figures 3 and 4. It is seen from the (3 $\times$ 3 $\mu$m$^{2})$ area scanned AFM surface images that the morphology of films is densely homogeneous and neither pores nor defects can be observed over the film surface. The as-grown films,presenting essentially different thicknesses,show different morphology of surface grains,which are dependent on the deposition temperature and In-doping. The three-dimensional analyzed images (3 $\times$ 3 $\mu$m$^{2})$ reveal a big density of broadened nano-grains which look like mountains,and have grown from the inner to the top along the $c$-direction with different heights ($z$ (nm)) and roughnesses. The root mean square (RMS) is expressed as follows^[11],

$\begin{equation} \label{eq4} {\rm RMS}=\sqrt {\frac{1}{N}\sum {(Z_{\rm m} -Z_{\rm i})^{2}} } , \end{equation}$

(4)

where $N$ is the number of deviations,$Z_{\rm i}$ is the height reference,and $Z_{\rm m}$ is the profile value. From the displayed values in Table 1,the film surfaces,produced at 400 $^\circ C$,are so smooth (RMS $>$ 17 nm),while In-doped samples are roughly smooth (RMS $\sim$ 23 nm) compared to the pure ZnO films (RMS $\sim$ 35 nm). Other papers have cited that the RMS value is the same order of magnitude for the ZnO films prepared at 700 $^\circ C$^[12]. Further,an average roughness below 10 nm was found by Zaleta et al. for the sprayed IZO films^[13]. It is observed in 2D-view that the film surface is formed by several micro/nano-domains,which have well boundaries as shown in the statistic particle image (two-dimensional view). The AFM scanned images demonstrate obviously the boundaries of nano-grains with different heights $z$ (nm) as tabulated below. Based on the X-ray pattern,the sprayed films are preferentially oriented along the $c$-axis. This latter is the ZnO (002) plane,which possesses a minimum surface energy. This result is in good agreement with AFM observations,which confirm roughly the pointed-type architecture of nano-grains. Gaikwad et al. reported a similar morphology of pure and boron-doped ZnO films prepared by a spray pyrolysis route^[14]. In general,the top of the grains are more broadened but become narrower at 400 $^\circ C$ and the grains have grown with the same shape. The statistical distribution of the data for pure and In-doped ZnO is respectively shown below the scanned AFM images. The 2D topography describes domains which look like small islands with sharp borders and the $z$ value is ranged within 34 to 82 nm.

3.4 Photoluminescence investigation

Room temperature photoluminescence (PL) spectra of the sprayed ZnO and IZO,is achieved with a He-Cd laser (15 mW) of wavelength 325 nm,as shown in Figure 5. Strong emissions,located at 2.11 and 2.8 eV,are observed as sketched in Figure 5(a). A less significant emission situated around 2.3 eV was also exhibited,corresponding to green luminescence,which might be emitted from defect centers. The first peak in PL,peaked at 2.81 eV,spectra corresponds to band to band transition and the emissions,ranged within 440-500 nm,are showing blue luminescence as can be seen in Figure 5(b). According to Fang's paper,this blue luminescence center ($\sim$ 462 nm) is ascribed to the transition of electrons from the defect level of Zn interstitial atoms to the top level of the valence band^[15],whereas Wang explained this behavior of blue luminescence of undoped ZnO as a response of oxygen vacancies,which could produce two defect donor levels (shallow donor level of energy of 2.8 eV)^[16].

As seen from the PL plots of the as-grown ZnO films,the intensity peak is observed at 2.11 eV ($\sim$587 nm),which shows yellow luminescence. In-doping slightly shifts the blue and yellow emissions to the shorter wavelength and further,it reduces the intensity of the peak located at 2.108 eV (see Figure 5(b)). It is well-known that the ZnO band gap ($E_{\rm g})$ is around 3 eV,these emissions whose energies are less than the obtained $E_{\rm g}$ values reveal the occurrence of defects and impurities levels in our synthesized films. Tomakin reported that Tauc's relation supposes free carrier behavior with no electron-hole interactions,while in PL spectra,the important electron-hole correlation is present due to the large excitonic binding energy of 60 meV of ZnO^{[17, 18, 19]}.

3.5 Electrical measurements

The electrical resistivity,bulk density and mobility of charge carriers at room temperature are determined by the Hall effect system measurements. The linear dependence plot of resistivity versus deposition temperature is depicted in Figure 6. The obtained values of electron bulk density ($n$),mobility ($\mu )$ and resistivity ($\rho )$ are listed in Table 1. In-doping reduces considerably the resistivity from 385 to 8.35 $\Omega \cdot$cm; this decrease in resistivity might be due to the interstitial sites of indium ions. However,the mobility is 0.26 cm$^{2}$/(V$\cdot$s) for undoped ZnO and increased up to 1.56 cm$^{2}$/(V$\cdot$s) by In-doping. It is obvious that the doped sample exhibits the highest mobility and lower bulk density of electrons. Based on the Hall measurement,ZnO and IZO are n-type semiconductors; similar results have already been cited^[20]. Here,the largest carrier concentration ($n$ $>$ 20 $\times$ 10$^{16}$ cm$^{-3})$ is obtained for pure ZnO at 250 $^\circ C$ and ($n$ $>$ 13 $\times$ 10$^{16}$ cm$^{-3})$ for IZO. The background carrier concentration differs a lot according to the quality of the films and to their deposition techniques but is frequently around 10$^{16}$ cm$^{-3}$. The highest reported n-type doping is $\sim$ 10$^{20}$ electrons cm$^{-3}$ and the highest reported p-type doping is $\sim$ 10$^{18}$-10$^{19}$ holes cm$^{-3}$ as reported in References ^{[21, 22]}. Furthermore,the sheet electron concentration $n_{\rm s}$ (10$^{11}$ cm$^{-2})$ and resistance $R_{\rm s}$ (10$^{6}$ $\Omega \cdot$cm$^{-2})$ are also investigated here,$n_{\rm s}$ goes from 1 to 24 and $R_{\rm s}$ increases from 4.5 to 40. We suppose that a decrease in $n_{\rm s}$ and an increase in $R_{\rm s}$ could be associated with the increasing temperature of the substrate,but the doping changes $n_{\rm s}$ up to 15 and $R_{\rm s}$ to 0.8 as shown in Table 1. In comparison to our results,Prasada et al. reported high resistive films,(350-450 $^\circ C$,$\rho$ $\sim$ 6 $\times$ 10$^{3}$ $\Omega\cdot$cm),which exhibit lower electron concentration ($n$ $\sim$ 7 $\times$ 10$^{15}$ cm$^{-3})$ and the lowest mobility ($\mu$ $\sim$~0.34 cm$^{2}$/(V$\cdot$s))^[10].

4. Conclusion

Nanostructures of ZnO and IZO are produced by an easy and low cost ultrasonic spray pyrolysis technique at various temperatures. X-ray exhibited the polycrystalline structure of films which grew along the (002) orientation. The sprayed films are highly transparent and wide band gap (3.21-3.31 eV) semiconductors. The smooth micro and nanosized structures are revealed by AFM micrographs (RMS ＜ 40 nm). Strong blue and yellow luminescences (2.11 and 2.80 eV) emitted from sprayed ZnO and IZO are observed. The linear dependence of resistivity on the substrate temperature of undoped ZnO films is confirmed and In-doping reduces considerably the resistivity from 384 to 8 $\Omega \cdot$cm.

Acknowledgments

This research is a part of the CNEPRU project N$^\circ$ D01920120039 supported by the ministry of high teaching and scientific research MESRS www.mesrs.dz. The first author gratefully acknowledges the Mexican and Turkish laboratory staff for their helpful contributions.