Indium tin oxide is widely used as a transparent conductor in solar cells, liquid crystal displays and so on. Transparent conductors are an essential technology to realize optoelectronic devices that are vital for the contemporary smart society. Indium is a kind of supply-constrained element, which is a precious and highly valuable element. Nowadays, ZnO has attracted much attention as a substitute for indium tin oxide. The advantage of the material is its low cost and environmentally friendly aspect at the same time. A full description of the properties can be found in the literature [1]. ZnO is basically a semiconductor with a band gap of 3.44 eV; the large band gap results in a transparent property for visible light. It also holds interesting properties such as large exciton binding energy, large piezoelectric constants, and strong non-linear optical coefficients as well. When Al is doped in ZnO, the electric conductivity is drastically enhanced to be a conductor.

Many structures of ZnO have been reported such as nanowall [2], nanoparticles [3] and microrods [4]. Even flower-shaped ZnO nanorods were fabricated with graphene oxide coating [5]. The ZnO:Al is expected to be a candidate for the substitution of indium tin oxide. It is then suitable as a material for window layers for solar cells [6]. The ZnO:Al system has also been widely investigated: heterojunctions with Si [7], homojunction of ZnO [8], and so on.

ZnO films are well deposited by means of sputtering methods that are widely used in semiconductor device fabrication. This method, however, incurs high cost and long lime because it requires a vacuum system. For the purpose of minimizing the cost and time for fabrication, a metalorganic decomposition method was applied in this article. This method is so simple and attractive; the starting liquid is dropped on a substrate, spread out and then annealed under the atmosphere. The method is a proper one for oxide materials such as TiO₂ [9], LaNiO₃ [10], BiVO₄ [11], Bi₄Ti₃O₁₂ [12] and so on. ZnO has not yet been well grown by means of the MOD method. Only a few studies have been reported [13].

The problem is the parameter control such as annealing temperature and time. Another problem is the low reproducibility owing to the use of liquid. So it is a good challenge to reveal the effects of conditions for the ZnO film growth by means of the MOD method.

In this article, only undoped ZnO films were grown as follows. Starting materials were a commercially available MOD liquid for ZnO (SYM-ZN20 by Kojundo Chemical Lboratory Co., Ltd) and a dilution solvent (290353 by Kojundo Chemical Lboratory Co., Ltd); 200 μl each. The solution was mixed well with the help of an ultrasonic cleaner for 5 minutes. The mixed solution of 40 μl was dropped onto a borosilicate glass (S1225 of Matsunami Glass Industry, Ltd.) that was cleansed with an ultrasonic cleaner for 5 minutes in distilled water and then dried by contact with ethanol vapor. Dimension of the glass was 25 × 75 mm. The mixed solution was spread by means of scanning a bar coater on the substrate. The scanning was performed for several seconds at a constant velocity. The nominal value of film thickness was 27 μm. The sample was dried on a hot plated at 120˚C for 5 minutes before annealing at 600˚C in an electric furnace for 40 minutes under the atmosphere environment. There was no ramp region for increasing temperature. During the annealing the sample was installed on a ceramic long bar so that the center part of the glass substrate was only contacted by a bar (sample A). Another sample (sample B) was installed on a ceramic plate where the whole face was contacted to the plate. There exists possibly the difference in the heat conduction to the glass substrate.

It was also performed an experiment how annealing temperature affects film area. The temperature was varied from 200˚C to 600˚C where annealing time was fixed at 50 minutes. All the other conditions were the same as sample B.

Scanning electron microscope (SEM) photos were taken with the microscope of JSM-6510LA (EOL Ltd. Japan). The acceleration voltage was 20 kV.

Figure 1 shows a photo of sample A, which consisted of two parts: a transparent region in the center and two line-shaped white lines in the upper and lower regions. The sample was transparent before the annealing. The annealing definitely brought a color change. Just after the sample was taken out of the furnace, the white parts looked lightly yellowish and then became white as it was cooled down naturally to room temperature. On the other hand, the transparent part at the center did not change during the cooling.

Figure 1.An optical photo taken from sample A. Two parts were observed: transparent part (center) and white parts (the upper and lower).

Figure 2.SEM photos from two parts of sample A: the edge part (left) and the center (right).

Figure 3.The photo of sample B. white lines exist randomly on the substrate.

The two different parts were separately observed by the SEM. Figure 2 shows the two SEM photos from the two parts. The left one is from the edge part and the right one from the center. By comparison between the two photos, both films have a porous structure and the grain size is bigger at the edge part than that at the center.

The photo of sample B is shown in Figure 3. The photo shows that the white part is not restricted to the edge of the glass substrate but is randomly situated on the plate. The especially bright white vertical line in the left area is the reflection of the fluorescent lamp on the ceiling of our laboratory. The white line was not apparently well controlled.

Figure 4.The largest area decreases with the annealing temperature.

The two photos of Figures 1-3 provide interesting differences, though they cannot be estimated quantitatively. It was also confirmed that the Zn was detected more in the white part by the energy dispersive x-ray spectrometry using the SEM. We discuss here the difference between the two cases. As for sample A (Figure 1), the edge part was not in contact with the stage bar. It follows that the temperature at the edge was lower than that at the center because heat flowed directly from the stage bar at the center. This temperature difference caused a Marangoni convection. The film spread on the glass apparently moved from the center to the edge. On the other hand, sample B (Figure 3) provided a random white line that suggests a Bénard-convection-like appearance instead. The Bénard convection usually makes a systematic pattern like a honeycomb. But sample B does not show the one apparently because the film was too viscous to form such a systematic pattern. From these two cases it is expected that the elaborate control of temperature dependent on the site of the glass realizes the better formation of the films although the liquid temperature control cannot be well performed even nowadays.

Next described is the dependence of a film area on annealing temperature. The annealed film often consisted in an island-like form. The area was determined by measuring the largest one on the glass. The result is shown in Figure 4. The support condition of the substrates was the same as sample B. The area decreases with temperature until 500˚C. As the temperature increases, the convection becomes active. The repellency of the liquid becomes noticeable and then each island shrinks in dimension. Based on the result, the lower temperature is desirable. From the chemical point of view, a high enough temperature is needed for the decomposition of the starting material; this creates a conflict between the two. An idea for surface tension control for hydrophilicity is anticipated. The area slightly increases at 600˚C, but this is considered to be a kind of experimental error.

Undoped ZnO films were deposited on glass substrates by means of the MOD method. It was found that the heating condition during annealing greatly affects the film structure. The results are well understood by considering the convection of the starting liquid. The obtained films comprise an archipelago structure. The dimension of the maximum island decreases as the annealing temperature increases till 500˚C. An idea of reducing surface tension for hydrophilic surfaces will be needed to improve the fabrication by this method.