1. Introduction

Transparent conducting oxides have attracted much attention in the growing market of flat panel displays and thin film solar cells^[1]. Zinc oxide (ZnO) is considered a promising material for a variety of electronic and optoelectronic applications, including blue and UV solid-state lighting devices^[2]. ZnO is a kind of non toxic^{[3, 4]}, low cost^[5] and abundant^{[3, 6]} material. With a wide direct bandgap of 3.37 eV^{[5, 7]} it shows an excellent transparency for the entire visible spectrum^[6]. In addition, ZnO devices are thermally stable^[3], and do not suffer from dislocation degradation during operation^[8]. It was found that ZnO thin films had a multitude of immensely important applications in electronic and optoelectronic devices such as transparent conductors, solar cell windows, gas sensors, surface acoustic wave (SAW) devices, and heat mirrors etc^[3]. ZnO thin films can be prepared by different methods such as chemical solution deposition (CSD)^[9], thermal evaporation^[10], DC magnetron sputtering^[11], spin coating^[12], and the spray pyrolysis technique (SP)^{[13, 14, 15]}. The spray pyrolysis used in this paper is a simple and low cost technique, which can intentionally and simply add the dopant into the precursor solution.

The electrical conductivity of ZnO is controlled by intrinsic defects, i.e., oxygen vacancies and/or zinc interstitials, which act as n-type donors^[1]. It can be doped both by n- and p-type^[9]. To get n-type ZnO, doping can be done with group III elements such as B, Al, Ga, or In. In annealing experiments up to 400 C, the conduction due to intrinsic defects is thermally unstable, whereas a better thermal stability is observed on layers doped extrinsically by indium or aluminum^[1].

Measurements of the electrical properties of a semiconductor or an insulator are strongly affected by the contacts used. Hence, it is very important for choosing the suitable material and geometry of the contacts before measurements. The linear current-voltage characteristic, particularly at low voltages, is a criterion for the perfection of the contacts^{[13, 15, 16]}. Ohmic contact resistance is strongly dependent on doping concentration because the tunneling process is more dominant at high doping levels^[17]. The fabrication of ohmic contacts frequently includes a high temperature step, so that the deposited metals can either alloy with the semiconductor, or the high-temperature anneals reduce the unintentional barrier at the interface^{[13, 15, 18]}.

The literature reveals that different types of contacts were used for ZnO. Some researchers used a single element to form ohmic contacts, such as gold^{[1, 6, 19]}, aluminum^{[9, 20]}, indium^[11], silver^{[3, 21]} and Ti^[9]. Among them, Ti was found to be a poor ohmic contact. Others used alloys for the same purpose, for example Al/Ti/Au^{[22, 23]}, and/or Ti/Au^[20] contacts, which have the lowest contact resistivity between Ti, Al and Ti-Au.

Aluminum, silver and copper are used to make the contacts of ZnO thin films in this paper, where all of them are found to be ohmic. Our aim is to make a comparison between these ohmic contacts for ZnO thin films, and to find the best one in the electrical properties. The importance of this study lies in the use of ZnO as a transparent conducting oxide or a window material in heterojunction solar cells. For this purpose, it is important to investigate the electrical properties of this material, and good contacts are necessary for getting confidential results.

2. Materials and methods

To produce ZnO thin films, a precursor solution is prepared by using 7.61 $\times$ 10$^{-3}$ mole of ZnCl$_{2}$, dissolved in 300 mL of distilled water. To prevent the formation of Zn(OH)$_{2}$, the PH of the solution was made equal to 3 by adding HCl. This solution was intermittently sprayed by using a home-made spray pyrolysis apparatus that was described in Reference ^[14]. The spraying was done vertically through a PVC nozzle on ordinary glass substrates placed on a heating block. The dimensions of the substrates are 2.5 $\times$ 6 $\times$ 0.1 cm$^{3}$ and the substrate temperature is $T_{\rm s}$ $=$ 450 C. Prior to deposition, the glass slides were cleaned by dipping in distilled water then by an ultrasonic cleaner, then soaked in methanol for 20-30 min, again soaked in distilled water and finally polished with lens papers. Nitrogen N$_{2}$ was used as the carrier gas. The solution spray rate was in the range of 3-5 mL/min. The optimum carrier gas pressure for this rate of solution flow was around 5 kg/cm$^{3}$. The nozzle to substrate distance was adjusted to get the largest uniform area about 30 cm in diameter.

The $I$-$V$ measurements are taken in the dark and in room light at room temperature by a system that consists of a Keithley 2400 Source Meter capable of measuring 10$^{-11}$ A. Aluminum, silver, and copper are used as the contact materials for the films. Two strips of each of these metals are deposited on the surface of the film by thermal evaporation. The strips are of 1 cm length, 2 mm width, with a separation of 3 mm. A number of samples were annealed in nitrogen ambient at 250 C for 45 min after depositing the contacts on the surfaces of the films.

The transmittance of the films was measured in the wavelength range $\lambda$ $=$ 290-1100 nm by using a double beam Shimadzu UV 1601 (PC) spectrophotometer with respect to a piece of glass similar to the substrates. The minima and maxima in the transmittance curves are used to estimate the film thickness, where films of thickness between 300 and 600 nm are produced. The morphology of the films is observed with a scanning electron microscope (SEM) (LEITZ-AMR 1000A) and their structure is measured with an X-ray diffraction (XRD) (Philips PW1840) compact X-ray diffractometer system by using Cu K$\alpha$ ($\lambda$ $=$ 1.5405 . The measurements are recorded at a diffraction angle 2$\theta$ in the range of 2$^\circ$- 81$^\circ$.

3. Results and discussion

The set of films under study are all deposited from the same precursor solution with the same deposition parameters. The morphology of the films is explored by SEM observations, where Figure 1 displays the SEM image for one of the as-deposited films. As shown in Figure 1, the film appears to be polycrystalline and fully covered with material.

The structure is investigated with XRD, and Figure 2 depicts the diffractogram of a film, where the figure shows a hexagonal (wurtzite) structure. The strong narrow peak corresponds to the reflection from the (002) plane, and it confirms that the preferred growth direction is along the $c$-axis perpendicular to the substrate. In addition, there are five weaker diffraction peaks, which correspond to (100), (101), (102), (103) and (004), respectively. The Bragg equation (2$d$sin$\theta$ $=$ $n \lambda )$ was used to calculate interplanar distance $d$, where $\theta$ is Bragg's angle, $n$ is the order which equals to 1, and $\lambda$ $=$ 1.5405 {\AA} is the wavelength of the incident X-ray (Cu K$_{\alpha }$ line). The interplanar distance $d$ was found to be 0.261 nm, and the crystal lattice constant $c$ was also calculated and found to be 0.522 nm. The grain size was estimated from the Scherrer formula and the line (002) in the XRD diffractogram.

$d=\frac{\lambda }{D\cos \theta },$

(1)

where $d$ is the grain size, $\lambda$ is the X-ray wavelength defined before, $D$ is the angular line width of the half-maximum intensity, and $\theta$ is the Bragg angle defined before. The grain size calculated from Equation (1) is about 30 nm, which means that the particulates in the SEM image are aggregates, which consist of smaller crystallites.

The three different contacts: Al, Ag and Cu are deposited on the surfaces of the films by thermal evaporation-one type of contact on each film. Films with Al contacts are annealed in a nitrogen atmosphere for 45 min after depositing the contacts. Some of the criteria for the perfection of the contacts are linear current-voltage characteristics particularly at low voltages, and the absence of a photovoltaic effect^[16]. The three aforementioned contacts are examined for achieving these criteria. For this purpose, the $I$-$V$ measurements are recorded in the forward and reverse directions at room temperature in the dark and room light, and displayed in Figure 3. As shown in the figure, the relations are linear, which means that the three materials form ohmic contacts for ZnO.

Linear fits which pass through the origin are performed, and the fit parameters $R^{2}$, standard deviation SD and $F$ are listed in Table 1. The resistance $R$ is calculated from the slope of each straight line and used to find the resistivity $\rho$ by using the relation $R$ $=$ $\rho L/A$, where $L$ is the length in the direction of the current and $A$ is the cross-sectional area perpendicular to the current. The estimated values of $R$ and $\rho$ are listed in Table 2.

From Table 1 first, both in the dark and in room light, the highest slope is obtained by using Al contacts, while the lowest one by using Cu contacts. Second, the smallest standard deviation (SD) is got with Cu contacts, while the largest SD with Ag contacts. Third, in the dark and in room light, the best linearity, which is represented by the value of $R^{2}$ is for Al contacts, which means that these contacts show the best ohmicity in the set under study. The largest $F$ value is for Al contacts and the smallest one is for Ag contacts. Hence the best ohmic contacts from the set under study are Al contacts. These results can be explained in the light of the comparison between the electron affinity of ZnO, which is 4.1 eV^[25] and the work functions of these contact materials, which are 4.74, 4.24, and 4.94 eV for Ag, Al and Cu respectively^[26]. So, the closest work function to the electron affinity of ZnO is Al, while the farthest one is Cu. Hence the best ohmicity is Al, and the worst one is Cu, although all of them produced ohmic contacts because the values of the work functions are comparable to the electron affinity of ZnO.

From Table 2, the smallest resistance and resistivity are obtained with Al contacts, while the largest ones are with Cu contacts both in the dark and room light. As mentioned before, these results can be explained in terms of the electron affinity of ZnO and the work functions of the metals used as contacts. The work function of Al which equals 4.24 eV^[26] is the closest to the electron affinity of ZnO which is 4.1 eV^[25], while the work function of Cu, which is 4.94 eV^[26], is the farthest.

The ratios of the dark resistivity $\rho_{\rm dark}$ to the light resistivity $\rho_{\rm light}$ are shown in Table 2. The Cu contacts are the most affected by light (highest $\rho_{\rm dark}$/$\rho_{\rm light)}$, then the Ag contacts. These results can be interpreted in terms of the decrease in the barrier height caused by light. Al contacts are nearly unaffected by light, or approximately the photovoltaic effect on them is absent, which confirms that they are the best ohmic contacts in the set under study.

Annealing is known to improve the ohmic behavior of the contacts. Figure 4 displays the $I$-$V$ plots for another two films with different thicknesses and Al contacts recorded in the dark before and after annealing. The relations are linear as before, and so linear fits that pass through the origin are performed. The fit parameters are listed in Table 3, and the estimated values of the resistance and resistivity and the ratios of the dark resistivity to the light resistivity are listed in Table 4. For both films there is a decrease in the resistivity after annealing. This is probably because more Al diffusion into the ZnO film occurs at high temperatures, which results in doping the region under the contacts and the formation of an alloy, and then a lower contact resistance. Comparing the resistivities of the thicker and thinner films after annealing, it is found that the resistivity of the thicker film is decreased by a factor of about 1.08, while that of the thinner one is decreased by a factor of about 2.44. Because grain size increases with film thickness^{[27, 28]}, the grains of the thicker film are larger, so a slight enlargement with annealing occurs and the decrease in resistivity is mainly due to the diffusion of aluminum into the film. In the case of the thinner film, the grain size is smaller, so more enlargement of the grains with annealing occurs, beside the diffusion of aluminum into the film. So, both of these factors contribute to decreasing the resistivity.

4. Conclusions

Polycrystalline ZnO thin films were prepared by the spray pyrolysis technique on glass substrates. A comparison between three ohmic contacts for ZnO thin films was performed by the use of the $I$-$V$ characterization. These contacts are aluminum, copper and silver. It is found that aluminum contacts are the best, because they resulted in the smallest resistivity, they contributed in doping through Al diffusion into the film, and they are less affected by light as required. Copper contacts are found to be the worst in the set, because they resulted in the largest resistivity and smallest ohmicity. Annealing improved the Al contact through improving the ohmicity, the interdiffusion, and alloying.

Acknowledgments

We want to thank Sameer Farrash at the University of Jordan for making the contacts by thermal evaporation. We also thank Marsil Imsais from the geology department in the University of Jordan for the XRD measurements and Mr. Khalil Tadros from the geology department in the University of Jordan for the SEM images.