Influence of Metallic Indium Concentration on the Properties of Indium Oxide Thin Films

Introduction

Transparent and conducting indium oxide thin films have a wide range of applications; among the most important are photovoltaic devices, liquid crystal displays, sensors, antireflection coatings, solar cells and electrochromic devices [1–3].

Indium oxide thin films have been deposited by different methods such as spray pyrolysis [4], pulsed laser deposition [5], oxygen-ion beam assisted deposition [6], reactive evaporation [7, 8], reactive direct-current magnetron sputtering [9], etc. Recently a lot of work has been reported on physical properties of indium oxide films [10–15]. The effect of presence of indium nanoparticles in indium oxide crystals on optical properties has been studied in detail [16]. The nanocomposite films have been fabricated by embedding palladium nanoparticles into an indium oxide matrix and the effect of palladium content in these nanocomposite films on thermoelectric response has been studied [17]. As seen from this reference, interest for such kind of cermets can be found in thermoelectric converters.

In the present study, indium-embedded indium oxide thin films were prepared by reactive evaporation of pure indium at various air pressures of the deposition and substrate temperatures. Detailed investigations were carried out to study the influence of metallic indium concentration on the properties of these films. Some possible conduction mechanisms are proposed.

Specifically, the conductivity and transparency values obtained from the thin films we have studied in this work indicate that these oxide films can be used in optoelectronic devices and windows which need high conductivity and low transparency.

Experimental details

The indium oxide thin films were prepared on glass substrates by reactive evaporation technique at different air pressures. The adopted cleaning procedure was as follows: the glass substrates were cleaned in a detergent solution using an ultrasonic cleaner; they were then taken out of the solution and rinsed with distilled water. Thereafter, they were wetted in hot chromic acid solution for 15 min and rinsed with distilled water. Finally, they were dried in the oven at 130°C. A tungsten basket was used as the source holder and heater for indium evaporation. Samples were produced by evaporating 99.99% pure indium on glass substrates at different temperatures in a vacuum system, initially held at 1 × 10⁻⁴ Torr and gradually raised admitting air by a leak needle valve control to the desired pressure varying between 10⁻²and 10⁻⁴ Torr. The source–substrate distance was kept constant at about 9 cm. Films of about 620–850 Å thickness were grown on Corning 7058 stage. Thin Ag electrodes with thickness of ~1000 Å were deposited onto the substrates by thermal evaporation under vacuum after the oxidation process had completed. The temperature of the substrate was varied between 50 and 125°C to an accuracy of ±2°C. A flat electric heater was used for substrate heating. The heater was controlled by a circuit of electronic thermostat.

Current–voltage measurements of films were carried out using a Schlumberger 1240 and a Unigor D 410 digital multimeter. The thickness of the thin films was determined by Fizeau’s method for fringes of equal chromatic order (FECO), using a Hilger and Watts N130 type interferometer.

Optical transmittance characteristics of the films were studied using a Varian series 634 UV-Vis spectrophotometer in the wavelength range 300–900 nm.

Results and discussion

It may be assumed that the evaporated films consist of a continuous amorphous indium oxide in which small crystallites of pure metallic indium spheres are uniformly dispersed. The structure of evaporated film is illustrated in Figure 1(a). To determine the metallic indium particle concentration in indium oxide thin films, the optical model is used. This model is based on the equivalence of optical absorption characteristics of an indium oxide thin film, containing uniformly distributed metallic indium particles and a pure metallic indium thin film [18]. We assumed that the optically equivalence of our samples can be represented as shown in Figure 1(b).

Figure 1:

(a) Indium-embedded indium oxide film on a glass substrate with the reflectivity and the transmittivity ρ andτ, respectively. (b) Optically equivalent film system. d_m is the equivalent metallic film thickness and d_in is the interferometrically measured thickness of the original sample. ρ_m and τ_m are the reflectivity and transmittivity of the thin metallic indium film, respectively.

The concentration of the metallic indium within the oxide is controlled by the deposition pressure, the evaporation rate and the distance of the substrate from the source. Thus, it is possible to change the amount of metallic indium within the oxide.

A cube full of touching identical spheres will represent 74% of the total volume. For metallic concentration C_opt higher than 74%, these spheres will be in contact with each other and the grain boundaries of the spheres are occupied by the indium oxide. For C_opt less than 74%, the spheres will be separated by the insulating barrier of amorphous oxide. The spacing between the metallic particles will thus depend on the diameter as well as the concentration of these particles.

If the interferometrically measured thickness of the original sample is d_in and that of the equivalent metallic film, having the same optical absorption characteristics as the original sample, is d_m, the concentration of the metallic indium particles within the indium oxide can be calculated from the following relation [18]:

where C_opt is the metallic indium concentration due to the optical model.

In the present study, the concentration calculated from eq. (1) was found to be less than 74%. Table 1 shows the concentration of the metallic indium within the indium oxide which was controlled by the deposition pressure. The thickness of the deposited films varied from 600 to 850 Å and was found to depend on the air pressure during deposition. The film thickness decreased with decreasing pressure.

The optical transmittance of the films as a function of wavelength for different concentrations is given in Figure 2. It can be seen that the films which have high concentration show low optical transmittance due to large amounts of randomly distributed excess metallic particles embedded in indium oxide. The curves in Figure 3 show the typical electrical characteristics of a sample containing a considerable amount of metallic indium within indium oxide. We observe that conductivity increases with increasing metallic indium concentration. This means that all samples show high electrical conductivity. In our samples, conductivity and transmittance depend upon the metallic indium concentration. As can be seen from Table 1, this concentration is nearly constant within the temperature region of interest. Therefore in our samples neither transmittance nor conductivity changes as the temperature varied at least in the temperature range (50–125°C) that we are interested. Due to these observations, the substrate temperature is not a parameter affecting the optical transmittance and the electrical conductivity of these samples.

Figure 2:

Optical transmittance of indium-embedded indium oxide films deposited at two different substrate temperatures for different C_opt concentrations. C_opt, metallic indium concentration.

Figure 3:

Log I vs log V at various C_opt concentrations for two different substrate temperatures.

Now, we want to discuss about the possible conduction mechanisms in our samples. To do this we first measured the current density vs electrical field and the results are shown in Figure 4. On the other hand, we know the following Schottky current formula

Figure 4:

Log J vs F^1/2 at various C_opt concentrations for two different substrate temperatures.

where ϕ is the barrier height in eV, T is the absolute temperature, k is the Boltzmann constant, Fis the electric field and the constant β is given by [19]

Here, e is the charge of electron, ε is the dielectric constant and ε₀ is the permittivity of free space. Fitting our experimental data, shown in Figure 4, with the Schottky current formula we have obtained ϕ.

High-field electrical conduction is quite complex because it is difficult to distinguish between the several possible involved mechanisms. These mechanisms are usually governed by a number of factors like the work function of the electrodes and that of the sandwich layer, the applied field and the structural defects. These defects act as potential trapping centers for current carriers. Keeping the evaporation rate constant and increasing the evaporation pressure, the radius of the metallic spheres may be decreased as the concentration decreases. Table 2 shows the barrier height ϕ which increases for decreasing concentration at two different substrate temperatures. The smaller the size of the particles, the greater is the activation energy. In both cases the conductivity is lowered [20]. The larger the distance between the particles, the lower is the tunneling probability.

Secondly, we have observed that the current exhibits a voltage dependence of the form I∝Vn, where n is found to be greater than 1 (Figure 3). Consequently, the conduction is found to be non-ohmic. In this case, there are three different types of conduction mechanisms: space charge limited current (SCLC), Schottky type and Poole–Frenkel type [21]. The conduction mechanisms in our samples cannot be SCLC, because the slope for this type of mechanism has been observed to be around 2. From Figure 4 it is evident that the slope is less than 2.

According to eq. (2) if either of the above-mentioned last two mechanisms are dominant in these samples, a plot of log J vs F1/2should yield a straight line. Our Figure 4 supports this. Therefore, the Schottky or Poole–Frenkel type of mechanism may be the possible conduction mechanisms in our samples [21–24]. Oxide films are generally amorphous and are expected to have a large number of traps. The samples are assumed to be amorphous, and the presence of metallic indium within the indium oxide is due to the evaporation process. In the present work, the indium particles act as trap centers in the indium oxide with the effective Fermi level moving up toward the conduction band. From the above observations, it may be inferred that the Poole–Frenkel type of conduction mechanism is the dominant one in our samples. According to the Poole–Frenkel effect, when the electric field interacts with the coulombic potential barrier of a trap, the height of the barrier is lowered.

Conclusion

The indium-embedded indium oxide thin films with Ag electrodes were prepared by reactive evaporation onto glass substrates. In particular, metallic indium concentration was found to be an important parameter affecting the film properties. The purpose of our work was to investigate mainly the effect of metallic indium concentration on the overall film conductivity and transparency. Within the temperature range, our results show that with increasing metallic indium concentration the conductivity of the films increases while the transparency decreases.

One other result of this work was that the exponential current increases with increasing voltage, which indicates a non-ohmic behavior at high electric field values. According to our experimental results, the nonlinear current–voltage behavior may be attributed to Poole–Frenkel type of conduction mechanism.