Effect of RF Power

Figure 4.1 illustrates the deposition rate of ZnO thin film as a function of applied

RF power. In general, the deposition rate increases nearly linearly with increasing RF power [80-81, 85-87, 159-160]. Increasing RF power increases the plasma number density in the bulk and ion bombardment energy onto the Zn target by increasing the ionization probability and the ion impact target surface probability, which sputters more Zn or ZnO particles into the plasma. More ZnO particles are thus formed by reacting with atomic oxygen in the discharge, and in turn, the deposition rate becomes larger.

To explain the argument, we must note that in this study the typical I-V curve was measured by the Langmuir probe. The detail analysis process is reported in Appendix A.

The floating potential (Vf) and plasma potential (Vs) were determined from the I-V curve as shown in Fig. 4.2. The ion bombardment energy (Vs – Vf) of sputtered particles bombarding the substrate is estimated by the difference between plasma potential and floating potential [81-82, 108], as a function of applied RF power, as shown in Fig. 4.3.

Due to RF power increasing, the input voltage and current were increased. Hence, the particles were able to obtain higher energy for the ionization process by absorbing the input voltage and current. For this reason, it was possible to observe that the floating potential has only slight increases with increasing RF power, but plasma potential and ion bombardment energy have more obvious increases.

Therefore, the moderation ion bombardment energy is likely to increase the surface mobility of adatoms and lead to the formation of dense layers during the deposition process [81]. Furthermore, the RF power increased with process temperature during deposition process. Then the higher process temperature induced the worse vacuum (high pressure). In order to maintain the constant working pressure, the pumping speed of the cryogenic pump was increased. Compared with low process temperature, particle concentration was decreased, but the mean free path (MFP) was increased. Meanwhile, particle energy dispersion was reduced with less collision probability and the

probability of impact substrate was increased. However, (Vs - Vf) generally increases with increasing applied RF power, except in the case with the applied power of 50W.

Influence of increasing (Vs - Vf) has three stages. First, more Zn atoms or ZnO particles are sputtered into the plasma bulk, and thus more ZnO particles are formed, because the ions bombard the target with higher impact energy. This causes the increase of deposition rate, as presented earlier. Second, it is highly likely that the energetic ions which bombard the substrate cause the substrate temperature to rise substantially at where they bombard. This in turn enhances the local surface mobility of adatoms and thus produces larger-grain ZnO layers during the deposition process [84]. Third, electrons obtain higher energy through the acceleration in the sheath with increasing (Vs - Vf). These higher-temperature electrons will enhance the ionization in the plasma bulk, which further increases the plasma number density, as shown in Fig. 4.4.

For this reason, electron temperature increases from 2.8 eV (at 50 W) to 4.2 eV (at 400 W), which is normal in highly vacuum sputtering plasmas. It should be noted that the increase of electron temperature from applied power of 50 W to 100 W becomes minimal due to the slightly decrease of (Vs - Vf).

Incidentally, Fig. 4.4 also indicated that electron number density (ne) and ion number density (ni) increase with RF power. Ion number density increases from 3.03 × 10¹⁷ #/m³ (at 50 W) to 6.15 × 10¹⁷ #/m³ (at 400 W), while electron number

density increases from 9.28 × 10¹⁶ #/m³ at 50W to 3.25 × 10¹⁷ #/m³ at 400 W.

Otherwise, ion flux could be calculated by following simple plasma theory:

e i

i

Ion Flux n KT

= m (4.1) Therefore, the calculation of ion flux is approximate to our analysis data because ion flux is proportional to ion number density. This is because greater ion number density will generate more atoms from the target, which then makes ion flux increase. In addition, the deposition rate increases about 11 times from 50 W to 400 W by this study (see Fig. 4.1). However, ion flux and ion number density only added about 2 times at the same condition, meaning that the amount of atoms deposited onto substrate increases when RF power increases. Therefore, we deduced that the sputtering yield (γ_sput) might increase 5.5 times because it is defined as the number of target atoms

ejected per incident ion [137]. Furthermore, we also drew a relation figure between ion flux and deposition rate as shown in Fig. 4.5. Results reveal that ion flux and deposition rate present a linear relationship when RF power increases. Actually, we believed that the sputtering yield multiplication of the ion flux should be equal to the deposition rate [141]. Likewise, note that

. .

sput

D R

Ion Flux∝γ (4.2) In other words, we conjectured that the sputtering yield increases 5.5 times while the RF power changed from 50 W to 400 W.

In order to explain this phenomenon, we used a simple plasma hypothesis. In

general, ion number density and potential (DC bias of target) increase 2 times when the power becomes 2 times for capacitively coupled plasma (CCP) type. Here, ion number density and potential (DC bias of target) could be assumed to be current (I) and voltage (V), respectively. Electrical power is calculated using Joule's law:

P I V= ⋅ (4.3) In this study, power and current (meaning ion number density) are equal to 8 (from 50 W to 400 W) and 2 (see Fig. 4.4), respectively. According to Joule’s law, we knew that the voltage increases 4 times. This means that the bombardment energy near sheath also increases. These results indicated the power increases not only increase ion number density, but also the potential of the bombardment target. Owing to the potential means of the capability of the bombardment target, the voltage increases accompany with sputtering yield increasing. Consequently, we thought that the sputtering yield increase of 5.5 times was reasonable, although ion number density only promoted to 2 times.

Due to this result, there is enough evidence to show that the RF power might enhance the number of particles that arrives at the substrate, so the deposition rate also increases during the sputtering process (see Fig. 4.1). Additionally, oxygen might produce negative charge particles (O^-, O2-) in the plasma environment, called negative plasma; hence the electron number density is not equal to ion number density. For pure

argon plasma, electron number density should be nearly equal to ion number density.

This is because the electron number density is defined by electron saturation in the I-V curve region of the Langmuir probe. When positive voltage was applied to the probe tip, it collected the negative charge and repelled the positive charge. The negative charge can be classified into electron and oxygen (O^-, O2-). These two charges of molecule weight had very large difference. Strictly speaking, the probe collected charge particles when the number of light charge particles (electron; e^-) was higher than heavy charge particle (O^-, O2-). Finally, the difference between these two particles represents approximately the total number density of negative charge species (O^-, O2-, etc.) in this electronegative (Ar/O2) plasma. In the present case, total number density of negative charged species amounts to approximately 3 × 10¹⁷ #/m³ at all power levels, except at 50 W (~2 × 10¹⁷ #/m³), in which the electric field is not strong enough to generate more negative charged species.