Cleaning Effect of Atmospheric-Plasma-Sprayed Y₂O₃ Coating Using Piranha Solution Based on Contamination Particle Measurement

Abstract

A Y₂O₃ coating was prepared using the atmospheric plasma spraying (APS) technique. On exposing the coating to CF₄/O₂/Ar plasma, a fluorine contamination layer (YO_xF_y) was formed, which was the main cause of process drift and contamination particle generation on the APS–Y₂O₃ coating surface. To remove the YO_xF_y layer on the coating surface, a piranha solution, which is a mixture of sulfuric acid and hydrogen peroxide, was employed for cleaning. The piranha solution was found to be an excellent medium for removing the YO_xF_y layer. The amount of contamination particle generated could be reduced by approximately 37% after cleaning with a 3:1 piranha solution compared with before cleaning.

1. Introduction

Plasma is used in various processes, such as etching, deposition, and cleaning, in the semiconductor/display industry. Since the etching process uses plasma composed of highly reactive carbon-, fluorine-, and chlorine-based compounds, it induces a chemical reaction with the internal components of the chamber. The by-products of this reaction formed on the surface are deposited on the surrounding parts, generating contamination particles, which ultimately affect the product yield [1,2,3,4,5]. Therefore, the chamber wall and internal parts are coated with plasma-resistant materials using various coating methods for the semiconductor etching process. Al₂O₃ and Y₂O₃ are widely used as ceramic coating materials exhibiting plasma corrosion resistance [6,7,8,9,10,11,12,13,14,15]. Conventionally, Al₂O₃ has been used as a representative plasma corrosion resistance material owing to its low price and easy sintering process. However, it lacks chemical stability to highly reactive halogen element radicals. Y₂O₃ is widely applied as a plasma corrosion resistance material owing to its lower etch rate and better chemical stability than Al₂O₃ for fluorine-based plasma during the semiconductor etching process. The atmospheric plasma spraying (APS) coating method is commonly used in the industry; this is attributed to the fewer restrictions on raw materials and coating conditions, ease of control of the coating thickness, realization of complex shapes, and coating application over a large area [14,15,16]. However, when an atmospheric-plasma-sprayed Y₂O₃ coating is exposed to highly reactive fluorine-based plasma for a long duration, there is surface corrosion, and a YO_xF_y layer, which is a fluorinated layer, is formed. The YO_xF_y layer is the main cause of process drift and contamination particle generation [8,10]. To remove the fluorine contamination layer after plasma exposure, efforts have been made to clean and reuse the coating surface [17,18]. Cleaning a plasma-resistant coating requires physically etching the surface using a surfactant, and the degree of cleaning should be evaluated again after visual observation. Despite this problem, there has been no technical research on the cleaning technology for plasma-resistant coatings.

In this study, a piranha solution was used for cleaning a APS-Y₂O₃ coating for the first time. A piranha solution is a mixture of sulfuric acid and hydrogen peroxide; it is a very strong oxidizing agent and has highly corrosive properties. It is used to remove organic residues and to dissolve metal oxides and carbonates [19,20,21,22]. We applied it to the cleaning of a plasma corrosion resistance coating owing to its potential for dissolving the fluorine contamination layer on the APS–Y₂O₃ coating surface. Currently, most studies on the use of the piranha solution have been conducted focusing on Si wafer cleaning, and research on the cleaning of plasma corrosion resistance coating, which is important in the semiconductor industry, is lacking. In this work, a APS–Y₂O₃ coating exposed to CF₄/O₂/Ar plasma was cleaned using piranha solutions prepared at different ratios (2:1, 3:1, 4:1, and 5:1). Before and after the cleaning process, the formation of a residual fluorine contamination layer on the APS–Y₂O₃ coating was confirmed, and the amount of contamination particle generated was analyzed and compared.

2. Experimental

We prepared an aluminum 6061-T6 alloy (HIGGLAB, Seoul, Republic of Korea) in the form of a circular sample with dimensions of 76.2 ø × 3 mm. The substrates were coated by APS (Axial III, Northwest Mettech, Surrey, BC, Canada), where Y₂O₃ was in a powder form (99.99%, 25–30 µm, Tokyo, Japan, Figure 1). The APS coating was applied with Ar sprayed as a carrier gas at 1500 Torr, and the plasma gun power was set to 980 W.

To evaluate the plasma resistance of the APS–Y₂O₃ coating, a vertical conductively coupled plasma-reactive ion etching (VCCP-RIE) system was used, as shown in Figure 2. The plasma density was enhanced by arranging a niobium magnet in 1000 G on the upper electrode. The RF power used was 13.56 MHz (Sizer Generator, Advanced Energy, Fort Collins, CO, USA), and an impedance matching network was employed (Navigator, Advanced Energy, Fort Collins, Denver, CO, USA). It was exposed to fluorine-based plasma (CF₄/O₂/Ar) used in the etching process;

Table 1. Details of the plasma etching parameters.

Plasma Etching Parameters
Gas	CF4/O2/Ar 20, 5, 15 sccm
Working pressure	250–262 mTorr
Power	200 W
Etch time	60 min

presents the detailed conditions. The contamination particles generated due to plasma exposure were measured in real time using an in situ particle monitor (ISPM; Stiletto, INFICON, Heidland, Switzerland). To compare the amount of contamination particle generated before and after cleaning with the piranha solution, the APS–Y₂O₃ coating was first exposed to CF₄/O₂/Ar plasma. Subsequently, the amount of contamination particle generated by re-exposure to the CF₄/O₂/Ar plasma without any cleaning was compared with the amount of contamination particle generated after the cleaning.

The piranha solution was prepared in different volume ratios by mixing sulfuric acid (extra pure grade 95%, DUKSAN, Seoul, Republic of Korea) and hydrogen peroxide (extra pure grade 30%, DUKSAN, Seoul, Republic of Korea).

Table 2. Details of the piranha solution ratio.

Piranha Solution Ratio
H2SO4 + H2O2	2:1	40 mL, 20 mL
	3:1	45 mL, 15 mL
	4:1	48 mL, 12 mL
	5:1	50 mL, 10 mL

presents the preparation conditions of the piranha solutions. The piranha solution was maintained at a temperature of 70 °C using a heating stage (Hotplates & Stirrer, MTOPS, Seoul, Republic of Korea). Figure 2 shows the schematic of a vertical conductively coupled plasma etching system. First, the APS–Y₂O₃ coating was exposed to CF₄/O₂/Ar plasma, and later dipped in a piranha solution for 5 min. After rinsing in DI water, it was cleaned 100 times using a #800 grit scouring pad. Figure 3 shows the experimental method of piranha solution cleaning. The morphologies of the APS–Y₂O₃ coating surface before and after cleaning with the piranha solution were analyzed by field emission scanning electron microscopy (FE-SEM, Secondary electron, S-4800, Hitachi, Tokyo, Japan). APS-Y₂O₃ coating was coated by Pt using the sputtering method. The residual components on the APS–Y₂O₃ coating surface were analyzed by X-ray photoelectron spectroscopy (XPS, Al K$\mathsf{\alpha }$ 1486.6 eV, 45° from sample surface, PHU 5000 VersaProbe, Ulvac-PHI, Hagisono, Japan). The surface roughness was analyzed by confocal microscopy.

3. Results

Figure 4 shows the surface images of the APS–Y₂O₃ coating before and after CF₄/O₂/Ar plasma exposure and after cleaning with the piranha solution. Figure 4a shows the surface image of pristine APS–Y₂O₃ before plasma exposure, and Figure 4b shows the image after plasma exposure. After exposure to fluorine-based plasma, a brown contamination pattern was formed on the coating surface. This is a fluorine contamination layer due to surface fluorination, which generates contamination particles [17,18]. Figure 4c–f shows that the APS–Y₂O₃ coating surface exposed to plasma was cleaned with piranha solutions prepared at ratios of 2:1, 3:1, 4:1, and 5:1, respectively. The fluorine contamination layer was removed regardless of the piranha solution ratio. Figure 5 shows the FE-SEM image for a more detailed surface analysis.

Figure 5a,b show the surface images before and after plasma exposure of the APS–Y₂O₃ coating, respectively. In Figure 5a, pores and nonmelted particles can be observed, which is common on an APS coating surface [15]. The APS coating powder has a size range of 30–50 $\mathsf{\mu }\mathrm{m}$, which is greater than those of other spray coatings [8,12]. In addition, the droplets were rapidly cooled as they collided with the substrate. Because of this, nonmelted particles were formed. Plasma erosion occurs from the edge of the nonmelted particles. Consequently, a mushroom-shaped etching trace could be observed, as shown in Figure 5b. Figure 5c–f shows the surface images after cleaning with piranha solutions prepared at ratios of 2:1, 3:1, 4:1, and 5:1, respectively. After cleaning with piranha solutions with ratios of 2:1, 4:1, and 5:1, a relatively rough surface with some pores was observed. In contrast, after cleaning with a piranha solution at a ratio of 3:1, a smooth surface with the least number of defects was observed.

Figure 6 shows the EDS mapping image, where the presence of a fluorine contamination layer on the APS–Y₂O₃ coating surface can be confirmed. The pristine APS–Y₂O₃ coating surface did not contain any fluorine, as shown in Figure 6a. Conversely, the APS–Y₂O₃ coating exposed to CF₄/O₂/Ar plasma showed a high fluorine content on the surface, as shown in Figure 6b. From this, it can be confirmed that the surface of the APS–Y₂O₃ coating was fluorinated by plasma. Figure 6c–f shows the APS–Y₂O₃ coating after cleaning with piranha solutions with ratios of 2:1, 3:1, 4:1, and 5:1, respectively. Regardless of the piranha solution ratio, the surface fluorine content decreased under all the conditions.

Figure 7 shows the XPS spectra and Ar ion sputtering depth profile of the APS–Y₂O₃ coating surface before and after CF₄/O₂/Ar plasma exposure and after piranha cleaning. Figure 7a shows an image of the APS–Y₂O₃ coating before plasma exposure. This process was performed in an atmospheric environment; therefore, the carbon present in the air penetrated the coating surface and existed therein. Therefore, carbonate bonding can be observed in Figure 7a. In the pristine APS–Y₂O₃ coating, the Y 3d_3/2 and Y 3d_5/2 peak positions were 157.6 and 155.5 eV, and two peaks had a difference of 2.1 eV with an intensity ratio of 3:2 in their binding energy [23,24]. After CF₄/O₂/Ar plasma exposure, the fluorine present on the surface was approximately 35%, and the Y–O bonds were replaced by Y–F bonds owing to surface fluorination. The Y–O and Y–F binding energies of Y 3d_3/2 were 161.2 and 159.0 eV and those of Y 3d_5/2 increased to 159.1 and 156.9 eV, respectively. This can be possibly attributed to the higher electronegativity of fluorine (4.0) atoms than oxygen (3.5) atoms [9,16]. Figure 7c–f shows XPS spectra and Ar ion sputtering depth profile, after cleaning with the piranha solution. The coating surface depth profile analysis shows that, after cleaning with the piranha solutions with ratios of 2:1, 3:1, 4:1, and 5:1, the fluorine content on the coating surface decreased to 2.49%, 2.35%, 3.18%, and 4.17%, respectively, decreasing within 5% regardless of the piranha solution ratio. Moreover, the binding energies of the Y 3d_3/2 and Y 3d_5/2 peaks were 158.0 and 155.9 eV, respectively. The binding energy was similar to that of the pristine APS–Y₂O₃ coating.

Figure 8 shows the surface roughness of the APS–Y₂O₃ coating before and after plasma exposure and after cleaning with the piranha solution. The pristine APS–Y₂O₃ coating shows a rough surface (Ra = 0.716 $\mathsf{\mu }\mathrm{m}$) in Figure 8a. The APS coating uses a large powder size of approximately 30 $\mathsf{\mu }\mathrm{m}$. It is melted by the plasma jet and hits the substrate. At the same time, it cools rapidly, and nonmelted particles and pores are formed. Figure 8b shows the surface roughness of the APS–Y₂O₃ coating exposed to the CF₄/O₂/Ar plasma; the surface roughness value is lower than that of the pristine APS–Y₂O₃ coating. This was caused by flattening under continuous Ar⁺ bombardment. Figure 8c–f shows the surface roughness after cleaning with the piranha cleaning. At a ratio of 2:1, the roughness was similar to that of the pristine APS–Y₂O₃ coating. In particular, at a ratio of 3:1, the surface was relatively smooth, consistent with the SEM image shown in Figure 6. The surface roughness values under the other conditions were Ra = 0.582 and 0.654 $\mathsf{\mu }\mathrm{m}$, respectively, reflecting a smoother surface than the pristine APS–Y₂O₃ coating. These results confirmed that cleaning with the 3:1 piranha solution helped remove the most amount of fluorine contamination layer on the APS–Y₂O₃ coating, while imparting the smoothest surface.

Figure 9 shows the amount of contamination particles generated when the APS–Y₂O₃ coating was exposed to CF₄/O₂/Ar plasma. In Figure 9a, the amount of contamination particles generation in the APS–Y₂O₃ pristine coating is 450 EA. When re-exposed to plasma without cleaning, this amount increased by approximately twofold to 867 EA, as shown in Figure 9b. This increase can be attributed to the surface corrosion caused by the plasma. On the other hand, the amounts of contamination particle generated after cleaning with piranha solutions at ratios of 2:1, 3:1, 4:1, and 5:1 were 857, 542, 663, and 757 EA, respectively. At a ratio of 3:1, the amount of contamination particle generation reduced by approximately 37% compared to that without cleaning; this is because the surface of the APS–Y₂O₃ coating had fewer surface defects and a denser microstructure, such as a smooth surface, compared to that observed for ratios of 2:1, 4:1, and 5:1, and the XPS analysis showed that the surface had the lowest fluorine content (2.35%). The fluorine contamination layer formed on the APS–Y₂O₃ coating surface was removed to the maximum extent, and the amount of contamination particles generated was the least.

4. Conclusions

We exposed an APS–Y₂O₃ coating to CF₄/O₂/Ar plasma and confirmed the formation of a fluorine contamination layer on the coating surface, which was the cause of contamination particle generation. Subsequently, the APS–Y₂O₃ coating was cleaned using piranha solutions prepared at ratios of 2:1, 3:1, 4:1, and 5:1, and the change in the coating surface properties and the amount of contamination particle generated after cleaning were investigated. Regardless of the piranha ratio, the fluorine contamination layer formed on the coating surface could be reduced to approximately 5%. In particular, after cleaning with a 3:1 piranha solution, the surface fluorine content was the lowest, and the surface was the smoothest. The amount of contamination particle generated from the pristine APS–Y₂O₃ coating was 450 EA, and when re-exposed to plasma without any cleaning, this amount increased by two folds to 867 EA compared with that of the pristine APS–Y₂O₃ coating. This increase is attributed to the corrosion caused by the plasma. Cleaning with the piranha solution reduced the amount of contamination particle generated compared to before cleaning. For a ratio of 3:1, the contamination particle generated reduced by approximately 37% compared with that before cleaning, and it was most similar to the pristine APS–Y₂O₃ coating. Therefore, we confirmed that cleaning with the 3:1 piranha solution produced a coating property similar to that of the pristine APS–Y₂O₃ coating.