Plasma Etching Behavior of SF6 Plasma Pre-Treatment Sputter-Deposited Yttrium Oxide Films

Yttrium oxyfluoride (YOF) protective materials were fabricated on sputter-deposited yttrium oxide (Y2O3) by high-density (sulfur fluoride) SF6 plasma irradiation. The structures, compositions, and fluorocarbon-plasma etching behaviors of these films were systematically characterized by various techniques. After exposure to SF6 plasma, the Y2O3 film surface was fluorinated significantly to form a YOF film with an approximate average thickness of 30 nm. X-ray photoelectron spectroscopy revealed few changes in the elemental and chemical compositions of the surface layer after fluorination, confirming the chemical stability of the YOF/Y2O3 sample. Transmission electron microscopy confirmed a complete lattice pattern on the YOF/Y2O3 structure after fluorocarbon plasma exposure. These results indicate that the SF6 plasma-treated Y2O3 film is more erosion resistant than the commercial Y2O3 coating, and thus accumulates fewer contamination particles.


Introduction
In recent years, semiconducting integrated circuits (ICs) have progressed toward various functionalities on single electrical chips, and have been downscaled to nanometer size [1]. Reducing the influence of particle contaminants in silicon wafers is becoming increasingly critical because such particles decrease the mass-production yield of ICs [2,3]. In particular, the corrosive gases (e.g., C 2 F 6 , CF 4 , CHF 3 , and C 4 F 6 ) used in semiconductor manufacturing processes generate high-density fluorocarbon plasma that bombards the IC chip and reacts with its inner chamber wall and ceramic parts (such as chamber windows, electrodes, showerhead cover baffles, and rings), generating contaminant particles [4][5][6]. Ceramics are desired as plasma-resistant materials on parts of plasma equipment, by virtue of their high hardness, high wear resistance, dielectric strength and chemical stability [7][8][9]. The plasma-facing inner wall of the chamber is often coated with yttrium oxide (Y 2 O 3 ), which provides superior plasma erosion resistance on silicon-based materials, and also extends their lifetime. However, the reaction between Y 2 O 3 and fluorine plasma generates thin fluorinated particles at the grain boundaries [10][11][12]. Recently, yttrium fluoride (YF 3 ) coatings have been proposed as potential alternative ceramic materials for Y 2 O 3 because they prevent the generation of fluoride particles. The standard enthalpy of the metal-oxygen bond is lower in YF 3 (−392 kJ mol -1 ) than in Y 2 O 3 (−318 kJ mol −1 ), indicating that YF 3 is more chemically stable than Y 2 O 3 [13][14][15]. Ceramic protective 2 of 9 coatings on chamber walls are typically fabricated by plasma spray techniques. However, although the rapid deposition forms thick films, spray methods tend to produce porous structures and rough surfaces, resulting in critical particle impurities [16]. These problems in the semiconductor plasma-etching process are believed to be resolved by vacuum coating techniques, such as radio frequency (RF) magnetron sputtering, but when YF 3 target is sputtered by these techniques, the stoichiometric variations in the Y/F ratio cause fluorine atom deficiencies from the target to the substrate, resulting in non-stoichiometric films with possibly degraded plasma erosion resistance [17,18]. Existing studies have focused on yttrium oxyfluoride (YOF) as a protective coating of the parts inside the manufacturing equipment chambers of semiconductor plasma processes [19,20]. In our previous study, we reported the formation of a YOF-altered thin layer on Y 2 O 3 and YF 3 coatings surfaces [21]. Furthermore, we also evaluated the characteristics of YOF coating fabricated by atmospheric plasma spraying. The YOF coating exhibited higher mechanical and electrical properties than those of the YF 3 coating [22]. The end result was fewer contaminants in the semiconductor IC products [23]. In the present paper, we develop an SF 6 plasma treatment that generates a YOF-altered layer on the Y 2 O 3 surface and evaluate the etching behaviors of the Y 2 O 3 and YOF/Y 2 O 3 layers under fluorocarbon plasma exposure.

Materials and Methods
Y 2 O 3 thin films were deposited on commercial c-plane sapphire substrate by RF magnetron sputtering (SSI-100, Shihsin, Tainan, Taiwan) in a vacuum chamber. Y 2 O 3 ceramic target (99.99% purity, 2 inch diameter, 3 mm thickness) was prepared as a source for film deposition. Prior to film deposition, the substrates were sequentially cleaned in acetone and alcohol, and then by ultrasonic cleansing in de-ionized water for 30 min. The cleaned substrates were blow-dried in nitrogen gas. The sputtering gas was high-purity argon (99.995%) maintained at a constant flow rate (~100 sccm). The sputtering process was performed in a base chamber pressure of approximately 1.5 × 10 −5 Torr, preserved with turbo molecular and oil diffusion pumps. Plasma generation was activated by RF power at 13.56 MHz. The target-substrate distance was 15 cm. To ensure a uniform film thickness, the substrate holder was rotated at 20 rpm during the deposition process. The substrate heating temperature was varied from 200 to 600 • C in steps of 200 • C. After the Y 2 O 3 thin-film deposition, the specimens were exposed to reactive ion etching (RIE). The plasma treatment time was 300 s, the process gas was SF 6 (30 sccm under a pressure of 100 Torr), and a RIE power (bias) of 100 W. The plasma etching of all samples was performed in an inductively-coupled plasma etcher (ICP, EIS-700, Elionix, Tokyo, Japan). The etching gases were mixed CF 4 and O 2 (25 sccm CF 4 and 5 sccm O 2 ). The crystallographic properties of Y 2 O 3 films were investigated via an X-ray diffractometer (XRD, XRD-6000, Shimadzu, Kyoto, Japan) with Cu Kα X-ray source (λ = 1.541874 Å) radiation. The crystallite size and full width at half maximum (FWHM) of the Y 2 O 3 film grown on the sapphire substrate were calculated using the Debye-Scherer equation. The surface morphologies, microstructures and compositions of the as-deposited Y 2 O 3 and SF 6 plasma-treated Y 2 O 3 films were analyzed by field emission scanning electron microscopy (FE-SEM, S-3000H, Hitachi, Tokyo, Japan), atomic force microscopy (AFM, DI-3100, Veeco, New York, NY, USA), high-resolution transmission electron microcopy (HRTEM, H-600, Hitachi, Tokyo, Japan) and X-ray photoelectron spectroscopy (XPS, PHI 5000 VersaProbe, ULVAC-PHI, Kanagawa, Japan) using a monochromatic Cu Kα X-ray source (λ = 1.541874 Å) at a passing energy of 20 eV with a spot size of 650 µm. sufficient energy to enhance the mobility of the atoms and to further improve the film crystallites. Subsequent analyses were performed on the sample prepared at 600 • C, which showed a dense structure in the XRD analysis result. Figure 1a shows the XRD patterns of the Y2O3 films deposited at different substrate temperatures. The polycrystalline nature of all deposited Y2O3 films was indexed to JCPDS card file 43-1036. Four clear diffraction peaks of Y2O3 (222), (400), (440) and (622) appeared on the substrate deposited at 600°C. Those of the Y2O3 (222) peak deposited at different substrate temperatures are shown in Figure 1b. Increasing the substrate temperature from 200 to 600 °C increased the crystalline size from 10.5 to 15.2 nm and narrowed the FWHM of the (222) peak. The higher substrate temperature supplied sufficient energy to enhance the mobility of the atoms and to further improve the film crystallites. Subsequent analyses were performed on the sample prepared at 600 °C, which showed a dense structure in the XRD analysis result.  Figure 2 shows the FE-SEM images of the as-deposited Y2O3 and SF6 plasma-treated Y2O3 film surface before and after fluorocarbon plasma etching. The plasma etching decreased the flake polygon size of the as-deposited Y2O3 films (Figure 2b), but did not significantly alter the film surface morphology (Figure 2c). This indicates that the SF6 plasma-treated Y2O3 was robust to irradiation with fluorocarbon plasma. The chemical stability of the SF6 plasma-treated Y2O3 might be attributable to the thin YOF layer formed on the Y2O3 surface (YOF/Y2O3). Figure 3 shows the surface morphologies of the as-deposited Y2O3 and SF6 plasma-treated Y2O3 films before and after fluorocarbon plasma etching, obtained by AFM. After fluorocarbon plasma etching, the root mean square (RMS) roughness values of the as-deposited Y2O3 films decreased from 10.7 to 4.97 nm, but those of the plasma-treated Y2O3 sample decreased only from 6.51 to 5.35 nm. The small RMS difference in the latter sample might be attributed to the YOF-altered layer formed on the plasma-treated Y2O3 surface. The Y-F bonds formed by YOF in the film might reduce the reactivity of YOF with fluorine radicals during the etching process [24].   (Figure 2b), but did not significantly alter the film surface morphology (Figure 2c). This indicates that the SF 6 plasma-treated Y 2 O 3 was robust to irradiation with fluorocarbon plasma. The chemical stability of the SF 6 plasma-treated Y 2 O 3 might be attributable to the thin YOF layer formed on the Y 2 O 3 surface (YOF/Y 2 O 3 ). Figure 3 shows the surface morphologies of the as-deposited Y 2 O 3 and SF 6 plasma-treated Y 2 O 3 films before and after fluorocarbon plasma etching, obtained by AFM. After fluorocarbon plasma etching, the root mean square (RMS) roughness values of the as-deposited Y 2 O 3 films decreased from 10.7 to 4.97 nm, but those of the plasma-treated Y 2 O 3 sample decreased only from 6.51 to 5.35 nm. The small RMS difference in the latter sample might be attributed to the YOF-altered layer formed on the plasma-treated Y 2 O 3 surface. The Y-F bonds formed by YOF in the film might reduce the reactivity of YOF with fluorine radicals during the etching process [24].

Results and Discussion
The effects of the as-deposited Y 2 O 3 films and SF 6 plasma-treated Y 2 O 3 films before and after fluorocarbon plasma etching were determined by XPS. The wide energy spectra of the films are shown in Figure 4. To generate the atomic signals for the XPS measurements, the samples were bombarded by an argon beam for 60 s. The XPS peaks in the as-deposited Y 2 O 3 film contained Y and O elements. After exposure to SF 6 plasma, an intense F1s peak appeared in the spectrum of the Y 2 O 3 film. This peak can be attributed to the penetration of fluorine radicals into the film. The fluorine radicals partially substituted the oxygen atoms, forming a fluorine-rich layer on the surface. By contrast, the XPS spectrum of the SF 6 plasma-treated sample was unaffected by fluorine-plasma etching. The quantitative XPS evaluations of F (fluorine), Y (yttrium), and O (oxygen) are listed in Table 1. Exposure to fluorocarbon plasma marginally altered the chemical composition of the SF 6 plasma-treated Y 2 O 3 film but largely changed the O 1 S and F 1 S contents of the as-grown Y 2 O 3 film. Similar results were reported in a previous study of fluorocarbon plasma etching [25].  The effects of the as-deposited Y2O3 films and SF6 plasma-treated Y2O3 films before and after fluorocarbon plasma etching were determined by XPS. The wide energy spectra of the films are shown in Figure 4. To generate the atomic signals for the XPS measurements, the samples were bombarded by an argon beam for 60 s. The XPS peaks in the as-deposited Y2O3 film contained Y and O elements. After exposure to SF6 plasma, an intense F1s peak appeared in the spectrum of the Y2O3 film. This peak can be attributed to the penetration of fluorine radicals into the film. The fluorine radicals partially substituted the oxygen atoms, forming a fluorine-rich layer on the surface. By contrast, the XPS spectrum of the SF6 plasma-treated sample was unaffected by fluorine-plasma etching. The quantitative XPS evaluations of F (fluorine), Y (yttrium), and O (oxygen) are listed in Table 1. Exposure to fluorocarbon plasma marginally altered the chemical composition of the SF6   The effects of the as-deposited Y2O3 films and SF6 plasma-treated Y2O3 films before and after fluorocarbon plasma etching were determined by XPS. The wide energy spectra of the films are shown in Figure 4. To generate the atomic signals for the XPS measurements, the samples were bombarded by an argon beam for 60 s. The XPS peaks in the as-deposited Y2O3 film contained Y and O elements. After exposure to SF6 plasma, an intense F1s peak appeared in the spectrum of the Y2O3 film. This peak can be attributed to the penetration of fluorine radicals into the film. The fluorine radicals partially substituted the oxygen atoms, forming a fluorine-rich layer on the surface. By contrast, the XPS spectrum of the SF6 plasma-treated sample was unaffected by fluorine-plasma etching. The quantitative XPS evaluations of F (fluorine), Y (yttrium), and O (oxygen) are listed in Table 1. Exposure to fluorocarbon plasma marginally altered the chemical composition of the SF6  Figure 5 shows the XPS spectra of yttrium atoms in the as-deposited Y 2 O 3 films and SF 6 plasma-treated Y 2 O 3 films before and after fluorocarbon plasma etching. In the curve-fitted XPS spectra of the Y 2 O 3 films, the Y3d peak was divisible into two peaks representing cations with Y3d 5/2 and Y3d 3/2 electrons. The intensity ratio of these peaks and the binding energy difference between the peaks werẽ 3:2 and~2.0 eV, respectively, consistent with the XPS analysis reported in the literature [26]. Before the as-deposited Y 2 O 3 films reacted with the fluorine plasma, the binding energies of Y3d 5/2 and Y3d 3/2 peaked at 157 and 159 eV, respectively (see Figure 5a). In the spectrum of the Y 2 O 3 films subjected to fluorine-plasma treatment, Y3d 5/2 and Y3d 3/2 deconvoluted into four binding energy peaks (Figure 5b). The two peaks at the low-binding energies (157 and 159 eV) correspond to Y-O bonding, whereas Coatings 2020, 10, 637 5 of 9 those at the higher binding energies (159.5 and 161.5 eV) are ascribed to Y-F bonding in the Y 2 O 3 films. The Y-F bonds form because fluorine atoms have a higher electronegativity than oxygen atoms (4.0 vs. 3.5). Higher electronegativity promotes electron transfer to fluorine, decreasing the electron density around the cation and hence increasing the binding energy [27,28]. In the SF 6 plasma-treated Y 2 O 3 films, the locations of the XPS peaks were less changed after exposure to fluorocarbon plasma (Figure 5c,d). This result agrees with a previous study on the fluorination mechanism of the YOF layer reaction with fluorine plasma [29].
Coatings 2020, 10, x FOR PEER REVIEW 5 of 10 plasma-treated Y2O3 film but largely changed the O 1S and F 1S contents of the as-grown Y2O3 film. Similar results were reported in a previous study of fluorocarbon plasma etching [25].   Figure 5 shows the XPS spectra of yttrium atoms in the as-deposited Y2O3 films and SF6 plasma-treated Y2O3 films before and after fluorocarbon plasma etching. In the curve-fitted XPS spectra of the Y2O3 films, the Y3d peak was divisible into two peaks representing cations with Y3d5/2 and Y3d3/2 electrons. The intensity ratio of these peaks and the binding energy difference between the peaks were ~3:2 and ~2.0 eV, respectively, consistent with the XPS analysis reported in the literature [26]. Before the as-deposited Y2O3 films reacted with the fluorine plasma, the binding energies of Y3d5/2 and Y3d3/2 peaked at 157 and 159 eV, respectively (see Figure 5a). In the spectrum of the Y2O3 films subjected to fluorine-plasma treatment, Y3d5/2 and Y3d3/2 deconvoluted into four binding energy peaks (Figure 5b). The two peaks at the low-binding energies (157 and 159 eV) correspond to Y-O bonding, whereas those at the higher binding energies (159.5 and 161.5 eV) are ascribed to Y-F bonding in the Y2O3 films. The Y-F bonds form because fluorine atoms have a higher electronegativity than oxygen atoms (4.0 vs. 3.5). Higher electronegativity promotes electron transfer to fluorine, decreasing the electron density around the cation and hence increasing the binding energy [27,28]. In the SF6 plasma-treated Y2O3 films, the locations of the XPS peaks were less changed after exposure to fluorocarbon plasma (Figure 5c,d). This result agrees with a previous study on the fluorination mechanism of the YOF layer reaction with fluorine plasma [29].    Figure 6a shows a cross-sectional TEM image of the as-deposited Y 2 O 3 films after surface irradiation by SF 6 plasma. The fluorine-concentrated altered layer was approximately 20-40 nm thick, and contained multiple phases of YOF (104) and (006). The significantly increased altered layer on the Y 2 O 3 surface is attributable to the large number of oxygen atoms in Y 2 O 3 and the many vacancy reactions with SF 6 plasma, which has a higher fluorine density than CF 4 /O 2 plasma [30]. Figure 6b,c shows the as-deposited Y 2 O 3 and the SF 6 plasma-treated Y 2 O 3 films after fluorocarbon plasma etching, respectively. After irradiation with fluorocarbon plasma, thin fluorinated altered layers were formed on the as-deposited Y 2 O 3 . This uneven fluorination layer with a disordered Moiré pattern and incomplete lattice fringes was observed near the Y 2 O 3 surface. Surface damaged was caused by chemical reactions with the fluorine plasma. The damage layer was probably responsible for the cracks and particle contamination observed in Figure 6b [31]. Meanwhile, the surface of the SF 6 plasma-treated Y 2 O 3 exhibited no fluorinated layer, because the reaction with fluorocarbon plasma was prevented by the Y-F bonds and the high concentration of fluorine in the original film (Figure 6c).
Consistent with the XPS results, the plasma-treated Y 2 O 3 films demonstrated no apparent chemical change after plasma irradiation. Therefore, the chemical erosion was slower on the SF 6 plasma-treated Y 2 O 3 films than on the as-deposited Y 2 O 3 film. Based on these observations, we proposed a mechanism explaining the fluorine-plasma etching behavior on the as-deposited Y 2 O 3 films and SF 6 plasma-treated Y 2 O 3 films (see Figure 7). The YOF on the Y 2 O 3 surface forms by decomposition of the Y-O bonds under sulfur fluoride plasma. Owing to the smaller bonding energy of S-O than of Y-O (549 kJ/mol versus 685 kJ/mol), the S-O bond reacts more efficiently with the sulfur fluoride deposited film than the Y-O bond. The SF 6 plasma-treated Y 2 O 3 film etched by CF 4 /O 2 plasma exhibited no notable chemical corrosion reactions. Therefore, the YOF/Y 2 O 3 layer effectively reduces particle generation from erosion caused by fluorocarbon plasma.

Conclusions
We compared the etching behaviors of sputter-deposited Y2O3 films and SF6 plasma-treated Y2O3 films after exposure to CF4/O2 plasma. Cross-sectional TEM observations revealed a thick YOF-altered layer on the Y2O3 surface after SF6 plasma irradiation. The surface roughness change on

Conclusions
We compared the etching behaviors of sputter-deposited Y 2 O 3 films and SF 6 plasma-treated Y 2 O 3 films after exposure to CF 4 /O 2 plasma. Cross-sectional TEM observations revealed a thick YOF-altered layer on the Y 2 O 3 surface after SF 6 plasma irradiation. The surface roughness change on the as-deposited Y 2 O 3 was probably attributable to the disordered layer after exposure to fluorocarbon plasma. As evidenced in the XPS results, fluorocarbon plasma induced fewer chemical reactions on the SF 6 plasma-treated Y 2 O 3 surface than on the as-prepared Y 2 O 3 films. The YOF film is expected to provide an excellent protective barrier against damage caused by the fluorine-plasma etching process.