Plasma Etching Behavior of YOF Coating Deposited by Suspension Plasma Spraying in Inductively Coupled CHF 3 / Ar Plasma

: Dense yttrium oxyﬂuoride (YOF) coating was successfully deposited by suspension plasma spraying (SPS) with coaxial feeding. After deposition for 6 min at a plasma power of 105 kW, the thickness of the YOF coating was 55 ± 3.2 µ m with a porosity of 0.15% ± 0.01% and the coating rate was ~9.2 µ m / min. The crystalline structure of trigonal YOF was conﬁrmed by X-ray di ﬀ ractometry (XRD). The etching behavior of the YOF coating was studied using inductively coupled CHF 3 / Ar plasma in comparison with those of the Al 2 O 3 bulk and Y 2 O 3 coating. Crater-like erosion sites and cavities were formed on the whole surface of the Al 2 O 3 bulk and Y 2 O 3 coating. In contrast, the surface of the YOF coating showed no noticeable di ﬀ erence before and after exposure to the CHF 3 / Ar plasma. Such high resistance of the YOF coating to ﬂuorocarbon plasma comes from the strongly ﬂuorinated layer on the surface. The ﬂuorination on the surface of materials was conﬁrmed by X-ray photoelectron spectrum analysis (XPS). Depth proﬁles of the compositions of Al 2 O 3 , Y 2 O 3 , and YOF samples by XPS revealed that the ﬂuorination layer of the YOF coating was much thicker than those of Al 2 O 3 and Y 2 O 3 . These results indicate that if the inner wall of the semiconductor process chamber is coated by YOF using SPS, the generation of contamination particles would be minimized during the ﬂuorocarbon plasma etching process.


Introduction
In the semiconductor manufacturing process, dry etching of dielectric SiO 2 and cleaning processes have increased due to multi-level interface connections [1]. Moreover, recently, as the incorporation of integrated circuits on wafers reaches its limit, three-dimensional vertical NAND (3D V NAND) technology, which stacks circuits in multi-layers, has been introduced and thereby the dry etching process is repeated many times in a chamber. As fluorocarbon plasma gases such as CF 4 , CHF 3 , and C 2 F 6 are typically used in the process, the inner wall of the chamber is exposed to those corrosive gases [2]. These fluorocarbon gases tend to etch the materials of the inner wall. As a result, many contamination particles are generated and the etching causes a process drift phenomenon [3]. These phenomena cause fatal problems such as a decrease in yield during mass production and the reduced operating efficiency of machines [4]. In order to prevent the etching phenomena, SiO 2 and Al 2 O 3 are used as plasma-resistant materials of the inner wall of the chamber. However, they turn out to become vulnerable to the fluorocarbon plasma gases as the number of drying etching cycles increases [1,[5][6][7][8].
Recently, Y 2 O 3 , which is more chemically stable to the fluorocarbon plasma than Al 2 O 3 and SiO 2 , has been used as a plasma-resistant material [9]. When Y 2 O 3 coating is exposed to fluorocarbon

Experimental Procedures
Commercially available Y 5 O 4 F 7 suspension (Nippon Yttrium Co., Ltd., Omuta, Fukuoka, Japan), which consists of Y 5 O 4 F 7 particles with an average size of 3 µm dispersed in deionized water with a solid concentration of 10 wt.%, was used as a feedstock material. A YOF coating was prepared using a suspension plasma spraying (SPS) system, which had triple anodes and cathodes with coaxial feeding (Mettech's Axial III, Northwest Mettech Corp., North Vancouver, BC, Canada). A plate of Al alloy 6061 (50 mm × 50 mm × 10 mm) was used as a substrate.
To improve the adhesion strength of the YOF coating, the substrate was sandblasted to have a surface roughness average (Ra) of~2.8 µm by alumina particles less than 254 µm in size. The Al alloy substrate was preheated with the plasma flame before the SPS coating process. The temperature of the substrate was found to be~380 K after the substrate was preheated, and was measured to be~550 K by a pyrometer (568 IR thermometer, Fluke, Washington, USA) after SPS coating. During the SPS process, the substrate was cooled by an air gun at a distance of 100 cm to prevent the substrate from melting in the high-temperature plasma flame. The axial III SPS system with axial feeding has a higher coating efficiency than other SPS systems with radial feeding.
As shown in Figure 1, the suspension flows directly through the plasma jet and thereby the heat generated from the plasma jet transfers efficiently to the suspension. The SPS coating was carried out under the following conditions; the flow rate of argon as a primary gas was set to 90 standard liters per minute (slm), and the flow rates of nitrogen and hydrogen as secondary gases were set to 54 and 36 slm, respectively, to induce the generation of a plasma arc with an arc current of 230 A. The role of gases is illustrated in our previous paper [11]. The feeding rate of the suspension was Coatings 2020, 10, 1023 3 of 11 45 standard cubic centimeters per minute (sccm), and the flow rate of nitrogen atomizing gas was 30 sccm. The atomizing gas, which is injected to the feeding flow and splits the suspension droplets into small ones, turned out to affect the quality of the YOF coating because it made the droplet size of the suspension smaller and thus, the particles in the droplet melted more easily [14,15]. The transverse speed of the plasma gun was 1000 mm/s, and coating cycles were repeated 20 times during the entire spraying process. The stand-off distance was 50 mm, which refers to the distance between the plasma gun and Al substrate. The details of the processing parameters of the YOF coating are shown in Table 1. The thickness of the YOF coating was about 55 µm, and the surface was polished to less than 0.1 µm of roughness for the test of fluorocarbon plasma etching. Polycrystalline Al 2 O 3 bulk and Y 2 O 3 coating deposited by SPS were prepared as a comparison group to investigate the etching behavior of the YOF coating and were polished to less than 0.1 µm surface roughness.
were 70% of the maximum allowable values of the ICP etching chamber (3000, 300 W). Dry and turbo molecular pumps were utilized to prepare the vacuum. The working pressure in the plasma etching chamber was 20 mTorr. The specimens were exposed to CHF3/Ar plasma for 60 min. The details of the plasma etching conditions are shown in Table 2.
The microstructures and surface morphologies were observed by field emission scanning electron microscopy (FE-SEM, SU-70, Hitachi, Tokyo, Japan) to analyze erosion behavior after fluorocarbon plasma etching. The crystal structure of the YOF coating by the SPS process was analyzed by high-resolution X-ray diffractometry (HR-XRD, SmartLab, Rigaku, Austin, TX, USA). In order to analyze the porosity of the cross-section of the YOF coating, an image analyzing program (ImageJ software (version 1.51k)) was used [16]. The hardness of the YOF coating was measured by a Vickers hardness tester (Duramin-40, Struers, Cleveland, USA) under a load of 200 gf. The etching depth of the specimens was measured by a noncontact three-dimensional surface profiler (NanoView-E1000, NanoSystem, Daejeon, Korea). High-resolution X-ray photoelectron spectrum (HR-XPS, AXIS SUPRA, Kratos, Manchester, UK) analysis was carried out by a monochromatic Al Kα X-ray source at a passing energy of 20 eV with a 700 µm × 400 µm spot size. The deconvolution of the photoelectron spectrum was performed by using a fitting program (Fitt-win software (version 1.3)) to analyze the spectrum of the core energy levels of the Y3d and Al2p states from the surface of YOF and Al2O3 after exposure to fluorocarbon plasma. In order to investigate the chemical compositions, depth profiling was performed with focused Ar + ions for sputtering of the etched surface of the Al2O3 bulk, Y2O3 coating, and YOF coating.   The plasma etching process was carried out by an inductively coupled plasma (ICP) etcher (Multiplex ICP, Surface Technology Systems (STS), Newport, UK) with the gases CHF 3 and Ar. Figure 2 shows a schematic of the STS multiplex ICP system. The YOF coating, Y 2 O 3 coating, and Al 2 O 3 bulk specimens were placed on a silicon wafer and loaded into the ICP etching chamber to compare their etching behavior using fluorocarbon plasma. The mixture of CHF 3 and Ar gases was injected into the ICP etching chamber for plasma generation, and the gases were supplied at a 6:1 ratio of 50 and 8.3 sccm, respectively, through a gas inlet. Helium gas was used to cool the specimens on the wafer to prevent thermal damage during the etching process and was injected into the gas inlet under the electrostatic chuck (ESC). The RF source power and bias power were set to 2100 and 210 W, respectively, which were 70% of the maximum allowable values of the ICP etching chamber (3000, 300 W). Dry and turbo molecular pumps were utilized to prepare the vacuum. The working pressure in the plasma etching chamber was 20 mTorr. The specimens were exposed to CHF 3 /Ar plasma for 60 min. The details of the plasma etching conditions are shown in Table 2.    Table 1. The FE-SEM image of the as-coated YOF surface morphology in Figure 3a shows both smooth and rough areas. After the solvent of the droplets coming out from the torch outlet is evaporated, the remaining particles will be melted by the plasma flame. The completely melted particles will be spread out onto a growing surface and produce smooth areas called splats [17]. The partially or incompletely melted particles produce rough areas.

Results and Discussion
The pores observed on the cross-section in Figure 3b came from partial or incomplete melting [15]. The porosity was 0.15% ± 0.01%, which was calculated by the image analyzing program and determined by averaging five randomly selected areas from the entire area of coating. The thickness  The microstructures and surface morphologies were observed by field emission scanning electron microscopy (FE-SEM, SU-70, Hitachi, Tokyo, Japan) to analyze erosion behavior after fluorocarbon plasma etching. The crystal structure of the YOF coating by the SPS process was analyzed by high-resolution X-ray diffractometry (HR-XRD, SmartLab, Rigaku, Austin, TX, USA). In order to analyze the porosity of the cross-section of the YOF coating, an image analyzing program (ImageJ software (version 1.51k)) was used [16]. The hardness of the YOF coating was measured by a Vickers hardness tester (Duramin-40, Struers, Cleveland, USA) under a load of 200 gf. The etching depth of the specimens was measured by a noncontact three-dimensional surface profiler (NanoView-E1000, NanoSystem, Daejeon, Korea). High-resolution X-ray photoelectron spectrum (HR-XPS, AXIS SUPRA, Kratos, Manchester, UK) analysis was carried out by a monochromatic Al Kα X-ray source at a passing energy of 20 eV with a 700 µm × 400 µm spot size. The deconvolution of the photoelectron spectrum was performed by using a fitting program (Fitt-win software (version 1.3)) to analyze the spectrum of the core energy levels of the Y3d and Al2p states from the surface of YOF and Al 2 O 3 after exposure to fluorocarbon plasma. In order to investigate the chemical compositions, depth profiling was performed with focused Ar + ions for sputtering of the etched surface of the Al 2 O 3 bulk, Y 2 O 3 coating, and YOF coating.  Table 1. The FE-SEM image of the as-coated YOF surface morphology in Figure 3a shows both smooth and rough areas. After the solvent of the droplets coming out from the torch outlet is evaporated, the remaining particles will be melted by the plasma flame. The completely melted particles will be spread out onto a growing surface and produce smooth areas called splats [17]. The partially or incompletely melted particles produce rough areas.

Results and Discussion
Coatings 2020, 10, x FOR PEER REVIEW 5 of 11 of the YOF coating in Figure 3b was 55 ± 3.2 µm and the coating rate was ~9.2 µm/min. The porosity of the YOF coating affects its hardness. The Vickers hardness of the YOF coating in Figure 3 was 553 ± 60 HV, which is much higher than 290 ± 30 HV for the YOF coating deposited by atmospheric plasma spraying (APS) reported recently by Lin et al. [18] and 69.34 ± 4 HV (0.68 ± 0.04 GPa) for the YOF coating fabricated by hot pressing reported by Tsunoura et al. [10].  Figure 4 shows the XRD peaks of the YOF coating. The major peaks indicated by the reverse triangles in Figure 4 represent the crystalline structure of trigonal YOF. The minor peaks indicated by squares and rhombuses represent, respectively, the crystalline phases of cubic and monoclinic Y2O3. Although cubic and monoclinic Y2O3 are unwanted phases, their formation was unavoidable to a certain extent. A possible scenario for the formation of Y2O3 would be as follows. First, Y5O4F7 particles in suspension would be transformed into YOF particles, while Y5O4F7 particles volatilize in the form of YF3 in the plasma jet region [19]. Then, some of the YOF would be transformed into Y2O3, while the YOF is mainly volatilized in the form of YF3 by sufficient heat energy in the plasma jet region [20]. The detailed mechanism is explained in our previous paper [11].  Figure 5 shows the FE-SEM images of the surface microstructure of the Al2O3 bulk, Y2O3 coating, and YOF coating before and after exposure to the CHF3/Ar plasma for 60 min. The Al2O3, Y2O3 coating, and YOF coating were polished to less than 0.1 µm of roughness, respectively. Furthermore, as shown in Figure 5a,c,e, they had a smooth surface before plasma exposure. After exposure to the fluorocarbon plasma, the erosion on the surface of the Al2O3 bulk was much more severe than that of the Y2O3 coating and the YOF coating. As shown in Figure 5b, large and small crater-like erosion sites were generated and contamination particles, which are shown as tiny dots in the figure, were observed on the Al2O3 surface after exposure to fluorocarbon plasma. As shown in Figure 5d, The pores observed on the cross-section in Figure 3b came from partial or incomplete melting [15]. The porosity was 0.15% ± 0.01%, which was calculated by the image analyzing program and determined by averaging five randomly selected areas from the entire area of coating. The thickness of the YOF coating in Figure 3b was 55 ± 3.2 µm and the coating rate was~9.2 µm/min. The porosity of the YOF coating affects its hardness. The Vickers hardness of the YOF coating in Figure 3 was 553 ± 60 HV, which is much higher than 290 ± 30 HV for the YOF coating deposited by atmospheric plasma spraying (APS) reported recently by Lin et al. [18] and 69.34 ± 4 HV (0.68 ± 0.04 GPa) for the YOF coating fabricated by hot pressing reported by Tsunoura et al. [10]. Figure 4 shows the XRD peaks of the YOF coating. The major peaks indicated by the reverse triangles in Figure 4 represent the crystalline structure of trigonal YOF. The minor peaks indicated by squares and rhombuses represent, respectively, the crystalline phases of cubic and monoclinic Y 2 O 3 . Although cubic and monoclinic Y 2 O 3 are unwanted phases, their formation was unavoidable to a certain extent. A possible scenario for the formation of Y 2 O 3 would be as follows. First, Y 5 O 4 F 7 particles in suspension would be transformed into YOF particles, while Y 5 O 4 F 7 particles volatilize in the form of YF 3 in the plasma jet region [19]. Then, some of the YOF would be transformed into Y 2 O 3 , while the YOF is mainly volatilized in the form of YF 3 by sufficient heat energy in the plasma jet region [20]. The detailed mechanism is explained in our previous paper [11].
Coatings 2020, 10, x FOR PEER REVIEW 5 of 11 of the YOF coating in Figure 3b was 55 ± 3.2 µm and the coating rate was ~9.2 µm/min. The porosity of the YOF coating affects its hardness. The Vickers hardness of the YOF coating in Figure 3 was 553 ± 60 HV, which is much higher than 290 ± 30 HV for the YOF coating deposited by atmospheric plasma spraying (APS) reported recently by Lin et al. [18] and 69.34 ± 4 HV (0.68 ± 0.04 GPa) for the YOF coating fabricated by hot pressing reported by Tsunoura et al. [10].  Figure 4 shows the XRD peaks of the YOF coating. The major peaks indicated by the reverse triangles in Figure 4 represent the crystalline structure of trigonal YOF. The minor peaks indicated by squares and rhombuses represent, respectively, the crystalline phases of cubic and monoclinic Y2O3. Although cubic and monoclinic Y2O3 are unwanted phases, their formation was unavoidable to a certain extent. A possible scenario for the formation of Y2O3 would be as follows. First, Y5O4F7 particles in suspension would be transformed into YOF particles, while Y5O4F7 particles volatilize in the form of YF3 in the plasma jet region [19]. Then, some of the YOF would be transformed into Y2O3, while the YOF is mainly volatilized in the form of YF3 by sufficient heat energy in the plasma jet region [20]. The detailed mechanism is explained in our previous paper [11].  Figure 5 shows the FE-SEM images of the surface microstructure of the Al2O3 bulk, Y2O3 coating, and YOF coating before and after exposure to the CHF3/Ar plasma for 60 min. The Al2O3, Y2O3 coating, and YOF coating were polished to less than 0.1 µm of roughness, respectively. Furthermore, as shown in Figure 5a,c,e, they had a smooth surface before plasma exposure. After exposure to the fluorocarbon plasma, the erosion on the surface of the Al2O3 bulk was much more severe than that of the Y2O3 coating and the YOF coating. As shown in Figure 5b, large and small crater-like erosion sites were generated and contamination particles, which are shown as tiny dots in the figure, were  coating, and YOF coating were polished to less than 0.1 µm of roughness, respectively. Furthermore, as shown in Figure 5a,c,e, they had a smooth surface before plasma exposure. After exposure to the fluorocarbon plasma, the erosion on the surface of the Al 2 O 3 bulk was much more severe than that of the Y 2 O 3 coating and the YOF coating. As shown in Figure 5b, large and small crater-like erosion sites were generated and contamination particles, which are shown as tiny dots in the figure, were observed on the Al 2 O 3 surface after exposure to fluorocarbon plasma. As shown in Figure 5d, although the Y 2 O 3 coating showed much less erosion than the Al 2 O 3 bulk, it had more cavities and deeper erosion than the YOF coating.
deeper erosion than the YOF coating.
As shown in Figure 5f, there was no noticeable difference in the surface of the YOF coating before and after exposure to the fluorocarbon plasma in comparison with those of the Al2O3 bulk and Y2O3 coating. Contamination particles were not observed. Although both the Y2O3 and YOF coatings were deposited by the SPS process under the same conditions, the YOF coating showed less erosion, which means that YOF is chemically more stable than Y2O3 against fluoride etching species, which is in agreement with the previous report by Yoshinobu et al. [5].
On the other hand, the plasma resistance of the YOF coating is closely related to its densification because the internal pores within the grains or at grain boundaries tend to be intensively eroded from their edges, forming crater-like erosion sites [21]. The crater-like erosion sites formed from internal pores were barely observed in Figure 5f, which would be due to the low porosity of 0.15% ± 0.01% of the YOF coating deposited by SPS with a uniaxial feeding system. Hence, in addition to the chemical stability of the YOF coating, it is further necessary to reduce the porosity of the YOF coating to improve its plasma resistance. The SPS process is suitable for reducing the porosity of the YOF coating.  Figure 6 shows the etched depth of the Al2O3, Y2O3, and YOF surfaces with the plasma exposure time. The etched depth was measured in nine areas of their surfaces and the average value was calculated. The etched depth of the Al2O3 surface steeply increased after 30 min in comparison with that of the Y2O3 and YOF surfaces and became more than 1025 nm after 60 min. As plasma etching progressed, crater-like erosion sites were formed. Once they were formed, the erosion seemed to be accelerated, spreading from the erosion sites as shown in Figure 6. On the other hand, the etched depths of the Y2O3 and YOF surfaces were, respectively, 289 and 142 nm after 60 min of plasma exposure and increased almost linearly with the plasma exposure time. Considering that the Al2O3, As shown in Figure 5f, there was no noticeable difference in the surface of the YOF coating before and after exposure to the fluorocarbon plasma in comparison with those of the Al 2 O 3 bulk and Y 2 O 3 coating. Contamination particles were not observed. Although both the Y 2 O 3 and YOF coatings were deposited by the SPS process under the same conditions, the YOF coating showed less erosion, which means that YOF is chemically more stable than Y 2 O 3 against fluoride etching species, which is in agreement with the previous report by Yoshinobu et al. [5].
On the other hand, the plasma resistance of the YOF coating is closely related to its densification because the internal pores within the grains or at grain boundaries tend to be intensively eroded from their edges, forming crater-like erosion sites [21]. The crater-like erosion sites formed from internal pores were barely observed in Figure 5f, which would be due to the low porosity of 0.15% ± 0.01% of the YOF coating deposited by SPS with a uniaxial feeding system. Hence, in addition to the chemical stability of the YOF coating, it is further necessary to reduce the porosity of the YOF coating to improve its plasma resistance. The SPS process is suitable for reducing the porosity of the YOF coating. Figure 6 shows the etched depth of the Al 2 O 3 , Y 2 O 3 , and YOF surfaces with the plasma exposure time. The etched depth was measured in nine areas of their surfaces and the average value was calculated. The etched depth of the Al 2 O 3 surface steeply increased after 30 min in comparison with that of the Y 2 O 3 and YOF surfaces and became more than 1025 nm after 60 min. As plasma etching progressed, crater-like erosion sites were formed. Once they were formed, the erosion seemed to be accelerated, spreading from the erosion sites as shown in Figure 6. On the other hand, the etched depths of the Y 2 O 3 and YOF surfaces were, respectively, 289 and 142 nm after 60 min of plasma exposure and increased almost linearly with the plasma exposure time. Considering that the Al 2 O 3 , Y 2 O 3 , and YOF surfaces were equally bombarded by ions of Ar plasma in the presence of highly corrosive fluorocarbon [13], the difference in the etched depth of the three samples in Figure 6 means that the YOF coating is more resistant to the CHF 3 plasma than the Y 2 O 3 coating, which is more resistant than the Al 2 O 3 bulk.
Coatings 2020, 10, x FOR PEER REVIEW 7 of 11 Y2O3, and YOF surfaces were equally bombarded by ions of Ar plasma in the presence of highly corrosive fluorocarbon [13], the difference in the etched depth of the three samples in Figure 6 means that the YOF coating is more resistant to the CHF3 plasma than the Y2O3 coating, which is more resistant than the Al2O3 bulk.  Figure 7 shows the XPS spectra of the Al2O3, Y2O3, and YOF surfaces after 60 min of CHF3/Ar plasma etching. The peak separation of XPS spectra of the three samples indicated that their surfaces were obviously fluorinated. The XPS spectra of the fluorinated surfaces, which are shown by the dotted lines in Figure 7, were fitted to two doublets, which are shown by the dashed and long dashed lines.
In Figure 7a of Al2O3 bulk, the two doublets of Al atoms from Al2O3 are Al2p3/2 and Al2p1/2 peaks, which represent 3/2 and 1/2 spins of the 2p orbital, respectively. It is reported in the handbook of XPS standards [22] that the peaks have an intensity ratio of 2:1 and peak position spacing in the binding energy of 0.4 eV. The two peaks of Al2p3/2 were located at 75.0 and 76.2 eV. The peak of lower binding energy of 75.0 eV would be for Al-O bonding of pure Al2O3 [23,24]. The peak at the higher binding energy of 76.2 eV would be for Al-F bonding, which was also confirmed by the peak of F1s located at 686.2 eV in the XPS spectrum of fluorine bonding [22].
On the other hand, the XPS spectra of Y atoms from the Y2O3 and YOF coatings with two doublets in Figure 7b,c consist of Y3d5/2 and Y3d3/2 peaks, respectively. The two peaks have an intensity ratio of 3:2 and peak position spacing in the binding energy of 2.05 eV [25]. The peaks of the lower binding energies of Y3d5/2 and Y3d3/2 from Y2O3 coating were located at 157.10 and 159.15 eV, and those of the higher binding energies of Y3d5/2 and Y3d3/2 were located at 159.40 and 161.45 eV in Figure 7b. The peaks at the lower binding energies at 157.10 and 159.15 eV correspond to the Y-O bonding, and those at the higher binding energies at 159.40 and 161.45 eV correspond to the Y-F bonding.
In Figure 7c of the YOF coating, the peaks at the lower binding energies of Y3d5/2 and Y3d3/2 at 158.35 and 160.40 eV correspond to the Y-O bonding, and those peaks at the higher binding energies of Y3d5/2 and Y3d3/2 at 159.55 and 161.60 eV correspond to the Y-F bonding. Fluorination was also confirmed at the peak of F1s located at 685.1 eV in the XPS spectrum of fluorine bonding. The peak position of the YOF coating was changed less than those of the Al2O3 bulk and Y2O3 coating after the plasma etching. The shift of the peaks of fluorine bonding to higher binding energy than those of oxygen bonding was attributed to the replacement of oxygen atoms by fluorine atoms around the  Figure 7 shows the XPS spectra of the Al 2 O 3 , Y 2 O 3 , and YOF surfaces after 60 min of CHF 3 /Ar plasma etching. The peak separation of XPS spectra of the three samples indicated that their surfaces were obviously fluorinated. The XPS spectra of the fluorinated surfaces, which are shown by the dotted lines in Figure 7, were fitted to two doublets, which are shown by the dashed and long dashed lines.
In Figure 7a of Al 2 O 3 bulk, the two doublets of Al atoms from Al 2 O 3 are Al2p 3/2 and Al2p 1/2 peaks, which represent 3/2 and 1/2 spins of the 2p orbital, respectively. It is reported in the handbook of XPS standards [22] that the peaks have an intensity ratio of 2:1 and peak position spacing in the binding energy of 0.4 eV. The two peaks of Al2p 3/2 were located at 75.0 and 76.2 eV. The peak of lower binding energy of 75.0 eV would be for Al-O bonding of pure Al 2 O 3 [23,24]. The peak at the higher binding energy of 76.2 eV would be for Al-F bonding, which was also confirmed by the peak of F1s located at 686.2 eV in the XPS spectrum of fluorine bonding [22].
On the other hand, the XPS spectra of Y atoms from the Y 2 O 3 and YOF coatings with two doublets in Figure 7b,c consist of Y3d 5/2 and Y3d 3/2 peaks, respectively. The two peaks have an intensity ratio of 3:2 and peak position spacing in the binding energy of 2.05 eV [25]. The peaks of the lower binding energies of Y3d 5/2 and Y3d 3/2 from Y 2 O 3 coating were located at 157. 10  from the higher electronegativity of fluorine-3.98 as compared to 3.44 for oxygen [26].
The relative intensity ratio of the Al-F to Al-O peaks on the Al2O3 surface was 0.19 and that of the Y-F to Y-O peaks on the Y2O3 coating was 2.70, whereas that of the Y-F to Y-O peaks on the YOF coating surface was 4.79. These values indicate that the YOF coating formed much stronger fluorination on the surface, which would be related with the excellent intrinsic chemical stability and strong erosion resistance, suppressing the generation of fluorine contamination particles after exposure to fluorocarbon plasma. Figure 7. High-resolution X-ray photoelectron spectra of (a) Al from the surface of Al2O3, (b) Y from the surface of the Y2O3 coating, and (c) Y from the surface of the YOF coating after exposure to CHF3/Ar plasma for 60 min. Figure 8 shows the XPS depth profiles of compositions as a function of the Ar + sputtering time from the surface of the Al2O3, Y2O3, and YOF samples after exposure to CHF3/Ar plasma for 60 min. Fluorination was confirmed on all the surfaces of the Al2O3, Y2O3, and YOF, while the fluorine content on the fluorinated layer of the YOF coating was higher than those of the Al2O3 and Y2O3. The percentages of F atoms reached the maximum values of 58.97% on the fluorinated layer of the Y2O3 coating ( Figure 8b) and 17.76% on the Al2O3 bulk (Figure 8a), while the percentage reached the highest value of 67.75% on the YOF coating surface (Figure 8c). Furthermore, the percentage of F atoms started to decrease from the maximum value to the original composition with sputtering time. Given that the fluorine content decreased from 17.76% and approached 0% with sputtering time in Figure 8a, the Al2O3 bulk surface appeared to be barely fluorinated. In addition, the fluorine content on the Y2O3 coating surface was rapidly decreased compared to that on the YOF coating surface (Figure 8b,c). These results indicated that the fluorination layer of the YOF coating was much thicker than that of the Y2O3 coating. This fluorination layer, with a higher concentration of fluorine on the In Figure 7c of the YOF coating, the peaks at the lower binding energies of Y3d 5/2 and Y3d 3/2 at 158.35 and 160.40 eV correspond to the Y-O bonding, and those peaks at the higher binding energies of Y3d 5/2 and Y3d 3/2 at 159.55 and 161.60 eV correspond to the Y-F bonding. Fluorination was also confirmed at the peak of F1s located at 685.1 eV in the XPS spectrum of fluorine bonding. The peak position of the YOF coating was changed less than those of the Al 2 O 3 bulk and Y 2 O 3 coating after the plasma etching. The shift of the peaks of fluorine bonding to higher binding energy than those of oxygen bonding was attributed to the replacement of oxygen atoms by fluorine atoms around the cations during the etching process [26]. Fluorine bonding being stronger than oxygen bonding comes from the higher electronegativity of fluorine-3.98 as compared to 3.44 for oxygen [26].
The relative intensity ratio of the Al-F to Al-O peaks on the Al 2 O 3 surface was 0.19 and that of the Y-F to Y-O peaks on the Y 2 O 3 coating was 2.70, whereas that of the Y-F to Y-O peaks on the YOF coating surface was 4.79. These values indicate that the YOF coating formed much stronger fluorination on the surface, which would be related with the excellent intrinsic chemical stability and strong erosion resistance, suppressing the generation of fluorine contamination particles after exposure to fluorocarbon plasma. Figure 8 shows the XPS depth profiles of compositions as a function of the Ar + sputtering time from the surface of the Al 2 O 3 , Y 2 O 3 , and YOF samples after exposure to CHF 3 /Ar plasma for 60 min.  (Figure 8c). Furthermore, the percentage of F atoms started to decrease from the maximum value to the original composition with sputtering time. Given that the fluorine content decreased from 17.76% and approached 0% with sputtering time in Figure 8a, the Al 2 O 3 bulk surface appeared to be barely fluorinated. In addition, the fluorine content on the Y 2 O 3 coating surface was rapidly decreased compared to that on the YOF coating surface (Figure 8b,c). These results indicated that the fluorination layer of the YOF coating was much thicker than that of the Y 2 O 3 coating. This fluorination layer, with a higher concentration of fluorine on the surface, plays an important role in preventing erosion from fluorocarbon plasma and would not easily vaporize because of its chemical stability during the etching process [1,27].
Coatings 2020, 10, x FOR PEER REVIEW 9 of 11 surface, plays an important role in preventing erosion from fluorocarbon plasma and would not easily vaporize because of its chemical stability during the etching process [1,27]. On the other hand, the carbon content sharply decreased from the surface to the depth direction with the sputtering time in all of the Al2O3, Y2O3, and YOF samples, which indicated that the fluorocarbon layer is very thin, as shown in Figure 8. The thin fluorocarbon layer formed on the surface of the samples was previously reported in Si-based materials etched after exposure to fluorocarbon plasma [28]. From these results, we could confirm that the fluorination layer of the YOF coating was much thicker than those of the Al2O3 and Y2O3 samples. This would be why the YOF coating exhibits excellent plasma-resistance to CHF3/Ar plasma etching.  On the other hand, the carbon content sharply decreased from the surface to the depth direction with the sputtering time in all of the Al 2 O 3 , Y 2 O 3 , and YOF samples, which indicated that the fluorocarbon layer is very thin, as shown in Figure 8. The thin fluorocarbon layer formed on the surface of the samples was previously reported in Si-based materials etched after exposure to fluorocarbon plasma [28]. From these results, we could confirm that the fluorination layer of the YOF coating was much thicker than those of the Al 2 O 3 and Y 2 O 3 samples. This would be why the YOF coating exhibits excellent plasma-resistance to CHF 3 /Ar plasma etching.

Conclusions
The plasma etching behavior of the YOF coating deposited by SPS was studied using inductively coupled CHF 3 /Ar plasma in comparison with those of Al 2 O 3 bulk and Y 2 O 3 coating. When the Al 2 O 3 bulk and Y 2 O 3 coating were exposed to fluorocarbon plasma, crater-like erosion sites and cavities were formed on the whole surface. In the case of the YOF coating, however, there was no obvious difference of the surface before and after the plasma etching. The etched depth of the Al 2 O 3 bulk was about 3.5 times deeper than that of the Y 2 O 3 coating, and the etched depth of the Y 2 O 3 coating was about twice that of the YOF coating. Such high erosion resistance of the YOF coating to fluorocarbon plasma comes from the formation of a strongly fluorinated layer on the surface. The YOF coating deposited by SPS is expected to minimize the generation of contamination particles by preventing erosion of the inner wall of the semiconductor process chamber by fluorocarbon plasma.