Change in Electrical/Mechanical Properties of Plasma Polymerized Low Dielectric Constant Films after Etching in CF4/O2 Plasma for Semiconductor Multilevel Interconnects

As semiconductor chips have been integrated to enhance their performance, a low-dielectric-constant material, SiCOH, with a relative dielectric constant k ≤ 3.5 has been widely used as an intermetal dielectric (IMD) material in multilevel interconnects to reduce the resistance-capacitance delay. Plasma-polymerized tetrakis(trimethylsilyoxy)silane (ppTTMSS) films were created using capacitively coupled plasma-enhanced chemical vapor deposition with deposition plasma powers ranging from 20 to 60 W and then etched in CF4/O2 plasma using reactive ion etching. No significant changes were observed in the Fourier-transform infrared spectroscopy (FTIR) spectra of the ppTTMSS films after etching. The refractive index and dielectric constant were also maintained. As the deposition plasma power increased, the hardness and elastic modulus increased with increasing ppTTMSS film density. The X-ray photoelectron spectroscopy (XPS) spectra analysis showed that the oxygen concentration increased but the carbon concentration decreased after etching owing to the reaction between the plasma and film surface. With an increase in the deposition plasma power, the hardness and elastic modulus increased from 1.06 to 8.56 GPa and from 6.16 to 52.45 GPa. This result satisfies the hardness and elastic modulus exceeding 0.7 and 5.0 GPa, which are required for the chemical–mechanical polishing process in semiconductor multilevel interconnects. Furthermore, all leakage-current densities of the as-deposited and etched ppTTMSS films were measured below 10−6 A/cm2 at 1 MV/cm, which is generally acceptable for IMD materials.


Introduction
As semiconductor chips have become more integrated, their pitch size has been reduced to improve their performance. This reduction in pitch size can decrease the gate delay in field-effect transistors [1,2]. However, the reduced pitch size causes an increase in the resistance-capacitance delay (RC delay), which results in the degradation of the performance of semiconductor chips. To resolve this problem, aluminum (Al) has been replaced by copper (Cu) because the resistivity of Cu (1.67 µΩ·cm) is smaller than that of Al (2.7 µΩ·cm). Traditional interlayer dielectric silicon oxide (SiO 2 ) has been replaced by low dielectric constant (low-k) materials with a relative dielectric constant below 3.9 in multilevel interconnects [3,4]. As a low-k material, SiCOH films have been widely employed by plasma-enhanced chemical vapor deposition (PECVD) of precursors composed of Si-O and hydrocarbon functional groups [5][6][7][8][9]. In our previous work, tetrakis(trimethylsilyoxy)silane (TTMSS, C 12 H 36 O 4 Si 5 ) was introduced as the precursor to fabricate SiCOH films [10][11][12][13]. It Materials 2023, 16, 4663 2 of 12 was reported that the SiCOH films showed low-k values in the range of 2.1 to 3.57 with good mechanical properties ranging from 7.12 to 41.4 GPa [11]. SiCOH films are referred to as plasma-polymerized tetrakis(trimethylsilyoxy)silane (ppTTMSS) films because plasma sources are used to generate a gas discharge that provides energy to activate or fragment molecular precursors and deposit polymer thin films. On the other hand, Cu dry etching is extremely challenging because it is difficult to remove byproducts (chlorides or fluorides) that are nonvolatile and have high boiling points above 1000 • C. Thus, the interconnect integration changed from metal (Cu) patterning followed by dielectric filling to dielectric patterning followed by metal filling with Cu. This is the so-called "damascene" process, which has raised the demand for low-k patterning by plasma etching. Unfortunately, the damascene process requires plasma exposure for etching, which can damage low-k films [14]. As a result, low-k films exhibit increased k values and leakage-current densities after plasma exposure [15]. In this study, the following investigations were performed to evaluate the suitability of ppTTMSS films as intermetal dielectric (IMD) materials, even after plasma exposure for etching. The effect of reactive ion etching (RIE) on ppTTMSS films was investigated to understand the changes in the chemical properties of the etched film surface. The RIE process was performed using plasma with CF 4 and O 2 gases. The etch rate was related to the chemical structures of the ppTTMSS films depending on the deposition plasma power. The refractive index, electrical properties including k value and leakage-current density, and mechanical properties including hardness and elastic modulus were also studied to evaluate the suitability of the IMD materials [16,17].

Materials and Methods
The ppTTMSS films were deposited using a capacitively-coupled plasma-enhanced chemical vapor deposition (PEVCD) system with a 13.56 MHz radio frequency (RF) power supply. A TTMSS precursor (Sigma Aldrich, St. Louis, MO, USA, 99% purity) was used. TTMSS has a cross-shaped structure composed of four oxygen atoms attached to a central silicon atom, and each oxygen atom is bonded to another silicon atom linked with three CH 3 molecules. Two types of 4-inch Si wafers were used as substrates: phosphorus-doped n-type Si (100) with a resistivity of 1-10 Ω·cm and highly boron-doped p++-type Si (100) with a resistivity of less than 0.005 Ω·cm. They were then cleaned by ultrasonication in acetone and ethanol for 5 min each. To vaporize the molecules from the precursors, a bubbler containing the TTMSS precursor was heated to 95 • C, and an argon (Ar) gas flow rate of 60 sccm with a purity of 99.999% was introduced into the bubbler as a carrier gas, which transported the vaporized precursors to the process chamber. The ppTTMSS films were deposited at 25 • C and a working pressure of 80 Pa. Under the same conditions, the chamber pressure, when only Ar carrier gas was supplied to the chamber at a flow rate of 60 sccm, was 50 Pa. The deposition plasma power was varied in the range of 20 to 60 W. Figure 1 illustrates the RIE process of ppTTMSS film. The carbon tetrafluoride/oxygen (CF 4 /O 2 ) gas injected into the chamber is activated by plasma to create radicals such as O and F and the active species react with the surface of the thin film to cause etching of the ppTTMSS film surface. Both the plasma reaction and chemical reaction in the chamber occur as indicated at the bottom of the figure. After the reaction, byproducts, such as SiF x and CH x , could be evacuated from the chamber through the pump. The ppTTMSS films were etched for 5 min using the RIE process with carbon tetrafluoride/oxygen (CF 4 /O 2 ) gas chemistry at gas flow rates of 5/5 sccm. The pressure in the RIE process chamber was 96 Pa. A capacitance-coupled plasma system was used for the RIE at 13.56 MHz. The bias power was maintained at 20 W.  The thickness and refractive index (n) of the ppTTMSS films were measured using an ellipsometer (M-2000, J.A. Woolam Co., Lincoln, NE, USA). The thicknesses of the ppTTMSS films were confirmed by field-emission scanning electron microscopy (FESEM; JSM-7600F, Jeol, Tokyo, Japan). To investigate the chemical composition of the ppTTMSS films, X-ray photoelectron spectroscopy (XPS; VG Microtech, ESCA2000, London, UK) was employed with a twin anode (13 kV) of Al-Kα (1486.6 eV) and Ma-Kα (1253.6 eV). The beam and filament currents of the X-ray source were 15 mA and 4.5 A, respectively, and the scanning step was 0.1 eV. Owing to the high beam energy, XPS measurements were conducted at a depth of ~10 nm from the surface without Ar sputtering. Fourier transform infrared (FTIR) spectroscopy (Bruker, VERTEX7, Billerica, MA, USA) was used to determine the chemical structures of the ppTTMSS films. FTIR absorption spectra were scanned 64 times with wavenumbers ranging from 4000 to 600 cm −1 with a resolution of 4 cm −1 . The peaks obtained from the FTIR and XPS spectra were deconvoluted into their constituent peaks using Gaussian peak fitting with OriginPro software. The analysis of chemical and optical properties mentioned above used n-type Si-wafer samples. To investigate the electrical properties of the ppTTMSS films, a metal/insulator/metallic silicon wafer (MIS) structure consisting of Al/ppTTMSS/p++-Si was used. To measure the k value of the ppTTMSS films, an inductance, capacitance, and resistance (LCR) meter (4287 A, Agilent, Santa Clara, CA, USA) was employed at an alternating current power source with a frequency of 1 MHz and voltage of 100 mV. The leakage-current densities of ppTTMSS films were measured using an electrometer (6617 B, Keithley, Cleveland, OH, USA). The hardness and elastic modulus were measured by a load-and depth-sensing indentation technique with a nanoindenter (NanoTest Vantage Platform, Micro Materials, Anaheim, CA, USA) using an n-type Si wafer sample. The nanoindenter was measured at a depth of 30 to 40% of the film thickness to avoid the influence of surface oxide film and Si wafer. Figure 2 shows the deposition and etch rates of the ppTTMSS films as a function of the deposition plasma power. As the deposition plasma power increased from 20 to 60W, the as-deposited samples were fabricated at 553, 542, and 438 nm, and the deposition rate decreased from 0.92 to 0.73 nm/s. After the RIE process, the thickness of the ppTTMSS films decreased to 513, 510, and 409 nm, and the etching rate decreased from 1.33 to 0.96 nm/s. An increase in the deposition plasma power can cause an increase in the density of The thickness and refractive index (n) of the ppTTMSS films were measured using an ellipsometer (M-2000, J.A. Woolam Co., Lincoln, NE, USA). The thicknesses of the ppTTMSS films were confirmed by field-emission scanning electron microscopy (FESEM; JSM-7600F, Jeol, Tokyo, Japan). To investigate the chemical composition of the ppTTMSS films, X-ray photoelectron spectroscopy (XPS; VG Microtech, ESCA2000, London, UK) was employed with a twin anode (13 kV) of Al-Kα (1486.6 eV) and Ma-Kα (1253.6 eV). The beam and filament currents of the X-ray source were 15 mA and 4.5 A, respectively, and the scanning step was 0.1 eV. Owing to the high beam energy, XPS measurements were conducted at a depth of~10 nm from the surface without Ar sputtering. Fourier transform infrared (FTIR) spectroscopy (Bruker, VERTEX7, Billerica, MA, USA) was used to determine the chemical structures of the ppTTMSS films. FTIR absorption spectra were scanned 64 times with wavenumbers ranging from 4000 to 600 cm −1 with a resolution of 4 cm −1 . The peaks obtained from the FTIR and XPS spectra were deconvoluted into their constituent peaks using Gaussian peak fitting with OriginPro software. The analysis of chemical and optical properties mentioned above used n-type Si-wafer samples. To investigate the electrical properties of the ppTTMSS films, a metal/insulator/metallic silicon wafer (MIS) structure consisting of Al/ppTTMSS/p++-Si was used. To measure the k value of the ppTTMSS films, an inductance, capacitance, and resistance (LCR) meter (4287 A, Agilent, Santa Clara, CA, USA) was employed at an alternating current power source with a frequency of 1 MHz and voltage of 100 mV. The leakage-current densities of ppTTMSS films were measured using an electrometer (6617 B, Keithley, Cleveland, OH, USA). The hardness and elastic modulus were measured by a load-and depth-sensing indentation technique with a nanoindenter (NanoTest Vantage Platform, Micro Materials, Anaheim, CA, USA) using an n-type Si wafer sample. The nanoindenter was measured at a depth of 30 to 40% of the film thickness to avoid the influence of surface oxide film and Si wafer. Figure 2 shows the deposition and etch rates of the ppTTMSS films as a function of the deposition plasma power. As the deposition plasma power increased from 20 to 60 W, the as-deposited samples were fabricated at 553, 542, and 438 nm, and the deposition rate decreased from 0.92 to 0.73 nm/s. After the RIE process, the thickness of the ppTTMSS films decreased to 513, 510, and 409 nm, and the etching rate decreased from 1.33 to 0.96 nm/s. An increase in the deposition plasma power can cause an increase in the density of ppTTMSS films by forming a SiO 2 -like structure [10,11]. The SiO 2 -like ppTTMSS film is developed by the competition between ablation and polymerization (CAP) mechanisms during the process of plasma polymerization [18,19]. Two processes occur simultaneously: ablation, which removes surface molecules, and polymerization, which deposits the surface monomer. These two processes are in competition, and the deposition rate can increase or decrease depending on whether the polymerization process or the ablation process is dominant. It was also found that the polymer deposition hardly occurred at a very low flow rate and the deposition rate decreased with increasing deposition plasma power [18,19]. Two major factors that influence the balance between polymer formation and ablation are considered to be the control of ablation due to reactive species by chemical reactions and plasma conditions, particularly the plasma energy density [18,19]. In this study, as the deposition plasma power increased from 20 W to 60 W, the deposition rate decreased. Thus, it is likely that the decreased deposition rate was caused by either a low flow rate or ablation in the plasma. However, the deposition plasma power could affect the changes in the chemical-bonding configurations of the films during deposition. The dissociation energy of chemical bonds depends on the bond type: C-H (3.5 eV), Si-C (4.7 eV), and Si-O (8.3 eV) [14,20]. More hydrocarbon-related molecules induced by C-H and Si-C were considered to appear in the films. However, as they increased, a large portion of the molecular precursors decomposed into small fragments of Si, O, and C, leading to SiO 2 -like films. It is possible that more SiO 2 -like structures, rather than carbon-related structures, induced lower etch rates at a higher deposition plasma power. ppTTMSS films by forming a SiO2-like structure [10,11]. The SiO2-like ppTTMSS film is developed by the competition between ablation and polymerization (CAP) mechanisms during the process of plasma polymerization [18,19]. Two processes occur simultaneously: ablation, which removes surface molecules, and polymerization, which deposits the surface monomer. These two processes are in competition, and the deposition rate can increase or decrease depending on whether the polymerization process or the ablation process is dominant. It was also found that the polymer deposition hardly occurred at a very low flow rate and the deposition rate decreased with increasing deposition plasma power [18,19]. Two major factors that influence the balance between polymer formation and ablation are considered to be the control of ablation due to reactive species by chemical reactions and plasma conditions, particularly the plasma energy density [18,19]. In this study, as the deposition plasma power increased from 20 W to 60 W, the deposition rate decreased. Thus, it is likely that the decreased deposition rate was caused by either a low flow rate or ablation in the plasma. However, the deposition plasma power could affect the changes in the chemical-bonding configurations of the films during deposition. The dissociation energy of chemical bonds depends on the bond type: C-H (3.5 eV), Si-C (4.7 eV), and Si-O (8.3 eV) [14,20]. More hydrocarbon-related molecules induced by C-H and Si-C were considered to appear in the films. However, as they increased, a large portion of the molecular precursors decomposed into small fragments of Si, O, and C, leading to SiO2-like films. It is possible that more SiO2-like structures, rather than carbon-related structures, induced lower etch rates at a higher deposition plasma power.  The absorption band in the range of 3100 to 2800 cm −1 corresponds to the C-Hx (x = 2, 3) stretching vibration mode. The peak at 1300 to 1260 cm −1 originated from the Si-CH3 bond. The absorption band centered at 1260 to 950 cm −1 was assigned to the Si-O-Si stretching vibration mode. The absorption band at 900 to 750 cm −1 corresponded to the Si-(CH3)x (x = 1, 2, 3) bending vibration [21][22][23][24]. The Si-O-Si stretching vibration mode had the highest intensity among these peaks because the ppTTMSS films were mostly composed of siloxane induced from the molecular structure of the TTMSS precursor. As the deposition plasma power increased, the Si-O-Si stretching mode became dominant, and the intensities of the three hydrocarbon-related peaks diminished. There were few changes in these peaks after the etching of the ppTTMSS films and their peak-area ratios were analyzed for comparison. Figure 3b,c represent the peak-area ratios of the Si-O-Si and hydrocarbonrelated peaks of the as-deposited and etched ppTTMSS films, respectively. For as-deposited ppTTMSS films, the peak-area ratio of Si-O-Si increased from 71.76 to 88.83%, and  Figure 3a shows the FTIR absorption spectra of the as-deposited and etched ppTTMSS films as a function of the deposition plasma power in the range of 4000 to 400 cm −1 . The absorption band in the range of 3100 to 2800 cm −1 corresponds to the C-Hx (x = 2, 3) stretching vibration mode. The peak at 1300 to 1260 cm −1 originated from the Si-CH 3 bond. The absorption band centered at 1260 to 950 cm −1 was assigned to the Si-O-Si stretching vibration mode. The absorption band at 900 to 750 cm −1 corresponded to the Si-(CH 3 )x (x = 1, 2, 3) bending vibration [21][22][23][24]. The Si-O-Si stretching vibration mode had the highest intensity among these peaks because the ppTTMSS films were mostly composed of siloxane induced from the molecular structure of the TTMSS precursor. As the deposition plasma power increased, the Si-O-Si stretching mode became dominant, and the intensities of the three hydrocarbon-related peaks diminished. There were few changes in these peaks after the etching of the ppTTMSS films and their peak-area ratios were analyzed for comparison. Figure 3b 88.83%, and those of hydrocarbon-related peaks, including Si-(CH 3 )x (x = 1, 2, 3), Si-CH 3 and C-Hx (x = 2, 3), decreased with an increase in the deposition plasma power from 20 to 60 W. Furthermore, the peak-area ratios of Si-(CH 3 )x, Si-CH 3 , and C-Hx decreased from 19.68 to 9.69%, from 4.28 to 0.42%, and from 4.28 to 1.06%, respectively. For etched ppTTMSS films, with an increase in deposition plasma power, trends in peak-area ratios, similar to those of the as-deposited films, were observed. As the deposition plasma power increased from 20 to 60 W, the peak-area ratio of Si-O-Si increased from 71.11 to 88.40% and those of Si-(CH 3 ), Si-CH 3 , and C-Hx decreased from 19.65 to 10.11%, from 4.02 to 0.37%, and from 4.22 to 1.12%, respectively. those of hydrocarbon-related peaks, including Si-(CH3)x (x = 1, 2, 3), Si-CH3 and C-Hx (x = 2, 3), decreased with an increase in the deposition plasma power from 20 to 60 W. Furthermore, the peak-area ratios of Si-(CH3)x, Si-CH3, and C-Hx decreased from 19.68 to 9.69%, from 4.28 to 0.42%, and from 4.28 to 1.06%, respectively. For etched ppTTMSS films, with an increase in deposition plasma power, trends in peak-area ratios, similar to those of the as-deposited films, were observed. As the deposition plasma power increased from 20 to 60 W, the peak-area ratio of Si-O-Si increased from 71.11 to 88.40% and those of Si-(CH3), Si-CH3, and C-Hx decreased from 19.65 to 10.11%, from 4.02 to 0.37%, and from 4.22 to 1.12%, respectively.   [21,25]. The peak located at approximately 1030 cm −1 was allocated to the vibration of the suboxide structure with a bonding angle lower than 144° [26][27][28]. The peak at 1070 cm −1 was assigned to the vibration of the network structure with a bonding angle of ~144° [26,28]. The peak at 1140 cm −1 was ascribed to the vibration of the cage structure with a bonding angle larger than 144° [29]. As the deposition plasma power increased, the suboxide structure became dominant, whereas the network structure became more recessive. The peak-area ratios of the deconvoluted Si-O-Si peaks for both the asdeposited and etched ppTTMSS films are shown in Table 1. As the deposition plasma power increased from 20 to 60 W, the peak-area ratios of suboxide structure for the asdeposited ppTTMSS films increased from 27 to 56%, whereas those of the network and cage structures decreased from 45 to 22% and from 28 to 22%, respectively. The peak-area ratios remained almost the same for the etched ppTTMSS films.   [21,25]. The peak located at approximately 1030 cm −1 was allocated to the vibration of the suboxide structure with a bonding angle lower than 144 • [26][27][28]. The peak at 1070 cm −1 was assigned to the vibration of the network structure with a bonding angle of~144 • [26,28]. The peak at 1140 cm −1 was ascribed to the vibration of the cage structure with a bonding angle larger than 144 • [29]. As the deposition plasma power increased, the suboxide structure became dominant, whereas the network structure became more recessive. The peak-area ratios of the deconvoluted Si-O-Si peaks for both the as-deposited and etched ppTTMSS films are shown in Table 1. As the deposition plasma power increased from 20 to 60 W, the peak-area ratios of suboxide structure for the as-deposited ppTTMSS films increased from 27 to 56%, whereas those of the network and cage structures decreased from 45 to 22% and from 28 to 22%, respectively. The peak-area ratios remained almost the same for the etched ppTTMSS films.

Suboxide Network Cage
As -deposited ppTTMSS   20  27  45  28  40  45  31  23  60  56  22  22   Etched ppTTMSS   20  27  45  28  40  45  31  23  60  56  22  22 The XPS spectra of the ppTTMSS films were studied to investigate the effect of CF4/O2 etching on their surface chemistry. Figure 5a,b are plotted as the normalized C1s peaks of the as-deposited and etched ppTTMSS films, respectively. Figure 5c,d show the normalized Si2p peaks of the as-deposited and etched ppTTMSS films, respectively. There were little changes in the C1s and Si2p peaks of the as-deposited films when the deposition plasma power increased from 20 to 60 W. For etched ppTTMSS films, a shoulder peak at approximately 287 eV was observed in the C1s peaks, and chemical shifts to a higher binding energy occurred with 104.03, 103.57, and 102.97 eV for 20, 40, and 60 W in the Si2p peaks of etched ppTTMSS films, respectively. The chemical shifts of the Si2p peaks could be caused by an increased link with oxygen and fluorine because oxygen and fluorine have more electron affinity than carbon and silicon [30]. The atomic concentrations and peak-area ratios of the deconvoluted C1s peaks of the as-deposited and etched ppTTMSS films are presented in Table 2. Consistent with the XPS spectra of the as-deposited ppTT-MSS films, there were little changes in the atomic concentrations for s ranging from 20 to 60 W. When the deposition plasma power increased from 20 to 60 W for as-deposited ppTTMSS films, the concentration of oxygen increased from 28.37 to 30.96%, whereas the concentrations of silicon and carbon decreased from 28.11 to 26.56% and 43.52 to 42.48%, The XPS spectra of the ppTTMSS films were studied to investigate the effect of CF 4 /O 2 etching on their surface chemistry. Figure 5a,b are plotted as the normalized C1s peaks of the as-deposited and etched ppTTMSS films, respectively. Figure 5c,d show the normalized Si2p peaks of the as-deposited and etched ppTTMSS films, respectively. There were little changes in the C1s and Si2p peaks of the as-deposited films when the deposition plasma power increased from 20 to 60 W. For etched ppTTMSS films, a shoulder peak at approximately 287 eV was observed in the C1s peaks, and chemical shifts to a higher binding energy occurred with 104.03, 103.57, and 102.97 eV for 20, 40, and 60 W in the Si2p peaks of etched ppTTMSS films, respectively. The chemical shifts of the Si2p peaks could be caused by an increased link with oxygen and fluorine because oxygen and fluorine have more electron affinity than carbon and silicon [30]. The atomic concentrations and peak-area ratios of the deconvoluted C1s peaks of the as-deposited and etched ppTTMSS films are presented in Table 2. Consistent with the XPS spectra of the as-deposited ppTTMSS films, there were little changes in the atomic concentrations for s ranging from 20 to 60 W. When the deposition plasma power increased from 20 to 60 W for as-deposited ppTTMSS films, the concentration of oxygen increased from 28.37 to 30.96%, whereas the concentrations of silicon and carbon decreased from 28.11 to 26.56% and 43.52 to 42.48%, respectively. For etched ppTTMSS films, however, a relatively higher concentration of oxygen was measured, which decreased from 53.67 to 42.24% as the deposition plasma power increased from 20 to 60 W. In comparison with the as-deposited ppTTMSS films, a dramatic reduction in carbon concentration was observed, which increased from 14.92 to 31.00% as the deposition plasma power increased from 20 to 60 W. The concentration of silicon did not change much compared to those of carbon and oxygen after the RIE process, decreasing from 29.17 to 24.74% as the deposition plasma power increased from 20 to 60 W. Additionally, the concentration of fluorine was measured in the range of 1.99 to 2.24% owing to the CF 4 /O 2 plasma etching. The decrease in carbon concentration and increase in oxygen concentration after the RIE process can be explained by the reaction between the CF 4 /O 2 plasma and the film surface. For etching ppTTMSS films using CF 4 plasma, it is desirable to form F-radicals through electron impacts in the plasma and volatile Si compounds such as SiFx. This requires breaking the Si-O, Si-C, and C-F bonds and the formation of Si-F bonds. By adding oxygen to the plasma, carbon can be removed via the formation of CO and CO 2 gases, which are easily pumped away. This decreases the amount of carbon available to form CFx radicals, increasing the relative F concentration in the plasma and the etch rate. Meanwhile, oxygen can create SiO 2 on the surface, resulting in slower etching in CF 4 chemistry. The etch rate is not strictly proportional to the concentration of F radicals because of the competition between F atoms and O atoms for the active Si surface site. crease in oxygen concentration after the RIE process can be explained by the reaction be tween the CF4/O2 plasma and the film surface. For etching ppTTMSS films using CF plasma, it is desirable to form F-radicals through electron impacts in the plasma and vol atile Si compounds such as SiFx. This requires breaking the Si-O, Si-C, and C-F bonds and the formation of Si-F bonds. By adding oxygen to the plasma, carbon can be removed via the formation of CO and CO2 gases, which are easily pumped away. This decreases the amount of carbon available to form CFx radicals, increasing the relative F concentration in the plasma and the etch rate. Meanwhile, oxygen can create SiO2 on the surface, result ing in slower etching in CF4 chemistry. The etch rate is not strictly proportional to the concentration of F radicals because of the competition between F atoms and O atoms fo the active Si surface site.   Figure 6a,b show the deconvoluted C1s peaks of the as-deposited and etched ppTT MSS films, respectively. Table 3 shows the peak-area ratios of the deconvoluted C1s peaks of the as-deposited and etched ppTTMSS films. Subpeaks C-Si, C-C/C-H, and C-O were defined by their associated binding energies of 283.8, 284.8, and 286.3 eV for as-deposited ppTTMSS films, and additional C-CF was observed at a binding energy of 287.0 eV fo etched ppTTMSS films [31]. The subpeak of C-H/C-C was predominant with peak-area ratios of 79 to 80%, followed by 12% of C-Si and 8-9% of C-O for as-deposited films. Afte   Figure 6a,b show the deconvoluted C1s peaks of the as-deposited and etched ppTTMSS films, respectively. Table 3 shows the peak-area ratios of the deconvoluted C1s peaks of the as-deposited and etched ppTTMSS films. Subpeaks C-Si, C-C/C-H, and C-O were defined by their associated binding energies of 283.8, 284.8, and 286.3 eV for as-deposited ppTTMSS films, and additional C-CF was observed at a binding energy of 287.0 eV for etched ppTTMSS films [31]. The subpeak of C-H/C-C was predominant with peak-area ratios of 79 to 80%, followed by 12% of C-Si and 8-9% of C-O for as-deposited films. After CF 4 /O 2 plasma etching, there was little change within the range of 12 to 14% in the peak-area ratios of the C-Si peak. However, the peak-area ratios of the oxygen-related C-O peak increased, ranging from 16 to 33%, which is consistent with the increase in oxygen concentration after CF 4 /O 2 plasma etching. The concentration of oxygen decreased from 53.67 to 42.24%, and the peak-area ratio of C-O decreased from 33 to 16% with an increase in the deposition plasma power from 20 to 60 W. The peak-area ratios of the carbon-related C-C/C-H peaks decreased, ranging from 45 to 69%, which is in accordance with the decrease in carbon concentration after CF 4 /O 2 plasma etching. The concentration of carbon increased from 14.92 to 31.00%, and the peak-area ratio of C-C/C-H increased from 45 to 69% with an increase in the deposition plasma power from 20 to 60 W. The C-CF peak also appeared owing to CF 4 chemistry and its peak-area ratios decreased from 8 to 2% as the deposition plasma power increased from 20 to 60 W. C-C/C-H peaks decreased, ranging from 45 to 69%, which is in accordance with the decrease in carbon concentration after CF4/O2 plasma etching. The concentration of carbon increased from 14.92 to 31.00%, and the peak-area ratio of C-C/C-H increased from 45 to 69% with an increase in the deposition plasma power from 20 to 60 W. The C-CF peak also appeared owing to CF4 chemistry and its peak-area ratios decreased from 8 to 2% as the deposition plasma power increased from 20 to 60 W.  As-deposited ppTTMSS 20 12 80 8  -40  12  79  9  -60  12  80  8  -Etched ppTTMSS   20  14  45  33  8  40  13  56  28  3  60 13 69 16 2 Figure 7a,b present the refractive index (n) and relative dielectric constant (k) of the as-deposited and etched ppTTMSS films, respectively. The n values of as-deposited ppTT-MSS films increased from 1.46 to 1.60 with an increase in the deposition plasma power from 20 to 60 W. After etching the ppTTMSS films, the n values were slightly lower than those of the as-deposited films, increasing from 1.45 to 1.58 when the deposition plasma power increased from 20 to 60 W. It has been reported that the n value is closely related   Figure 7a,b present the refractive index (n) and relative dielectric constant (k) of the asdeposited and etched ppTTMSS films, respectively. The n values of as-deposited ppTTMSS films increased from 1.46 to 1.60 with an increase in the deposition plasma power from 20 to 60 W. After etching the ppTTMSS films, the n values were slightly lower than those of the as-deposited films, increasing from 1.45 to 1.58 when the deposition plasma power increased from 20 to 60 W. It has been reported that the n value is closely related to the density of the films [32,33]. It is also known that the refractive index is affected by the carbon content of the film; an increase in the refractive index could result from a decrease in the carbon content. This is consistent with the fact that carbon concentration decreased from 43.52 to 42.48% when the deposition plasma power increased from 20 to 60 W. It was noted that the refractive index did not change significantly for etched ppTTMSS films, although the carbon content decreased from 14.92 to 31.00%. This could be caused by the change in the surface chemistry resulting from the increased oxygen and fluorine content. The k values of the as-deposited ppTTMSS films increased from 2.33 to 3.76 and those of etched ppTTMSS films increased from 2.31 to 3.65 as the deposition plasma power increased from 20 to 60 W. Pore formation is known to be derived from the cage structure in ppTTMSS films and contributes to the reduction in k values [21]. The peak-area ratio of cage structures, which could be related to pores inside the ppTTMSS film, reduced from 28 to 22% for both as-deposited and etched ppTTMSS films as the deposition plasma power increased from 20 to 60 W. This behavior can be explained by an increase in the fraction of cage structures, resulting in decreased k values [21]. It is generally accepted that the refractive index is directly proportional to the k value [34][35][36]. In accordance with this behavior, both the refractive index and the k value increased as the deposition plasma power increased. The k values remained almost the same after the ppTTMSS films were etched. This explains why etching did not significantly affect the electrical properties, which is beneficial for the stability of the ppTTMSS films. It was assumed that the etching process mostly affected the surface chemistry without causing any significant damage to the film.
increased from 20 to 60 W. Pore formation is known to be derived from the cage structure in ppTTMSS films and contributes to the reduction in k values [21]. The peak-area ratio of cage structures, which could be related to pores inside the ppTTMSS film, reduced from 28 to 22% for both as-deposited and etched ppTTMSS films as the deposition plasma power increased from 20 to 60 W. This behavior can be explained by an increase in the fraction of cage structures, resulting in decreased k values [21]. It is generally accepted that the refractive index is directly proportional to the k value [34][35][36]. In accordance with this behavior, both the refractive index and the k value increased as the deposition plasma power increased. The k values remained almost the same after the ppTTMSS films were etched. This explains why etching did not significantly affect the electrical properties, which is beneficial for the stability of the ppTTMSS films. It was assumed that the etching process mostly affected the surface chemistry without causing any significant damage to the film. To evaluate whether the ppTTMSS films are suitable as IMD materials, their mechanical properties, including hardness and elastic modulus, were analyzed, as shown in Figure 8. Increased density has been shown to enhance mechanical strength [37]. With an increase in the deposition plasma power, the hardness and elastic modulus increased from 1.06 to 8.56 GPa and from 6.16 to 52.45 GPa, respectively. For a higher deposition plasma power of 60 W, more SiO2-like structures were observed, leading to a higher density and mechanical strength. Enhanced mechanical properties could lead to a smaller etch rate because films with high mechanical strength are resistant to physical etching due to ion collisions. It has been reported that the hardness and elastic modulus exceeding 0.7 and 5.0 GPa, respectively, are required to ensure chemical-mechanical polishing stability [38]. The measured hardness and elastic modulus met this requirement, which allowed for IMD application. To evaluate whether the ppTTMSS films are suitable as IMD materials, their mechanical properties, including hardness and elastic modulus, were analyzed, as shown in Figure 8. Increased density has been shown to enhance mechanical strength [37]. With an increase in the deposition plasma power, the hardness and elastic modulus increased from 1.06 to 8.56 GPa and from 6.16 to 52.45 GPa, respectively. For a higher deposition plasma power of 60 W, more SiO 2 -like structures were observed, leading to a higher density and mechanical strength. Enhanced mechanical properties could lead to a smaller etch rate because films with high mechanical strength are resistant to physical etching due to ion collisions. It has been reported that the hardness and elastic modulus exceeding 0.7 and 5.0 GPa, respectively, are required to ensure chemical-mechanical polishing stability [38]. The measured hardness and elastic modulus met this requirement, which allowed for IMD application.  Figure 9a,b show the leakage-current densities of the as-deposited and etched ppTT-MSS films, respectively. The leakage-current densities for the as-deposited and etched ppTTMSS films were lower than 10 −6 A/cm 2 at 1 MV/cm, which is generally required for the IMD material [17]. At the electric field below 3 MV/cm, a breakdown was not observed for any of the ppTTMSS films. The ppTTMSS films in this study are potentially suitable IMD materials because of their sufficient electrical insulating properties.  Figure 9a,b show the leakage-current densities of the as-deposited and etched ppTTMSS films, respectively. The leakage-current densities for the as-deposited and etched ppTTMSS films were lower than 10 −6 A/cm 2 at 1 MV/cm, which is generally required for the IMD material [17]. At the electric field below 3 MV/cm, a breakdown was not observed for any of the ppTTMSS films. The ppTTMSS films in this study are potentially suitable IMD materials because of their sufficient electrical insulating properties.  Figure 9a,b show the leakage-current densities of the as-deposited and etched ppTT-MSS films, respectively. The leakage-current densities for the as-deposited and etched ppTTMSS films were lower than 10 −6 A/cm 2 at 1 MV/cm, which is generally required for the IMD material [17]. At the electric field below 3 MV/cm, a breakdown was not observed for any of the ppTTMSS films. The ppTTMSS films in this study are potentially suitable IMD materials because of their sufficient electrical insulating properties. According to the technical roadmap for semiconductor devices and systems, good mechanical strength of low-k materials along with a reduction in the k-value are the key factors in the interconnects with the continuous scaling down [39]. The ppTTMSS films used in this study showed good mechanical strength to resist the etching process and to meet the chemical-mechanical polishing stability and maintained their k-values below 4.0 and their leakage-current densities lower than 10 −6 A/cm 2 at 1 MV/cm after the etching process; thus, they can be applied as IMD materials to solve various problems in the scaledown challenge.

Conclusions
In this study, ppTTMSS films were deposited by PECVD with deposition plasma powers ranging from 20 to 60 W. Subsequently, the ppTTMSS films were etched using the RIE technique. The k values of as-deposited and etched ppTTMSS films ranged from 2.33 to 3.76 and from 2.31 to 3.65, respectively. The n values of the as-deposited ppTTMSS films were in the range of 1.46 to 1.60, similar to the n values of the etched ppTTMSS films (1.45 to 1.58). As the deposition plasma power increased, the density of the ppTTMSS films increased due to ion bombardment. As a result, the hardness and elastic modulus of the