Synthesis of thin-film materials using nonthermal plasma at a higher degree of dissociation

Received: 3 February 2023 | Revised: 8 March 2023 | Accepted: 14 March 2023 DOI: 10.1002/ppap.202300019 RESEARCH ARTICLE Synthesis of thin‐film materials using nonthermal plasma at a higher degree of dissociation Vladimir Cech 1 |Martin Branecky 2 1 Institute of Materials Chemistry, Faculty of Chemistry, Brno University of Technology, Brno, Czech Republic 2 Applied and Integrated Photonics, Institute of Scientific Instruments of the CAS, Brno, Czech Republic Correspondence Vladimir Cech, Institute of Materials Chemistry, Faculty of Chemistry, Brno University of Technology, Purkynova 118, CZ‐612 00 Brno, Czech Republic. Email: [email protected] Funding information Technology Agency of the Czech Republic, Grant/Award Number: TA01010796; Grantová Agentura Ceské Republiky, Grant/Award Number: 16‐09161S; Ministerstvo Školství, Mládeže a Telovýchovy, Grant/Award Number: LM2015056 Abstract Lower flow rates of precursor molecules are favorable for the synthesis of thin‐ film materials using nonthermal plasma at a higher degree of dissociation and sufficiently high deposition rate. These deposition conditions can be used for both continuous wave (CW) and pulsed plasmas and result in higher consumption of precursor molecules, which is beneficial for industrial applications due to cost reduction. A wider range of power can be used to control the chemical and physical properties of thin‐film materials based on power‐dependent plasma chemistry. Hydrogenated amorphous silicon carbide films deposited in CW and pulsed plasma are used as an example. The different kinetics of film growth and the role of self‐bias voltage in both types of plasma are discussed. KEYWORDS degree of dissociation, nonthermal plasma, organosilicon precursors, plasma‐enhanced chemical vapor deposition (PECVD), sticking coefficient, thin films 1|INTRODUCTION Low‐pressure plasma‐enhanced chemical vapor deposition (PECVD) is known as a thin‐film technology to synthesize materials with variable chemical and physical properties using nonthermal plasma. The principle of the technology was described in detail by Jansen. [1] This plasmachemical process is controlled by many parameters such as power, flow rate of precursor molecules, process pressure, precursor molecule chemical character, substrate temperature, and reactor type. Among these process parameters, power and flow rate appear to be very important. Let us see how these parameters affect the mechanical properties of the synthesized material. Hydrogenated amorphous carbon (a‐C:H, diamond‐ like carbon) was synthesized from cyclohexane (200 sccm) diluted with hydrogen gas (50 sccm), and increasing the power density from 0.10 to 0.36 W cm −2 Plasma Process Polym. 2023;20:e2300019 www.plasma-polymers.com | 1of11 https://doi.org/10.1002/ppap.202300019 This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2023 The Authors. Plasma Processes and Polymers published by Wiley‐VCH GmbH. led to an increase in Young's modulus from 120 up to 170 GPa. [2] The flow rate is given in standard cubic centimeters per minute (sccm). Similarly, a‐C:H films deposited from pure methane (100 sccm) showed an increase in hardness from 11.9 to 16.5 GPa with increased radio frequency (RF) power of 100–300 W. [3] Hydrogenated amorphous silicon carbonitride (a‐ SiCN:H) films were prepared from trimethylsilane in a mixture of nitrogen and 20 sccm argon by Peter et al. [4] The total flow rate of these precursors was 180 sccm. Increasing the trimethylsilane flow rate from 10 to 50 sccm at constant discharge power had no effect on Young's modulus of approximately 135 GPa. Frischmuth et al. [5] pointed to hydrogenated amorphous silicon carbide (a‐SiC:H) films deposited from methane (13.5 sccm), silane (6.5 sccm), and argon (50 sccm), where Young's modulus (137–196 GPa) of the films was controlled by power (750–2000 W). A Young's modulus ranging from 113 to 126 GPa due to increasing RF power (100–1600 W) was determined for silicon‐rich a‐SiC:H films deposited from a Si 2 H 6 /CH 4 /He/Ar mixture of 4400 sccm. [6] The above‐mentioned studies used flow rates of tens to thousands of sccm and therefore changes in power allowed mechanical properties to be changed only by tens of percent. Significant changes in mechanical properties can be achieved by changing the power while using a much lower flow rate. Using 3.8 sccm tetravinylsilane in a pulsed plasma, Young's modulus of a‐ SiC:H films was increased from 10 to 143 GPa, while increasing the effective power in the range of 2–150 W. [7] The Yasuda parameter, W/FM [J g −1 ](W,F, and Mare the power, flow rate, and molecular weight of the precursor), represents the input energy delivered to the plasma per unit mass of the precursor. [8] This parameter is expected to be proportional to the concentration of the activated plasma species used for the synthesis of the thin film material and the deposition rate is therefore increased by increasing the power at a constant flow rate. [9] This idea applies to higher flow rates, where only about 1% of the precursor molecules are activated by electrons to form chemically reactive fragments. [1] Deposition conditions characterized by limited fragmentation of precursor molecules are called by Inagaki precursor‐sufficient conditions, [9] in which the process pressure approximately corresponds to the preset pressure before the ignition of the plasma. Inagaki pointed out that the increase in the deposition rate with increasing power (Yasuda parameter) is followed by its decrease due to the lack of precursor molecules with respect to more pronounced precursor fragmentation, and he called these deposition conditions precursor‐ deficient conditions. However, when using a low flow rate, the process pressure drops from the preset value even for the power range where the deposition rate increases, indicating a higher degree of precursor fragmentation. [7] A low flow rate of precursor molecules means a low concentration of gas molecules in the plasma process, if a flow system is used, where the plasma reactor is continuously pumped. In this study, the drop in process pressure is analyzed and related to the degree of dissociation of the precursor molecules. A continuous wave (CW) plasma of tetravinylsilane at a low flow rate was used to deposit a‐SiC:H films. Here, the plasma species and the chemical and physical properties of the thin films are characterized as a function of the power supplied to the discharge. The plasma and thin film characteristics are compared with those obtained in a previous study [7] for pulsed plasma under similar deposition conditions. The different kinetics of film growth in CW and pulsed plasma is also discussed. 2|EXPERIMENTAL SECTION 2.1 |Thin film deposition Capacitively coupled nonthermal plasma was used for the deposition of thin films in a deposition system equipped with asymmetric planar parallel electrodes as previously described. [10] The RF (13.56 MHz) plasma was operated in CW mode. Double‐sided polished (100) silicon wafers (0.8 × 10 × 10 mm 3 ; ON Semiconductor) were loaded into the smaller powered electrode. The plasma reactor was evacuated to a base pressure of 2×10 −4 mTorr and the silicon substrate was pretreated with argon plasma (10 sccm, 43 mTorr, 5 W) for 10 min to remove physisorbed molecules from its surface, followed again by evacuation to the base pressure. Tetravinylsilane (TVS; Si(–CH═CH 2 ) 4 ; Sigma‐Aldrich) as a precursor was introduced into the reactor at a flow rate of 3.8 sccm and a pressure of 20 mTorr was set by the butterfly control valve (VAT 615 DN 63). The plasma reactor is equipped with Leybold Vacuum CTR90 and CTR91 capacitive sensors. A power range of 10–70 W was used to deposit a‐ SiC:H films. The lower limit is determined by the minimum output power of the RF generator (CESAR 1310; Advanced Energy) and the upper limit corresponds to the maximum power resulting in a stable and reproducible plasma. A matching network was used to reduce the reflected power to zero. After thin film deposition, the TVS vapor was removed from the deposition system with argon gas (10 sccm, 43 mTorr) for 60 min. The sample was then left under the base pressure for another 12 h before being transferred to atmospheric pressure. 2of11 | CECH and BRANECKY 16128869, 2023, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ppap.202300019 by Technical University In Brno, Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2.2 |Analytical techniques The HPR‐30 Process and Residual Gas Analyzer (Hiden Analytical) made it possible to check low levels of residual gases (oxygen, nitrogen, and water molecules) in the deposition chamber before and during the plasma process to avoid contamination of the a‐SiC:H films with oxygen and nitrogen atoms. The remote location of the mass spectrometer input ensures that only neutral species enter the mass spectrometer. During the deposition process, the mass spectrometer enables the monitoring of neutral substances in the deposition chamber depending on the applied RF power. The default electron energy of 70 eV was used. The system can analyze ions from 0.4 to 510 amu (atomic mass unit) with a minimum step of 0.01. The Tandetron 4130 MC (HVEE) is a linear accelerator that was used for Rutherford backscattering spectrometry and elastic recoil detection analysis (ERDA) measurements to characterize the elemental composition in the bulk of the film. The silicon atoms in the films were analyzed with 2.73 MeV alpha particles, while the lighter carbon atoms were measured with 2.4 MeV protons; both types of particles incident perpendicular to the film surface and those scattered particles were detected at an angle of 170°. For the characterization of hydrogen atoms, an ERDA with an incident beam of 2.75 MeV alpha particles at 75° to the normal of the film surface was used. Hydrogen atoms recoiled at 30° were detected by a surface barrier detector covered with a 12 µm thick Mylar stopping foil. Fourier‐transform infrared spectroscopy (FTIR) using a VERTEX 80v (Bruker Optics) was employed at 160 Pa to characterize the chemical structure of the a‐SiC:H films after 12 h under the base vacuum. Two hundred fifty‐six scans were collected with a scan resolution of 4cm −1 to obtain transmission spectra in the range of 400–4000 cm −1 . The infrared spectrum of the film was determined by subtracting the spectrum of the bare silicon wafer and interference effects [11] from the recorded spectrum. The Young's modulus and hardness of the a‐SiC:H films were investigated using a 2D TriboScope TS75 (Bruker Hysitron) attached to an NTegra Prima Scanning Probe Microscope (NT‐MDT). For cyclic nanoindentation, [12] a three‐sided pyramidal Berkovich indenter with a radius of curvature of 50 nm was used to determine the depth profile of the mechanical properties up to 20% of the film thickness. Depth profiles of Young's modulus and hardness were extrapolated to zero contact depth (film surface) to determine the correct mechanical parameters of the film unaffected by the silicon substrate. [12] The thickness and density of the a‐SiC:H films were analyzed using high‐resolution X‐ray reflectometry with a conventional copper X‐ray tube, a parabolic multilayer mirror collimator, and a Ge 220 Bartels‐type monochromator. The intensity pattern was measured by a scintillation detector using a slit. The optical properties and thickness of the deposited films were evaluated by a phase‐modulated spectroscopic ellipsometer UVISEL (Horiba Scientific). The measurements were performed at an incidence angle of 70° and a spot size of 100 × 300 μm 2 using a wavelength of 250–830 nm in steps of 5 nm. The dispersion dependence of the dielectric function was fitted using the five‐ parameter Tauc–Lorentz formula, which was derived for the parameterization of the optoelectronic response of amorphous dielectrics. [13] The deposition rate was calculated as the ratio between the film thickness determined by ellipsometry and the deposition time. 3|RESULTS AND DISCUSSION 3.1 |Relative pressure drop By introducing the precursor into the deposition chamber at a given flow rate, a steady‐state pressure is reached after a certain time depending on the pumping speed. The pumping speed of the vacuum pump system is typically limited by the butterfly control valve that allows the pressure of the precursor molecules to be preset to a given value, p off , before the plasma is ignited (plasma off regime). The general gas equation can be converted into the form pnkT= , B off off (1) where the pressure, p off , is related to the concentration of precursor molecules, n off ,k B is the Boltzmann constant, and Tis the gas temperature. Analogous to Equation (1), after the plasma is ignited (plasma on regime), the process pressure stabilizes at p on , which is related to the concentration of plasma species, n on , that can be expressed as nn nnn=+++ , on undis bp int rad (2) where n undis is the concentration of precursor molecules that were not dissociated, n bp is the concentration of by‐ products, n int is the concentration of intermediates, and n rad is the concentration of radicals; all these components are constant at the steady‐state plasma for a given RF power. During the plasma process, precursor molecules are fragmented into free radicals in the processes of CECH and BRANECKY | 3of11 16128869, 2023, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ppap.202300019 by Technical University In Brno, Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License electron‐impact dissociation and dissociative electron‐ impact ionization. As a result of recombination processes in the bulk plasma (association of plasma species into oligomers) and on the surface of the growing film (chemisorption), those chemically reactive species (radicals) form by‐products (e.g., H 2 ,CH 4 ) or intermediates containing a double or triple bond (e.g., H 2 C═CH 2 , HC≡CH) or are chemically bound to the surface of the growing film. At the same time, it is true that nn n=+ , off undis dis (3) where n dis is the concentration of dissociated molecules. Higher flow rates, which result in only about 1% of dissociated precursor molecules, [1] do not significantly affect the process pressure, and thus p on ≈p off since n on ≈n undis ≈n off ; concentrations, n dis ,n bp ,n int , and n rad , are small compared to n undis . However, at a degree of precursor fragmentation higher than 1%, a change in process pressure can be expected. Every dissociated molecule is split into two or more fragments, meaning that n on >n off and the process pressure increases after the plasma is ignited. Plasma deposition is a dynamic process, where reactive plasma species are generated but also recombined, and the steady‐state concentration of plasma species, n on , is controlled by the generation and recombination rates. The resulting process pressure, p on , can therefore be higher, lower, or, in a special case, equal to the preset pressure, p off . If the recombination process is fast enough, the process pressure is lower than the preset pressure and the relative pressure drop can be expressed as ∆ppp p nn n =−=− . r off on off off on off (4) The degree of dissociation, f dis , is expressed as the fraction of the precursor molecules that have dissociated, thus with respect to Equation (3) f n nn nn n =+=− . dis dis dis undis off undis off (5) Since n on >n undis and using Equations (2), (4), and (5) we get ∆pf< . rdis (6) The relative pressure drop, therefore, expresses the lower limit of the degree of dissociation and multiplied by 100% indicates the minimum consumption of the precursor in percent. This means that the actual degree of dissociation is higher than this value, Δp r . The last relation (Equation 6)is valid for neutral as well as charged (ionic) plasma species, although the charged species may have a higher process temperature, T on >T off . As an example, the process pressure (black symbol) dependent on the RF power for CW plasma (10–70 W) and pulsed plasma (2–150 W) [7] at a TVS flow rate of 3.8 sccm and a preset pressure of 20 mTorr, as shown in Figure 1. Effective power is used for the pulsed plasma, detailed in the studybyBraneckyetal. [7] The corresponding relative pressure drop is shown as a red symbol (Figure 1). A smoothing B‐spline (Origin; OriginLab Corporation) was used to fit the data only to capture the dependence trend. It is evident that the degree of dissociation is higher for a CW plasma than for a pulsed plasma at a given RF power. While for the pulsed plasma, the relative pressure drop increases from 0.04 to 0.33, in the case of the CW plasma, the relative pressure drop varies from 0.41 to 0.57. Even higher pressure dropsofupto0.81werefoundforatubularplasmasystem using a TVS flow rate ranging from 1.4 sccm (preset pressure 29–114 mTorr) to 7.0 sccm (29 mTorr). [14] Several TVS flow rates were tested to observe the pressure drop in the plan‐ parallel plasma system [10] used in this study. A pressure drop was observed for a flow rate of 1.9 sccm and a preset pressure of 20–162 mTorr and 3.8 sccm and 20–81 mTorr, but no pressure drop was noted for a flow rate of 7.6 sccm and a preset pressure of 41–81 mTorr. In the next section, we will look at how the CW plasma leading to a higher degree of dissociation affects the deposition process. 3.2 |Plasma chemistry The power‐dependent (CW mode, 10–70 W) mass spectra of neutral species in TVS plasma together with the spectrum corresponding to zero power (plasma off) are FIGURE 1 Power‐dependent process pressure and the corresponding relative pressure drop for continuous wave (CW) and pulsed plasma. RF, radio frequency. 4of11 | CECH and BRANECKY 16128869, 2023, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ppap.202300019 by Technical University In Brno, Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License FIGURE 2 Mass spectra of the neutral species in TVS plasma corresponding to 10–70 W operated in continuous wave mode and zero power (plasma off). FIGURE 3 (a) Cumulative power dependence of carbon species (m/z25–28) and silicon‐containing species (m/z53–57, 80–83, and 105– 109) for continuous wave (solid line) and pulsed (dashed line) plasma. (b) Power dependence of individual carbon species (m/z25–28) for continuous wave (CW) (solid line) and pulsed (dashed line) plasma. RF, radio frequency plotted in Figure 2. No oxides were caused by residual water molecules in the deposition chamber and no oligomers were detected by mass spectrometry. Mass spectra can be quantitatively analyzed analogously to TVS pulsed plasma. [15] This analysis is based on the idea that changes in plasma chemistry occurring in the deposition chamber are reflected in corresponding changes in the mass spectrometer due to similar fragmentation patterns for electron impact dissociation and dissociative electron impact ionization, which was demonstrated by strong correlations between the most produced cations recorded by the mass spectrometer and the film chemistry. [15] The partial pressure of a given cation is related to its concentration as described in Equation (1). The most produced cations include carbon species (m/z25–28), silicon‐containing species (m/z 53–57 with one vinyl group, 80–83 with two vinyl groups, 105–109 with three vinyl groups), which are the building blocks of the growing film and determine its elemental composition and chemical structure, [15] and also include a hydrogen molecule (m/z2) as a by‐ product; m/zmeans the ratio of ion mass (m) to charge valence (z). The power dependence of the above species for the CW plasma (solid line) is compared with the distribution for the pulsed plasma (dashed line) in Figure 3a. The carbon species dominate over the silicon‐containing species for all powers in the 10–70 W range and include ethynyl ion (m/z25), acetylene ion (m/ z26), vinyl ion (m/z27), and ethylene ion (m/z28) and their power dependences are shown in Figure 3b again for CW and pulsed plasma. It is typical for all power dependences that the concentration of cations is lower in CW plasma than in pulsed plasma. 3.3 |Sticking coefficient The growth rate of the film is significantly influenced by the density of free‐binding sites on the surface of the growing film. As explained in detail in a previous CECH and BRANECKY | 5of11 16128869, 2023, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ppap.202300019 by Technical University In Brno, Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License study, [15] the concentration of free radicals (H, CH 3 , C 2 H, and C 2 H 3 ) in the plasma is sufficient for very efficient hydrogen removal (hydrogen abstraction) [16] from the film surface, allowing the formation of free binding sites. Free radicals and multiple bond intermediates then bind to these binding sites and contribute to the growth of the film. The deposition rate of a‐SiC:H films prepared from the TVS precursor using CW plasma is lower than that of pulsed plasma for a given power, as seen in Figure 4a.Themaximum deposition rate for CW plasma is 210 nm min −1 (20 W) but 319 nm min −1 (25 W) for pulsed plasma. A lower deposition rate was also confirmed for a‐SiC:H films deposited from the SiH 4 /CH 4 mixture using CW plasma compared to pulsed plasma. [17] The deposition rate is related to the number of plasma species that impinge on a unit surface per unit time (molecular incidence rate). However, not every radical or intermediate is covalently bound to the surface of the growing film due to steric and collision orientation effects. [18,19] The probability that an atom or molecule is chemisorbed can be expressed using the sticking coefficient, [20] which ranges from 0 to 1. The growth of the film also depends on the size of the attached fragments. The deposition rate (R) can therefore be expressed as the sum of the contributions of individual types of reactive plasma species  RγV=Φ , i N iii =1 (7) where γ i is the sticking coefficient, Φ i is the molecular incidence rate, and V i is the volume corresponding to the ith fragment type. The molecular incidence rate for a given fragment of mass (m i ) is determined by its partial pressure (p i ) [21] p πmk T Φ=2 , ii iB (8) and the corresponding volume of the bound fragment is given by FIGURE 4 (a) Deposition rate as a function of radio frequency (RF) power. (b) Molecular incidence rate of carbon species and silicon‐ containing species depending on RF power. (c) Power dependence of the sticking coefficient for carbon species in the case of continuous wave (CW) (solid line) and pulsed (dashed line) plasma. 6of11 | CECH and BRANECKY 16128869, 2023, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ppap.202300019 by Technical University In Brno, Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License V M Nρ = , ii A(9) where M i is the molar mass of the fragment, N A is the Avogadro's constant, and ρis the film density. Hydrogen abstraction is most efficient due to hydrogen atoms, leading to the release of hydrogen molecules. [16] In the case of pulsed plasma, [15] the period consists of an RF power‐on part (1 ms) and an RF power‐ off part (1–7 ms), and since the lifetime of the ions is only 10 −6 s, [22] shortly after the power is turned off, there are no more electrons available to dissociate the neutral plasma species. This means that H 2 production increases due to hydrogen abstraction during the time when the power is turned off, as the lifetime of the radicals can be longer than 1 h. [8] However, the H 2 production and as well as the production of other plasma species are dissociated continuously in time in CW plasma, leading to their lower concentrations than in the case of pulsed plasma (Figure 3). Thus, hydrogen abstraction may be more effective in CW plasma than in pulsed plasma and may thus affect the magnitude of the stickiness coefficient. Cumulative molecular incidence rates in Figure 4b were calculated (Equation 8) for carbon and silicon‐ containing species based on the partial pressures from Figure 3a for CW (solid line) and pulsed (dashed line) plasma. The molecular incidence rate of carbon species dominates (10–70 W) and furthermore, the sticking coefficient for silicon‐containing species is at least 10 times lower than that for carbon species, as found in a previous study, [15] and thus the contribution of silicon‐ containing species to the deposition rate can be neglected. The simplified Equation (7) can be used to estimate the sticking coefficient for the carbon species represented mainly by the acetylene molecule (Figure 3b) with a molar mass of 26.04 g mol −1 . The film density as assessed by X‐ray reflectometry increased from 1.35 g cm −3 (10 W) to 1.84 g cm −3 (70 W), reducing the volume per fragment (Equation 9). The molecular incidence rate of the carbon species in the deposition chamber, which was used in Equation (7), was determined from the partial pressure proportional to the process pressure (Figure 1). The estimated sticking coefficient as a function of power for CW (solid line) and pulsed (dashed line) plasma is given in Figure 4c. The sticking coefficient for the CW plasma increased from 3 × 10 −5 (10 W) to 6 × 10 −5 (70 W), indicating an increased effectivity of hydrogen abstraction with enhanced power. The sticking coefficient for pulsed plasma ranges between values (3–4) × 10 −5 and is like the values for CW plasma at lower powers (10–25 W). These values of the sticking coefficient correspond to the values published for the CH 3 radical in the case of a‐C:H film growth. [23] The maximum deposition rate for the CW plasma is lower than for the pulsed plasma (Figure 4a) due to the lower molecular incidence rate, as the sticking coefficient and fragment volume are very similar for both types of plasma. The generation and recombination of plasma species in both types of plasma differs due to the different kinetics of plasmachemical processes during the on and off regime of the pulsed plasma. Precursor molecules are dissociated continuously in time in CW plasma, in contrast to pulsed plasma, where dissociation occurs only when the plasma is turned on for 1 ms, resulting in a higher degree of dissociation and higher consumption of precursor molecules in CW plasma, which is indicated by a higher decrease in the process pressure. When the plasma is switched off within 1–7 ms, the concentration of plasma species (H 2 ,CH 4 ,C 2 H 2 , and C 2 H 4 ) due to hydrogen abstraction and no ongoing dissociation increases and, together with the concentration of the main radicals (C 2 H, C 2 H 3 ), are thus higher on average over the period for the pulsed plasma than for the CW plasma, resulting in higher deposition rate. 3.4 |Chemical properties The elemental composition of a‐SiC:H films, which is controlled by the power for CW (solid line) and pulsed (dashed line) plasma, is shown in Figure 5a. The concentration of silicon atoms is similar for both types of plasma and approximately independent of power. The concentration of carbon increases with enhanced power at the expense of hydrogen also in both types of plasma, but in the case of CW plasma the concentration of hydrogen atoms is higher at lower powers (10–25 W) than in pulsed plasma. As‐deposited films are oxygen‐ free materials, which is consistent with the results of the mass spectra analysis. McCurdy et al. [24] also found that when CW plasma was used, there was greater incorporation of hydrogen into a‐SiC:H films deposited from a SiH 4 /CH 4 mixture than when pulsed plasma was used. The FTIR spectra shown in Figure 5b provide insight into the chemical structure of the oxygen‐free a‐CSi:H films modified by the applied power used for the CW plasma. The figure also includes the assignment of the main absorption peaks to the vibrations of specific chemical groups. [25,26] The concentration of vinyl groups (1404, 1007, and 953 cm −1 ) bound to the carbon‐silicon network decreases with enhanced power, as already known in the case of pulsed plasma. [7] Also, the area of the absorption bands for the CH x (2905–2895 cm −1 ) and SiH x (2116–2122 cm −1 ) vibrations decreases as hydrogen is eliminated from the a‐CSi:H films at higher powers. CECH and BRANECKY | 7of11 16128869, 2023, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ppap.202300019 by Technical University In Brno, Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License This means that reducing the hydrogen concentration in the carbon‐silicon network results in a more crosslinked and denser material, [7] see the increase in film density with power in the previous chapter. In addition, the denser carbon–silicon network contains a higher concentration of carbon with a double bond (1350–1650 cm −1 ), also observed in the case of pulsed plasma. [7] The hydrogen of higher concentration in films deposited in CW plasma at lower power of 10 W appears to be bound in the vinyl groups (1404 cm −1 ) as well as in the carbon‐silicon network (CH 2 in sp 3 configuration, 1454 cm −1 ) as shown in the inset graph (Figure 5b). 3.5 |Physical properties The mechanical properties of a‐CSi:H films, Young's modulus (black symbol), and hardness (red symbol), deposited in CW plasma (solid line) are compared with those obtained for pulsed plasma (dashed line) in Figure 6a. Both parameters increase significantly with enhanced power for the CW plasma, namely, Young's modulus from 12 GPa (10 W) to 81 GPa (70 W) and the hardness from 0.8 GPa (10 W) to 8.8 GPa (70 W). For pulsed plasma, the mechanical properties were found to increase due to increased crosslinking of the carbon‐ silicon network characterized by an increased film density of 1.5–1.9 g cm −3 . [7] The mechanical properties corresponding to the CW plasma are similar to those of the pulsed plasma (Figure 6a), probably due to the similar degree of network crosslinking at a given power. Power‐dependent dispersion curves for the refractive index and the extinction coefficient (optical properties) in the case of CW plasma are very similar to those obtained for a‐CSi:H films deposited in pulsed plasma. [7] Thus, the optical properties of a‐CSi:H films deposited using CW (solid line) and pulsed (dashed line) plasma were compared only for a specific wavelength of 633 nm (He–Ne laser) in Figure 6b. The power dependences for the refractive index (black symbol) and the extinction coefficient (red symbol) are again similar for both types of plasma. The refractive index increases with enhanced power for CW plasma from 1.7 to 2.1 depending on the optical density of the material described by the Clausius–Mossotti relation. [27] The strong correlation between the refractive index and Young's modulus demonstrated in Figure 6c proves that both parameters are controlled by the degree of network crosslinking for both types of plasma. The power dependence of the extinction coefficient in Figure 6b shows that the a‐CSi:H films deposited at 10 and 20 W are transparent to a wavelength of 633 nm, corresponding to the power‐ dependent band gap in Figure 6d. The band gap decreased from 2.0 to 1.2 with enhanced power for the CW plasma consistent with the data for the pulsed plasma. This reduction was correlated with a reduced concentration of nanovoids (vinyl groups) in the material and an increased sp 2 content in the carbon‐silicon network, which are responsible for band gap modification. [7] 3.6 |Self‐bias voltage In asymmetric capacitively coupled RF plasma systems with a smaller powered electrode and a larger grounded electrode, a negative DC self‐bias voltage occurs on the powered electrode due to the higher mobility of electrons than ions. [28] This negative sheath voltage accelerates the positive ions that bombard the powered electrode with a FIGURE 5 (a) Power‐dependent elemental composition of a‐CSi:H films deposited using continuous wave (CW) (solid line) and pulsed (dashed line) plasma. (b) Infrared spectra of a‐CSi:H films deposited using CW plasma at 10–70 W, together with a detail of the difference spectra between CW and pulsed plasma for 10 W in the inset graph. RF, radio frequency. 8of11 | CECH and BRANECKY 16128869, 2023, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ppap.202300019 by Technical University In Brno, Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License loaded substrate, which in principle can affect the growth of the film. However, the sputtering effect is monitored with copper cations (m/z63.5) due to the sputtering of copper atoms from the powered electrode, and no such cations were observed in CW (10–70 W) and pulsed (2–150 W) plasma. As analyzed and discussed in the previous study, [15] the neutral plasma species dominate the growth of the film and no effect of ions on the chemical and physical properties of the deposited films was demonstrated. Catherine and Couderc [29] found that the self‐bias voltage (V B ) is proportional to the power (W) and the process pressure (p on )as ≈ V W p . B on (10) This dependence is plotted in Figure 7for CW and pulsed plasma. The mean value of the self‐bias voltage was used for the pulsed plasma because the bias oscillates. Individual power values are shown directly for the given dependency, for CW plasma in blue and for pulsed plasma in red. The self‐bias voltage for a given power is higher for CW plasma than for pulsed plasma except for 10 W (Figure 7), for example, 929 V (70 W) for CW plasma and 555 V (75 W) for pulsed plasma. From FIGURE 6 (a) Power dependence of Young's modulus and hardness for a‐CSi:H films deposited in continuous wave (CW) (solid line) and pulsed (dashed line) plasma. (b) Comparison of refractive index and extinction coefficient at 633 nm for CW and pulsed plasma. (c) Correlation of refractive index with Young's modulus for CW and pulsed plasma. (d) Power‐dependent band gap for CW and pulsed plasma. RF, radio frequency. FIGURE 7 Self‐bias voltage corresponding to marked power versus (W/p on ) 1/2 for continuous wave (CW) (solid line) and pulsed (dashed line) plasma. CECH and BRANECKY | 9of11 16128869, 2023, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ppap.202300019 by Technical University In Brno, Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License