Effect of Substrate Negative Bias on the Composition and Structure of nc-Cu/a-C:H Nanocomposite Films Deposited by FCVA

Abstract

Copper-doped hydrogenated amorphous carbon (Cu-doped a-C:H) films were synthesized using copper as the cathode and C₂H₂ as the precursor. The result shows that the negative bias voltage can affect the composition and microstructure of nanocomposite films. With bias voltage increasing, Cu content first increases in the range of 50~300 V and then declines with higher voltage, while the deposition rate decreases continuously. The stress and sp³ content present a similar trend with the bias voltage, increasing during the range from 50 V to 200 V and then decreasing with higher voltage.

1. Introduction

Amorphous carbon films with hydrogen (a-C:H), a prominent class of diamond-like carbon (DLC) material, have attracted significant interest due to their outstanding mechanical properties, such as high hardness, low friction coefficient, and chemical inertness [1,2,3,4,5,6,7,8]. However, a-C:H films still face challenges such as high intrinsic compressive stress and environmental sensitivity (e.g., friction degradation in humid conditions), limiting their applications in precision optoelectronic devices and thick coatings [9,10,11,12]. To address these limitations, many studies have focused on embedding metallic nanoparticles into the a-C:H matrices, forming nanocomposite systems that combine the advantages of both components [4,13].

Nanocomposite DLC films can be categorized into two types: those formed with reactive metals, such as titanium (Ti), tungsten (W) and molybdenum (Mo), which can react with carbon to create carbide structures, and those containing non-reactive metals, such as silver (Ag), platinum (Pt), and nickel (Ni), which do not form carbides. These nc-MeC/a-C:H films and nc-Me/a-C:H films have demonstrated remarkable improvements in stress relief and film structure [13,14,15,16]. For instance, Ti-doped DLC films, synthesized via reactive-biased target ion beam deposition, form cubic TiC nanocrystallites at 4.2 at.% Ti, enhancing mechanical hardness and biocompatibility for bone implant coatings [17]. Wang et al. from the Chinese Academy of Sciences further revealed that substrate negative bias voltage critically regulates Ti content and carbide formation kinetics, highlighting the pivotal role of ion energy in tailoring nanocomposite structures [14]. The presence of TiC nanoparticles can affect the sp²/sp³ hybridization ratio, altering the microstructure of the carbon matrix and ultimately influencing the hardness, electrical conductivity, and other properties of the films. Non-reactive metals like Cu, Ag, Pt, and Ni can also offer distinct advantages, as they do not engage in carbide formation, allowing for the preservation of the carbon matrix’s integrity. Ag-doped DLC films, as demonstrated by researchers from the Technical University of Munich in Germany, exhibit excellent corrosion resistance and antibacterial properties [18]. The incorporation of Ni can be used to modulate the degree of graphitization, friction coefficient, and corrosion resistance of the nanocomposite films. Research conducted by a team from India has focused on the optical and electrical properties of Ni-doped DLC films [16], and our team has investigated the wear resistance and corrosion resistance of Ni-doped films [15]. Incorporating Cu into DLC films can significantly enhance their electrical and optical properties. This topic has been explored by multiple research teams, including Shahd Bakhet from Lithuania [19], Canyan Che from Dalian University of Technology in China [20], and Neeraj Dwivedi’s group from India [21]. Despite the promising potential of these non-reactive metals, the mechanisms of incorporation and the modulation of properties in nanocomposite films containing these metals remain largely underexplored in the existing literature. This gap presents an opportunity for further research into the unique benefits and applications of non-reactive metal-doped DLC films.

Building on this foundation, our group developed nc-Cu/a-C:H nanocomposite films using a filtered cathodic vacuum arc (FCVA) system, where chemically inert Cu nanoparticles are embedded within a-C:H matrices. Previous work by Zhang et al. [22] systematically explored the impact of C₂H₂ flow rate on film composition and microstructure, demonstrating that increased carbon precursor flux enhances sp³ bonding but reduces Cu incorporation. Despite these advances, the influence of substrate negative bias—a key parameter governing ion energy and plasma interactions—on the structural evolution and compositional characteristics of Cu-doped a-C:H films remains unexplored [23].

Compared with previous metal-doped DLC studies [14,21,22,23], this work provides the first systematic investigation of bias voltage effects on copper–carbon co-deposition dynamics in FCVA systems. We reveal how ion energy governs Cu grain size evolution, sp³ bonding content and residual stress. This study provides the first comprehensive framework for tailoring stress in Cu-DLC coatings via bias voltage control, addressing a critical gap in non-reactive metal-doped DLC research. Our findings elucidate the critical interplay between ion bombardment energy and re-sputtering mechanisms, providing actionable insights for tailoring nanocomposite architectures in energy-sensitive deposition systems.

2. Experimental Details

2.1. Processing of Coatings

The film deposition was performed with a self-developed filtered cathodic vacuum arc (FCVA) system, comprising three main components: a cathode vacuum arc source, a 90–bent magnetic filtering duct (180 mm inner diameter), and a vacuum deposition chamber [22]. The FCVA technique employs arc discharge to generate plasma, while magnetic filtering is used to remove macroparticles and neutral atoms, resulting in a highly ionized beam suitable for film deposition. Prior to deposition, the substrates made of single crystal Si (100) were ultrasonically cleaned in ethanol to ensure a clean surface. A high-purity Cu cathode (99.99% pure, Φ100 mm) served as the arc source.

The base pressure was maintained at 2.5 × 10⁻³ Pa. The arc current at the cathodic arc source was set to 90 A. A positive bias (20 V) was added to the bending tube to focus the ion beam. The deposition process commenced with a high negative bias, gradually decreasing from 800 V to 400 V, applied to the substrate for 1 min to perform sputter cleaning and remove the oxide layer on the substrate surface. A 2 min copper interlayer was deposited to boost adhesion between the film and substrate. Subsequently, C₂H₂ gas was introduced into the vacuum chamber at 80 sccm. Different samples were deposited by varying the negative bias during the deposition process.

2.2. Analyzing Methods

The thickness and curvature radius of the films were assessed using a surface topography device, Talysurf 5P-120 (Taylor Hobson, Berwyn, IL, USA). The film residual stress was calculated by Stoney formula [22], which is provided below:

$$\sigma ={\displaystyle \frac{E_s}{6(1-{\upsilon }_s)}}\times {\displaystyle \frac{t_s^2}{d}}\times ({\displaystyle \frac{1}{r}}-{\displaystyle \frac{1}{r_0}})$$

(1)

where E_s, ν_s and t_s represent the elastic modulus, Poisson’s ratio and substrate thickness, respectively. d is the film thickness; r₀ and r denote the radius of curvature before and after deposition, respectively.

The composition and microstructure of the nc-Cu/a-C:H films were analyzed by X-ray photoelectron spectroscopy (XPS) (ThermoFisher, Cambridge, UK) and X-ray diffraction (XRD) (PANalytical B.V., Almelo, The Netherlands). XPS measurements were conducted with an ESCALAB 250 system to analyze the elemental composition. For XRD analysis, the crystallinity and phase composition were examined using an X’Pert PRO MPD system.

3. Results and Discussion

3.1. Film Composition

The influence of substrate negative bias on the composition of nanocomposite films (nc-Cu/a-C:H) was examined using XPS. Figure 1 displays the relationship between Cu content and substrate negative bias. The specific values have also been listed in

Table 1. Deposition parameters, composition, and characterization of Cu-doped DLC films deposited at different bias voltages.

Negative Bias (V)	Cu Content (at.%)	Deposition Rate (nm/s)	sp3 Content (%)	Stress (GPa)	Grain Size of Cu (nm)
50	3.89	7.2	49	1	/
100	4.73	5.2	56.9	2.2	/
200	7.94	4.6	57.3	4.6	10.6
300	9.45	4.1	56.7	1.3	13
500	8.62	3.7	49.4	0.5	12.3

. As illustrated in Figure 1, the Cu content initially increases with increasing substrate negative bias, reaching a maximum at a bias of 300 V, after which it begins to decrease. This non-monotonic trend contrasts with the behavior observed in Ti-doped DLC films reported in Ref. [14] where Ti content exhibited continuous increase with rising negative bias. The distinct behaviors may originate from different metal–carbon interaction mechanisms: in the case of Ti doping, the formation of stable TiC compounds through chemical reaction between Ti and a carbon matrix likely suppresses metal sputtering even at higher biases, whereas Cu remains in metallic state without forming carbides in the DLC matrix. An appropriate increase in negative bias can promote the diffusion and effective incorporation of Cu atoms and excessively high negative bias results in increased kinetic energy, making Cu more susceptible to sputtering compared to TiC, which leads to a reduction in Cu content.

The XPS Cu 2p fine spectra of films prepared at different substrate negative biases are presented in Figure 2. Two distinct peaks are observed at 932.8 eV and 952.6 eV, which are indicative of the Cu 2p₃/₂ and Cu 2p₁/₂ characteristic peaks, respectively. With the increase in the substrate negative bias from 20 V to 300 V, the intensities of both peaks significantly enhance, indicating an increase in copper content within the films. However, when the negative bias is further increased to 500 V, a noticeable decrease in the peak intensities of Cu 2p₃/₂ and Cu 2p₁/₂ is observed. This trend aligns with the variation in copper content depicted in Figure 1, suggesting that the initial increase in copper incorporation is followed by a reduction at higher biases. The observed behavior can be attributed to the interplay between the increasing ion energy at elevated negative biases, which enhances the incorporation of copper into the DLC films, and the potential effects such as sputtering of copper at very high biases. This dynamic ultimately influences the overall copper content in the films. These results indicate that the adjustment of substrate negative bias can effectively control the Cu content within the films.

The film deposition rate was determined by calculating the ratio of the film thickness, measured using a step profiler, to the deposition time. Figure 3 presents the correlation between the deposition rate of nanocomposite films (nc-Cu/a-C:H) and the substrate bias voltage. The data indicate that the deposition rate initially increases with rising bias voltage but subsequently decreases at higher bias values. This behavior aligns with the observations made by Peng Wang et al., who reported a similar pattern regarding the titanium (Ti) concentration in doped a-C:H films produced via the FCVA technique [14]. The initial increase in deposition rate can be attributed to enhanced ion energy under moderate bias conditions. Higher ion energy facilitates surface diffusion and nucleation, thereby promoting film growth. However, at elevated bias levels, intensified ion bombardment may lead to secondary sputtering of the film [23]. The etching rate escalates with temperature, which itself rises with increasing bias voltage due to enhanced plasma energy. Consequently, the net deposition rate declines when the bias exceeds an optimal threshold. These results highlight the delicate balance between ion-influenced growth and etching-dominated processes during film growth.

Figure 4a presents the XPS C 1s spectra of nc–Cu/a-C:H nanocomposite films prepared under different substrate negative bias conditions. The C 1s spectra can be deconvoluted into three peaks corresponding to binding energies of approximately 284.6 eV, 285.2 eV, and 286.0 eV, which are attributed to sp² C, sp³ C, and C-O bonds, respectively. Figure 4b illustrates the deconvoluted XPS C 1s spectrum of the film deposited at a negative bias of 200 V. By calculating the area percentage of sp³ C, it is found that the sp³ C content in the films initially increases and then decreases with increasing bias voltage, as shown in Figure 5. It should be noted that the quantitative accuracy of XPS analysis may be influenced by systematic errors, including calibration uncertainties in instrumental sensitivity factors, signal deviations caused by surface roughness or contamination, and subjectivity in peak deconvolution processes (e.g., baseline selection and peak area determination). These factors contribute to systematic errors rather than random errors, which is why error bars—typically used to represent random errors—are not included in the sp³ content curve in Figure 5, derived from the XPS data. When the bias voltage is less than 200 V, the sp³ C content increases with increasing bias voltage, reaching a maximum of 57.3% at 200 V. However, when the bias voltage exceeds 200 V, the sp³ C content decreases. The variation trend of sp³ content with negative bias in Cu-doped DLC is similar to that [14] in Ti-doped DLC. This behavior is primarily attributed to the growth process of the nc–Cu/a-C:H nanocomposite films and the high-temperature, high-pressure environment induced by the thermal spike effect.

The film growth involves two key processes: adsorption on the film surface and the physical shallow-layer injection effect within the subsurface of the film. These processes are influenced by the energy of the ion beam during deposition. At lower bias levels, the energy of the ions is relatively low, resulting in most ions being insufficient to penetrate the growth surface of the film; instead, they are simply adsorbed onto the surface. This promotes the formation of a loosely structured film with graphitic-like bonds, where both the sp³ content and internal stress are minimal. As the bias increases to 200 V, the ions gain sufficient energy to penetrate the subsurface of the film, leading to the generation of compressive stress and densification of the film, which in turn increases the sp³ bonding. At this point, both the sp³ content and internal stress reach their maximum values. The thermal peak effect plays a crucial role in this process, as the localized high-temperature and high-pressure environment generated by ion collisions facilitates the rearrangement of carbon atoms and the formation of sp³ bonds.

When the bias continues to increase beyond 200 V, the observed decrease in sp³ C content and the reduction in internal stress indicate the presence of a stress release mechanism. On one hand, excessively high ion energy may disrupt the ordered arrangement of atoms within the film, inhibiting the formation of sp³ bonds. On the other hand, the dissipation of excess heat generated by ion collisions may relieve high internal stress, leading to the formation of a more loosely structured carbon network and a subsequent reduction in sp³ content, favoring the development of graphitic-like bonds. The interplay between ion energy, sputtering effects, and atomic diffusion influences both the internal stress and the mechanical properties of the films. High stress can negatively impact the adhesion of the films to the substrate, leading to potential delamination or cracking. Conversely, an optimal balance of stress can enhance adhesion.

3.2. Film Structure

Figure 6 presents the XRD spectra of nc-Cu/a-C:H nanocomposite films fabricated with varying substrate negative biases. When the bias is relatively low (50–100 V), the Cu content in the film is relatively small, as can be seen from Figure 1, and no distinct Cu characteristic peaks are observed in the XRD patterns. With the bias increases from 100 V to 200 V, the Cu content in the film rises from 4.7 at.% to 7.9 at.%, and the characteristic peak of Cu (111) plane begins to appear in the XRD patterns. When the bias reaches 300 V, the Cu content is at its highest, and the characteristic peaks are the strongest. Subsequently, as the bias continues to increase, the peak intensity weakens.

Cu grain sizes were calculated using Scherrer’s equation, as shown below [24]:

$$D={\displaystyle \frac{k\lambda }{B\cos \theta }}$$

(2)

where D is the grain size, the constant k = 0.89, B is the full width at half maximum (FWHM), λ is the wavelength, and θ is the diffraction Angle. Figure 7 illustrates the relationship between the grain size of the nc-Cu/a-C:H nanocomposite films and the substrate negative bias, which ranges from 200 V to 500 V. The grain size increases with the substrate negative bias, peaking at 13 nm when the bias is 300 V. At 200 V, the grain size is recorded at 10 nm. As the bias continues to rise to 500 V, the grain size decreases to 12 nm. Cu nanoparticles can enhance fracture toughness by deflecting microcracks, while smaller grains improve hardness via grain boundary strengthening, as per the Hall–Petch relationship. This dual effect enables tunable hardness–toughness trade-offs in nc-Cu/a-C:H films. When the negative bias is below 300 V, as the bias increases, the enhanced ion energy improves the surface mobility of the deposited atoms, promoting the surface diffusion and aggregation of Cu nanoparticles, which in turn facilitates grain growth. However, when the bias exceeds 300 V, excessive ion bombardment causes some of the formed grains to be sputtered off (i.e., re-sputtering effect), resulting in a reduction in grain size. At the same time, the intense ion bombardment may also lead to the amorphization of the grain boundary regions, thereby limiting further grain growth. Additionally, as shown in Figure 1, the Cu content reaches its maximum at 300 V, providing sufficient material for grain growth. At higher biases, the reduction in Cu content (possibly due to intensified re-sputtering) further suppresses grain growth. Therefore, the grain size exhibits a trend of increasing and then decreasing with bias, and the underlying mechanism can be attributed to the competition between ion energy promoting nucleation and the re-sputtering effect.

4. Conclusions

This study demonstrates that the substrate negative bias significantly influences the composition and microstructure of nc-Cu/a-C:H nanocomposite films. As the substrate negative bias increases, the Cu content initially rises and then decreases, reaching its maximum at a bias of 300 V. Concurrently, the deposition rate of the films gradually declines with increasing bias. The sp³ content and internal stress exhibit a similar trend, increasing from 50 V to 200 V before subsequently decreasing. Additionally, the grain size of Cu is minimized at a bias of 200, measuring 10 nm. As the bias voltage increases, the enhanced ion energy promotes better film densification and improves bonding, which can lead to increased hardness and overall mechanical strength. However, at excessively high biases, the adverse effects of re-sputtering and amorphization may counteract these benefits, resulting in a decrease in mechanical performance. These findings highlight the complex relationships between substrate bias, material composition, structural properties, and mechanical characteristics in nc-Cu/a-C:H films, suggesting that careful optimization of bias voltage is crucial for tailoring the desired properties of these nanocomposite coatings.