Room-temperature sputter deposition of gold-colored TiN assisted by niobium bombardment from a bipolar HiPIMS source Open Access

Manufacturing lightweight and durable components is becoming increasingly crucial for fast-growing supply chains in the aerospace, energy, defense, and automobile industries.^1,2 The disadvantages of the light metals used in many of these applications are their low wear resistance and susceptibility to corrosion, especially in applications that prioritize performance.^3,4 One of the solutions to this problem has been the use of chromium electroplating.^5,6 While chromium plating produces surfaces with various colors, including blue, black, brown, green, gray, and gold, its significant drawback is the associated production of toxic chemical waste, the most hazardous component of which is hexavalent chromium, potentially more carcinogenic than diesel exhaust.⁷ 

The European Union has proactively regulated hazardous substances, including hexavalent chromium, restricted under the EU’s REACH (Registration, Evaluation, Authorization and Restriction of Chemicals) regulation. Scientists and engineers are, therefore, actively looking for alternatives to chromium plating.^8,9

Physical vapor deposition (PVD) is a versatile, durable, and environmentally friendly alternative to conventional electroplating. Available substitute materials for chromium plating using PVD processes are titanium nitride,¹⁰ zirconium nitride,¹¹ chromium nitride,¹² and titanium carbonitride.¹³ Titanium nitride, TiN, has been popular and is sought after because of its golden color and impressive mechanical properties.^14,15 The deposition of TiN by PVD methods using direct current magnetron sputtering (dcMS) or cathodic arc evaporation is an established technology.¹⁶ However, the substrate must be held at elevated temperatures, with heat usually supplied additionally. Cathodic arc evaporation can be carried out at a lower substrate temperature than sputtering because of the higher ionization states of the depositing flux, which unfortunately delivers a large amount of thermal energy that may adversely affect temperature-sensitive substrates. Another major disadvantage is the presence of macroparticles in the coating, which induce surface roughness.¹⁷ Plastics such as ABS typically demand that the substrate not exceed the temperature of 50 °C, approximately during the deposition process, which requires frequent interruptions that may result in a long and commercially nonviable process. Without the application of heating to the substrates, TiN coatings tend to lose their unique golden color and impressive mechanical properties. Any alternative process in which the supply of thermal energy is reduced should provide compensation in the form of enhanced energetic impact of the depositing species.

Titanium has a high ionization probability in the plasma because of its large collision cross section for electrons and low ionization energy. High-power impulse magnetron sputtering (HiPIMS) can supply a high fraction of ionized sputtered material¹⁸ and, combined with the application of bias potential to the substrate, has been utilized to tailor and improve the properties of deposited titanium nitride^19–21 and titanium carbonitride films.²² However, applying bias voltage on the substrate is impractical for nonconducting ABS and plastic substrates, and therefore, driving the depositing ions with energy provided by another means is the only viable solution. RF bias is an option, but it is expensive and unsuitable for larger substrates to produce uniform color coatings. Although the application of a time-synchronized magnetic field through the current-carrying coil in conventional HiPIMS has been demonstrated to provide some additional energy through ambipolar acceleration,^23–26 the electrostatic repulsion of ions away from the target in the novel hybrid process known as bipolar HiPIMS^27,28 appears to be a practical solution for industrial applications where more than one cathode will be operating. Bipolar HiPIMS uses a positive pulse after a conventional HiPIMS negative pulse to repel positive ions toward the substrate. This results in improved film properties, such as reduced stress, higher densification, and better coverage of 3D complex parts.^28–31 Alternatively, time-synchronized bias voltage pulses can produce promising results, but challenges remain with more considerably larger target-to-substrate distances to allow for metal ion acceleration and operation of the HiPIMS process using multiple cathodes simultaneously. Therefore, bipolar HiPIMS remains the favored process when either single or multiple cathodes are under operation.²⁹ 

The appearance of coatings to the eye is the key feature for decorative applications. Porosity has a strong influence on color. Gold (570–590 nm) and Rose Gold^TM are relatively highly reflecting bright coatings (with luminosity “L” > 60) deposited with cathodic arc technology, which achieves dense, low-porosity coatings because of energetic ion bombardment. Darker coatings such as gunmetal and deep black are composites of hydrogenated carbon and chromium deposited by sputtering technology with plasma-enhanced CVD, which delivers relatively high porosity dark coatings with a columnar microstructure that assists in light absorption. Black coatings for solar-absorbing surfaces, containing also TiAlNO,³² are preferably deposited using a reactive dcMS system operated in argon–acetylene gas mixtures. The color coordinates L, a, and b* are used to quantify colors for human perception. The “L” represents the lightness of the color, while “a” and “b” represent the color on green-to-red and blue-to-yellow axes, respectively. Bright coatings are created by arc evaporation of Ti and Zr targets to give L values higher than 70, while the sputtering of chromium is preferred for darker DLC colors, particularly for the colors with an L value of less than 40. The disadvantages of cathodic arc technology could potentially be mitigated by developing a sputtering technology with higher ionization and more energetic ions in plasma. Using bombardment by creating energetic ions from the sputtering plasma would be a critical step in lowering the porosity and increasing the density of sputtered coatings to produce more reflective bright colors.

The hybrid bipolar HiPIMS method in this work uses a first magnetron operating on a titanium target in non-HiPIMS dcMS mode combined with a second magnetron operating on a niobium target in bipolar HiPIMS mode. The hybrid unipolar HiPIMS mode refers to the operation of a niobium target in conjunction with the dcMS operating on titanium in a unipolar HiPIMS configuration. Niobium was chosen as a secondary material for assisting the deposition because of its higher atomic mass than titanium, which gives it a range and a momentum that enables it to densify the surface region of the growing film. The energy of the incident niobium is sufficient but not too large to allow it to be incorporated on lattice sites without causing substantial lattice disruption, resulting in a compressive stress that is not excessive. Helmut Holleck calculated the phase diagram for Ti–Nb–N where he showed that NbN and TiN can form a solid solution over the entire composition range. Nb fits very well into the TiN lattice and would not be detrimental by forming a grain boundary or second phase.³³ Although such a hybrid approach has been reported before,^30,31 such an approach has not been used to achieve decorative coatings. This novel hybrid approach tunes the densification of the growing film through the power applied to the bipolar HiPIMS, which, in turn, governs the flux and energy of the Nb ions. This mode of energy delivery eliminates the need for external heating of substrates. The recoiling Nb ions densify the growing titanium nitride films with minimal niobium incorporation and without compromising the gold color. This method also reduces the influence of the target-to-substrate distance on metal ion acceleration, making it suitable for industrial batch-type coaters and the deposition onto large-area temperature-sensitive substrates with complex shapes. The work presented here shows the use of a dcMS source to achieve smooth titanium nitride films with a high deposition rate.

An industrial batch coater (PVT XPro 2500) was operated with two rectangular magnetrons for the film synthesis and a laboratory deposition system (AJA International USA) was operated with two circular magnetrons for time-resolved ion energy measurements (IEDF; see Fig. 1). Two systems were used because the carousel set up in the industrial system intruded on the space required to mount the time-averaging IEDF analyzer during the deposition processes. We made sure that the scale-down measures were taken and the plasma-operating conditions made identical for both chambers, such as deposition pressure 3.5 × 10⁻³mbar, argon–nitrogen gas ratio, and the sputter power density.

In both systems, a Grade 2 (99.99% purity) Ti target was selected for the deposition of a TiN reference film deposited by dcMS. All target materials were procured from Avaluxe International, Fuerth, Germany. A Gencoa rectangular magnetron (400 mm × 100 mm) with a balanced magnetic field configuration resulting in a very low bombardment to the depositing film was used for the TiN deposition using a Grade 2 Ti target. The second magnetron (400 mm × 100 mm rectangular magnetron) in the industrial system with an unbalanced magnetic field configuration was operated in bipolar HiPIMS to generate and accelerate the Nb metal ions from a Grade 2 Nb target. A 10 kW HipV power supply (HipV, Spain) was used in unipolar and bipolar modes to drive the HiPIMS Nb cathode, while an AE Pinnacle power supply was used to drive the dcMS with the Ti cathode. A second HipV power supply was employed to apply the low-voltage DC Bias (−30 V) to the substrate table. The substrate table consists of three spindles of 150 mm diameter, each connected to a 350 mm rotating table. Each substrate is clamped in a different spindle and subjected to twofold rotation during deposition. The laboratory system contains 75.6 mm-diameter circular magnetrons (balanced for the Ti Grade 2 target and unbalanced for the Nb Grade 2 target) kept adjacent to each other. The laboratory deposition system was fitted with a Hiden Analytical quadrupole energy-resolved mass spectrometer detector (Warrington, UK) to obtain IEDF of the singly ionized species of argon, niobium, and titanium during the operation.

The film characterization was carried out using a set of substrates made from three different materials: Silicon (001) wafers were used for the thickness, morphology, and composition measurements in SEM-EDX (Hitachi 4800); electropolished 304L stainless steel (50 × 50 mm²) for the Rockwell adhesion test (Accud AC-HR150A), defined by the VDI 3198 guideline; finally, M2 high-speed steel (HSS) coupons (10 mm diameter) were used for the nanoindentation hardness measurement.

A series of (Ti)_1−xNb_xN films were produced with various values of x in the range of 0.012–0.135. Hybrid unipolar HiPIMS-deposited films have the lowest value of Nb content, whereas hybrid bipolar HiPIMS deposited films have the higher values of Nb content. Reference samples of TiN made by dcMS and unipolar HiPIMS were prepared for comparison with the properties of the hybrid dcMS/bipolar HiPIMS films. A constant voltage bias of −100 V was applied to the substrate table for the deposition process either by dcMS or by unipolar HiPIMS. This was done to compare TiN films deposited on the biased substrate by standard processes with the TiN:Nb films deposited on the floating substrate by hybrid processes.

No external heating was applied during the process, which commenced with a base pressure of 6 × 10⁻⁶ mbar. A pulsed argon glow discharge (450 V and 150 kHz) was applied for 30 min prior to film deposition to remove the substrate contaminants and native oxide, followed by a thin Ti bonding layer deposited for 5 min by conventional dcMS (a constant average power of 3 kW, 398 V–7.56 A) and an argon flow of 56 SCCM. The TiN matrix was deposited with the same dcMS conditions (a constant average power of (3 kW, 433 V–6.92 A), adding a nitrogen flow of 25 SCCM, for a total pressure of P(Ar + N₂) = 5.7 × 10⁻³ mbar. The rotation speed in the industrial system was maintained at 5 rpm.

The average power applied to the niobium HiPIMS target was varied in the range of 0.5–2 kW by adjusting the pulsing frequency, which varied between 200 and 850 Hz, while maintaining a peak current of 160 A (0.40 A/cm² peak current density). The pulsing times were kept constant at 40 μs (negative part of the pulse) and 100 μs (positive part of the pulse). The delay between the negative and the positive parts of the pulse was kept constant at 3 μs. The amplitude of the positive component of the pulse (V_pos) was varied and selected to be either 0 V (unipolar HiPIMS) or +300 V. A (Ti)_1−xNb_xN film prepared with hybrid unipolar HiPIMS was used as a reference.

The selection of pulse parameters (for both the positive and negative parts of each HiPIMS pulse) was based on the experiments published in Refs. 31 and 34, in which it was demonstrated that the application of a short negative pulse followed by a larger positive pulse maximizes the fraction of positive ions accelerated during the positive phase. In this work, we report that the duration of the negative pulse is longer than the time-of-flight of the metal ions so that ions from the negative part of the HiPIMS pulse reach the substrate without being accelerated. This parameter selection depends on the target-to-substrate distance, which ranges between 8 and 15 cm in the present system configuration. The dcMS-TiN deposition rate was approximately 17 nm/min, increasing as the Nb average power was increased from 0.5 to 2 kW (17 to 23.5 nm/min). A further increase in the power values for unipolar HiPIMS compromised the final color (L, a, and b values) and was also affected by the arcing emerging from the HiPIMS cathode. Hence, for this experimental work, we decided to limit the bipolar HiPIMS power values to 2 kW.

The color coordinates were measured by using the TS8500 Benchtop Grating Spectrophotometer (3nh, China). The TS8500 supports two light sources—a xenon pulsed source and an LED source. The repeatability of ΔE*ab is within 0.015, and the interinstrument error is within 0.2. The Lab* color space is commonly used in color science, color management, and image processing to quantify colors in a way that is intended to correspond more closely to human perception than other color spaces such as RGB or CMYK.

The quality of adhesion of the dcMS-TiN films with Nb bipolar HiPIMS onto the stainless-steel substrate was evaluated using a Rockwell indenter (Ernst, NR3D), a reliable and accurate method. To conduct the destructive test, we followed the guidelines of the VDI 3198 standard and applied a load of 100 kg for 10 s using a sphere–conical diamond tip with 120° angular sides. The optical micrographs of the indentation sites allowed us to confidently evaluate the coating failure.

A compositional analysis was performed using glow discharge optical emission spectrometry in a GD Profiler 2 from HORIBA. A complementary compositional analysis was performed using 2 MeV He+ion Rutherford backscattering spectrometry (RBS) to determine the film density. RBS spectra obtained for the samples were analyzed by RUMP simulation software.

The phases and textures in the coatings were examined using a PAN-analytical x-ray diffractometer in the Bragg–Brentano configuration using CuKα radiation from a Cu anode operated at 45 kV and 40 mA and a Ni filter. During data treatment, the background was subtracted, and each spectrum was normalized so that the highest peak belonging to the coating material and substrate peaks were aligned to a steel diffraction pattern to evaluate peak shifts. The hardness values were estimated from the nanoindentation curves that were obtained by nanoindentation using a Hysitron TI950 triboindenter equipped with a Berkovich-shaped diamond tip, whose area function was calibrated using multiple indents on a standard fused silica sample. Penetration depths were kept below 10% of the total film thickness to minimize substrate effects. At least 10 indents were made per sample, with maximum loads of 12 mN. The Stoney formula was used to calculate the residual stress in the coatings directly from the measured curvature of the substrates before and after the deposition process.³⁵ The density of the coatings was measured using RBS with 2 MeV He ions, and the results were analyzed with RUMP simulation software.

The hybrid bipolar HiPIMS process uses a first magnetron operating on a titanium target in non-HiPIMS dcMS mode combined with a second magnetron operating on a niobium target in the bipolar HiPIMS mode. On the other hand, the hybrid unipolar HiPIMS process uses a first magnetron operating on a titanium target in the dcMS mode combined with a second magnetron operating on a niobium target in the unipolar HiPIMS mode.

Figure 2 shows the voltage applied to the Nb cathode (operated with HiPIMS), the discharge current, and the net ion current collected at the substrate (bias current) as a function of time. The data were obtained using the laboratory system for unipolar and two bipolar HiPIMS hybrid processes with different positive pulse voltages, together with a Ti target operated with dc reactive magneton sputtering in a nitrogen atmosphere.

For bipolar HiPIMS, during the negative part of the HiPIMS pulse, the net substrate current is dominated by electrons, whereas ions dominate during the positive part of the pulse. The ion current at the substrate in the positive part of the bipolar HiPIMS is much larger than the ion current at the corresponding phase in unipolar HiPIMS, and it is sustained throughout the positive part of the pulse, which is the result of ions created in the dc sputtering discharge being propelled to the substrate by the electric field. The large ion current to the target during the negative part of the HiPIMS pulse declines to zero when the unipolar HiPIMS pulse is switched off, whereas in bipolar HiPIMS, the target draws a large electron current during the positive pulse.

Figure 3 shows the color coordinates for coatings obtained using all the deposition conditions. The coating that most closely approaches the international standard for gold was TiN deposited by unipolar HiPIMS with substrate bias, followed by the hybrid mode combining dcMS-TiN and bipolar niobium HiPIMS operated at 2 kW power but without substrate bias.

The deposition rate for the various processes as a function of power applied to the Nb target (where applicable) is shown in Fig. 4, together with the compressive stress, hardness, and density of the TiN coatings. Without operating the Nb target, the highest deposition rate is obtained with dc magnetron sputtering rather than HiPIMS. When the niobium target is operated, the deposition rate increases with power supplied to the niobium target for both hybrid bipolar and hybrid unipolar HiPIMS. The highest deposition rate is obtained for the hybrid bipolar HiPIMS process. As the Nb target power in hybrid bipolar HiPIMS is increased, the density progressively increases. The coatings produced at the highest deposition rate have hardness and density values comparable to those for TiN coatings prepared with non-hybrid unipolar HiPIMS with substrate bias. The compressive stress increases with niobium power but remains below the values obtained for hybrid unipolar HiPIMS. These results, combined with those presented in Fig. 3, demonstrate that the use of hybrid bipolar HiPIMS produces coatings with required L,a,b* values and with hardness and density comparable to TiN coatings prepared with nonhybrid unipolar HiPIMS (without dcMS), while the deposition rate is much higher and the compressive stress is lower.

The SEM cross-sectional image in Fig. 5(a) shows the formation of a dense columnar structure for a TiN film grown by nonhybrid unipolar HiPIMS with substrate bias, whereas open-voided tapered fibrous structures were obtained for pure dcMS-TiN coatings produced without HiPIMS [Fig. 5(b)]. As the power applied to the Nb cathode in the hybrid bipolar HiPIMS is increased, the open-voided tapered fibrous structures transform to dense columnar structures [Figs. 5(c)–5(e)]. The lower right image (5e) shows that the columnar structure becomes more compact, and when the Nb bipolar deposition power is increased to 2 kW, the surface becomes smoother [Fig. 5(g)]. As the Nb deposition power in hybrid bipolar HiPIMS is increased, Nb concentration in the TiN films also increases [Fig. 5(f)]. The films deposited by the hybrid bipolar HiPIMS process have only a moderately high Nb content compared to the films deposited by the hybrid unipolar HiPIMS process.

Figure 5(h) compares the XRD patterns of pure TiN deposited by unipolar HiPIMS with substrate bias with the hybrid bipolar HiPIMS processes that use a niobium target in the bipolar HiPIMS mode at two different powers (0.5 and 2 kW). As we increase the power to the bipolar target, the peak corresponding to the (111) lattice plane shifts toward lower 2θ values, while the (200) peak moves toward a higher 2θ region. The diffraction peak corresponding to the film deposited by unipolar TiN HiPIMS with substrate bias has sharp peaks with the lowest 2θ value for (111) and the highest 2θ value for (200). By increasing the power applied to the Nb target in bipolar HiPIMS, the intensities of the peaks (111) and (200) decrease and the peaks become broader.

To evaluate the adhesion of TiN coatings to the substrate, we used the HF grading scheme, where HF1 is the highest level of adhesion and HF6 represents the lowest level.³⁶ The Rockwell indentation test (VDI 198) reveals the highest adhesion of HF1 on the TiN layers coated using the hybrid mode, which combines dcMS-TiN and bipolar niobium HiPIMS operated at 2 kW. In contrast, the HF2 result for the unipolar TiN films with substrate bias indicates that the surface exhibited a lower adhesion (Fig. 6). The HF1 adhesion level is observed for the TiN layers coated by the hybrid mode operated at all powers [figures not shown as they appear the same as the micrograph shown in Fig. 6(a)].

The time-averaged IEDF measured for singly ionized ions of titanium, niobium, and argon is shown in Fig. 7. In the hybrid unipolar HiPIMS mode of operation at 2 kW, the energies of all ions are contained within 40 eV at half peak flux intensity. On average, the niobium ions are the most energetic, followed by argon and then titanium ions.

During the application of bipolar HiPIMS, the positive part of the applied potential forces the plasma to a large positive potential, propelling some of the depositing species to an energy of more than 100 eV. At the constant voltage of 300 V applied during the positive part of the pulse, the higher the power (through an increased repetition rate), the higher the transfer of energy and momentum between the species. This is explicitly observed in the increased intensities and energies of singularly ionized titanium and niobium as a function of increased applied power (Fig. 8). It is important to note that the intensities of argon ions are now low relative to the intensities of depositing Nb ions, and this may contribute to the lower compressive stress relative to TiN coatings deposited using HiPIMS operating on the titanium target.

The results of this study show that the color hue (L,a,b*) and mechanical properties such as the residual stress of TiN films are modified favorably by the application of a hybrid bipolar HiPIMS process compared to hybrid unipolar HiPIMS process, non–hybrid unipolar HiPIMS TiN with substrate bias process and pure dcMS TiN process. The trends toward increasing hardness, density, and compressive stress that occur with increasing HiPIMS power (Fig. 4), which we show causes increased bombardment energy to the niobium ions (Fig. 8), agree with those of previously reported work³⁷ in which substrate bias applied during cathodic arc deposition improves the mechanical properties of titanium nitride. The substrate bias also increases the energy of the bombarding ions. The significance of these results is that we can achieve the deposition of bright coatings with the desired gold color onto temperature-sensitive substrates with outstanding adhesion and free of macroparticles. The coatings achieved by cathodic arc deposition have the disadvantage of a relatively high content of macroparticles.³⁸ Removing macroparticles is possible by using a magnetic filter; however, this dramatically reduces the deposition rate. Our hybrid bipolar HiPIMS films, the TiN dcMS films deposited with Nb bipolar HiPIMS, offer gold-colored coatings with a smooth surface and HF1-level adhesion [Fig 5(c)] at a high deposition rate.

Our findings highlight the advantages of hybrid bipolar HiPIMS as a means of providing energetic ions to improve the mechanical properties of films without the need for substrate bias. The increase in energy that can be seen when comparing niobium ion energies at the same power and source-to-substrate distance between hybrid HiPIMS operated either in the unipolar (Fig. 7) or in the bipolar mode (Fig. 8) are consistent with the hypothesis that in bipolar HiPIMS, the positive bias pulse or “kick pulse” applied to the target repels ions already in the plasma and in the sheath formed in the immediately preceding negative pulse. In this way, hybrid bipolar HiPIMS facilitates the energetic impact of depositing ions onto nonconducting substrates. In contrast, in case of applying substrate bias, either synchronized or continuous, only the ion fluxes that have already reached the substrate sheath will be accelerated. Substantial advantages in total ion flux and acceleration of ionic species accelerated by bipolar HiPIMS lead to the observed changes in the mechanical properties with a substantial increase in the deposition rate. Note that applying bias voltage on the substrate is impractical for nonconducting ABS and plastic substrates, and therefore, driving the depositing ions with high energy provided by another means is the only viable solution.

Before continuing our discussion on selective ion acceleration, it should be pointed out that dcMS-TiN samples deposited with unipolar HiPIMS and without substrate bias possess a slightly higher density than the TiN samples deposited by dcMS alone. A possible explanation for this difference is that in the presence of the HiPIMS discharge, electrically floating substrates acquire a negative potential that typically reaches up to −16 to −25 V,³⁹ something that the grounded substrates series never do, even when bipolar HiPIMS is used. When using the Nb ion assistance of hybrid unipolar HiPIMS, the energy tail of the niobium ions driven from the Nb target extends to 30 eV at half of its intensity (Fig. 7). Therefore, the energy profile of the Nb ions in the plasma contributes energy in addition to that derived from the self-bias. The cumulative energetic impact is sufficient to overcome the surface lattice displacement threshold and enhance the density of TiN coatings in comparison with the coatings obtained by pure dcMS [Fig. 4(d)].

It is of interest to correlate the evolution of the deposition rate, density and mechanical properties such as hardness, and compressive stress, film cross-sectional morphology, and most importantly, the color coordinates L, a* and b* with the increasing energies and intensities of ions in the plasma volume, which are modulated by the magnitude of the bipolar positive voltage kick pulse. The increase in the deposition rate shown in panel 4a enabled by bipolar HiPIMS of Nb can be explained by the increase in the flux of depositing titanium and niobium ions accelerated toward the substrate as the power applied to the bipolar HiPIMS plasma is increased (Fig. 8). As the Nb deposition power in hybrid bipolar HiPIMS is increased, the hardness [Fig 4(b)] and density [Fig. 4(d)] progressively increase, which can be explained by the elevation of energy of depositing titanium and niobium ions in the plasma volume that are more strongly accelerated toward the substrate as the power applied to the bipolar HiPIMS plasma is increased. Two processes take place: first, the film void percent is decreased (increasing the number density of atoms), and then, Nb content is increased, increasing the average mass density by increasing the average atomic weight. A closer packing of the columns tends to give a brighter color, represented by high L values [Fig. 3(a)].

The compressive stress was highest for unipolar HiPIMS-TiN films deposited with the substrate bias, while it is lowest for the dcMS-TiN films [Fig. 4(c)]. Compressive stress values for Nb:TiN films made with the hybrid mode using bipolar HiPIMS show a clear trend to increase with bipolar HiPIMS power but remain below the level of compressive stress shown by the films made by unipolar HiPIMS with substrate bias. According to Fig. 5(f), these bipolar films have only a moderately high Nb content than the unipolar films. Figure 4(c) indicates that bipolar films have a slightly higher compressive stress than the unipolar ones. Assuming the atom and ion radii of both Ti, Nb, and their masses, the stress formation may not be related much by the solid solution formation, but rather by the ion impact. We attribute this reduction to the lower flux and energy of argon ions in the plasma volume (Fig. 8) relative to the flux and the energy of the depositing metal ions. The Nb:TiN films made with bipolar hybrid mode have properties close to those made with unipolar TiN HiPIMS with substrate bias [Fig. 4(a)]. By increasing the power applied to the Nb target in bipolar HiPIMS, the intensities of the XRD peaks (111) and (200) decrease and the peaks become broader [Fig. 5(h)], confirming the decrease in the film densities as confirmed by the RBS results [Fig. 4(d)]. The superior HF-1 level film adhesion for the Nb:TiN films by hybrid mode can be explained by the lower levels of compressive stress with respect to unipolar TiN HiPIMS [Fig. 4(c)]. We consider that argon ions are especially effective at generating the highest levels of compressive stress, as previously suggested.^25,40 Nb was used to give a selective acceleration of depositing ions in the plasma volume, whereas the use of titanium in bipolar HiPIMS leads to very high stress and poor adhesion, as explained in our previous work on the influence of the size and mass of the ions and their contribution of energetic impacts on the overall compressive film stress.⁴¹ 

The enhancement in the energy of impact of Nb ions on the growing TiN films has effects on the film structure observed in the SEM cross sections. The images show an open-voided tapered fibrous structure for dcMS-TiN films [Fig 5(b)] that becomes denser in the film deposited with hybrid bipolar HiPIMS at a power of 2 kW [Fig. 5(e)]. The close packing of structures gives a stronger reflection of light and therefore a brighter color, which leads to the increase in L values (Fig. 3).

This work demonstrates the use of a hybrid dcMS and bipolar HiPIMS process to produce smooth Nb:TiN coatings with the sought-after gold color and with a hardness and density comparable to TiN coatings deposited with unipolar HiPIMS equipped with substrate bias. The hybrid process maintains a high deposition rate, while keeping the compressive stress relatively low, to avoid adhesion problems. The high deposition rate is the contribution from the dcMS, while the deposition temperature is maintained low on sensitive substrates, since external substrate heating is eliminated and replaced with the impact of energetic ions accelerated by the positive voltage “kick” from the bipolar HiPIMS. The process is suitable for temperature sensitive insulating substrates as it requires no substrate bias or heating.

This project was partially supported by the Australian Research Council. We thank Masoud Zhianmanesh from The University of Sydney for the schematic diagram. We also thank Joerg Patscheider for the valuable discussion and scientific inputs.

The authors have no conflicts to disclose.

Ivan Fernandez-Martinez: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Funding acquisition (equal);Investigation (equal); Methodology (equal); Project administration (equal); Resources (equal). Rajesh Ganesan: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Behnam Akhavan: Investigation (equal); Supervision (equal); Visualization (equal); Writing – review & editing (equal). David T. A. Matthews: Conceptualization (equal); Formal analysis (equal); Supervision (equal); Writing – review & editing (equal). Michael Stueber: Conceptualization (equal); Data curation (equal); Investigation (equal); Supervision (equal); Validation (equal); Writing – review & editing (equal). Marcela M. M. Bilek: Conceptualization (equal); Formal analysis (equal); Supervision (equal); Writing – review & editing (equal). David. R. McKenzie: Conceptualization (equal); Data curation (equal); Supervision (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal).

The data that support the findings of this study are available from the corresponding authors upon reasonable request.