A comparison of plasma generation, plasma transport, and film formation for a DC vacuum arc source with Ti–X compound cathodes (X = W, C, Al, and Si)

Vacuum arc generation of a multi-component plasma from a compound cathode is today a common method to deposit decorative and functional thin layers.¹ A wide range of elemental combinations within the cathode and, as a result, in the deposited films, allow improved and tuned film properties to meet industrial requirements.² Ti is a commonly used element in industrial coatings, with properties enhanced by the addition of different elements. For example, the addition of Al into a Ti cathode and arcing from the compound cathode in a reactive N2 atmosphere, results in a significantly increased oxidation threshold temperature of the resulting (Ti, Al)N coatings of up to 800 °C, compared to 500 °C for a TiN coating.^3,4 However, the added elements affect the properties of the generated plasma. For instance, the addition of Al leads to a decrease in the average charge state and average kinetic energy of the generated ions;^5,6 while the addition of, e.g., Si leads to the opposite effect.⁷ The average ion charge state and the average ion kinetic energy directly determine the energy provided by the plasma to a substrate, which is an important parameter for the structural and compositional evolution during film deposition.^1,8,9 Moreover, there are also reports of different effects of the addition of Al/Si on the intensity of the macroparticle generation from the arc spot, resulting in variable roughness of the deposited coatings.^5,7,10,11 Despite the range of Ti-based compound cathodes, e.g., Ti–Al, Ti–Si, Ti–B, Ti–Cu, and Ti–Hf, for which selected features of the arcing process have been reported,^12–15 there is still a limited knowledge of fundamental changes of the plasma generation caused by modification of the cathode surface through the addition of an element into a Ti cathode. Since the arc spot directly determines the plasma generation process and, therefore, influences the deposition of the thin films, further investigation of arcing from Ti-based compound cathodes is required, both for fundamental understanding and industrial applicability.

Examples of elements diverging from the characteristics of Ti, but still being important for applied coatings, are C and W. The first notable discrepancies are the melting and boiling temperatures, which for Ti are 1941 and 3560 K, and for W are 3695 and 6203 K, respectively. Hence, the boiling temperature of Ti is lower than the melting temperature of W. As previously reported in Ref. 16, such a discrepancy can have a strong influence on both plasma generation and film formation. As shown for arcing from a Mo–Cu cathode, an addition of a relatively small amount of Cu (5–15 at. %) drastically changes the charge states, the kinetic energies of the generated ions, the intensity of the ion flux, and the intensity of and mechanism behind macroparticle generation. Altogether this leads to Cu drastically affecting the resulting film properties.¹⁶ In turn, C does not have a melting and/or boiling temperature, but its sublimation temperature is 3915 K, which is still higher than the evaporation temperature of Ti. The atomic masses of the elements are also very different, with the mass of a C atom/ion being significantly lower than for Ti, which, in turn, is much lighter than W. The “velocity rule” states that the peak velocities of different ion species in plasma generated from a compound cathode are equal, and, therefore, the difference in the peak ion energies of those species is determined by the difference in their mass.¹⁷ This suggests clearly diverging properties of plasmas from Ti–C and Ti–W cathodes. In turn, the “cohesive energy rule” establishes a correlation between plasma properties and the cohesive energy of the cathode,¹⁸ suggesting that, due to the discrepancy in the cohesive energies of Ti, C, W and their possible intermetallic compounds,¹⁹ the plasma properties will be dependent on the constituting elements as well as on their relative amount within the cathode.

Coatings based on both Ti–W and Ti–C find areas of industrial applicability. For instance, Ti–W thin films are proposed to be used as diffusion barriers for metal contacts.²⁰ W is supposed to serve as an interlayer diffusion barrier, while Ti can reduce oxidation and act as a barrier for grain boundary diffusion.²¹ Further, in the machining technology industry, Ti–C-based compounds have been extensively investigated and are characterized by excellent performance for cutting applications.^22,23 Though analyses of plasma and films generated with and deposited by a DC vacuum arc from Ti–C and Ti–W cathodes of different compositions have not yet been reported.

Here, we investigate the influence of Ti–C and Ti–W cathode composition on cathode, plasma, and basic film properties for a DC vacuum arc deposition. The results are compared with previous work on Ti–Al and Ti–Si compound cathodes, performed in a similar deposition system for the same experimental conditions. All experiments were performed at base pressure, without any macroparticle filtering. Consequently, already established effects of gas and external magnetic field on the plasma properties could be excluded, which allows us to provide a base for an improved understanding and for a potential expansion of the application area for DC arc plasma generation from compound cathodes for thin film deposition.

The experiments were performed using a deposition system equipped with an industrial-scale DC arc source (Ionbond) for 63-mm-diameter cathodes. Ti, C, W, Ti1–xCx (x = 0.05, 0.10, 0.25), and Ti1–xWx (x = 0.05, 0.25, 0.5, 0.7) cathodes were used (x represents atomic percent), as produced by powder metallurgy.²⁴ The Ti–W cathodes are produced from a sintered mix of Ti and W powder. In turn, Ti–C cathodes are made from a sintered mix of Ti and TiC powders. To achieve steady state conditions, the cathodes were used for at least 10 min prior to any measurements. The arc source was operated at a 90 A arc current. The base (operational) pressure of the system was around 5 × 10⁻⁶ Torr. A mass-energy analyzer (MEA, Hiden Analytics model EQP) was placed in front of the arc source with the orifice (50 μm diameter) about 35 cm from the cathode surface. For each cathode, the plasma was characterized through mass-scans at a fixed ion energy and energy scans at a fixed mass-to-charge ratio for all detected ions. The energy scans were recorded in steps of 0.25 eV/charge up to 250 eV/charge to capture the entire ion energy distribution (IED).²⁵ The raw data were also treated with consideration of the specific features of EQP diagnostics of multiply charged ions, in line with Refs. 26 and 27. The presence of isotopes in the ion flux and their influence on the relative ratios of the measured IEDs were evaluated in line with previous work.^5,27 Each IED was recorded at least three times to ensure the consistency of the data. Over time, the MEA orifice may be coated and the recorded intensity reduced as an effect of reduced orifice size. To prevent such an effect, the inlet channel was cleaned after finalizing the analysis from each cathode. To determine the plasma composition, the IEDs were integrated to obtain areas proportional to the number of ions of each species. The average ion energies and the ratios between the total intensities of IEDs of different ions in one cathode were found to be reproducible within 5%.

Films were deposited by fixing a Si (100) substrate in a position equivalent to the front end of the plasma analyzer, at grounded potential. Temperature calibration showed a substrate temperature not exceeding 250 °C. The surface morphology of the films and the cathodes and the composition of Ti–W films were characterized using a LEO 1550 scanning electron microscope (SEM) equipped with energy dispersive x-ray spectroscopy (EDS). The Ti–C film composition was determined through time-of-flight energy elastic recoil detection analysis (TOF-E ERDA), using a 36 MeV¹²⁷I⁹⁺ ion beam at 22.5° incidence angle relative to the surface and a 45° recoil angle.²⁸ The resulting time-of-flight vs recoil energy spectra were evaluated using the CONTES code.²⁹ 

Visual observation of the macroparticle generation was performed using a Nikon D800 digital camera.³⁰ A Nikkor 50/1.8 objective was used with an aperture of 8; sensitivity was set equal to ISO 100. Photos were taken at approximately 80 cm from the cathode surface, through a glass window.

Figure 1 shows ion energy distributions obtained from the elemental C, Ti, and W cathodes at base pressure (5 × 10⁻⁶ Torr).

Figure 1 clearly demonstrates a discrepancy in the properties of the generated plasmas. Integrating the IEDs shows that the total number of ion counts is very different, 6 × 10⁷ for C; 3 × 10⁷ for Ti; and 1 × 10⁷ for W. The presented measurements were performed with an EQP equipped with an electron multiplier, giving a signal which is proportional to the number of ions registered and independent of the ion charge state.²⁵ Therefore, the dissimilarity of the measured total ion counts with the used discharge current of 90 A could be an effect from a corresponding change in the ratio of the generated ion flux current and the full arc current (the so-called “normalized ion current”).¹ The present results are consistent with Ref. 31, where the normalized ion current from the C, Ti, and W cathodes are 19%, 9.7%, and 5%, respectively. However, plasmas from the C, Ti, and W cathodes consist of different numbers of highly charged ions. Therefore, for the same ion flux current, plasmas from different cathodes can consist of a different total number of ions, which would influence our recorded signal. A similar correlation between the total ion count and the normalized ion current was previously also confirmed for Ti–Al and Ti–Si cathodes.^5,7

A comparison of the IEDs generated from the Ti and W cathodes suggests that these plasmas are in line with the cohesive energy rule, which states that an element with a higher cohesive energy (8.99 eV for W and 4.85 eV for Ti)¹ will have a higher ion charge state and ion kinetic energy. For the plasma generated from the C cathode (the cohesive energy is 7.37 eV),¹ the cohesive energy rule used in a comparison with Ti and/or W plasmas does not work. This may be due to the special features of arcing from a C cathode, and the operation of so-called third-type arc spots, which cannot be directly compared with metallic plasmas.¹ It should be noted that in the previously reported Ti–Al and Ti–Si studies, the cohesive energy rule for the elemental cathodes is confirmed without any exception.^5,7,10,11

The IEDs measured for the C plasma seem to be the narrowest ones, having the lowest peak value (22 eV for C; 52 eV for Ti; 160 eV for W), while the W plasma consists of the most energetic ions with the widest kinetic energy distributions. Notably, the W plasma is also characterized by the highest level of ionization, with the presence of 5+ ions, while the Ti plasma has ion charges up to 3+, and the C plasma up to 2+ only. This is in line with previous work^5,7 and the correlation reported in Ref. 27, where it is experimentally established that for plasma from a heavier element cathode, the generated ions typically have a higher kinetic energy, a wider ion energy distribution, and a higher ionization level.

The IEDs from the elemental C, Ti, and W cathodes allow the evaluation of possible changes in the plasma properties with arcing from Ti-rich (Ti–C and Ti–W) cathodes. Figure 2 shows IEDs from compound Ti–C and Ti–W cathodes.

As a first observation of Fig. 2, there is a significant increase in the peak energy of the Ti ions from the Ti0.75C0.25 cathode compared to plasma from an elemental Ti cathode, cf. Fig. 1, while the peak energy is similar for Ti ions from the Ti0.75W0.25 cathode and from the Ti elemental cathode. Furthermore, for the C ions, the peak kinetic energies are the same for plasma from both the elemental C cathode and the Ti0.75C0.25 cathode.

It is known that the peak kinetic energies are given by the “velocity rule,” which states that peak velocities of ions from different elements from one compound cathode are close to the same.¹⁷ For the elemental cathodes, the peak velocities, obtained from the peak energies in Fig. 1, are 18.8 km/s for C, 14.3 km/s for Ti, and 13.1 km/s for W. For the Ti0.75C0.25 compound cathode in Fig. 2, the velocities are 19.2 km/s for C and 19 km/s for Ti. For the Ti0.75W0.25 cathode, the ion velocities are 14 km/s for W and 14.3 km/s for Ti. Taking all these velocities into account, one may conclude that the addition of an element, which is faster in its single element cathode, into a mixed element cathode, leads to the acceleration of all ions in the plasma from the compound cathode. It is also possible that a significantly higher amount of Ti in the Ti0.75W0.25 cathode may result in Ti causing an increase in the resulting velocity (and, therefore, also the energy) of the W ions. However, an explanation which is based on a discrepancy in the amount of the cathode elements cannot be used for the Ti0.75C0.25 cathode, where the significantly lower amount of C seems to be the determining factor of the velocity. Notably, the Ti–W cathodes were manufactured with elemental W and Ti powder, while manufacturing of the Ti–C cathodes was done with Ti and TiC powder. The cohesive energy of the TiC (7.4 eV) is similar to the elemental C (7.37 eV). This similarity and generation of C ions from a TiC compound makes the kinetic energy of C very similar for the elemental (C) and compound (Ti0.75C0.25) plasmas. Whether the relatively high amount of TiC (50 at. %) in the Ti0.75C0.25 cathode is responsible for the resulting high velocity of the Ti ions remains a question that requires further, detailed investigation, depending on the actual phase content in the manufactured sintered cathodes.

It should be noted that the total intensities of the plasmas from the compound cathodes are changed compared to the elemental cathodes. For plasma from the Ti0.75C0.25 cathode, the total intensity is around 2 × 10⁷ counts, while for the plasma from the Ti0.75W0.25 cathode, the total intensity is slightly lower than 1 × 10⁷ counts, to be compared to an intensity of 3 × 10⁷ for elemental Ti (see Fig. 1). This reduction, compared to the elemental Ti cathode, can be caused by a change in the erosion rate of the cathode surface with an increase in the cohesive energy of the operational surface of the cathode.

Further analysis of plasma generated from the different cathodes at various operating conditions is done by the calculation of the average ion energies and average ion charge states from the measured IEDs (see Ref. 5). Figure 3 presents the average charge states and average ion energies as a function of the corresponding cathode composition; the x axis displays the relative amount of C/W (present work) and Al/Si (previously reported) incorporated in a Ti cathode.

Figure 3 demonstrates that changes in the average ion charge state and in the average ion energy for most cathode compositions generally display the same trends. However, for the Ti–C system, averages of the Ti and C ions have somewhat diverging trends. The addition of C into the Ti cathode seems to increase both the energy and the charge state of the Ti ions, while for the C ions, there are less significant changes, especially for the average ion energy. Such discrepancy in the ion characteristics may be understood based on the generation of C ions from a TiC compound, as discussed above. That is, the similarity of the cohesive energy of the TiC compound and elemental C, makes the C ions only weakly dependent on the amount of the elemental Ti in the cathode.

A higher charge state of C from a Ti–C cathode, compared to from an elemental C cathode, cannot be explained by the cohesive energy rule.¹⁸ As mentioned above, the cohesive energies of C and TiC are similar. Still, the presence of Ti atoms, which have lower ionization energies than C and therefore lose electrons more easily, can lead to a change in the density of the electron cloud at the arc spot. This would result in ionization, recombination, and charge-exchange processes there, which may lead to more energetic ionization of C in the vicinity of the cathode spot. Moreover, the electron temperature within the generated plasma can attain a weighted average of the pure element electron temperatures of the contributing materials.¹⁵ The electron temperature for Ti and C in pure form is 3.2 and 2.0 eV, respectively.¹ An average between these values would justify higher charge states for C.

For the Ti–W system, there are less significant changes in the average ion energy and ion charge state with a change in cathode composition compared to the Ti–C system. The kinetic energy of Ti and/or W ions and their charge states vary only in the range of up to ±10%. Such behavior, if one considers the ions separately, cannot be explained by applying the cohesive energy rule. The significant discrepancies in the Ti and W cohesive energies (4.85 eV and 8.99 eV respectively) and in their electron temperatures (3.2 eV and 4.3 eV) would suggest significantly different ion properties with a change in the relative amount of Ti and W in the cathode composition. Though not discussed further here, the explanation is obviously more complex. Only the change from the elemental Ti cathode to the elemental W cathode is in line with the cohesive energy rule, i.e., the ion energy increases with an increase in the cohesive energy of the cathode material.¹ 

For plasmas from the Ti–Al and Ti–Si cathodes, as seen in Figs. 3(b) and 3(d), the cohesive energy seems to be a contributing factor. For the Ti–Al case the difference between Ti (4.85 eV) and Al (3.39 eV) cohesive energies, as well as between cohesive energies of their intermetallic compounds, seems to result only in a slight reduction in the averages. However, for more than 50 at. % Al in the cathode, the cohesive energy rule seems to not be applicable, and other mechanisms must come more into play. A possible explanation is that when Al becomes the main cathode component, the “velocity rule”¹⁷ and the gas dynamic mechanism of ion acceleration,³² in combination with the lower mass of Al, leads to an increase in the energies of both elements.⁵ For the average charge state of the Ti–Al plasma, no significant differences for the elemental ions are seen. However, the overall average ion charge state is decreasing with the addition of Al into the cathode, which is in line with the cohesive energy rule. For the Ti–Si plasma, an increase in the relative amount of Si increases both averages, which is in line with the corresponding increase in the cohesive energy of the operational surface of the cathode (see details in Ref. 7).

Summarizing the integrated areas of the IEDs for all ions in the plasma and applying the corresponding correction coefficients based on the natural isotope distributions⁵ allows investigation of the discrepancy between the plasma and the cathode composition (see Fig. 4).

Figure 4 shows that the ionized phase of the plasma from the Ti–C cathodes consists of less C than in the corresponding cathodes. The results show the opposite for the Ti–W cathodes, with significantly more W ions in the generated plasma compared to in the cathode.

For a DC arc operated in the steady state regime, the total material flux from the cathode has to be the same as the cathode composition. However, the analysis performed here is done directly in front of the cathodes, along the surface normal of the cathode. Therefore, possible discrepancies in spatial distributions of the elements may explain diverging cathode–plasma compositions. The present results are in line with Ref. 33, where it was suggested that the generation of the ion flux within the arc spot leads to a situation where heavier elements tend to be at the center of the ion flux. The lack of ions of a lighter element at the center of the ion flux, normally on the cathode surface, has also been seen in the plasmas generated from the Ti–Al and Ti–Si cathodes (see Fig. 4, right).^5,7 Moreover, it is well known that material flux from the arc spot consists of ions, neutrals, and macroparticles.¹ Therefore, diverging cathode and plasma (ion) composition may at least in part be due to the presence of neutrals and macroparticles of a composition different from the cathode composition.

To investigate the effect of cathode and plasma composition on the materials synthesis, thin films were deposited. The Ti/W ratio in the films was determined by EDS, and the Ti/C ratio was determined by ERDA. Figure 5 (left graph) presents the relative amount of C/W within the deposited films as a function of the cathode composition.

Even though the observed trends in Fig. 4 are found also in Fig. 5, there is only a slight depletion of C and a slight excess of W within the films, compared to the cathode composition. This may be explained by the above-suggested discrepancy in the element ratios within the ion flux, the neutral flux, and the macroparticles, and possible variations in the angular distributions of the elements in the generated plasma. It is noteworthy that for the Ti–Al and Ti–Si cathodes (Fig. 5, right), the compositions of the deposited films were found to be consistent with the cathode compositions.^5,7

The EQP mass-energy analyzer does not allow analysis of the ratio between neutral species and the ion flux.²⁵ Still, the macroparticle generation can be evaluated. Figure 6 shows SEM images of the surface of the deposited films.

Figure 6 shows that with an addition of a high-melting-point material (W) to a Ti cathode, the intensity of the macroparticle generation is significantly reduced. The reduction is possible due to a corresponding reduction in the volume of the molten bath within the arc spot, caused by a higher average melting point of the cathode material.^34,35 In addition, the number of macroparticles at the film surface can also be reduced by a faster solidification (for higher-melting-point materials) during the flight between the cathode and the film surface, resulting in macroparticles bouncing off the substrate. This is exemplified for the C cathode, where there is an intensive flux of visually observed macroparticles (see Fig. 7), while the resulting film surface is smooth (see Fig. 6). It should be noted that none of the other cathodes investigated in the present study produce a macroparticle flux that can be visually observed.

Analyzed macroparticles from Ti–C and Ti–W cathodes were all found to be Ti-rich (compared to the cathode composition), while macroparticles from the Ti–Al and Ti–Si cathodes were of a composition close to the cathode composition.^5,7 An explanation for these observations can be found in the difference in melting and/or evaporation temperatures of the phases found in the cathodes (see Fig. 8). For the Ti–W and Ti–C cases, elemental Ti has the lowest melting/evaporation temperature in comparison to elemental W, C, or possible compounds with these elements. Hence, melting of Ti-rich grains on the cathode surface is suggested to be preferred, contributing to macroparticle formation.

For the Ti–W films, Ti-rich macroparticles can contribute to the similar cathode and film composition (see Fig. 4), as the ion phase composition of the plasma has significantly less Ti ions (see Fig. 5). Though the low number of macroparticles on the Ti0.75W0.25 and Ti0.5W0.5 films suggests that the macroparticles in the film are not enough to account for the plasma-film composition difference and, therefore, an excess of Ti in the neutral flux is suggested.

For the Ti–C system, Ti-rich macroparticles on the film surface would lead to a cathode-film composition difference even larger than that from the cathode–plasma composition difference (see Fig. 4). However, as observed, the Ti–C film composition (Fig. 5) is close to the corresponding cathode composition. Melting and evaporation temperatures of TiC are 3430 and 5090 K, much higher than for elemental Ti. Therefore, as suggested for the Ti–W system, an intensive melting-evaporation of Ti can be expected. However, it is also well known that the ionization degree of carbon resulting from an arc spot operating on an elemental carbon cathode is around 75% only,³⁶ while for arcing from elemental metallic cathodes, the ionization degree can be close to 100%.¹ We, therefore, expect a significant flux of C neutrals, contributing to the resulting film composition.

Figure 8 shows the effect of arcing on the operational cathode surface for the Ti0.5W0.5 and Ti0.75C0.25 cathodes. The XRD shows clear peaks only for the elemental Ti, W, and TiC with no detectable presence of any other phases. Notably, for the Ti–C cathode, elemental Ti forms the hexagonal α structure. For the Ti–W cathode, Ti in the as-received cathode is in the cubic β form, with a slight presence of the hexagonal α Ti within the Ti0.95W0.05 cathode. The β structure of Ti is expected from the production and sintering process of the Ti–W cathode, where the Ti and W mixture is heated to at least 800 °C, which according to the phase diagram results in the formation of the β phase of Ti.¹⁹ For the Ti–C compositions, the binary phase diagram does not indicate a change in the Ti structure.¹⁹ 

A comparison of intensities of the peaks related to Ti and to W and/or TiC for the as received and used cathodes in Fig. 8 confirms the reduction of elemental Ti on the cathode surface after arcing. For the steady state mode of operation, the composition of the total flux (macroparticles, ions, and neutrals) must be consistent with the composition of the cathode body. However, the composition of the top-most layer of the operational cathode surface could locally have a diverging composition to meet the optimal requirements for the formation of ions, neutrals, and macroparticles. For example, Ti melts and evaporates easier than W from a Ti–W cathode, from the liquid bath surrounding the spot, from which also the macroparticles are formed. Locally, there may, therefore, be a lower amount of Ti, and consequently less ionization of Ti. Still, when all open Ti-rich regions have been preferentially eroded/evaporated from the surface, new Ti-rich regions will be opened only after arcing from W-rich regions. This forces the total material flux to be the same as the cathode body. Similar arguments can be used for the Ti–C cathode, considering a mixture of elemental Ti and TiC.

Compared to previously investigated Ti–Al and Ti–Si cathodes, where the arc spot glows smoothly and runs over the whole operational surface of the cathode,^5,7 the present work on Ti–W and Ti–C cathodes shows different features of the arc spot tracks (see Fig. 9). Moreover, a correlation between the smoothness of the cathode surface and the stability of the discharge has been detected, as previously shown for the Mo–Cu system.¹⁶ 

Figure 9 shows that the operational cathode surface after at least one hour of plasma generation is very different for different cathode compositions. For the Ti–C case, there are very smooth surfaces, except for the Ti0.75C0.25 and C cathode. For the Ti–W cathodes, the surfaces are smooth below 50 at. % of W, but rougher for a higher relative content of W, and smooth again for a pure W cathode. For both the Ti0.75C0.25 cathode and the Ti0.50W0.50 cathode, there are similar-sized voids/holes on the surfaces. The holes may be explained by the difference in the melting and evaporation temperatures of the powders used for cathode manufacturing; Ti, W, and TiC (1943 and 3560 K for Ti; 3695 and 5825 K for W; 3260 and 5093 K for TiC). Moreover, the evaporation temperature of Ti is lower than the melting temperature of W and very close to the melting temperature of TiC. As previously reported in Ref. 16, a large discrepancy in the melting/evaporation points can lead to damage of the operational surface and even prevent ignition of the discharge. In the present work, such challenges in arcing have been detected for the Ti0.3W0.7 cathode, suggesting that the general mechanism is the same as in a previous report,¹⁶ i.e., explosive-like evaporation of one of the cathode elements and formation of postdischarge craters that prevent a direct view of the system anode. It should also be noted that no large macroparticles were detected, as for the Mo–Cu system (see Ref. 37).

In the present study, we have investigated the influence of cathode compositions on the cathode, plasma, and basic film characteristics for Ti–C and Ti–W cathodes. The results were compared to previous work on Ti–Al and Ti–Si cathodes, performed under very similar operating conditions. A direct correlation between the average ion energy/charge and the cohesive energy of cathode compounds was demonstrated, with the only exceptions being the behavior of the average charge state with an increase in the relative amount of C in the Ti–C system, and the observed increase in the ion kinetic energy with an Al content higher than 50 at. % in Ti–Al cathodes. Hence, for an improved understanding of the observed plasma properties, also the “velocity rule” and equal/diverging electron temperatures need to be considered. The performed experiments show that a change in cathode composition allows tuning of the energy provided to the substrate by the ions via changes in the average charge state and the average ion energy. Also, the intensity of macroparticle generation can be changed via an accompanying change in the average melting temperature of the cathode. Dependence of the spatial distribution of the generated ions on their mass is suggested as one of the main reasons for the discovered discrepancies between the cathode and plasma/film compositions. The discrepancy between the composition of the cathode and the deposited film is found to be significantly lower than between the cathode and plasma compositions. Although small, a perceptible contribution of neutrals and/or macroparticles to the final film composition is suggested. The presented results contribute to the fundamental understanding of vacuum arc processes and are of importance for the further development of arc-deposited coatings from Ti-based compound cathodes.

We acknowledge financial support from The Knut and Alice Wallenberg Foundation for a project grant (The Boride Frontier, KAW 2015.0043) and a Scholar grant and the Swedish Government Strategic Research Area in Materials Science on Advanced Functional Materials at Linköping University (Faculty Grant SFO-Mat-LiU No. 2009-00971).

The authors have no conflicts to disclose.

Igor Zhirkov: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Project administration (equal); Validation (equal); Visualization (equal); Writing – original draft (equal). Peter Polcik: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Funding acquisition (equal); Resources (equal). Andrejs Petruhins: Data curation (equal); Formal analysis (equal); Funding acquisition (equal); Methodology (equal). Szilard Kolozsvári: Conceptualization (equal); Formal analysis (equal); Funding acquisition (equal); Resources (equal). Johanna Rosen: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Funding acquisition (equal); Methodology (equal); Project administration (equal); Resources (equal); Supervision (equal); Validation (equal); Writing – review & editing (equal).

The data that support the findings of this study are available within the article and from the corresponding author upon reasonable request.