Reactive High-Power Impulse Magnetron Sputtering of Chromium-Carbon Films

: Chromium-carbon ﬁlms were deposited by utilizing reactive high-power impulse magnetron sputtering at di ﬀ erent mixture ratios of ethyne and argon atmosphere, and di ﬀ erent substrate bias voltages and deposition temperature, with the same pulse frequency, duty cycle, and average power. The microstructure and mechanical properties of the obtained ﬁlms were compared. The ﬁlms consist of amorphous or nanocrystalline chromium carbide, hydrogenated amorphous carbon, and minor α -chromium phase. Decreasing the fraction of ethyne increases the content of the α -chromium phase but decreases hydrogenated amorphous carbon phase. The ﬁlm’s hardness increases by enhancing the negative substrate bias and raising the deposition temperature, which could be attributed to the increase of ﬁlm density and the Hall–Petch strengthening e ﬀ ect induced by the nanoscale crystallization of the amorphous

High-power impulse magnetron sputtering (HiPIMS) provides high power densities in pulses with a low duty cycle [25,26] on a magnetron sputtering source cathode, which enables high-density plasma

Materials and Methods
The Cr-C films were deposited in a vacuum chamber with one unipolar HiPIMS power and one unipolar substrate bias power. The details of the power supplies, the target, the size of substrate samples, and the cleaning of substrate samples can be found in our previous work of chromium film [31]. Three different substrate materials had been polished and were simultaneously used, including an AISI M35 high-speed steel disc with a size of φ 30 × 1 (h) mm 2 , a SUS 304 stainless steel mirror plate, and silicon wafer chips. After placing samples into the deposition chamber, vacuuming, and cleaning with argon plasma, the substrate temperature raised to about 190 • C without using additional substrate heating. For a higher deposition temperature, of 300 • C, the additional substrate heating would be applied during the argon plasma cleaning. The deposition parameters for Cr-C films are shown in Table 1. Different mixture ratios of argon/ethyne-1:1, 6:1, 8:1, 12:1-were respectively introduced to keep a deposition pressure 0.8 Pa for Cr-C film depositions. The applied target HiPIMS voltage pulses were near to a rectangle shape with a width of 60 µs and a frequency of 332 Hz. The pulse voltage and corresponding peak current on the target floated between −518 to −612 V due to the varying target conditions and the set deposition atmosphere. The pulse frequencies were manually set with the same pulse duty cycle, 2.5%, to maintain a constant target average power at 3 kW for all depositions.
As reported in the previous work [31], applying a pulsed substrate bias synchronized with the HiPIMS discharge pulse is helpful to selectively attract ions and the charged species with a stable bias Coatings 2020, 10, 1269 3 of 10 voltage and enable the target ion bombardment for a stronger film adhesion. The Cr ion bombardment was carried out before the Cr-C layer deposition. As the target HiPIMS pulse began, the substrate bias voltage pulse of −1000 V with a 1.75-fold pulse width of HiPIMS pulses was simultaneously initiated in order to accelerate Cr ions to bombard on the substrate. To minimize the roughness and temperature change on the substrate surface, the Cr ion bombardment was restricted to 40 s. After the bombardment, the substrate bias voltage switched to 0 V (grounded), pulsed −60 V, or pulsed −150 V directly without any gradient substrate biasing, to deposit different Cr-C films.
The microstructures of deposited films are observed by using a scanning electron microscope (SEM, S-4800 Cold Field Emission Scanning Electron Microscope, Hitachi, Tokyo, Japan). The film's thickness was estimated from cross-sectional SEM micrographs, and the crystallinity was characterized by utilizing an X-ray diffractometer (D8 Discover X-ray diffractometer, Bruker, Billerica, MA, USA). A Raman spectrometer (PROTRUSTECH, Tainan, Taiwan) was used to determine the presence of oxide and carbon phases in the Cr-C films. The nanoindentations (A TTX-NH3 nanoindentation tester, Anton Paar, Graz, Austria) were carried out to evaluate the film's hardness. Settings of analysis instruments were as the same as described in the references [30,31].

The Influences of Different Ar/C 2 H 2 Flow Ratios
The Raman spectra of the Cr-C films deposited at different Ar/C 2 H 2 flow ratios and the pressure of 0.8 Pa for 10 min are shown in Figure 1. Only two major peaks, the G (~1580 cm −1 ) and D (~1350 cm −1 ) bands of carbon phases, and a minor peak of Cr 2 O 3 (~541 cm −1 ) phase can be identified. The peaks of chromium carbides are very weak. This indicates that the microstructure of deposited Cr-C films could be a composite of amorphous carbide, amorphous carbon phases, and metallic chromium phases. The intensities of D and G decrease with the fraction of C 2 H 2 . However, the estimated I D /I G values of four samples are close to each other, with a difference of less than 2%, which means that the ratios of sp3 and sp2 bonds of the amorphous carbon phase of the four samples could be approximate. These results imply that the content of amorphous carbon phase decreases with the fraction of C 2 H 2 . A similar trend was also found in the direct current magnetron sputtering deposited Cr-C films [24].
Coatings 2020, x, x FOR PEER REVIEW 3 of 10 As reported in the previous work [31], applying a pulsed substrate bias synchronized with the HiPIMS discharge pulse is helpful to selectively attract ions and the charged species with a stable bias voltage and enable the target ion bombardment for a stronger film adhesion. The Cr ion bombardment was carried out before the Cr-C layer deposition. As the target HiPIMS pulse began, the substrate bias voltage pulse of −1000 V with a 1.75-fold pulse width of HiPIMS pulses was simultaneously initiated in order to accelerate Cr ions to bombard on the substrate. To minimize the roughness and temperature change on the substrate surface, the Cr ion bombardment was restricted to 40 s. After the bombardment, the substrate bias voltage switched to 0 V (grounded), pulsed −60 V, or pulsed −150 V directly without any gradient substrate biasing, to deposit different Cr-C films.
The microstructures of deposited films are observed by using a scanning electron microscope (SEM, S-4800 Cold Field Emission Scanning Electron Microscope, Hitachi, Tokyo, Japan). The film's thickness was estimated from cross-sectional SEM micrographs, and the crystallinity was characterized by utilizing an X-ray diffractometer (D8 Discover X-ray diffractometer, Bruker, Billerica, MA, USA). A Raman spectrometer (PROTRUSTECH, Tainan, Taiwan) was used to determine the presence of oxide and carbon phases in the Cr-C films. The nanoindentations (A TTX-NH3 nanoindentation tester, Anton Paar, Graz, Austria) were carried out to evaluate the film's hardness. Settings of analysis instruments were as the same as described in the references [30,31].

The Influences of Different Ar/C2H2 flow ratios
The Raman spectra of the Cr-C films deposited at different Ar/C2H2 flow ratios and the pressure of 0.8 Pa for 10 min are shown in Figure 1. Only two major peaks, the G (~1580 cm −1 ) and D (~1350 cm −1 ) bands of carbon phases, and a minor peak of Cr2O3 (~541 cm −1 ) phase can be identified. The peaks of chromium carbides are very weak. This indicates that the microstructure of deposited Cr-C films could be a composite of amorphous carbide, amorphous carbon phases, and metallic chromium phases. The intensities of D and G decrease with the fraction of C2H2. However, the estimated ID/IG values of four samples are close to each other, with a difference of less than 2%, which means that the ratios of sp3 and sp2 bonds of the amorphous carbon phase of the four samples could be approximate. These results imply that the content of amorphous carbon phase decreases with the fraction of C2H2. A similar trend was also found in the direct current magnetron sputtering deposited Cr-C films [24].  The X-ray diffractograms of the Cr-C films deposited at different Ar/C 2 H 2 flow ratios are shown in Figure 2. There is no obvious diffraction peak in the diffractograms of films of Ar/C 2 H 2 flow ratios of 1:1 and 6:1, which indicates that films deposited in this range of mixture ratio are amorphous or partial nanocrystalline. Weak peaks of α-chromium and chromium carbide phases appeared in the films of 8:1, and the peak intensity of the α-chromium phase became stronger in the film of 12:1. This indicates that the fraction of C 2 H 2 may not be sufficient to react with the chromium species ejected from the target when the Ar/C 2 H 2 flow ratio is higher than 8:1. The XRD results confirm the previous surmise about amorphous carbon and partly nanocrystalline carbide phases, which are similar to the reported Cr-C films by using sputtering [10,24,34]. The intensity around 40 • slightly increases with decreasing C 2 H 2 . This trend is also similar to the dc magnetron sputtering deposited Cr-C films [24].
Coatings 2020, x, x FOR PEER REVIEW 4 of 10 The X-ray diffractograms of the Cr-C films deposited at different Ar/C2H2 flow ratios are shown in Figure 2. There is no obvious diffraction peak in the diffractograms of films of Ar/C2H2 flow ratios of 1:1 and 6:1, which indicates that films deposited in this range of mixture ratio are amorphous or partial nanocrystalline. Weak peaks of α-chromium and chromium carbide phases appeared in the films of 8:1, and the peak intensity of the α-chromium phase became stronger in the film of 12:1. This indicates that the fraction of C2H2 may not be sufficient to react with the chromium species ejected from the target when the Ar/C2H2 flow ratio is higher than 8:1. The XRD results confirm the previous surmise about amorphous carbon and partly nanocrystalline carbide phases, which are similar to the reported Cr-C films by using sputtering [10,24,34]. The intensity around 40° slightly increases with decreasing C2H2. This trend is also similar to the dc magnetron sputtering deposited Cr-C films [24]. The top and cross-sectional SEM SE images of the Cr-C films deposited at the pressure of 0.8 Pa with different Ar/C2H2 flow ratios are shown in Figure 3. The film of 1:1 presents a loosen structure of stacking clusters, which is similar to the reported sputtering-deposited Cr-C films with a high carbon content [10]. By increasing the fraction of argon, the film structure becomes denser and more columnar, without any granular edge, which is similar to the reported sputtering and HiPIMSdeposited Cr-C films with lower carbon contents [10,12]. These features of amorphous film are consistent with the results of XRD. The estimated deposition rates for Ar/C2H2 ratios of 1:1, 6:1, 8:1, and 12:1 are 70, 153, 100, and 122 nm/min, respectively. The partial pressure of argon with an Ar/C2H2 ratio of 1:1 might be too low to sputter chromium atoms efficiently. On the other hand, the films of 1:1 and 6:1 were too fragile for hardness and adhesion tests. For these reasons, only the ratio of 8:1 were used in the following experiments of different substrate bias voltage and deposition temperature. The deposition time was also extended to 190 min. The top and cross-sectional SEM SE images of the Cr-C films deposited at the pressure of 0.8 Pa with different Ar/C 2 H 2 flow ratios are shown in Figure 3. The film of 1:1 presents a loosen structure of stacking clusters, which is similar to the reported sputtering-deposited Cr-C films with a high carbon content [10]. By increasing the fraction of argon, the film structure becomes denser and more columnar, without any granular edge, which is similar to the reported sputtering and HiPIMS-deposited Cr-C films with lower carbon contents [10,12]. These features of amorphous film are consistent with the results of XRD. The estimated deposition rates for Ar/C 2 H 2 ratios of 1:1, 6:1, 8:1, and 12:1 are 70, 153, 100, and 122 nm/min, respectively. The partial pressure of argon with an Ar/C 2 H 2 ratio of 1:1 might be too low to sputter chromium atoms efficiently. On the other hand, the films of 1:1 and 6:1 were too fragile for hardness and adhesion tests. For these reasons, only the ratio of 8:1 were used in the following experiments of different substrate bias voltage and deposition temperature. The deposition time was also extended to 190 min.

Influences of the Synchronized Substrate Bias and Deposition Temperature
The X-ray diffractograms of the Cr-C films deposited at the pressure 0.8 Pa with the same Ar/C2H2 flow ratio, 8:1, but different substrate bias voltages are shown in Figure 4. All samples of the different substrate bias voltages are amorphous with minor α-chromium phases. The intensity around 40° is relatively stronger in films deposited with negative biases. A similar phenomenon was also found in Cr-C films prepared by using vacuum arc [6,22], sputtering [12,23], and HiPIMS [12,18] depositions in an Ar/carbon hydride atmosphere. This suggests the formation of some nanocrystalline CrC with applying negative substrate bias due to the kinetic energy enhancement of incident ions [27][28][29]. A contrary phenomenon in Figure 4, which has never been reported in other

Influences of the Synchronized Substrate Bias and Deposition Temperature
The X-ray diffractograms of the Cr-C films deposited at the pressure 0.8 Pa with the same Ar/C 2 H 2 flow ratio, 8:1, but different substrate bias voltages are shown in Figure 4. All samples of the different substrate bias voltages are amorphous with minor α-chromium phases. The intensity around 40 • is relatively stronger in films deposited with negative biases. A similar phenomenon was also found in Cr-C films prepared by using vacuum arc [6,22], sputtering [12,23], and HiPIMS [12,18] depositions in an Ar/carbon hydride atmosphere. This suggests the formation of some nanocrystalline CrC with applying negative substrate bias due to the kinetic energy enhancement of incident ions [27][28][29]. A contrary phenomenon in Figure 4, which has never been reported in other similar researches, is that Coatings 2020, 10, 1269 6 of 10 the intensity of the α-chromium phase decreases by enhancing the substrate bias. Since the relative intensity of α (200) does not change, it is unlike the preferred orientation change resulting from the substrate biasing [31]. Besides, the −60 V substrate bias is not strong enough to cause an obvious re-sputtering of light elements [22]. One possible reason is that the negative substrate bias promotes the formation of carbide phases [22], and therefore reduces the fraction of α-chromium phases.
Coatings 2020, x, x FOR PEER REVIEW 6 of 10 similar researches, is that the intensity of the α-chromium phase decreases by enhancing the substrate bias. Since the relative intensity of α (200) does not change, it is unlike the preferred orientation change resulting from the substrate biasing [31]. Besides, the −60 V substrate bias is not strong enough to cause an obvious re-sputtering of light elements [22]. One possible reason is that the negative substrate bias promotes the formation of carbide phases [22], and therefore reduces the fraction of αchromium phases. The cross-sectional SEM SE images of the Cr-C films deposited at the pressure of 0.8 Pa and the Ar/C2H2 flow ratio 8:1 but with different synchronized substrate bias voltages are shown in Figure 5. The film becomes more solid, and the fracture surface becomes smoother and more brittle, with enhancing the negative substrate bias. This also evidences the increase of kinetic energy incident ions and adatoms to form a denser structure. The nanoindentation results of the Cr-C film deposited with different substrate bias voltages shown in Figure 6 indirectly proves the densified film structure. The hardness of films increases by enhancing the negative substrate bias, which is similar to the results of vacuum arc deposited and the HiPIMS-deposited Cr-C films [6,12].  The cross-sectional SEM SE images of the Cr-C films deposited at the pressure of 0.8 Pa and the Ar/C 2 H 2 flow ratio 8:1 but with different synchronized substrate bias voltages are shown in Figure 5. The film becomes more solid, and the fracture surface becomes smoother and more brittle, with enhancing the negative substrate bias. This also evidences the increase of kinetic energy incident ions and adatoms to form a denser structure. The nanoindentation results of the Cr-C film deposited with different substrate bias voltages shown in Figure 6 indirectly proves the densified film structure. The hardness of films increases by enhancing the negative substrate bias, which is similar to the results of vacuum arc deposited and the HiPIMS-deposited Cr-C films [6,12].
Coatings 2020, x, x FOR PEER REVIEW 6 of 10 similar researches, is that the intensity of the α-chromium phase decreases by enhancing the substrate bias. Since the relative intensity of α (200) does not change, it is unlike the preferred orientation change resulting from the substrate biasing [31]. Besides, the −60 V substrate bias is not strong enough to cause an obvious re-sputtering of light elements [22]. One possible reason is that the negative substrate bias promotes the formation of carbide phases [22], and therefore reduces the fraction of αchromium phases. The cross-sectional SEM SE images of the Cr-C films deposited at the pressure of 0.8 Pa and the Ar/C2H2 flow ratio 8:1 but with different synchronized substrate bias voltages are shown in Figure 5. The film becomes more solid, and the fracture surface becomes smoother and more brittle, with enhancing the negative substrate bias. This also evidences the increase of kinetic energy incident ions and adatoms to form a denser structure. The nanoindentation results of the Cr-C film deposited with different substrate bias voltages shown in Figure 6 indirectly proves the densified film structure. The hardness of films increases by enhancing the negative substrate bias, which is similar to the results of vacuum arc deposited and the HiPIMS-deposited Cr-C films [6,12].  large measurement error could therefore be attributed to the microscale nonuniformity in the film structure. There are several phases, including a-CrxCy, a-C:H, Cr, Cr3C2, and droplets. The domain size of different phases could vary from several nanometers to a micrometer [10,22,23]. The reported hardness value of HiPIMS-deposited a-C:H could be higher than 30 GPa, and the hardness value of HiPIMS-deposited Cr could be lower than 7 GPa [12,[18][19][20]31]. Such great difference between different phases might be the reason of the high measurement error. The Cr-C film deposited at 300 °C and the same substrate bias of −150 V in the same atmosphere present a cross-sectional microstructure similar to that of the film deposited at 200 °C. The hardness of the film deposited at 300 °C is 14.46 ± 5.24 GPa, which is about 10% higher than the one at 200 °C. The higher deposition temperature enables a higher mobility of adatoms, to form a denser and crystalline structure which usually results in more hardness. The X-ray diffractograms of the Cr-C films deposited at 200 °C and 300 °C are compared in Figure 7. Peaks of stoichiometric carbide phases slightly increase and peaks of amorphous carbide phase slightly weaken by raising the deposition temperature. This indicates that a higher temperature may promote the formation of stoichiometric carbide phases rather than an amorphous carbide phase. According to the Hall-Petch effect, the appearance of nanoscale grains may provide a grain boundary and strengthen the hardness of materials [18,35]. This could be one of the reasons for the higher hardness of Cr-C films deposited at 300 °C.
Deposition at a higher temperature is not recommended due to two reasons. One reason is that a higher temperature induces the conversion from hydrogenated amorphous carbon to nanocrystalline graphite [36], which usually results in the decrease of film hardness. Another reason depends on the heating method and consequent target conditions. The heating region should be restricted to the substrate. Heating the target surface should be avoided to minimize the target poisoning in such reactive sputtering [26]. In particular, the target for the HiPIMS process requires abundant cooling to release the heat generated by the high-density power input during the pulse on time. The carbides formed on the target surface due to poisoning are potential initiation points of the arc discharge and accompanying ejection of droplets during the high-power pulse on time. The droplets embedded in the Cr-C films with a size of several micrometers are detrimental to the film uniformity, density, adhesion, and consequent performances. Since most of the obtained films in this work are thinner than 2 µm, the indentations were carried out with a max load of 5 mN and a 10-s pause to avoid the influences of substrate materials. The large measurement error could therefore be attributed to the microscale nonuniformity in the film structure. There are several phases, including a-Cr x C y , a-C:H, Cr, Cr 3 C 2 , and droplets. The domain size of different phases could vary from several nanometers to a micrometer [10,22,23]. The reported hardness value of HiPIMS-deposited a-C:H could be higher than 30 GPa, and the hardness value of HiPIMS-deposited Cr could be lower than 7 GPa [12,[18][19][20]31]. Such great difference between different phases might be the reason of the high measurement error.
The Cr-C film deposited at 300 • C and the same substrate bias of −150 V in the same atmosphere present a cross-sectional microstructure similar to that of the film deposited at 200 • C. The hardness of the film deposited at 300 • C is 14.46 ± 5.24 GPa, which is about 10% higher than the one at 200 • C. The higher deposition temperature enables a higher mobility of adatoms, to form a denser and crystalline structure which usually results in more hardness. The X-ray diffractograms of the Cr-C films deposited at 200 • C and 300 • C are compared in Figure 7. Peaks of stoichiometric carbide phases slightly increase and peaks of amorphous carbide phase slightly weaken by raising the deposition temperature. This indicates that a higher temperature may promote the formation of stoichiometric carbide phases rather than an amorphous carbide phase. According to the Hall-Petch effect, the appearance of nanoscale grains may provide a grain boundary and strengthen the hardness of materials [18,35]. This could be one of the reasons for the higher hardness of Cr-C films deposited at 300 • C.
Deposition at a higher temperature is not recommended due to two reasons. One reason is that a higher temperature induces the conversion from hydrogenated amorphous carbon to nano-crystalline graphite [36], which usually results in the decrease of film hardness. Another reason depends on the heating method and consequent target conditions. The heating region should be restricted to the substrate. Heating the target surface should be avoided to minimize the target poisoning in such reactive sputtering [26]. In particular, the target for the HiPIMS process requires abundant cooling to release the heat generated by the high-density power input during the pulse on time. The carbides formed on the target surface due to poisoning are potential initiation points of the arc discharge and accompanying ejection of droplets during the high-power pulse on time. The droplets embedded in the Cr-C films with a size of several micrometers are detrimental to the film uniformity, density, adhesion, and consequent performances. The above results indicate that applying the reactive HiPIMS deposition to synthesized Cr-C films can lead to a film structure, composition, and hardness similar to the one obtained by using direct current magnetron sputtering. Although the disadvantages of reactive direct current magnetron sputtering, such as the target poisoning and hysteresis, also occur in the reactive HiPIMS deposition, these can nevertheless be reduced by controlling the HiPIMS pulse frequency and duty cycle [26]. The results in this work indicate that applying the synchronized substrate bias can easily alter the film structure and film density at a relatively lower substrate temperature, due to the high ionization ratio in HiPIMS deposition.

Conclusions
This research preliminarily investigates the influences of reactive gas C2H2 ratio, synchronized substrate bias voltage, and deposition temperature on the Cr-C films deposited by using reactive high-power impulse magnetron sputtering with the same average power, pulse frequency, and pulse duty cycle at a constant pressure of 0.8 Pa. All the Cr-C films obtained in this work are mainly amorphous. The films consists of amorphous or nanocrystalline chromium carbide, hydrogenated amorphous carbon, and a minor α-chromium phase. Increasing the Ar/C2H2 ratio increases the content of the α-chromium phase but decreases the hydrogenated amorphous carbon phase. Increasing the Ar/C2H2 ratio also promotes the film structure to change from loose stacking clusters into a solid columnar feature. Applying synchronized negative substrate bias also densifies the film structure and enhances the film hardness. Same enhancements can also be achieved by raising the substrate temperature, but not over 300 °C . The nanocrystallization of amorphous carbide phases promoted by negative substrate bias and substrate temperature induces the Hall-Petch strengthening effect and results in a higher film hardness.  The above results indicate that applying the reactive HiPIMS deposition to synthesized Cr-C films can lead to a film structure, composition, and hardness similar to the one obtained by using direct current magnetron sputtering. Although the disadvantages of reactive direct current magnetron sputtering, such as the target poisoning and hysteresis, also occur in the reactive HiPIMS deposition, these can nevertheless be reduced by controlling the HiPIMS pulse frequency and duty cycle [26]. The results in this work indicate that applying the synchronized substrate bias can easily alter the film structure and film density at a relatively lower substrate temperature, due to the high ionization ratio in HiPIMS deposition.

Conclusions
This research preliminarily investigates the influences of reactive gas C 2 H 2 ratio, synchronized substrate bias voltage, and deposition temperature on the Cr-C films deposited by using reactive high-power impulse magnetron sputtering with the same average power, pulse frequency, and pulse duty cycle at a constant pressure of 0.8 Pa. All the Cr-C films obtained in this work are mainly amorphous. The films consists of amorphous or nanocrystalline chromium carbide, hydrogenated amorphous carbon, and a minor α-chromium phase. Increasing the Ar/C 2 H 2 ratio increases the content of the α-chromium phase but decreases the hydrogenated amorphous carbon phase. Increasing the Ar/C 2 H 2 ratio also promotes the film structure to change from loose stacking clusters into a solid columnar feature. Applying synchronized negative substrate bias also densifies the film structure and enhances the film hardness. Same enhancements can also be achieved by raising the substrate temperature, but not over 300 • C. The nanocrystallization of amorphous carbide phases promoted by negative substrate bias and substrate temperature induces the Hall-Petch strengthening effect and results in a higher film hardness.