- Email: [email protected]
Effect of magnetic field structure near cathode on the arc spot stability of filtered vacuum arc source of graphite
Surface and Coatings Technology 124 (2000) 135–141 www.elsevier.nl/locate/surfcoat
Effect of magnetic field structure near cathode on the arc spot stability of filtered vacuum arc source of graphite Jong-Kuk Kim a,b, Kwang-Ryeol Lee a, *, Kwang Yong Eun a, Kie-Hyung Chung b a Thin Film Technology Research Center, Korea Institute of Science and Technology, PO Box 131, Cheongryang, Seoul 130-650, South Korea b Department of Nuclear Engineering, Seoul National University, 56-1, Shinrim-Dong, Kwanak-Ku, Seoul 151-742, South Korea Received 18 April 1999; accepted in revised form 4 November 1999
Abstract Amorphic diamond films deposited by the filtered vacuum arc (FVA) method have attracted much attention due to their superior mechanical and optical properties. However, the instability of the arc limits the continuous operation of the FVA source, resulting in a poor productivity. In the present work, we investigated the effects of the cathode shape and the structure of the magnetic field near the cathode on the erosion behavior by both computer simulations and experimental studies. Arc instability in the configuration of parallel magnetic polarities of the source magnet and the extraction could be suppressed by placing a permanent magnet of opposite polarity behind the cathode. We show further that oscillation of the current of the source magnet was effective in extending the area of the arc spot movement. A tapered cathode exhibited a more stable arc than a cylindrical cathode, as confirmed by the time variation of the beam current. By using the oscillating current of the source magnet and a tapered cathode of diameter 80 mm, a continuous operation for 2000 min with an arc current of 60 A was obtained, at which more than 90% of the cathode volume could be used. A stable beam current of about 350 mA was obtained under the present operating conditions. © 2000 Elsevier Science S.A. All rights reserved. Keywords: Amorphic diamond films; Filtered vacuum arc; Magnetic field structure
1. Introduction Amorphic diamond films deposited by the filtered vacuum arc ( FVA) method have attracted much attention due to their superior mechanical and optical properties [1–3]. The vacuum arc process has many advantages against other CVD or PVD processes. Plasmas generated from the vacuum arc exhibit high ionization ratios of more than 90%. Hence, the effect of ion energy on the film structure appears significant. Moreover, the ion energy of the plasma jet generated from a vacuum arc is in the range of 10–100 eV, within which carbon films of excellent physical and chemical properties can be obtained [4]. Although the macroparticles generated from the arc spots have been considered as the main drawback of the vacuum arc process, magnetic filtering technology in the FVA process can efficiently remove the macroparticles, resulting in a smooth film surface [5]. It therefore becomes possible to consider carbon * Corresponding author. Tel.: +82-2-958-5494; fax: +82-2-958-5509. E-mail address: [email protected] ( K.-R. Lee)
films deposited by the FVA process as the protective layer on next-generation hard disks, for which extremely smooth and wear resistant thin films are required [6 ]. However, several problems concerned with the graphite cathode erosion must be solved. In contrast to arc spots on metal cathodes, arc spots on graphite cathodes have a point shape [7]. Hence, the vacuum arc on graphite cathodes shows relatively large instabilities in the arc spot motion and localized cathode erosion under some operating conditions. Instability of the arc spot motion limits continuous operation of the FVA source, which results in poor productivity. In addition, the localized erosion degrades the efficiency of the cathode usage. Efficient erosion of the cathode by a stable arc spot motion is thus a prerequisite for the industrial application of amorphic diamond coatings by the FVA process. Instability of the arc spot motion is known to be closely related to the arc spot shift to the edge of the cathode or an abnormal increase in the arc voltage caused by the increased resistance of the plasma when the electrical current flows across the magnetic field lines
0257-8972/00/$ - see front matter © 2000 Elsevier Science S.A. All rights reserved. PII: S0 2 5 7- 8 9 7 2 ( 9 9 ) 0 0 64 3 - X
136
J.-K. Kim et al. / Surface and Coatings Technology 124 (2000) 135–141
Fig. 1. Schematic of the filtered vacuum arc deposition system used in the present work: (1) Nd magnet, (2) cathode yoke, (3) cathode, (4) source magnet, (5) anode, (6) striker, (7) extraction magnet, (8) bending magnet, (9) outlet magnet, (10) deflection magnet, (11) plasme duct, (12) baffle, (13) substrate holder, (14) vacuum chamber.
Fig. 2. Magnetic field structure near cathode without permanent magnet and yoke for various currents of the source magnet of (a) 1 A, (b) 3 A and (c) 5 A. The current of the extraction magnet was fixed at 3 A.
[8,9]. The former reason of the instability can be minimized by controlling the arc spot motion. The movement of the arc spot is strongly influenced by the existing magnetic field [10–12]. The arc spot tends to move in the retrograde direction given by −J×B and towards
the opening of the acute angle formed by the field line and the cathode surface [13,14]. In order to stabilize the arc spot motion, it would thus be essential to design the magnetic field to confine the arc spot within the top surface of the cathode. Enhanced productivity and
J.-K. Kim et al. / Surface and Coatings Technology 124 (2000) 135–141
137
Fig. 3. Magnetic field structure near cathode with permanent magnet only for various currents of the source magnet of (a) 1 A, (b) 3 A and (c) 5 A. The current of the extraction magnet was fixed at 3 A.
effective cathode erosion thus necessitate a systematic analysis of the magnetic field structure near the cathode in a certain equipment configuration. The abnormal increase in arc voltage is the main reason for the arc instability when using a mirror magnetic configuration where the polarities of the source and the extraction magnet are parallel. However, the cusp magnetic configuration where the magnetic field diverts to the anode effectively suppresses the instability caused by the increasing arc voltage. The cusp configuration is thus used by many researchers [15–17]. However, the magnetic field diverts electrons to the duct walls so that a very small central section can be available for transporting electrons through the magnetic filter [15]. The cusp configuration makes the arc spot move around the edge of the cathode, which further degrades the efficiency of the beam transport. In contrast, the mirror magnetic configuration has an advantage of efficient ion beam transport. In the present work, we investigated the effect of the magnetic field structure on the cathode erosion behavior in the mirror magnetic configuration. The structure of the magnetic fields and the shape of the cathode were optimized from the view point of arc stability by both computer simulations and experimental studies. Using an oscillating current in the source magnet combined with a permanent magnet of opposite polarity behind the cathode, we could obtain a stable arc in the mirror magnetic configuration and increase the lifetime of the cathode. The cathode erosion was so uniform that more
than 90% of the cathode volume could be used during the lifetime.
2. Experimental Fig. 1 shows the schematic of the FVA system used in the present work. The FVA source was composed of a graphite cathode of diameter 80 mm, a ring-type anode of inner diameter 90 mm, a 60° bending plasma duct for magnetic filtering and four solenoid magnets for extraction, bending, outlet and deflection, respectively. The plasma duct was electrically isolated from the ground so that a bias voltage could be applied. The solenoid magnets could generate a center magnetic field of 250 Gauss at a maximum current of 5 A. Baffles in the duct were designed to prevent macroparticles from rebounding from the duct wall. A 115 mm diameter substrate holder was placed at a distance of 130 mm from the exit of the FVA source. In order to measure the beam current and apply bias voltage on the substrate, the substrate holder was electrically isolated from the reaction chamber. When measuring the beam current, the substrate holder was biased to −100 V to exclude the contribution of electrons. A solenoid source magnet (maximum center magnetic field of 150 Gauss) was installed around the cathode. We also placed a magnetic yoke (diameter 50 mm, thickness 10 mm) and a permanent Nd–Fe–B magnet (diameter 15 mm, thickness 10 mm, surface magnetic
138
J.-K. Kim et al. / Surface and Coatings Technology 124 (2000) 135–141
strength 4000 Gauss) at 5 and 17 mm from the back surface of the cathode, respectively. The magnetic field near the cathode was simulated using a Poisson code for various configurations of the permanent magnet, yoke and the current of the solenoid source magnet [18]. The self-magnetic field due to the arc spot was ignored in the simulation. The current of the extraction magnet was fixed at 3 A. Erosion behaviors were investigated by observing the eroded cathode after operating the FVA source until the stable ignition of the arc by the trigger could not be obtained (as defined by lifetime). The arc source was operated in an Ar environment at a pressure of 17 mPa. Two shapes of the cathode (cylindrical and tapered) were tested to assess the cathode erosion behavior and the stability of the beam current.
(a)
3. Results and discussion In our equipment, we used the mirror configuration where the magnetic polarities of the source and the extraction magnets were in the same direction to increase the efficiency of the beam transport [15,17]. In the present configuration of the magnets, the arc spot became unstable with increasing magnetic strength of the source magnet. Fig. 2 shows the magnetic field structures for various strengths of the source magnet. By increasing the current of the source magnet, i.e. increasing the magnetic strength, the opening of the acute angle formed by the magnetic field line on the cathode surface was directed outward. The arc spot tended to move to the edge of the cathode, resulting in the arc instability. We could not obtain a stable cathode arc erosion in this configuration. This instability can be suppressed by placing a permanent magnet of opposite polarity to that of the solenoid source magnet on the back side of the cathode. It can be seen in Fig. 3 that the opening of the acute angle was directed much more to the center of the cathode at the same current of the source magnet as in Fig. 2. However, the angle between the magnetic field line and the cathode surface also increased as the current of the source magnet increased. Erosion behaviors of the cathode were in good agreement with the simulation results. Fig. 4(a) shows the eroded cathode when the current of the source magnet was 1 A and the arc was ignited at the center of the cathode. Deep erosion in the center region showed that the arc spot moved around mainly in the center of the cathode. When the current of the source magnet increased to 3 A, the arc spot appeared to move in a wider region of the cathode, as shown in Fig. 4(b). This behavior can be easily understood by considering that the larger opening of the acute angle occurred with the higher current of the source magnet, as shown in Fig. 3(a) and (b). It must be noted that the erosion behavior in this
(b)
(c) Fig. 4. Morphologies of the used cathodes eroded with the permanent magnet only. The source magnet current was (a) 1 A and (b) 3 A. (c) Eroded cathode when arc was ignited at the edge of the cathode and the current of source magnet was 1 A.
magnetic configuration is dependent on the ignition position. Fig. 4(c) shows the eroded cathode when the arc was ignited at the edge of the cathode (top of the figure) and the current of the source magnet was 1 A. Even if the acute angle of the magnetic field was directed toward the center of the cathode, as can be seen in Fig. 3(a), the eroded cathode showed that the arc spot was not moved to the center of the cathode. Frequent ignition was required to maintain the arc under this operating condition. This behavior can be understood by considering that the initial erosion of the cathode
J.-K. Kim et al. / Surface and Coatings Technology 124 (2000) 135–141
139
Fig. 5. Magnetic field structure near cathode with both permanent magnet and yoke for various currents of the source magnet of (a) 1 A, (b) 3 A and (c) 5 A. The current of the extraction magnet was fixed at 3 A.
(a)
(b) Fig. 6. Morphologies of the used cathodes eroded with both the permanent magnet and the yoke. The source magnet current was ocillated in the range of (a) 0–1 A and (b) 0–1.5 A.
edge immediately changes the angle between the magnetic field and the cathode surface. In the present case, the magnetic field line becomes parallel to the eroded surface when the arc was ignited at the edge of the cathode. Another possible reason is the interaction between the distorted electric field near the edge and the self-magnetic field due to arc spot generation. The effect of a permanent magnet can be reduced by placing a yoke between the cathode and the permanent magnet. Fig. 5 shows the simulated magnetic field line structures for various currents of source magnet with the yoke and the permanent magnet. A comparison of Fig. 5 with Fig. 3 shows that the yoke increased the opening of the acute angle between the magnetic field line and the cathode surface under the same magnetic configuration. The larger acute angle can widen the arc spot movement, which was confirmed by the morphology of the eroded cathode at the current of the source magnet of 1 A. A combination of the yoke and the permanent magnet can increase the efficiency of the cathode by about 20%, as can be estimated from the eroded volume of the cathode over its life. A more efficient erosion of the cathode could be obtained by oscillating the current of the source magnet. Oscillating the magnetic strength of the source magnet can dynamically change the magnetic field structure near the cathode. By oscillating the current of the source magnet, the acute angle between the magnetic field line and the cathode surface was periodically varied. The arc spot can thus move around in a wider region of the
140
J.-K. Kim et al. / Surface and Coatings Technology 124 (2000) 135–141
cathode surface. Fig. 6 shows the morphology of the eroded cathode when the current of the source magnet oscillated from 0 to 1 A (a) and from 0 to 1.5 A (b). In contrast to deep erosion in the center region under a static source magnetic field [Fig. 4(a)], a much wider erosion was observed by the current oscillation, as can be seen in Fig. 6(a). The eroded area was proportional to the maximum current of the source magnet, as can be seen from a comparison between Fig. 6(a) and (b). Even if a uniform erosion of the cathode could be obtained by oscillating the current of the source magnet, the arc was frequently extinguished when a large oscillating current amplitude was used. Fig. 7(a) shows the carbon beam current variation for 300 s when using a cathode of cylindrical shape. The current of the source magnet was periodically varied from 1 to 3 A (upper data of the graph). A fluctuating and lower beam current was observed when the arc spot moved around the edge [arrow in Fig. 7(a)]. The arc spot in the edge region could not be recovered to the center of the cathode but was extinguished. During this experiment, therefore, two
ignition times of 70 and 210 s were required. A more stable arc could be obtained by changing the shape of the cathode from cylindrical to tapered. Fig. 7(b) shows the beam current when using the tapered cathode. In contrast to Fig. 7(a), a more stable beam current was obtained. It can be further noted that the arc spot near the edge moved to the top surface of the cathode, as can be seen in the beam current change at about 160 s [arrow in Fig. 7(b)]. This stabilizing action can easily be understood by considering that the large driving force for the arc spot movement to the center region could be applied when the arc spot was placed in the tapered side of the cathode [14]. Fig. 8(a) and (b), respectively, show the tapered cathode before and after operating the arc source for 2000 min at an arc current of 60 A. More than 90% of the cathode volume could be used in an optimum erosion condition by oscillating the source magnetic strength from 1 to 3 A. Figs. 7 and 8 clearly show that uniform erosion of the cathode and a stable beam current could be obtained in the present magnetic configuration by using an oscillating current of the source magnet and the tapered cathode. Fig. 9 shows the time-averaged beam current for 60 s for various values of the extraction magnet current. For this experiment, the currents of bending, deflection and outlet magnets were fixed at 3 A. During the measurement, the plasma duct was
(a)
(b) Fig. 7. Time variation of carbon beam current for (a) cylindrical cathode and (b) tapered cathode. The current of the source magnet varied from 1 to 3 A at 0.1 Hz. The insert of each graph shows the shape of the cathode.
Fig. 8. Morphologies of tapered cathode (a) before and (b) after use for 2000 min with both permanent magnet and yoke. The current of the source magnet varied from 1 to 3 A at 0.1 Hz.
J.-K. Kim et al. / Surface and Coatings Technology 124 (2000) 135–141
Fig. 9. Dependence of the carbon beam current on the extraction magnet current.
biased to 50 V to further enhance the transport of the plasma beam [19]. The beam current increased by increasing the current of the extraction magnet from 0 to 3 A. However, the beam current saturated at higher values of the extraction magnet current because the center of the plasma beam was displaced from that of the plasma duct. Higher beam current could be obtained when using a higher arc current. We could obtain about 350 mA when the arc current was 60 A and the extraction magnet current was 3 A. Amorphic diamond films were deposited on Si (100) wafers for various substrate bias voltages ranging from 0 to −250 V. By increasing the bias voltage, the hardness measured by nanoindentation and the residual compressive stress measured by the beam deflection method decreased from 65 to 45 GPa and 6.4 to 3.1 GPa, respectively. The content of graphitic component was increased by increasing the negative bias voltage, as observed in EELS and Raman spectroscopy. Details of the structure and the mechanical properties of deposited films will be reported elsewhere.
4. Conclusions The most significant result of the present work is that arc instability in the mirror magnetic configuration could
141
be suppressed by placing a permanent magnet of opposite polarity behind the cathode. This result means that the present magnetic configuration is suitable for the filtered vacuum arc system of high efficiency of the beam transport and the cathode erosion. We further showed that oscillating the current of the source magnet was effective in extending the area of the arc spot movement. A tapered cathode exhibited a more stable arc than a cylindrical cathode, as confirmed by the time variation of the beam current. By using the oscillating current of the source magnet and a tapered cathode of diameter 80 mm, continuous operation for 2000 min at an arc current of 60 A was obtained, at which more than 90% of the cathode volume could be used. A stable beam current of about 350 mA was obtained under the present operating conditions.
References [1] R. Lossy, D.L. Pappas, R.A. Roy, J.J. Cuomo, J. Bruley, J. Appl. Phys. 77 (1995) 4750. [2] R.L. Boxman, V. Zhitomirsky, B. Alterkop, E. Gidalevich, I. Beilis, M. Keidar, S. Goldsmith, Surf. Coat. Technol. 86–87 (1996) 243. [3] N. Mustapha, R.P. Howson, Surf. Coat. Technol. 92 (1997) 29. [4] H.C. Miller, J. Appl. Phys. 52 (1981) 4523. [5] I.I. Aksenov, S.I. Vakula, V.G. Padalka, V.M. Khoroshikh, Sov. Phys. Tech. Phys. 25 (1980) 1164. [6 ] C.S. Bhatia, S. Anders, K. Bobb, R. Hsiao, D.B. Bogy, I.G. Brown, Proc. Int. Conf. Micromechatronics for Information and Precision Equipment, Tokyo (1997). [7] J.E. Daalder, J. Phys. D: Appl. Phys. 12 (1979) 1769. [8] V.N. Zhitomirsky, R.L. Boxman, S. Goldsmith, Surf. Coat. Technol. 68–69 (1994) 146. [9] V.N. Zhitomirsky, R.L. Boxman, S. Goldsmith, J. Vac. Sci. Technol. 13 (1995) 2233. [10] M.G. Drouet, IEEE Trans. Plasma Sci. 13 (1985) 235. [11] J. Vyskovcil, J. Musil, Surf. Coat. Technol. 43–44 (1990) 299. [12] B.Y. Moizhes, V.A. Nemchinsky, J. Phys. D: Appl. Phys. 24 (1991) 2014. [13] P.D. Swift, J. Appl. Phys. 67 (1990) 1720. [14] B.F. Coll, D.M. Sanders, Surf. Coat. Technol. 81 (1996) 42. [15] M.M.M. Bilek, Y. Yin, D.R. McKenzie, IEEE Trans. Plasma Sci. 24 (1996) 1165. [16 ] S. Falabella, US Patent No. 5,468,363, 1995. [17] S. Anders, A. Anders, I.G. Brown, J. Appl. Phys. 75 (1994) 4895. [18] K. Halbach, R.F. Holsinger, Particle Accelerators 7 (1976) 213. [19] A. Anders, S. Anders, I.G. Brown, J. Appl. Phys. 75 (1994) 4900.
Effect of magnetic field structure near cathode on the arc spot stability of filtered vacuum arc source of graphite Jong-Kuk Kim a,b, Kwang-Ryeol Lee a, *, Kwang Yong Eun a, Kie-Hyung Chung b a Thin Film Technology Research Center, Korea Institute of Science and Technology, PO Box 131, Cheongryang, Seoul 130-650, South Korea b Department of Nuclear Engineering, Seoul National University, 56-1, Shinrim-Dong, Kwanak-Ku, Seoul 151-742, South Korea Received 18 April 1999; accepted in revised form 4 November 1999
Abstract Amorphic diamond films deposited by the filtered vacuum arc (FVA) method have attracted much attention due to their superior mechanical and optical properties. However, the instability of the arc limits the continuous operation of the FVA source, resulting in a poor productivity. In the present work, we investigated the effects of the cathode shape and the structure of the magnetic field near the cathode on the erosion behavior by both computer simulations and experimental studies. Arc instability in the configuration of parallel magnetic polarities of the source magnet and the extraction could be suppressed by placing a permanent magnet of opposite polarity behind the cathode. We show further that oscillation of the current of the source magnet was effective in extending the area of the arc spot movement. A tapered cathode exhibited a more stable arc than a cylindrical cathode, as confirmed by the time variation of the beam current. By using the oscillating current of the source magnet and a tapered cathode of diameter 80 mm, a continuous operation for 2000 min with an arc current of 60 A was obtained, at which more than 90% of the cathode volume could be used. A stable beam current of about 350 mA was obtained under the present operating conditions. © 2000 Elsevier Science S.A. All rights reserved. Keywords: Amorphic diamond films; Filtered vacuum arc; Magnetic field structure
1. Introduction Amorphic diamond films deposited by the filtered vacuum arc ( FVA) method have attracted much attention due to their superior mechanical and optical properties [1–3]. The vacuum arc process has many advantages against other CVD or PVD processes. Plasmas generated from the vacuum arc exhibit high ionization ratios of more than 90%. Hence, the effect of ion energy on the film structure appears significant. Moreover, the ion energy of the plasma jet generated from a vacuum arc is in the range of 10–100 eV, within which carbon films of excellent physical and chemical properties can be obtained [4]. Although the macroparticles generated from the arc spots have been considered as the main drawback of the vacuum arc process, magnetic filtering technology in the FVA process can efficiently remove the macroparticles, resulting in a smooth film surface [5]. It therefore becomes possible to consider carbon * Corresponding author. Tel.: +82-2-958-5494; fax: +82-2-958-5509. E-mail address: [email protected] ( K.-R. Lee)
films deposited by the FVA process as the protective layer on next-generation hard disks, for which extremely smooth and wear resistant thin films are required [6 ]. However, several problems concerned with the graphite cathode erosion must be solved. In contrast to arc spots on metal cathodes, arc spots on graphite cathodes have a point shape [7]. Hence, the vacuum arc on graphite cathodes shows relatively large instabilities in the arc spot motion and localized cathode erosion under some operating conditions. Instability of the arc spot motion limits continuous operation of the FVA source, which results in poor productivity. In addition, the localized erosion degrades the efficiency of the cathode usage. Efficient erosion of the cathode by a stable arc spot motion is thus a prerequisite for the industrial application of amorphic diamond coatings by the FVA process. Instability of the arc spot motion is known to be closely related to the arc spot shift to the edge of the cathode or an abnormal increase in the arc voltage caused by the increased resistance of the plasma when the electrical current flows across the magnetic field lines
0257-8972/00/$ - see front matter © 2000 Elsevier Science S.A. All rights reserved. PII: S0 2 5 7- 8 9 7 2 ( 9 9 ) 0 0 64 3 - X
136
J.-K. Kim et al. / Surface and Coatings Technology 124 (2000) 135–141
Fig. 1. Schematic of the filtered vacuum arc deposition system used in the present work: (1) Nd magnet, (2) cathode yoke, (3) cathode, (4) source magnet, (5) anode, (6) striker, (7) extraction magnet, (8) bending magnet, (9) outlet magnet, (10) deflection magnet, (11) plasme duct, (12) baffle, (13) substrate holder, (14) vacuum chamber.
Fig. 2. Magnetic field structure near cathode without permanent magnet and yoke for various currents of the source magnet of (a) 1 A, (b) 3 A and (c) 5 A. The current of the extraction magnet was fixed at 3 A.
[8,9]. The former reason of the instability can be minimized by controlling the arc spot motion. The movement of the arc spot is strongly influenced by the existing magnetic field [10–12]. The arc spot tends to move in the retrograde direction given by −J×B and towards
the opening of the acute angle formed by the field line and the cathode surface [13,14]. In order to stabilize the arc spot motion, it would thus be essential to design the magnetic field to confine the arc spot within the top surface of the cathode. Enhanced productivity and
J.-K. Kim et al. / Surface and Coatings Technology 124 (2000) 135–141
137
Fig. 3. Magnetic field structure near cathode with permanent magnet only for various currents of the source magnet of (a) 1 A, (b) 3 A and (c) 5 A. The current of the extraction magnet was fixed at 3 A.
effective cathode erosion thus necessitate a systematic analysis of the magnetic field structure near the cathode in a certain equipment configuration. The abnormal increase in arc voltage is the main reason for the arc instability when using a mirror magnetic configuration where the polarities of the source and the extraction magnet are parallel. However, the cusp magnetic configuration where the magnetic field diverts to the anode effectively suppresses the instability caused by the increasing arc voltage. The cusp configuration is thus used by many researchers [15–17]. However, the magnetic field diverts electrons to the duct walls so that a very small central section can be available for transporting electrons through the magnetic filter [15]. The cusp configuration makes the arc spot move around the edge of the cathode, which further degrades the efficiency of the beam transport. In contrast, the mirror magnetic configuration has an advantage of efficient ion beam transport. In the present work, we investigated the effect of the magnetic field structure on the cathode erosion behavior in the mirror magnetic configuration. The structure of the magnetic fields and the shape of the cathode were optimized from the view point of arc stability by both computer simulations and experimental studies. Using an oscillating current in the source magnet combined with a permanent magnet of opposite polarity behind the cathode, we could obtain a stable arc in the mirror magnetic configuration and increase the lifetime of the cathode. The cathode erosion was so uniform that more
than 90% of the cathode volume could be used during the lifetime.
2. Experimental Fig. 1 shows the schematic of the FVA system used in the present work. The FVA source was composed of a graphite cathode of diameter 80 mm, a ring-type anode of inner diameter 90 mm, a 60° bending plasma duct for magnetic filtering and four solenoid magnets for extraction, bending, outlet and deflection, respectively. The plasma duct was electrically isolated from the ground so that a bias voltage could be applied. The solenoid magnets could generate a center magnetic field of 250 Gauss at a maximum current of 5 A. Baffles in the duct were designed to prevent macroparticles from rebounding from the duct wall. A 115 mm diameter substrate holder was placed at a distance of 130 mm from the exit of the FVA source. In order to measure the beam current and apply bias voltage on the substrate, the substrate holder was electrically isolated from the reaction chamber. When measuring the beam current, the substrate holder was biased to −100 V to exclude the contribution of electrons. A solenoid source magnet (maximum center magnetic field of 150 Gauss) was installed around the cathode. We also placed a magnetic yoke (diameter 50 mm, thickness 10 mm) and a permanent Nd–Fe–B magnet (diameter 15 mm, thickness 10 mm, surface magnetic
138
J.-K. Kim et al. / Surface and Coatings Technology 124 (2000) 135–141
strength 4000 Gauss) at 5 and 17 mm from the back surface of the cathode, respectively. The magnetic field near the cathode was simulated using a Poisson code for various configurations of the permanent magnet, yoke and the current of the solenoid source magnet [18]. The self-magnetic field due to the arc spot was ignored in the simulation. The current of the extraction magnet was fixed at 3 A. Erosion behaviors were investigated by observing the eroded cathode after operating the FVA source until the stable ignition of the arc by the trigger could not be obtained (as defined by lifetime). The arc source was operated in an Ar environment at a pressure of 17 mPa. Two shapes of the cathode (cylindrical and tapered) were tested to assess the cathode erosion behavior and the stability of the beam current.
(a)
3. Results and discussion In our equipment, we used the mirror configuration where the magnetic polarities of the source and the extraction magnets were in the same direction to increase the efficiency of the beam transport [15,17]. In the present configuration of the magnets, the arc spot became unstable with increasing magnetic strength of the source magnet. Fig. 2 shows the magnetic field structures for various strengths of the source magnet. By increasing the current of the source magnet, i.e. increasing the magnetic strength, the opening of the acute angle formed by the magnetic field line on the cathode surface was directed outward. The arc spot tended to move to the edge of the cathode, resulting in the arc instability. We could not obtain a stable cathode arc erosion in this configuration. This instability can be suppressed by placing a permanent magnet of opposite polarity to that of the solenoid source magnet on the back side of the cathode. It can be seen in Fig. 3 that the opening of the acute angle was directed much more to the center of the cathode at the same current of the source magnet as in Fig. 2. However, the angle between the magnetic field line and the cathode surface also increased as the current of the source magnet increased. Erosion behaviors of the cathode were in good agreement with the simulation results. Fig. 4(a) shows the eroded cathode when the current of the source magnet was 1 A and the arc was ignited at the center of the cathode. Deep erosion in the center region showed that the arc spot moved around mainly in the center of the cathode. When the current of the source magnet increased to 3 A, the arc spot appeared to move in a wider region of the cathode, as shown in Fig. 4(b). This behavior can be easily understood by considering that the larger opening of the acute angle occurred with the higher current of the source magnet, as shown in Fig. 3(a) and (b). It must be noted that the erosion behavior in this
(b)
(c) Fig. 4. Morphologies of the used cathodes eroded with the permanent magnet only. The source magnet current was (a) 1 A and (b) 3 A. (c) Eroded cathode when arc was ignited at the edge of the cathode and the current of source magnet was 1 A.
magnetic configuration is dependent on the ignition position. Fig. 4(c) shows the eroded cathode when the arc was ignited at the edge of the cathode (top of the figure) and the current of the source magnet was 1 A. Even if the acute angle of the magnetic field was directed toward the center of the cathode, as can be seen in Fig. 3(a), the eroded cathode showed that the arc spot was not moved to the center of the cathode. Frequent ignition was required to maintain the arc under this operating condition. This behavior can be understood by considering that the initial erosion of the cathode
J.-K. Kim et al. / Surface and Coatings Technology 124 (2000) 135–141
139
Fig. 5. Magnetic field structure near cathode with both permanent magnet and yoke for various currents of the source magnet of (a) 1 A, (b) 3 A and (c) 5 A. The current of the extraction magnet was fixed at 3 A.
(a)
(b) Fig. 6. Morphologies of the used cathodes eroded with both the permanent magnet and the yoke. The source magnet current was ocillated in the range of (a) 0–1 A and (b) 0–1.5 A.
edge immediately changes the angle between the magnetic field and the cathode surface. In the present case, the magnetic field line becomes parallel to the eroded surface when the arc was ignited at the edge of the cathode. Another possible reason is the interaction between the distorted electric field near the edge and the self-magnetic field due to arc spot generation. The effect of a permanent magnet can be reduced by placing a yoke between the cathode and the permanent magnet. Fig. 5 shows the simulated magnetic field line structures for various currents of source magnet with the yoke and the permanent magnet. A comparison of Fig. 5 with Fig. 3 shows that the yoke increased the opening of the acute angle between the magnetic field line and the cathode surface under the same magnetic configuration. The larger acute angle can widen the arc spot movement, which was confirmed by the morphology of the eroded cathode at the current of the source magnet of 1 A. A combination of the yoke and the permanent magnet can increase the efficiency of the cathode by about 20%, as can be estimated from the eroded volume of the cathode over its life. A more efficient erosion of the cathode could be obtained by oscillating the current of the source magnet. Oscillating the magnetic strength of the source magnet can dynamically change the magnetic field structure near the cathode. By oscillating the current of the source magnet, the acute angle between the magnetic field line and the cathode surface was periodically varied. The arc spot can thus move around in a wider region of the
140
J.-K. Kim et al. / Surface and Coatings Technology 124 (2000) 135–141
cathode surface. Fig. 6 shows the morphology of the eroded cathode when the current of the source magnet oscillated from 0 to 1 A (a) and from 0 to 1.5 A (b). In contrast to deep erosion in the center region under a static source magnetic field [Fig. 4(a)], a much wider erosion was observed by the current oscillation, as can be seen in Fig. 6(a). The eroded area was proportional to the maximum current of the source magnet, as can be seen from a comparison between Fig. 6(a) and (b). Even if a uniform erosion of the cathode could be obtained by oscillating the current of the source magnet, the arc was frequently extinguished when a large oscillating current amplitude was used. Fig. 7(a) shows the carbon beam current variation for 300 s when using a cathode of cylindrical shape. The current of the source magnet was periodically varied from 1 to 3 A (upper data of the graph). A fluctuating and lower beam current was observed when the arc spot moved around the edge [arrow in Fig. 7(a)]. The arc spot in the edge region could not be recovered to the center of the cathode but was extinguished. During this experiment, therefore, two
ignition times of 70 and 210 s were required. A more stable arc could be obtained by changing the shape of the cathode from cylindrical to tapered. Fig. 7(b) shows the beam current when using the tapered cathode. In contrast to Fig. 7(a), a more stable beam current was obtained. It can be further noted that the arc spot near the edge moved to the top surface of the cathode, as can be seen in the beam current change at about 160 s [arrow in Fig. 7(b)]. This stabilizing action can easily be understood by considering that the large driving force for the arc spot movement to the center region could be applied when the arc spot was placed in the tapered side of the cathode [14]. Fig. 8(a) and (b), respectively, show the tapered cathode before and after operating the arc source for 2000 min at an arc current of 60 A. More than 90% of the cathode volume could be used in an optimum erosion condition by oscillating the source magnetic strength from 1 to 3 A. Figs. 7 and 8 clearly show that uniform erosion of the cathode and a stable beam current could be obtained in the present magnetic configuration by using an oscillating current of the source magnet and the tapered cathode. Fig. 9 shows the time-averaged beam current for 60 s for various values of the extraction magnet current. For this experiment, the currents of bending, deflection and outlet magnets were fixed at 3 A. During the measurement, the plasma duct was
(a)
(b) Fig. 7. Time variation of carbon beam current for (a) cylindrical cathode and (b) tapered cathode. The current of the source magnet varied from 1 to 3 A at 0.1 Hz. The insert of each graph shows the shape of the cathode.
Fig. 8. Morphologies of tapered cathode (a) before and (b) after use for 2000 min with both permanent magnet and yoke. The current of the source magnet varied from 1 to 3 A at 0.1 Hz.
J.-K. Kim et al. / Surface and Coatings Technology 124 (2000) 135–141
Fig. 9. Dependence of the carbon beam current on the extraction magnet current.
biased to 50 V to further enhance the transport of the plasma beam [19]. The beam current increased by increasing the current of the extraction magnet from 0 to 3 A. However, the beam current saturated at higher values of the extraction magnet current because the center of the plasma beam was displaced from that of the plasma duct. Higher beam current could be obtained when using a higher arc current. We could obtain about 350 mA when the arc current was 60 A and the extraction magnet current was 3 A. Amorphic diamond films were deposited on Si (100) wafers for various substrate bias voltages ranging from 0 to −250 V. By increasing the bias voltage, the hardness measured by nanoindentation and the residual compressive stress measured by the beam deflection method decreased from 65 to 45 GPa and 6.4 to 3.1 GPa, respectively. The content of graphitic component was increased by increasing the negative bias voltage, as observed in EELS and Raman spectroscopy. Details of the structure and the mechanical properties of deposited films will be reported elsewhere.
4. Conclusions The most significant result of the present work is that arc instability in the mirror magnetic configuration could
141
be suppressed by placing a permanent magnet of opposite polarity behind the cathode. This result means that the present magnetic configuration is suitable for the filtered vacuum arc system of high efficiency of the beam transport and the cathode erosion. We further showed that oscillating the current of the source magnet was effective in extending the area of the arc spot movement. A tapered cathode exhibited a more stable arc than a cylindrical cathode, as confirmed by the time variation of the beam current. By using the oscillating current of the source magnet and a tapered cathode of diameter 80 mm, continuous operation for 2000 min at an arc current of 60 A was obtained, at which more than 90% of the cathode volume could be used. A stable beam current of about 350 mA was obtained under the present operating conditions.
References [1] R. Lossy, D.L. Pappas, R.A. Roy, J.J. Cuomo, J. Bruley, J. Appl. Phys. 77 (1995) 4750. [2] R.L. Boxman, V. Zhitomirsky, B. Alterkop, E. Gidalevich, I. Beilis, M. Keidar, S. Goldsmith, Surf. Coat. Technol. 86–87 (1996) 243. [3] N. Mustapha, R.P. Howson, Surf. Coat. Technol. 92 (1997) 29. [4] H.C. Miller, J. Appl. Phys. 52 (1981) 4523. [5] I.I. Aksenov, S.I. Vakula, V.G. Padalka, V.M. Khoroshikh, Sov. Phys. Tech. Phys. 25 (1980) 1164. [6 ] C.S. Bhatia, S. Anders, K. Bobb, R. Hsiao, D.B. Bogy, I.G. Brown, Proc. Int. Conf. Micromechatronics for Information and Precision Equipment, Tokyo (1997). [7] J.E. Daalder, J. Phys. D: Appl. Phys. 12 (1979) 1769. [8] V.N. Zhitomirsky, R.L. Boxman, S. Goldsmith, Surf. Coat. Technol. 68–69 (1994) 146. [9] V.N. Zhitomirsky, R.L. Boxman, S. Goldsmith, J. Vac. Sci. Technol. 13 (1995) 2233. [10] M.G. Drouet, IEEE Trans. Plasma Sci. 13 (1985) 235. [11] J. Vyskovcil, J. Musil, Surf. Coat. Technol. 43–44 (1990) 299. [12] B.Y. Moizhes, V.A. Nemchinsky, J. Phys. D: Appl. Phys. 24 (1991) 2014. [13] P.D. Swift, J. Appl. Phys. 67 (1990) 1720. [14] B.F. Coll, D.M. Sanders, Surf. Coat. Technol. 81 (1996) 42. [15] M.M.M. Bilek, Y. Yin, D.R. McKenzie, IEEE Trans. Plasma Sci. 24 (1996) 1165. [16 ] S. Falabella, US Patent No. 5,468,363, 1995. [17] S. Anders, A. Anders, I.G. Brown, J. Appl. Phys. 75 (1994) 4895. [18] K. Halbach, R.F. Holsinger, Particle Accelerators 7 (1976) 213. [19] A. Anders, S. Anders, I.G. Brown, J. Appl. Phys. 75 (1994) 4900.