Thin Solid Films 390 2001.

192 196

Anode mode in cathodic arc deposition apparatus with various cathodes and ambient gases
R. Miyano, T. Saito, K. Kimura, M. Ikeda, H. TakikawaU , T. Sakakibara
Department of Electrical and Electronic Engineering, Toyohashi Uni ersity of Technology, Toyohashi, Aichi 441-8580, Japan

Abstract The anode mode of a vacuum arc in a cathodic arc deposition apparatus was observed as a function of ambient gas pressure ranging from 0.01 to 300 Pa. The chamber 400 mm in diameter and 600 mm in length. made of stainless steel SUS304. acted as the anode. The arc was operated at a relatively low constant current of 50 A. The cathode materials used were Al, Ti, Fe, Ni, and Cu, and ambient gases were He, Ne, Ar, H 2 , N2 , O 2 , and CH 4 . The principal results are as follows. 1. As the pressure was increased, the anode mode changed from diffuse-arc to footpoint to plane luminous to anode-spot mode. 2. The anode mode and resultant arc voltage increase were strongly dependent on gas species, and weakly on the cathode material. 3. Comparing diatomic and polyatomic H 2 , N2 , O 2 , and CH 4 . with mono-atomic molecule gases He, Ne, and Ar., the onset pressure of the anode mode transition in the former was lower, the arc voltage higher, and the footpoints more numerous, smaller, and clearer. Both the dependence of the ambient pressure and the inuence of the cathode materials and gas species on the anode mode changes were explained by the ion deciency theory. 2001 Elsevier Science B.V. All rights reserved.
Keywords: Cathodic arc deposition apparatus; Anode mode; Pressure dependence; Cathode material; Ambient gas

1. Introduction The cathodic vacuum arc is known to be the simplest metallic or carbon ion source and is industrially applied to the preparation of thin solid lms w1x. A variety of lms of metals, nitrides, oxides, carbides and carbonaceous materials can be fabricated using a cathodic vacuum arc. The technique is called vacuum arc deposition, cathodic arc deposition, or vacuum. arc ion plating. The authors have prepared various nitrides and oxides by this method to date w2x. On the other hand, a vacuum arc can also be observed in the vacuum bulb of an electrical circuit breaker. Such arcs are known to present a variety of anode appearances, depending on arcing conditions. In order to understand the basic characteristics of an arc in the vacuum bulb, and to improve bulb performance, a

number of studies on anode phenomena of vacuum arcs have been carried out and the results are summarized in some reviews w3 5x. The anode mode is usually mapped as a function of the arc current and electrode gap length, and sometimes of the arc current and pressure. The anode mode is characteristically classied as follows: 1. Diffuse-arc mode: the anode is inert, acting merely as a collector of particles emitted from cathode spot. 2. Footpoint mode: one or more luminous points exist on the anode surface. 3. Anode-spot mode: one large or several small anode spots appear on the anode surface. The anode spot is active and erodes the surface. 4. Intense-arc mode: the anode spot is very active and severely erodes the surface. However, the anode mode at currents as low as that used for cathodic vacuum arc deposition has not been

Corresponding author. E-mail address: takikawa@eee.tut.ac.jp H. Takikawa..

0040-6090r01r$ - see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S 0 0 4 0 - 6 0 9 0 0 1 . 0 0 9 1 8 - X

R. Miyano et al. r Thin Solid Films 390 (2001) 192 196

193

studied very well. The authors have observed the footpoint and the plane luminous modes in cathodic arc deposition apparatus for TiN lm preparation at a relatively high pressure w6,7x. Therefore, it is important to reveal the inuences of the cathode material and ambient gas species on the anode mode. In the present study, the appearance of the anode surface in a cathodic vacuum arc deposition apparatus was observed for various cathode materials and ambient gases as pressure ranged from medium to soft vacuum. The arc voltage was also measured simultaneously. 2. Experimental Fig. 1 depicts the vacuum arc deposition apparatus. A cylindrical vacuum chamber 400 mm in diameter and 600 mm in length. made of stainless steel SUS304. serves as a grounded anode. A cathode 64 mm in diameter. is placed at the end. The substrate is usually located at the center of the chamber. An insulation plate made of ceramic Zr2 O 3 . is placed around the cathode in order to expand the plasma in the substrate direction. The arc is ignited by a mechanical triggering system. In the present study, Al, Ti, Fe, Ni, and Cu were used as cathode materials, and the ambient gases used were He, Ne, Ar, H 2 , N2 , O 2 , and CH 4 . The chamber was rst pumped down to less than 0.01 Pa. Then the gas, at a ow rate of 40 mlrmin, was introduced into the chamber, and the main valve of the pumping system was closed. The arc was started at 0.01 Pa. The anode surface was observed, and the arc voltage was measured with a pen recorder, while the pressure was gradually increased to 300 Pa. A DC power supply with a constant voltage 500 V. was used, and the arc current was regulated at a constant of 50 A using a variable resistor bank. 3. Results 3.1. Aspects of the anode surface Fig. 2 shows photographs of the arcs with a Ti cathode in N2 gas at various pressures. The angle recorded was illustrated in Fig. 1. The diffuse-arc mode was observed at lower pressure. A typical image of the diffuse-arc is shown in Fig. 2a, indicating that the anode is inert. A few small luminous footpoints appeared on the anode surface at approximately 30 Pa. As the pressure was increased, the number of footpoints increased and then coalesced, so that a plane luminous layer appeared on the anode surface w6x. At higher pressure, the plasma stream was pinched, and a highly luminous position appeared on the anode surArc voltages for various cathodes and gases as a function of pressure are shown in Fig. 4. With a Ti cathode as shown in Fig. 4b, the arc voltage was approximately 20 V for N2 gas at low pressure. As the pressure was increased, the voltage increased, with the ongoing appearance of footpoints. The voltage abruptly increased at approximately 60 Pa, corresponding to the transition to the anode-spot mode. For other gases, the voltage was also 20 V in diffuse-arc mode. The pressures at which the footpoints rst appeared were 70 " 10, 60 " 10, 70 " 10, 7 " 2, 30 " 10, 5 " 2, and 7 " 2 Pa for He, Ne, Ar, H 2 , N2 , O 2 , and CH 4 , respectively. The onset pressure of mode-change and the voltage increase obviously occurred at lower pressure for diatomic and polyatomic than for mono-atomic molecule gases. In addition, the arc voltage had uctuations of

Fig. 1. Experimental apparatus for the vacuum arc.

face, which is considered to be the anode-spot mode, as shown in Fig. 2d. Such a transition in anode mode relative to the pressure diffuse-arc, footpoint, plane luminous, and anode-spot modes. was the same for other gases. However, the mode-change pressures were different, according to the gas species. Some unique and interesting anode aspects of a Ti cathode arc are shown in Fig. 3. For a mono-atomic molecule gas, the footpoints are relatively large, few in number and blurred, as shown in Fig. 3a,b for He and Ar, respectively. For a diatomic molecule gas, the footpoints tended to be small, numerous and clear. Especially for H 2 , small, brilliant footpoints with a sharp outline were distributed in a beautifully uniform array at regular intervals, as shown in Fig. 3c. For O 2 , very ne footpoints with strong radiation were observed, as shown in Fig. 3d. These gas effects were the same for other cathode materials. 3.2. Arc oltage

194

R. Miyano et al. r Thin Solid Films 390 (2001) 192 196

Fig. 2. Photographs of arc between the cathode and anode for Ti cathode and N2 gas: a. 5 Pa diffuse mode.; b. 40 Pa footpoint mode.; c. 60 Pa plane luminous mode.; and d. 150 Pa anode-spot mode..

approximately "2, "5 and "10 V for the diffuse, footpoint and anode-spot modes, respectively, for all gases.

For other cathodes, the trend mentioned above was the same, although the mode-change pressures were slightly different, as shown in Fig. 4a,c e.

Fig. 3. Photographs of the unique anode aspects of Ti cathode arc with various gases: a. He 100 Pa.; b. Ar 90 Pa.; c. H 2 30 Pa.; and d. O 2 20 Pa..

R. Miyano et al. r Thin Solid Films 390 (2001) 192 196

195

Fig. 4. Pressure dependence of arc voltage for various cathode materials and gas species: a. Al; b. Ti; c. Fe; d. Ni; and e. Cu.

4. Discussions 4.1. Pressure dependence of anode mode In general, provided that the anode is inert, the voltage of a vacuum arc is nearly equal to the cathode drop, because the anode and arc column drops are almost zero. When the anode surface becomes active, the arc voltage increases. The mode changes, as well as the arc voltage increases, are explained based on the formation of an ion deciency region near the anode w3 5,8x. The ion deciency region is formed when ions emitted from a cathode spot collide with other particles and do not reach the anode surface. Then the ion deciency must be compensated for by ions provided from the anode andror ambient gas near the anode.

This compensation causes an anode drop andror electric eld, thus increasing the arc voltage. The electric eld occurrence in a footpoint mode arc was actually conrmed in the same experimental apparatus by means of Langmuir probe measurement w6x. Therefore, when the pressure rises, it is readily understandable that the anode mode changes. 4.2. Inuence of gas species and cathode material on anode mode As shown in Section 3.2, the anode mode strongly depended on gas species and slightly on the cathode material. This fact may be qualitatively understood by the occurrence of ion deciency, i.e. banishment of ions emitted from the cathode spot. The collision

196

R. Miyano et al. r Thin Solid Films 390 (2001) 192 196

cross-sections of metal ion gas molecule are evaluated based on the polarization potential of the gas molecule. The order of cross-section is Ne - He - Ar- N2 ,O 2 . - CH 4 - H 2 approx., 1.0= 10y1 9 , 1.3= 10y19 , 1.5= 10y1 9 , 1.7= 10y1 9 , 1.7= 10y1 9 , 2.6= 10y1 9 , and 3.7= 10y1 9 m2 , respectively, for all metal ions in the present study.. Thus, qualitatively, the mode-change pressures are lower for diatomic and polyatomic than for monoatomic molecule gases. However, the order of modechange pressure is not strictly consistent with the cross-section order. We suppose that the process of gas dissociation and ionization may be further involved in the gas species dependency. With regard to cathode material dependency, it is considered as follows. The diameters of Al, Ti, Fe, Ni, and Cu atoms are similar 0.294, 0.286, 0.248, 0.230, and 0.256 nm, respectively., so that the collision crosssection between these metal ions and the gas molecules is not varied. Therefore, it would appear that the ion deciency forms at a similar pressure, and thus the anode mode changes at a similar pressure. Nonetheless, it is a fact that there is small dependence on the cathode material. This is considered to be due to not only the size of the ions, but also their charge, energy and density. 5. Conclusions

2. The anode mode and resultant voltage increase were strongly dependent on gas species and weakly on the cathode material. 3. Comparing diatomic and polyatomic H 2 , N2 , O 2 , and CH 4 . with mono-atomic molecule gases He, Ne, and Ar., the onset pressure of the anode mode transition in the former was lower, the arc voltage higher, and the footpoints were more numerous, smaller, and clearer. Both the dependence of the ambient pressure and the inuence of the gas species and cathode materials on the anode mode changes can be qualitatively interpreted by the ion deciency theory and the difference of collision cross-section. However, further investigation is required for qualitative and precise understanding. In addition, new and unique features of the regular formation of footpoints for diatomic molecule gases are also a subject to be investigated. Acknowledgements The present work was partly supported by a Grantin-Aid for Scientic Research from The Ministry of Education, Science, Sports and Culture of Japan. References

In the present study, the pressure dependence of anode modes was revealed at the low current of 50 A, using a cathodic vacuum arc deposition apparatus. The anode aspect was observed and arc voltage was measured, while the pressure was increased from 0.01 to 300 Pa for various cathodes Al, Ti, Fe, Ni, and Cu. and ambient gases He, Ne, Ar, H 2 , N2 , O 2 , and CH 4 .. The principal results are summarized as follows. 1. The anode mode changed from diffuse-arc to footpoint to plane luminous to anode-spot mode as the pressure was increased.

w1x R.L. Boxman, P.J. Martin, D.M. Sandars Eds.., Handbook of Vacuum Arc Science and Technology, Noyes Publications, 1995. w2x H. Takikawa, K. Shinsako, T. Sakakibara, Thin Solid Films 316 1998. 73. w3x R.L. Boxman, S. Goldsmith, A. Greenwood, IEEE Trans. Plasma Sci. 255 1997. 1174. w4x H.C. Miller, IEEE Trans. Plasma Sci. PS-13 1985. 242. w5x C.W. Kimblin, IEEE Trans. Plasma Sci. PS-10 1982. 322. w6x H. Takikawa, T. Fujishima, T. Sakakibara, Trans. IEE Jpn. 114A 1994. 123 in Japanese. w7x H. Takikawa, T. Fujishima, T. Sakakibara, Trans. IEE Jpn. 113A 1993. 486 in Japanese. w8x V.A. Nemchinskii, Sov. Phys. Technical. Phys. 28 1983. 146.