Tribological properties of nitrogen implanted diamond-like carbon

Tribological properties of nitrogen implanted diamond-like carbon

Nuclear Instruments and Methods in Physics Research B 148 (1999) 659±663 Tribological properties of nitrogen implanted diamond-like carbon S. Miyagaw...

495KB Sizes 1 Downloads 7 Views

Recommend Documents

No documents
Nuclear Instruments and Methods in Physics Research B 148 (1999) 659±663

Tribological properties of nitrogen implanted diamond-like carbon S. Miyagawa a

a,* ,

M. Ikeyama a, S. Nakao a, K. Saitoh a, Y. Miyagawa a, K. Baba b, R. Hatada b

National Industrial Research Institute of Nagoya, Hirate-cho, Kita-ku, Nagoya 462, Japan b Technology Center of Nagasaki, 2, Ikeda, Ohmura, Nagasaki 856, Japan

Abstract Diamond-like carbon (DLC) ®lms prepared by an ion plating method using C6 H6 plasma were implanted with 25 keV nitrogen 15 N ions to a total dose of 5 ´ 1017 ions/cm2 at temperatures up to 900°C. The atomic concentration of the implanted layer was measured by nuclear reaction analysis (NRA) and elastic recoil detection (ERD). Macro- and micro-tribological properties of the implanted layer were examined by a reciprocating tribotester and atomic force microscope (AFM), respectively. It was found that the nitrogen concentration in DLC ®lms implanted at RT saturated at 23%, and a groove was formed by micro-wear on the surface at a low load of 150 nN, whereas the groove was not formed on the unimplanted surface. At 900°C implantation, the maximum concentration of nitrogen was 10%, and nitrogen implanted at high dose di€used into the substrate. Ó 1999 Elsevier Science B.V. All rights reserved. PACS: 79.20.Rf; 68.35.Gy Keywords: DLC; Micro-tribology; Nitrogen; Ion implantation

1. Introduction Nanometer-scale friction, wear and lubrication between materials in contact are not only of fundamental importance in practical application, e.g. micromachines, magnetic recording head media, but are also of great interest as a physical phenomenon. The use of diamond-like carbon (DLC) thin ®lms as a mechanical protective coating material is well known. DLC ®lms with properties varying between those of diamond (sp3 ) and

* Corresponding author. Tel.: +81 52 911 2111; fax: +81 52 916 2802; e-mail: [email protected]

graphite (sp2 ) have been investigated in the past two decades [1]. A variety of methods has been employed to deposit DLC ®lms which may contain hydrogen or may be hydrogen free. Surface modi®cation using ion beams seems to be a useful method to improve micro-tribological properties of these ®lms. It is known that superior atomic-scale lubricities can be obtained by reducing the shear strength between sliding surfaces [2]. Nitrogen implantation is suitable for forming an amorphous layer on hard materials such as diamond. Recently, polished CVD diamond ®lms were implanted with nitrogen ions and the e€ects of nitrogen implantation into diamond ®lms on tribological properties were examined [2,3]. In this

0168-583X/98/$ ± see front matter Ó 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 5 8 3 X ( 9 8 ) 0 0 7 9 0 - 3

660

S. Miyagawa et al. / Nucl. Instr. and Meth. in Phys. Res. B 148 (1999) 659±663

paper, the depth pro®les of nitrogen implanted in DLC and the tribological properties of the implanted layer were investigated using nuclear reaction analysis (NRA), elastic recoil detection (ERD) and atomic force microscope (AFM). 2. Experimental DLC ®lms deposited on a Si wafer with thickness of 1 lm using an ion plating method employing a C6 H6 plasma were used [4]. Nitrogen implantation was performed with mass-analyzed 15 ‡ N2 ion beams at 50 keV, and the implanted area was 8 mm in diameter. A raster scanned beam was used. The beam current during hot implantation was measured using a Faraday cup in front of the target holder at intervals of a ®xed period. The beam current density of 15 N2 ions was in the range of 4±6 lA/cm2 , and the total dose was between 1 ´ 1016 and 5 ´ 1017 ions/cm2 for 25 keV 15 N ions. Samples held on an Mo plate were heated with electron beam bombardment on the back side of the sample holder using an electron beam gun to 900°C. The substrate temperature during implantation was monitored by an infrared thermometer. The implantation chamber was evacuated to 8 ´ 10ÿ9 Torr and the residual vacuum was about 5±8 ´ 10ÿ7 Torr during the hot implantation. After nitrogen implantation, depth pro®les of 15 N implanted into DLC were measured using the 15 N (p, ac) 12 C nuclear reaction at a resonance energy of 429 keV. The proton energy incident on the sample surface was regulated by supplying directly a retardation potential between 0 and +30 kV on the sample holder. 4.43 MeV c-rays emitted by the nuclear reaction were detected with an NaI(Ti) scintillation counter. The depth resolution was estimated to be about 4 nm at the surface [5]. The amount of hydrogen in DLC was determined using ERD method. A 2.8 MeV He beam incident at 75° was used, and the recoiled hydrogen atoms were collected at 30° through a mylar absorber foil of 6 lm thickness. Macro-tribological tests were performed with a reciprocating tribometer in air, where the friction couple consists of the upper diamond tip or alumina ball and the lower plate specimen. A prede-

termined load (5±80 gf) was applied with a weight, and then the plate specimen was oscillated repeatedly at a predetermined stroke amplitude of 5 mm and a constant friction speed of 6 mm/s. Frictional force was measured using a strain gage. Wear tests were carried out at a temperature of 28o C and a humidity of 55%. AFM (Seiko, SPI3800) with a very sharp diamond tip (radius 0.15 lm) was used for micro-wear test. The implanted surface was scanned repeatedly with the diamond tip with a load of 150 nN in the range of 1 lm ´ 1 lm. The square wear marks formed by the repeated scan were measured by the AFM with the same diamond tip with a load of 1 nN. 3. Results and discussion Nitrogen pro®les implanted in DLC with high dose were calculated by a Monte Carlo simulation using the dynamic SASAMAL code [6]. Fig. 1 shows the calculated depth pro®les of projected nitrogen. The implantation energy of nitrogen 15 N is 25 keV, and the ion doses are in the range of 0.1± 5 ´ 1017 ions/cm2 . The density of DLC is assumed to be 1.8 g/cm3 [4] and the DLC contains hydrogen of 10% as will be shown later. A detailed description of the simulation code is given in Ref. [6]. In the calculation, redistribution of the pro®les caused by di€usion of accumulated nitrogen is neglected. In Fig. 1, the maximum concentration of nitrogen exceeds 50% at an ion dose of 5 ´ 1017

Fig. 1. Ion dose dependence of 25 keV nitrogen (15 N) pro®les in carbon (1.8 g/cm3 ) calculated by SASAMAL simulation code.

S. Miyagawa et al. / Nucl. Instr. and Meth. in Phys. Res. B 148 (1999) 659±663

661

ions/cm2 , and the energy density deposited by nuclear collision is high near the surface. The peak shift of the range pro®les to the surface with increasing ion dose is ascribed to the recession of the surface by sputtering and the accumulation of nitrogen in shallower layers. Fig. 2 shows the c-ray excitation curves due to the nuclear reaction obtained from DLC implanted with di€erent doses of 25 keV 15 N at room temperature and 900o C. The horizontal axis shows the excitation energy of the protons. The surface of the DLC ®lm is at 429 keV. The vertical axis on the right side shows the nitrogen concentration (%) in DLC ®lm. As can be seen, the nitrogen concentration near the surface for RT implantation increases with implantation dose and saturates at about 23 at.% . The saturated concentration agrees well with the value obtained for glassy carbon implanted with 12 and 40 keV nitrogen, 25%, by Link et al. [7]. In the case of 900°C implantation, the pro®le at low dose implantation is similar to that for RT implantation, but the nitrogen concentration for high dose implantation is relatively low at the surface and saturates at 10%. In the latter case, most of the implanted nitrogen di€uses into the substrate. DLC ®lms prepared with C6 H6 plasma contain hydrogen [4]. ERD spectra of DLC unimplanted and implanted with 25 keV nitrogen ions at RT and 900°C are shown in Fig. 3. Simulated results

for the hydrogen pro®le obtained using RBX code [8] are also shown. Hydrogen concentration was calibrated with a hydrogen implanted Si sample. As can be seen, DLC prepared by C6 H6 plasma contains up to 9.5 at.% hydrogen, and it decreases with increasing implantation temperature down to 4.5 at.% at 900°C Surface roughness of DLC ®lms implanted with nitrogen was measured by AFM. Fig. 4(a), (b), and (c) shows AFM images of a DLC surface implanted with 25 keV nitrogen ions at a dose of 5 ´ 1017 ions/ cm2 at di€erent temperatures, (a) RT, (b) 500°C, and (c) 900°C. Unimplanted surfaces (not shown) at RT and at 500°C were very ¯at and similar to (a) in this magni®cation. At 900°C, the unimplanted surface got a little rough. In the case of 500°C implantation, a grain structure appears with implantation dose and the grain size reaches a maximum around a dose of 2 ´ 1017 ions/cm2 . At 5 ´ 1017 ions/cm2 , the surface was covered with grains with a diameter of 30±40 nm. At 900°C, a rectangular structure (40 ´ 70 nm) appeared as shown in (c). The ion dose dependence of the average surface roughness Ra obtained from AFM images at different temperatures is shown in Fig. 5. Ra on unimplanted surface increases slightly at 900°C, which originates in the dissociation of hydrogen. The increase of Ra at 2 ´ 1017 ions/cm2 in hot implantation is due to appearance of the grains. For

Fig. 2. Ion dose dependence of nitrogen pro®les in DLC implanted with 25 keV 15 N at room temperature and at 900°C. Horizontal axis shows excitation energy of the proton. The surface of DLC is at 429 keV. Vertical axis on the right side shows nitrogen concentration in DLC.

Fig. 3. ERD spectra of DLC implanted by 25 keV nitrogen ions at RT and 900°C. Thin solid line (±) shows the results of a simulation.

662

S. Miyagawa et al. / Nucl. Instr. and Meth. in Phys. Res. B 148 (1999) 659±663

Fig. 5. Ion dose dependence of surface roughness Ra of DLC.

higher doses the surface is ¯attened with increasing nitrogen dose, which is attributed to sputtering. To study macro-tribological properties, the dependence of the friction coecient of DLC ®lms on the load of diamond tip or alumina ball was measured. Sliding tests were performed with a reciprocating tribometer, where the friction couple consists of an upper diamond tip or alumina ball and a DLC ®lm. The results are shown in Fig. 6.

Fig. 4. AFM images of DLC implanted with 25 keV nitrogen ions at a dose of 5 ´ 1017 ions/cm2 . (a) RT. (b) 500°C. (c) 900°C.

Fig. 6. Dependence of the friction coecient of DLC on the load of diamond tip and alumina ball.

S. Miyagawa et al. / Nucl. Instr. and Meth. in Phys. Res. B 148 (1999) 659±663

663

4. Summary

Fig. 7. AFM image of micro-scale wear groove of DLC implanted at a dose of 5 ´ 1017 ions/cm2 at RT.

The friction coecients l for the diamond tip are expressed approximately as l / Wa , where W is load [3]. a ˆ 0.14 for unimplanted, and 0.08 for 900°C implantation. On the other hand, a ˆ )0.42 for alumina ball. Low friction coecient of 900°C implantation is caused by the formation of graphite layer on DLC surface by nitrogen implantation. These features agree with previous results obtained for diamond ®lms [3]. To evaluate micro-tribological property of nitrogen implanted DLC ®lms, the implanted surface was rubbed with diamond tip of an AFM. The atomic force between the specimen and diamond tip was estimated using the atomic force curve. Fig. 7 shows a micro-wear groove which was formed by repeated scans with a load of 150 nN on the surface implanted at RT with a dose of 5 ´ 1017 ions/ cm2 . A square wear mark of 1 lm ´ 1 lm was formed by repeated scan of 20 times, and the average depth of the groove was 3.5 nm. The rate of increase of wear per scan of nitrogen implanted DLC surface corresponds to about 0.18 nm. On unimplanted surface no wear grooves were formed.

DLC ®lms prepared by an ion plating method employing a C6 H6 plasma were implanted with nitrogen ions at room temperature, 500°C and 900°C. 1. At room temperature, the nitrogen concentration in DLC saturates at 23%, and the implanted layer is amorphised. A groove by micro-wear can be formed on the implanted surface at a low load of 150 nN, whereas no wear groove was formed on unimplanted surfaces. 2. At 900°C implantation, the saturation level of nitrogen is as low as 10%, and excess nitrogen di€uses into the substrate. The hydrogen concentration in DLC decreases with implantation temperature and a rectangular structure is formed above 2 ´ 1017 ions/cm2 , which size decreases with implantation dose.

Acknowledgements This study was performed through Special Coordination Funds for Promoting Science and Technology of the Science and Technology Agency of the Japanese Government. References [1] Y. Lifshitz, Diamond and Related Materials 5 (1996) 388. [2] R.Kaneko, T. Miyamoto, E. Hamada, in: B. Bhushan (Ed.), Handbook of Micro/Nano Tribology, CRC Press, Boca Raton, FL, 1995, p. 183. [3] S. Miyake, T. Miyamoto, R. Kaneko, Nucl. Instr. and Meth. B 108 (1996) 70. [4] K. Awazu, Y. Funada, K. Shimamura, M. Iwaki, in: Proc. 12th Symposium on Surface Modi®cation by Ion Implantation, 1996, p. 85. [5] Y. Miyagawa, K. Saitoh, M. Ikeyama, S. Nakao, S. Miyagawa, Nucl. Instr. and Meth. B 118 (1996) 209. [6] Y. Miyagawa, M. Ikeyama, K. Saitoh, G. Massouras, S. Miyagawa, J. Appl. Phys. 70 (1991) 7289. [7] F. Link, H. Baumann, A. Markwitz, E.F. Krimmel, K. Bethge, Nucl. Instr. and Meth. B 113 (1996) 235. [8] E. Kotai, Nucl. Instr. and Meth. B (1994) 588.