Thin Solid Films 392 Ž2001. 11᎐15
Influences of optical emission settings on wear performance of metal-doped diamond-like carbon films deposited by unbalanced magnetron sputtering夽 Da-Yung WangU , Chi-Lung Chang Department of Materials Engineering, National Chung-Hsing Uni¨ ersity, Taiwan, PR China Received 24 April 2000; accepted 16 March 2001
Abstract We deposited amorphous diamond-like carbon ŽDLC. films using an unbalanced magnetron sputtering method on M2 tool steels. The deposition process was controlled using a closed-loop optical emission monitor ŽOEM., which regulated the flow of reactive gases of N2 and C 2 H 2 via a fast-responding piezo valve. By tuning the OEM settings for N2 and C 2 H 2 , we were able to deposit a compound DLC coating consisting of Ti, TiN, TiCN, TiC and a-C:HrTi in sequence. Excellent mechanical and wear performance was achieved. Microstructure and tribological properties of DLC coatings were characterized using transmission electron microscopy, electron probe microanalysis, Raman spectroscopy, Vickers, and wear tests. The Raman intensity ratio I DrIG of the characteristic G and D bands decreases and the G line position moves toward 1550 cmy1 with the decrease of OEM settings, corresponding to a higher sp 3 content and higher microhardness. Wear tests demonstrate that the average friction coefficient of DLC films reduces from 0.33 to 0.14, and wear life increases from 900 to 24 000 m with the decrease of OEM settings. The Ti content, corresponding to enhanced wear properties, in DLC increases with OEM settings. Finally, we discovered that the phase transformation from TiC to DLC is strongly influenced by OEM settings, and is demarcated at an OEM setting of 20%. 䊚 2001 Elsevier Science B.V. All rights reserved. Keywords: Diamond-like carbon; Optical emission monitor; Raman; Wear
1. Introduction Diamond-like carbon films ŽDLC, designated as a metal-doped and hydrogenated a-C:HrTi film in this study. contain carbon atoms in a short-range mixture of bonding coordinates, i.e. four-fold-coordinated sp 3 carbon bonds of diamond characteristic and threefold-coordinated sp 2 carbon bonds of graphite characteristic w1x. Metal atoms are incorporated within the carbon network, serving the purpose of stress reducU
Corresponding author. Tel.: q886-4-23381042; fax: q886-423367010. E-mail address: [email protected]
ŽD. Wang.. 夽 Presented at ICMCTF 2000, San Diego, California.
tion. DLC films exhibit many advanced properties such as high hardness w2x, low friction coefficient w3,4x, high wear resistance, high mass density, and chemical inertness, promising for variety of industrial applications. DLC coatings can be synthesized by using electron cyclotron resonance chemical vapor deposition ŽECRCVD. w5,6x, d.c.- and rf-plasma enhanced chemical vapor deposition ŽPECVD. w7,8x, laser ablation w9x, and rf magnetron sputtering w10,11x. Various studies have attempted to improve the inherent adhesion problems between DLC and its substrates by applying barrier layers such as B and Ti w12᎐14x. Other developments in enhancing the mechanical properties were also reported such as inclusion of TiN, TiCN and TiC layers in DLC matrix w15x. Monaghan et al. designed a DLC
0040-6090r01r$ - see front matter 䊚 2001 Elsevier Science B.V. All rights reserved. PII: S 0 0 4 0 - 6 0 9 0 Ž 0 1 . 0 0 9 7 6 - 2
D.-Y. Wang, C.-L. Chang r Thin Solid Films 392 (2001) 11᎐15
film, which was graded in both composition and microstructure w16,17x and resulted in compound films with significant improvements in film properties. A combined physical vapor deposition ŽPVD.rPECVD process would deposit an interface layer of Ti and a transition layer of TiNrTiC x N y , which provide a supporting foundation for the DLC top-layer. The topology eliminates the compatibility problem between DLC and steel substrates successfully. The transformation from TiC x N y to DLC via a poisoned cathode is controlled using a closed-loop optical emission monitor ŽOEM. w18᎐20x. To compensate the fluctuation of pumping efficiency, target poisoning and sputtering parameters, the flow rate of reactive gases can be rapidly and accurately regulated via a set of piezo valves and optical emission monitors, where OEM readings reveal the exact proportion of reactive gas consumed by the metal plasma. The main objective of present investigation was to examine the effect of OEM settings on the properties and wear performance of DLC films deposited by unbalanced magnetron sputtering. 2. Experimental details In this study, DLC films Ža-C:HrTi. were synthesized on M2 tool steels Ž 60 mm. using an unbalanced magnetron sputtering system UDP-900 manufactured by Teer Coatings Ltd. A closed-loop optical emission monitor ŽOEM. controlled the dynamic flow of reactive gases of N2 and C 2 H 2 . The OEM detected the optical emission of the metal plasma excited by magnetron sources during the reactive sputtering process. The emission measurement proceeded with the aid of a fiber optics system and a matched photomultiplier. An electronic package connected to the measuring unit controlled a piezo valve as a process-specific regulating element for reactive gas admission. The OEM reading indicated the ratio of the plasma intensity of Ti 2q ions Žemission wavelength of 454 nm. emitted during the reactive sputtering process with respect to its full emission intensity in metallic sputtering mode. Target poisoning was resolved using a d.c. power supply linked to a 2-kHz medium-frequency pulse generator. A pulsed d.c. power supply, with 20᎐100 kHz variable frequency was applied to the substrate arcing and radical excitation during the final DLC formation. The background pressure in the sputtering chamber was pumped below 1.3= 10y3 Pa before reactive gases were introduced. The target-to-substrate distance was 13 cm. The substrate surfaces were sputter-cleaned and heated to 200⬚C before deposition. During deposition, the argon pressure was 0.2 Pa with a pulsed biasing voltage of y70 V. The target current was 4 A. To start the deposition procedure, an interface layer of Ti and a transition layer of TiNrTiC x N y were sputtered sequentially from Ti targets as the under-layer for the subse-
quent deposition of DLC by OEM setting decreasing. The film thickness was approximately 2 m. The tribological behaviors were studied using a ballon-disk tester ŽCSEM.. The measurements were carried out at room temperature with a relative humidity of approximately 75% RH. A 52 100 steel ball, 6 mm in diameter, was selected for counter wear. The sliding speed was kept constant for all tests at 0.3 mrs with a load of 10 N. The microstructure and mechanical properties of DLC coatings were analyzed using transmission electron microscopy ŽTEM., electron probe microanalysis ŽEPMA., Raman spectroscopy, and Vickers hardness test. 3. Results and discussion 3.1. Microstructure e¨ olution of compound DLC ¨ ia OEM control The microstructure of the compound DLC film is demonstrated in Fig. 1, in which the deposition sequence of Ti, TiN, TiC x N y , and DLC layers can be clearly seen. The selected area diffraction ŽSAD. pattern from the DLC layer reveals the existence of diamond Ž220., graphite Ž110., TiC Ž220. and Ž200. microcrystallines w21,22x within DLC matrix, but the crystallinity was poor. The DLC top-layer possesses an amorphous lattice structure. Experimental results indicate that microcrystalline diamond, graphite, and TiC were co-deposited with DLC and were embedded in the DLC matrix. The subsequent transition from TiC to DLC is primarily governed by reactive gas control
Fig. 1. Cross-section TEM micrograph and SAD pattern, showing the multilayered structure of compound DLC film.
D.-Y. Wang, C.-L. Chang r Thin Solid Films 392 (2001) 11᎐15
Fig. 2. Influence of OEM settings on the microhardness of DLC coatings.
via OEM. The cross-section TEM micrograph in Fig. 1 reveals smooth transitions across each interface. The graded and multilayered transition from Ti to DLC guarantees an improved film adhesion, which eliminates the abrupt variation in physical properties between steels and carbon films. 3.2. Influences of OEM settings on microhardness The influence of reactive gas control on microhardness is conducted with OEM settings between 10% and 30%. At lower OEM settings, magnetron becomes unstable due to excessive cathode poisoning. At higher OEM settings, the PECVD synthesis of DLC becomes unfavorable due to the competing process of TiC formation. The Vickers microhardness measurements of DLC synthesized at various OEM settings ŽFig. 2. demonstrate that the microhardness decreases with OEM settings. The decreased microhardness of DLC synthesized at OEM settings of 20᎐25% could be attributed to the large proportion of TiC retained within the carbon matrix. The lower microhardness of approximately 2800 kgfrmm2 measured from DLC synthesized at 25% OEM setting resembles the microhardness of TiCN, which lies underneath of DLC. At this low level of C 2 H 2 , DLC formation is unlikely. Fig. 3 depicts the Raman spectra of diamond-like carbon with various OEM settings. A standard Raman spectrum of pure DLC possesses two characteristic broad bands near 1550 and 1350 cmy1 , generally called the G Žgraphitic. and D Ždisordered. bands, respectively w23x. The distinct DLC band structure disappears at OEM settings greater than 20%. This result agrees with the previous microhardness measurements that the microhardness enhancement is achieved only in films deposited at OEM settings lower than 20%, where DLC starts to form. The intensity ratio of Raman
Fig. 3. Raman spectra of DLC coatings deposited at different OEM settings.
bands I D rIG was calculated by fitting two Gaussian curves and linear background in the region 1100᎐1800 cmy1 , yielding I D rIG values of 0.884, 0.997, 1.19 and 4.146 for films deposited at OEM settings of 10, 15, 20 and 25%, respectively. Intensity ratios I D rIG as function of G line peak position for various types of amorphous carbon films w24,25x including current work are demonstrated in Fig. 4. With the decrease of OEM settings, the intensity ratio I D rIG shifts to lower values and the G-line peak position moves toward the standard G band position of 1550 cmy1 . Tamor et al. indicate that the relative intensity ratio I D rIG can be used as an index for sp 3 bond content w24,26,27x. A smaller value of I D rIG ratio corresponds to higher sp 3 bond content. As a comparison, the DLC films deposited by unbalanced magnetron ŽUBM. sputtering posses higher I D rIG ratio and lower sp 3 bonds than that of Arc-deposited DLC films. This can be at-
Fig. 4. Intensity ratio I D rIG as a function of G-line position.
D.-Y. Wang, C.-L. Chang r Thin Solid Films 392 (2001) 11᎐15
Fig. 5. Influence of OEM settings on friction coefficient and wear life of DLC coatings.
tributed to the hydrogen and metal inclusions in the UBM-deposited DLC films Ža-C:HrMe.. The sp 3 network linkage is less completed than in the Arcdeposited DLC, where a-C films were synthesized from solid graphite cathodes. Furthermore, the lesser degree of ion bombardment in UBM-deposited DLC films leads to a deviation from properties of pure DLC films. From practical point of view, however, the added advantages of higher fracture toughness and enhanced film adhesion in compensation of reduced microhardness of the UBM-synthesized, metal-doped DLC films are more promising for industrial applications. 3.3. Influences of OEM settings on wear beha¨ iors We apply ball-on-disk tribometers to evaluate the wear performance of the metal-doped DLC coatings. The results of wear tests Ždepicted in Fig. 5 and Table 1. demonstrate that with the decrease of OEM settings, the wear life of DLC coating increases remarkably from 900 to 24 000 m of wear. The average friction coefficient of DLC coatings is reduced from 0.33 to 0.14. These results agree with the previous microhardness and Raman measurements. Films deposited at 10% and 15% OEM settings possess similar results of friction coefficients and wear lives, which reveal the
low-friction and wear-resistant nature of DLC coatings. At OEM settings higher than 20%, the deposited films possess an average friction coefficient between 0.25 and 0.35, which resembles that of TiC thin films. The reduced friction coefficients of DLC deposited at lower OEM settings affect the counter-wear of the steel ball significantly, as demonstrated in Table 1. The benefit of the reduced friction coefficient and prolonged wear life of DLC coatings is obvious. In addition to the formation of subsequent interlayers of the compound DLC coating, OEM settings also determine the amount of metal doping Žin forms of metals and metal carbides. in DLC. EPMA analyses depict the elementary distribution of DLC films deposited with various OEM settings ŽTable 2.. Results indicate an increase of Ti inclusion with OEM settings. Carbon content decreases with OEM settings. Most Ti contents exist in form of titanium carbides. The nitrogen signals are most likely picked up from the underlying nitride layers. In summary, the OEM setting plays an important role in UBM synthesis of metal-doped DLC coatings, not only to regulate the reactive sputtering of metal and metal nitridercarbide interlayers, but also to control the amount of metal inclusion in the final DLC deposition. This control mechanism provides the compound DLC coating enhanced friction and wear performance. 4. Summary By means of a set of optical emission monitor and piezo valve, compound DLC coatings consisting of a series of metal and metal nitridercarbide interlayers and a final a-C:HrTi carbon film was deposited. The evolved microstructure is an amorphous matrix with inclusion of diamond Ž220., graphite Ž110., and TiC Ž200. Ž220. microcrystallines. Raman intensity ratios I D rIG decrease and the G line positions move toward 1550 cmy1 with the decrease of OEM settings, which indicates an increase of sp 3 bond contents in DLC. Microhardness measurement agrees with this result. Wear behavior of DLC compound coatings is strongly influenced by OEM settings, especially during the final DLC deposition. With the decrease of OEM settings,
Table 1 Wear performance of DLC by ball-on-disk tests OEM
Friction coefficient Žave .
Revolution number Žlaps.
Wear distance Žkm.
10% 15% 20% 25% 30%
0.142 0.154 0.208 0.224 0.325
638 092 533 515 55 952 6245 8663
24.1 19.7 2.1 1.64 0.98
D.-Y. Wang, C.-L. Chang r Thin Solid Films 392 (2001) 11᎐15 Table 2 Quantitative elementary analysis by EPMA OEM
10% 15% 20% 25% 30%
9 11 16 22 25
86 83 79 72 68
5 6 5 6 7
the wear life of DLC coating increases remarkably from 900 to 24 000 m of wear. The average friction coefficient of DLC coatings is also reduced from 0.33 to 0.14. The OEM setting plays an important role in UBM synthesis of metal-doped DLC coatings, not only to regulate the reactive sputtering of metal and metal nitridercarbide interlayers, but also to control the amount of metal inclusion in the final DLC deposition. Acknowledgements The authors wish to thank Dr Wei-Yu Ho from Surftech Corp. for generously providing the UBM deposition system to accomplish all the experiments. Also, we would like to thank the partial financial support from the National Science Council of Taiwan under Research Contract NSC89-2216-E-005-010. References w1x W.S. Bacsa, J.S. Lannin, D.L. Pappas, J.J. Cuomo, Phys. Rev. B 47 Ž1993. 10931. w2x A. Grill, V. Patel, B.S. Meyerson, in: Y. Yzeng, M. Yoshikawa, A. Feldman ŽEds.., Applications of Diamond Films and Related Materials, Elsevier, 1991, p. 683. w3x A. Erdemir, M. Switala, R. Wei, P. Wilbur, Surf. Coat. Technol. 50 Ž1991. 17. w4x E.I. Melets, A. Erdemir, G.R. Fenske, Surf. Coat. Technol. 73 Ž1995. 39.
w5x M. Diani, A. Mansur., L. Kubler, J.B. Bischoff, D. Bolmont, Diamond Rel. Mater. 3 Ž1994. 264. w6x A. Bousetta, M. Lu, A. Bensaoula, A. Schultz, Appl. Phys. Lett. 65 Ž1994. 696. w7x J. Schwan, W. Dworschak, K. Jung, H. Khrhardt, Diamond Rel. Mater. 3 Ž1994. 1034. w8x S. Veprek, J. Weidemann, F. Glatz, J. Vac. Sci. Technol. A 13 Ž1995. 2914. w9x J. Seth, R. Padiyath, S.V. Baba, Diamond. Rel. Mater. 3 Ž1994. 210. w10x T. Okada, S. Yamada, Y. Takeuchi, T. Wada, J. Appl. Phys. 78 Ž1995. 7416. w11x N. Axen, G.A. Botton, R.E. Somekh, I.M. Hutchings, Diamond Rel. Mater. 5 Ž1996. 163. w12x M.D. Bentzon, K. Mogensen, J. Bindslev Hansen, C. BarholmHansen, C. Trholt, P. Holiday, S.S. Eskildsen, Surf. Coat. Technol. 68r69 Ž1994. 651. w13x E.R. Kupp, W.R. Drawl, K.E. Spear, Surf. Coat. Technol. 68r69 Ž1994. 378. w14x P.S. Weiser, S. Prawer, R.R. Manory, A. Hoffman, P.J. Evans, P.J.K. Paterson, Surf. Coat. Technol. 71 Ž1995. 167. w15x A.A. Voevodin, C. Rebholz, J.M. Schneider, P. Stevenson, A. Matthews, Surf. Coat. Technol. 73 Ž1995. 185. w16x D.P. Monaghan, D.G. Teer, P.A. Logan, I. Efeoglu, R.D. Arnell, Surf. Coat. Technol. 60 Ž1993. 525. w17x D.P. Monaghan, K.C. Laing, P.A. Logan, D.G. Teer, Surf. Eng. June Ž1993. 347. w18x H. Dimigen, H. Hubsch, US Patent 4 525 417 Ž1985.. w19x H. Dimigen, H. Hubsch, R. Memming, Appl. Phys. Lett. 50 Ž1987. 1056. w20x H. Dimigen, C.-P. Klages, Surf. Coat. Technol. 49 Ž1991. 543. w21x D.Y. Wang, C.L. Chang, W.Y. Ho, Surf. Coat. Technol. 120᎐121 Ž1999. 138. w22x D.Y. Wang, C.L. Chang, W.Y. Ho, Thin Solid Films 355᎐356 Ž1999. 246. w23x R.O. Dillon, J.A. Woollam, V. Katkanant, Phys. Rev. B 30 Ž1984. 3482. w24x M.A. Tamor, W.C. Vassell, J. Appl. Phys. 76 Ž1994. 3823. w25x B.K. Tay, X. Shi, H.S. Tan, H.S. Yang, Z. Sun, Surf. Coat. Technol. 105 Ž1998. 155. w26x J. Schwan, S. Ulrich, V. Batori, H. Ehrhardt, S.R.P. Silva, J. Appl. Phys. 80 Ž1996. 440. w27x M. Yoshikawa, G. Katagiri, H. Ishida, A. Ishitani, J. Appl. Phys. 64 Ž1988. 6464.