Diamond-like carbon-coated Ti6Al4V: influence of the coating thickness on the structure and the abrasive wear resistance

Diamond-like carbon-coated Ti6Al4V: influence of the coating thickness on the structure and the abrasive wear resistance

Wear 249 (2001) 489–497 Diamond-like carbon-coated Ti6Al4V: influence of the coating thickness on the structure and the abrasive wear resistance A. D...

621KB Sizes 147 Downloads 800 Views

Wear 249 (2001) 489–497

Diamond-like carbon-coated Ti6Al4V: influence of the coating thickness on the structure and the abrasive wear resistance A. Dorner a,∗ , C. Schürer b , G. Reisel c , G. Irmer d , O. Seidel a , E. Müller a a

c

Institute of Ceramic Materials, Freiberg University of Mining and Technology, Freiberg, Germany b Institute of Physical and Mechanical Technologies, Chemnitz, Germany Institute of Composite Materials and Surface Technology, Technical University of Chemnitz, Chemnitz, Germany d Institute of Theoretical Physics, Freiberg University of Mining and Technology, Freiberg, Germany Received 15 September 2000; received in revised form 28 March 2001; accepted 2 April 2001

Abstract Diamond-like carbon (DLC) coatings of 0.7, 1.5 and 3.0 ␮m thickness were deposited on the titanium-alloy Ti6Al4V by a direct current discharge using benzene as gaseous precursor. Scanning electron microscopy, Raman spectroscopy, atomic force microscopy were utilised for gaining information about the micro-structural features and the composition. Hardness measurements, scratch and simple abrasion wear tests (ball-cratering apparatus with an abrasive fluid) were carried out for describing the tribological performance of the DLC. The fractures cross-sections of the DLC seems to be featureless and smooth. The universal hardness (HU) amounts to values of between approximately 12.8 and 22.2 GPa and the Young’s modulus of between nearly 133 and 213 GPa. Raman spectroscopy points microstructural changes in dependence of the coating thickness. Although, the cohesive failure of the DLC is dependent on the coating thickness, the DLC spallation at the lateral edges of the scratch occurs at approximately 35 N for all coatings. The worn surfaces exhibit a very heterogeneous DLC damage due to the abrasive loading which also points to the micro-structural differences within the DLC film. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Diamond-like carbon; Abrasive wear; Titanium-alloy; Biomaterial; Raman spectroscopy; Atomic force microscopy

1. Introduction Diamond-like carbon coatings are suitable materials for the surface modification of biomaterials like titanium-alloy Ti6Al4V (Fig. 1). DLC exhibits a bioinert behaviour which can be modified by the addition of appropriate elements [1,2]. The high chemical resistance and the low permeability make DLC coatings potential candidates as corrosion protection on metallic implants [3–5]. For example, these coatings are discussed as possibility for the prevention of the metal ion release which could effect inflammation of the surrounding organic tissue. Low friction coefficient in various environments coupled with high wear resistance are features of DLC which also obtain relevance for the surface finishing of implants like artificial joints [6]. Especially, the formation of wear particles needs to be prevented because they could enforce or release critical reactions of the organic tissue. Therefore, a surface modification of the implant by the deposition of optimised DLC coatings can be considered ∗ Corresponding author. Tel.: +49-3731-39-2203; fax: +49-3731-39-3662. E-mail address: [email protected] (A. Dorner).

as an useful finishing step. The following contribution discusses the influence of different DLC coating thickness on structural and tribological properties. 2. Experimental 2.1. DLC processing For deposition of the diamond-like carbon, a direct current discharge was used as described in detail elsewhere [7,8]. Before the coating deposition, the Ti6Al4V substrates were polished to a 1 ␮m finish with diamond paste and cleaned ultrasonically in ethanol. After an additional in situ cleaning by bombarding with argon ions, the generation of the amorphous carbon structure took place. The substrate is sufficiently negative biased in relation to the plasma in order to achieve ion bombardment during the DLC growth. Due to the direct current discharge under a pressure of between 0.005 and 0.05 Pa a gaseous precursor (methane) is decomposed and the built fragments were accelerated to the Ti6Al4V substrates. The deposition temperature on the substrates remains below 100◦ C due to a water cooling.

0043-1648/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 0 4 3 - 1 6 4 8 ( 0 1 ) 0 0 5 8 7 - 7

490

A. Dorner et al. / Wear 249 (2001) 489–497

whereas the indentation depth of the Vickers shaped diamond did not reach 1/10 of the coating thickness in the case of 1.5 and 3.0 ␮m thick DLC. 2.5. Atomic force microscopy (AFM) Information about surface topography was gained by AFM measurements using a microscope of the type Topo Metrix 2010, USA. Images with an area of 10 ␮m × 10 ␮m were recorded on each surface. 2.6. Scratch test Fig. 1. Uncoated and DLC-coated Ti6Al4V femoral heads.

Hydrogen from the decomposition of benzene is incorporated into the coating, and actually, is crucial for many properties of this material. It influences the DLC structure, passivates the dangling bonds in the amorphous structures and effects the internal stresses in the coating [3]. With the described deposition technique DLC coatings with a thickness of 0.7, 1.5 and 3.0 ␮m were generated. 2.2. Scanning electron microscopy (SEM) The SEM investigations were made with a DSM 960 apparatus of the company Carl Zeiss Oberkochen, Germany. The fractured surfaces of the DLC-coated Ti6Al4V were covered with a thin layer of carbon. SEM investigations took place with a voltage of 15 keV. By observing the coatings fractured cross-section information about the adherence of DLC can be gained. Also the appearance of the fractured DLC can be characterised and compared with the fracture morphology of other amorphous diamond-like carbon films. 2.3. Raman spectroscopy Raman spectroscopy is a commonly applied method for the description of structural features of DLC. Raman spectra were obtained using a Raman spectrometer of the type Jobin Yvon T 64000, France. The 488 nm line of an argon ion laser was focused to spots of about 2 ␮m in diameter. Hereby, the samples were exposed to a laser power of 2 mW. Three measurements were carried out on each DLC specimen.

The scratch test gives information about the coating adhesion, the damage behaviour of the coating and to a certain extent about residual stresses. A diamond cone with a Rockwell-C-profile (radius of the cone top: 200 ␮m, opening angle of the cone: 120◦ ) is loaded by a continuously increasing normal force. The diamond cone is moved with a constant speed over the surface of the sample. Due to the materials damage the emission of acoustic signals occurs and is recorded. In addition, the frictional force is evaluated. The investigations were carried out with the Revetest scratch tester of the company CSEM Instruments, Switzerland. The diamond cone was moved over the DLC-coated Ti6Al4V samples with a constant speed of 10 mm min−1 and the load was increased at a rate of 1.67 N s−1 . A maximum normal load of 50 N was applied. On every sample, at least five scratches were undertaken. Normal forces which effect materials damage are called critical normal loads LC . Different coating damages can be distinguished. Dependent on the kind of the damage, for example, cohesive defects or spallation, different critical loads (LC1 to LCx ) are defined. 2.7. Wear tests The tribological behaviour of the uncoated and DLC-coated Ti6Al4V was investigated by applying a ball-cratering apparatus (Figs. 2 and 3) of the company CSEM Instruments, Switzerland (Calowear tester). The abrasive wear of the

2.4. Hardness The measurement of the universal hardness HU was carried out according to DIN 50359-1 by using a Fischerscope H100, Germany. The determination of the hardness takes place under load. Hereby, elastic and plastic deformation contributes to the measured data. A final load of 10 mN was applied during 24 s in 25 steps. For 0.7 ␮m thick DLC the indentation depth was below 1/6 of the coating thickness,

Fig. 2. Schematic of the apparatus used for measuring the abrasive wear (CSEM Instruments, Switzerland).

A. Dorner et al. / Wear 249 (2001) 489–497

491

Fig. 3. Schematic of the wear process and the formation of wear crater with the diameter b.

samples is caused by an abrasive fluid which consists of distilled water and SiC particles with an average diameter of about 4 ␮m. This fluid is constantly supplied and dripped on the balls surface. Due to the rotation of the ball (diameter: 30 mm, sliding speed: 0.05 m s−1 ), the abrasive fluid forms a continuous film on it. The weight and the rotation of the ball transfer a normal and a tangential force on the SiC particles which are able to penetrate into the worn surface, scratching it and removing material. For every measurement the sliding distance amounts to 1.50 ± 0.06 m. The acting normal force is approximately 400 mN and can be determined by a loading cell beneath the sample (see Fig. 2). The geometry of the formed spherical wear crater is measured and used for the calculation of the wear coefficient k. On every sample at least five measurements were carried out. As proven in the scanning electron microscope, under the chosen test parameters the DLC coatings are not perforated and Eq. (1) can be utilised for calculating the wear coefficient k [9]: k=

π b4 32LdFN

Fig. 4. A 0.7 ␮m thick DLC on Ti6Al4V (SEM).

(1)

where k is the wear coefficient (mm2 N−1 ), d the diameter of the rotating ball (mm), b the diameter of the spherical wear crater (mm), L the wear way (mm), and FN is the normal force (N). 3. Results and discussion 3.1. Fracture morphology Figs. 4–6 exhibit the fractured cross-sections of the DLC-coated Ti6Al4V. As known from previous experiments, the DLC shows a rather featureless appearance without any remarkable fracture characteristics. From the qualitative optical observation, a good coating adherence can be noted.

Fig. 5. A 1.5 ␮m thick DLC on Ti6Al4V (SEM).

3.2. Raman spectroscopy Differences of the DLC structure in dependence on the coating thickness are recognisable by applying Raman

492

A. Dorner et al. / Wear 249 (2001) 489–497

amorphous structure of DLC [11]. In that case the quenching of the amorphous network is partly prevented and a short range order may be promoted. 3.3. Hardness

Fig. 6. A 3.0 ␮m thick DLC on Ti6Al4V (SEM).

spectroscopy. DLC coatings exhibit two typical Raman signals. The so-called D peak is located at about 1380 cm−1 , and the so-called G peak is positioned at approximately 1550 cm−1 . The intensity ratio ID /IG is growing with increasing DLC thickness. The 0.7 ␮m thick amorphous carbon films show an average intensity ratio of 1.08 (Fig. 7). The average ID /IG value of the 1.5 ␮m thick carbon coatings amounts to 1.22. Measurements on the 3.0 ␮m DLC gave an ID /IG ratio of 1.75. Jui et al. [10] correlated the increase of the ID /IG ratio with the growth of crystallites which appear as islands in the amorphous carbon matrix in size and number. Taking this explanation into account it could be possible that the amount of areas with short range order of sp2 - or sp3 -hybridized carbon is increasing with the coating thickness. A probable cause for that may be a change of the cooling conditions due to the growing DLC coating. A thicker amorphous carbon coating could influence the heat transfer and reduce the quenching effect which is believed to be one reason for the

The hardness of the DLC depends strongly on the content of sp3 -hybridized carbon and residual stresses [3,12,13] and amounts to values of between 5 and 50 GPa. Table 1 contains the results of the hardness measurements. As can be seen from Table 1, the measured hardness increases with the DLC thickness. With decreasing coating thickness the influence of the substrate increases because the loading of the Vickers indenter was kept constant. As mentioned before, the indentation depth of the indenter is below 1/6 of the coating thickness in the case of the 0.7 ␮m DLC and below 1/10 of the coating thickness in the case of 1.5 and 3.0 ␮m DLC. The hardness values of the 3.0 ␮m DLC are the most reliable data. For the 0.7 ␮m thick DLC a substrate influence on the measured hardness should be taken into account. Information about the elastic material behaviour of the thin coatings are gained by the determination of the Young’s modulus and the relative elastic recovery. In the case of 3.0 ␮m DLC on Ti6Al4V, the elastic recovery reaches over 67%. Exhibiting a Young’s modulus of about 200 GPa, the DLC is situated in the lower range of the known literature data [3,12]. The considerable elastic recovery and the relative low Young’s modulus point out a high hydrogen content of the coatings. The hydrogen occupies dangling bonds. Therefore, less bonds are available for the formation of the carbon–carbon network. Due to this phenomenon, the Young’s modulus which is dependent on the binding energy between the atoms decreases and the elastic recovery increases. 3.4. Atomic force microscopy

Fig. 7. Typical Raman spectrum of a 0.7 ␮m thick DLC coating on Ti6Al4V.

In general, DLC coatings which are deposited by the described generation technique replicate the substrate surface roughness very well. Therefore, it is important to place emphasis on the surface preparation of the substrate. However, defect formation within the DLC is a known phenomenon and depends on the substrate as well as on the deposition technique [7,14,15]. Geometrical anomalies include tiny craters called pinholes or small hills with different shapes and gradient angle. Fung et al. [14] report that the formation of hillocks strongly depends on the energy of the ions or fragments. Higher energies are more effective at suppressing hillock formation. Certain impurities can be removed and the coating surface is smoothed in dependence on the kinetic energy of the ions. Li et al. [16] also explain that a smoother surface corresponds to a higher surface atomic mobility of the DLC generating ions and fragments. Under these conditions voids, impurities and rough particles can be eliminated during the coating growth. Traces of scratches

A. Dorner et al. / Wear 249 (2001) 489–497

493

Table 1 Universal hardness under loada according to DIN 50359-1 Sample

HU (GPa)

HUplast (GPa)

Relative elastic recovery (%)

Young’s modulus (GPa)

DLC (0.7 ␮m) DLC (1.5 ␮m) DLC (3.0 ␮m)

12.8 ± 0.5 19.5 ± 0.8 22.2 ± 1.3

18.6 ± 1.9 39.3 ± 4.0 41.4 ± 2.6

57.12 ± 0.97 69.14 ± 1.20 67.37 ± 1.45

133 ± 12 193 ± 14 213 ± 10

a

Load of 10 mN applied in 24 s and 25 steps.

from previous metallographic preparation steps are still visible after polishing to a 1 ␮m finish (Fig. 8). The DLC coating grows homogeneous on the Ti6Al4V substrate and the surface topography still is reproduced until a sufficient coating thickness is reached. The scratched Ti6Al4V surface is still influencing the DLC surface topography at a coating thickness of 0.7 and 1.5 ␮m, but in the case of the 1.5 ␮m coating the scratch depth is reduced. Finally, DLC coatings with a thickness of 3.0 ␮m reveal no scratch pattern. As recognisable in Figs. 9–11, not only the scratch pattern but also other features of the surface morphology depend on the coating thickness. The DLC with a thickness of 0.7 ␮m exhibits a surface that shows a considerable quantity of hillocks with different sizes and shapes. There are big hillocks with an average diameter of between 300 and 700 nm and smaller ones of about 100 nm. Also the coating surface of 1.5 ␮m thick DLC reveals hillocks of these two groups, but the amount seems to be reduced. The appearance of DLC with a thickness of 3.0 ␮m is changed in comparison to the 0.7 and 1.5 ␮m coating. It is smoother and only some hillocks of a medium size of about 200 nm in diameter are remaining. One possible reason for the low hillock density on top of the thicker coatings may be the minor or completely lost influence of the Ti6Al4V substrate on the growing conditions of the DLC.

Fig. 8. AFM image of the uncoated Ti6Al4V.

Fig. 9. AFM image of the 0.7 ␮m DLC coating.

3.5. Scratch test The investigation of the DLC-coated titanium-alloys by the scratch test confirms a sufficient adherence of the amorphous carbon coatings on this substrate. There are two critical loads measured on all samples (Fig. 12). The closer investigation of the scratches in the SEM in combination with the acoustic signals during the scratching leads to more

Fig. 10. AFM image of the 1.5 ␮m DLC coating.

494

A. Dorner et al. / Wear 249 (2001) 489–497

Fig. 13. Scratch track within a 0.7 ␮m thick DLC coating (diamond load: 40 N; SEM). Fig. 11. AFM image of the 3.0 ␮m DLC coating.

detailed information about the material damages. The first critical load LC1 correlates with cohesive coating cracking and is dependent on the coating thickness. The further enhancement of the normal force leads at the second critical load LC2 to the spallation of the DLC at the lateral edges of the scratch. The critical load LC2 was measured on all coatings to be about 37 N and is independent on the DLC thickness. The spallation of the coating starts at the edges of the scratch followed by adhesive failure between DLC and Ti6Al4V at the scratch bottom. In dependence on the DLC thickness, differences are visible concerning the extend of the cohesive damage. The size of de-adhered areas becomes larger with increasing DLC thickness. Also, the crack-net, which is recognisable on the bottom of the scratch trace before the spallation, exhibits a larger mesh-size on thicker coatings. Figs. 13 and 14 show the geometrical differences of the crack configuration for scratches on the 0.7 ␮m and

Fig. 12. Results of the scratch test.

the 3.0 ␮m thick DLC coating. The failure behaviour at the end of the scratch track shows a dependence on the DLC thickness. The 0.7 ␮m thick coatings do not exhibit any kind of spallation before the scratch tip (Fig. 15). Whereas, the 1.5 ␮m thick DLC already reveals signs of adhesive failure to some extent. Finally, the 3.0 ␮m thick amorphous carbon spalls extensively before the scratch tip (Fig. 16). The coating is loaded by tension beneath the diamond cone and by compression before it during the scratch test [17]. In addition, DLC coatings already reveal high residual compressive stress of between 0.5 and 8.5 GPa due to the processing [3,18]. The ions, atoms, fragments and clusters which were built due to the plasma decomposition of the gaseous precursor receive a high energy during the plasma deposition which may amount to numbers of between 0.1 and 10 keV. The negative voltage of the Ti6Al4V substrates effects the movement and the acceleration of the charged and excited particles from the plasma on the samples surface. These sites on the surface are compressed in the dimension of several hundred atom diameters, rapidly heated up and changed in the short range order. Furthermore, there is a

Fig. 14. Scratch track within a 3.0 ␮m thick DLC coating (diamond load: 40 N; SEM).

A. Dorner et al. / Wear 249 (2001) 489–497

495

There is extensive research about the wear resistance and mechanisms of DLC-coated materials especially

concerning tribological tests under sliding wear conditions [20–22]. Under these tribological conditions, the combination of different wear mechanisms like adhesive, abrasive or fatigue loading is possible. The formation of a lubricating transfer layer with a dimension of some nanometre is one important reason for the outstanding tribological resistance of DLC during sliding wear conditions [23]. This lubricating surface layer is formed due to the graphitisation of certain surface areas caused by the energy input during the wear, for example, friction heat. However, the formation of a graphitised surface film is not probable under simple abrasive conditions. In that case, the material loss is caused by the penetration of the abrasive parts and their following tangential movement within the sample surface which both may cause elastic and/or plastic material response like deformation or cracks. Of considerable importance for the abrasive wear resistance of a material are its hardness and strength as well as its fracture toughness. The loosening of wear particles from the surface is reduced, if the material is able to locally deform without damage, hereby dissipating the impact of the energy input during the wear process. DLC coatings are characterised by a high hardness as well as Young’s modulus. Therefore, a good resistance against simple abrasive wear may be expected. Fig. 17 shows the quantitative results from the tribological wear tests of the DLC-coated and the uncoated Ti6Al4V. As visible, the wear resistance of the DLC-coated Ti6Al4V is increased by 8–10 times in comparison to the uncoated metal. Apparently, the different thickness of the DLC do not influence the wear resistance. Figs. 18 and 19 present worn surfaces of the uncoated Ti6Al4V substrate. There are spiky holes and craters visible which are caused by the attack of the SiC particles. The polished metallic surface extensively plastically deforms under the abrasive action of the hard particles. The hole formation and the plastic deformation of the metallic surface is relatively homogeneous. On the contrary, the DLC coatings exhibit a very heterogeneous materials damage due to the tribological loading. Fig. 20 reveals the DLC surface

Fig. 16. Scratch tip within a 3.0 ␮m thick DLC coating on Ti6Al4V (SEM).

Fig. 17. Resistance of DLC-coated and uncoated Ti6Al4V against abrasive wear (Calowear test, abrasive fluid: SiC particles).

Fig. 15. Scratch tip within a 0.7 ␮m thick DLC coating on Ti6Al4V (SEM).

extremely high cooling rate forced due to the small dimension of these sites which leads to the quenching and solidification of metastable structures [19]. This atomic hammering as well as the inclusion of additional elements into the network, which may considerably differ from the solubility of these elements under stable equilibrium conditions, effects the stated before compressive stresses within the DLC. The external and the residual stresses are added if the coating is loaded by the Rockwell diamond. Because of the described high residual compressive stresses, the stresses which are necessary to cause the failure of the material are reached already under relatively low loads and coating damage or spallation occur. As possible reason for the larger failures with the increased coating thickness, higher residual compressive stresses in the thicker coatings may be considered. 3.6. Wear behaviour

496

A. Dorner et al. / Wear 249 (2001) 489–497

Fig. 18. Worn surface of the uncoated Ti6Al4V (SEM).

Fig. 19. Plastic deformed worn surface with spiky holes of the Ti6Al4V (SEM).

before the wear test. It is again visible that the amorphous carbon coating is growing homogeneously and dense on the Ti6Al4V surface. Defects like “pinholes” [7] were not found. Figs. 21 and 22 present the DLC coating after the wear test. There is only local damage of the amorphous

Fig. 20. DLC surface after deposition (SEM).

Fig. 21. Worn surface of the DLC (SEM).

carbon occurring. Minor scratches and network-like agglomerations of holes with a maximum dimension of some micrometers are visible. On other locations the DLC surface still seems to be smooth and undamaged. The damage is only cohesive in nature. Under the chosen wear parameters no adhesive failure is observed. However, it is remarkable that damage of the DLC takes place heterogeneously. This phenomenon may be attributed to structural differences within the amorphous carbon. Within the DLC different hybridisations of the carbon exist, while the long range order is absent and only a short range order is established. Clusters of the different carbon hybridisations may be present within a completely amorphous DLC matrix as interconnecting or not connected islands. Harder DLC regions with a high fraction of sp3 -hybridised carbon are expected to be more resistant against the penetration of the abrasive SiC particles in comparison to surface regions with a high content of sp2 -hybridised carbon. Robertson [24] modelled the nano- and microstructure of diamond-like carbon as a random network of covalently bonded carbon atoms of different hybridisations with a degree of medium range order in the 1 nm scale. In addition, structural heterogeneity in

Fig. 22. Heterogeneous material damage within the DLC (SEM).

A. Dorner et al. / Wear 249 (2001) 489–497

the diamond-like coating also could be in the 100 nm or micrometer scale. For example, Maharizi et al. [25] proved the cluster-like nanostructure of DLC by measurements of nanohardness and using atomic force microscopy. They confirm two types of granulation with different nanohardness and different average height profile (AHP) inside cluster shaped structures with a size of some hundred nanometres. They suggest that the clusters with increased AHP and nanohardness have a higher sp3 content than those of low AHP. As another probable reason for the different sensitivity of the DLC surface against tribological loading also residual stresses are conceivable. These may be established, for example, due to different cooling rates which also could cause structural heterogeneity within the growing DLC. 4. Summary and conclusions Although, under the chosen processing parameters a DLC thickness of between 0.7 and 3.0 ␮m has an influence on the structure, surface morphology and cohesive damage, no effect on the adherence and resistance against abrasive wear were found. Raman spectroscopy points out a growth of crystallites in the amorphous carbon matrix which could affect the mechanical and tribological properties of the DLC. The surface becomes smoother and defects like hillocks vanish with increasing coating thickness. This may be of importance for the contact behaviour of the DLC with organic fluid or tissue. The DLC reveals a good adherence to the titanium-alloy and coating spallation at the scratch edges was detected at about 35 N during the scratch test for all coating thickness. The susceptibility against cohesive damage also is of outstanding interest for the biological acceptance of the DLC. Early cohesive failure, for example, during wear needs in any case to be prevented because of the possible formation of wear particles. Under the chosen conditions for proving the resistance against abrasive wear a heterogeneous damage of the DLC coating was found which may be caused by microstructural heterogeneity within the amorphous carbon. The results from the scratch tests mark the 0.7 ␮m thick DLC as the most suitable coating. Further influences, for example, the DLC structural diversity on living tissue need to be proven in appropriate future tests.

497

References [1] C. Du, X.W. Su, F.Z. Cui, X.D. Zhu, Biomaterials 19 (1998) 651–658. [2] A. Schroeder, G. Francz, A. Brunenk, R. Hauert, J. Mayer, E. Wintermantel, Biomaterials 21 (2000) 449–456. [3] A. Grill, Diam. Relat. Mater. 8 (1999) 428–434. [4] A. Olborska, M. Swider, R. Wolowiec, P. Niedzielski, A. Aylski, S. Mitura, Diam. Relat. Mater. 3 (11994) 899–901. [5] E. Mitura, S. Mitura, P. Niedzielski, Z. Has, R. Wolowiec, A. Jakubowski, J. Szmidt, A. Sokolowska, P. Louda, J. Masciniak, B. Koczy, Diam. Relat. Mater. 3 (1994) 896–898. [6] Schultrich, B.H.-J. Scheibe, in: Proceedings of the Workshop on Anwendung moderner Oberflächentechnologien in der Medizintechnik, Dresden, Förderverein “Dünne Schichten” e.V., 10 May 1996. [7] A. Dorner, B. Wielage, C. Schürer, Thin solid films 351 (1999) 1–5. [8] B. Wielage, A. Dorner, C. Schürer, H. Podlesak, Metalloberfläche 51 (1996) 163–167. [9] K.L. Rutherford, I.M. Hutchings, J. Test. Eval., JTEVA 25 (2) (1997) 250–260. [10] J.-T. Jiu, H. Wang, C.-B. Cao, H.-S. Zhu, J. Mater. Sci. 34 (1999) 5205–5209. [11] B. Bliznakovska, Ch. Meixner, Progress in diamond and diamond-like coatings processing, in: Forschungszentrum Jülich GmbH, Science Series of the International Bureau, Vol. 23, 1994. [12] J. Robertson, Surf. Coat. Technol. 50 (1992) 185–209. [13] S. Logothetidis, C. Charitidis, M. Gioti, in: G.-M. Chow, I.A. Ovid’ko, T. Tsakalakos (Eds.), Nanostructured Films and Coatings, NATO Science Series, Kluwer Academic Publishers, Dordrecht, 2000, pp. 297–308. [14] M.K. Fung, K.H. Lai, H.L. Lai, C.Y. Chan, N.B. Wong, I. Bello, C.S. Lee, S.T. Lee, Diam. Relat. Mater. 9 (2000) 815–818. [15] M. Collins, R.C. Barklie, J.V. Anguita, J.D. Carey, S.R.P. Silva, Diam. Relat. Mater. 9 (2000) 781–785. [16] D. Li, Y.W. Chung, M.S. Wong, W.D. Sproul, Tribol. Lett. 1 (1995) 87–90. [17] H. Jehn, Charakterisierung dünner Schichten, Hrsg.: Deutsches Institut für Normung e.V., DIN-Fachbereicht 39, Beuth Verlag GmbH, 1993, pp. 210–334. [18] J. Deng, M. Braun, Diam. Relat. Mater. 5 (3–5) (1996) 478–482. [19] B. Rother, J. Vetter, Plasmabeschichtungsverfahren und Hartstoffschichten, Dt. Verlag für Grundstoffindustrie GmbH, Leipzig, 1992, pp. 36–38. [20] Y. Lui, A. Endemir, E.I. Meletis, Surf. Coat. Technol. 82 (1/2) (1996) 48–56. [21] H. Ronkainen, S. Varjus, K. Holmberg, Wear 22 (2) (1998) 120–128. [22] E.-S. Yoon, H. Kong, K.R. Lee, Wear 217 (1998) 262–270. [23] A. Endemir, C. Bindal, J. Pagan, P. Wilbur, Surf. Coat. Technol. 36/37 (1995) 692–697. [24] J. Robertson, Adv. Phys. 35 (1996) 317–322. [25] M. Maharizi, O. Segal, E. Ben-Jacob, Y. Rosenwaks, T. Meoded, N. Croitoru, A. Seidman, Diam. Relat. Mater. 8 (1999) 1050–1056.