Fresenius J Anal Chem (1997) 358 : 278–280 – © Springer-Verlag 1997

T. Niebuhr · H. Bubert · H.-D. Steffens · D. Haumann · K. Kauder · U. Dämgen

Examination of wear mechanisms of hard coatings Received: 25 July 1996 / Accepted: 17 November 1996 Abstract PVD-CrN and -TiN coatings are abraded in a wear test equipment which simulates the sliding-rolling friction of a screw rotor used in screw compressors. The coatings show significant differences concerning their durabilities. The worn-out coatings as well as the debris particles produced by abrasion are analyzed by means of X-ray induced photoelectron spectroscopy and Auger electron spectroscopy. The failure mechanism of TiN coatings can be explained by spalling off small debris particles, whereas the failure mechanism of CrN coatings has to be attributed to tribooxidation, because in this case the analyzed particles exclusively consist of chromium oxide. Nevertheless, CrN coatings show the lowest wear rate. When changing the environment (inert atmosphere or water lubrication) in the test equipment, the tribooxidation of CrN can be reduced or totally stopped, but a decrease in wear rate cannot be achieved. In this case tribooxidation leads to a better wear resistance.

1 Introduction The demand for environmentally friendly production, better quality, and low production costs for tribologically stressed components is increasingly fulfilled by the use of composites or coatings. Low-cost and easily machinable materials can often be applied if coated with friction-reducing and wear-resistant materials. Such coatings can also be used for gaining wear resistance of geometrically complex components such as screw rotors. With regard to wear resistance good results have often been achieved by a trial and error procedure concerning material and coating parameters. This is due to the lack of general understanding of the wear process [1]. By means of surface analysis, information on the tibological process can be obtained from the wear-induced modification of the surface-near layers and from the debris particles formed during wear. From this information more insight is gained into tribochemical reactions taking place and into wear mechanisms prevailing at the contact points. Because of the unsufficiently known tribological behaviour of screw rotors a test equipment has been set up simulating the sliding-rolling motion in order to test PVD-CrN and PVD-TiN coatings. The original and the stressed surfaces as well as the

T. Niebuhr · H. Bubert (Y) Institut für Spektrochemie und Angewandte Spektroskopie (ISAS), Bunsen-Kirchhoff-Strasse 11, D-44139 Dortmund, Germany

collected wear particles have been analyzed chemically using X-ray induced photoelectron spectroscopy (XPS) and Auger electron spectroscopy (AES).

2 Experimental The nitride coatings were deposited onto nitriding steel (standard no. 1.8591, 31 CrMoV 9) by means of Arc-PVD-technology. The thicknesses of these coatings were between 6 to 10 µm. The coatings are stressed in a test equipment consisting of two cylindrical rolls, which simulates the material load of the screw rotors (Fig. 1). One roll rotates with a circumferential speed of 30 m/s, while the other runs with 27 m/s. The rolls have a line contact length of 12 mm and are pressed against each other with a force of 24 N. The test equipment can also be encapsulated in order to run under water lubrication or protective gas. In this study one stage of a Batelle-type impactor [2] is used for active sampling of particles which were abraded from the coatings under dry conditions. When applying lubrication, a part of the lubricant water is dropped on a silicon wafer, dried under red light and the debris particles remaining on the wafer surface are analyzed. For depth profiling small pieces of about 5 × 5 mm2 are cut from the stressed roll surfaces, rinsed with water and ethanol and also dried under red light. Depth analyses of the debris particles and the worn-out surfaces were performed using an Auger electron spectrometer S 820 and a photoelectron spectrometer AXIS HS (Kratos, Manchester, UK). The Auger measurements were obtained at a primary electron energy of 3 keV, a beam current of 10 nA and an angle of incidence of 30° with respect to the surface normal. XPS analyses have been carried out with A1 Kα radiation, the X-ray tube was operated under 10 mA and 15 kV. At both instruments sputtering was carried out with 2 keV Ar+ ions rastered over 1 × 1 mm2 with an angle of incidence of 45° with respect to the surface normal. Under these conditions sputtering rates of 4 nm/min for AES and 1 nm/min for XPS were obtained in standard Ta2O5 layers.

3 Results and discussion 3.1 Wear under atmospheric conditions PVD-coatings normally have smooth surfaces which are only disturbed by a rough substrate or droplets from the coating process itself. Therefore, these coatings are most suitable for sliding conditions. Depth profiling of the coatings – as manufactured and stored in laboratory atmosphere – shows oxidized surfaces of less than the inelastic mean free path of the electrons of approximately 5 nm (after Seah et al. [3]). Wear tests under atmospherical conditions have been carried out with both coatings. The most wear-resistant coating is CrN with a wear rate of 0.5 µm/h while TiN is already worn out with 5 µm/h.

H.-D. Steffens · D. Haumann Lehrstuhl für Werkstofftechnologie, Fakultät Maschinenbau, Universität Dortmund, Otto-Hahn-Strasse 6, D-44227 Dortmund, Germany K. Kauder · U. Dämgen Fachgebiet Fluidenergiemaschinen, Fakultät Maschinenbau, Universität Dortmund, Otto-Hahn-Strasse 18, D-44227 Dortmund, Germany

Fig. 1 Principal design of the test equipment

279

a

Fig. 4 Auger spectra taken from the debris particles of the CrN coating tested under atmosphere, (a) original surface, (b) after sputtering for 2 min

b shiny. The collected particles have a diameter of at least 1.5 µm due to the cut-off of the impactor stage. Smaller particles than the cut-off are only collected as conglomerates. The largest debris particles found are somewhat larger than 20 µm which are either conglomerates or flat particles because of the layer thicknesses of 6 to 10 µm. In Fig. 2 b a SEM image of the debris particles from the CrN coating is given. This type of flat particles are only found here, while abrasive particles from TiN are smaller and more irregular. Auger depth profiles of the debris particles of TiN coatings show that TiN is worn out unchanged in stoichiometry. In contrast to this finding, CrN coatings oxidize during wear. Cr LMM and O KLL spectra obtained from the original surface of the particle and after 2 min sputtering are given in Fig. 4. No nitrogen is observed in any sputter cycle, which means that the whole particles are oxidized to chromium oxide. Fig. 2 a, b SEM images of (a) worn-out surface of the CrN coating and (b) an abraded particle 3.2 Wear under inert gas Having concluded that tribooxidation is the predominant wear process of CrN coatings in atmosphere, additional experiments were carried out under inert gas protection. For this, the test equipment was encapsulated and flushed with argon gas with a flow rate of 10 L/min. Auger depth profiles of the collected particles show that the oxidation cannot completely be stopped because of insufficient sealing of the capsule, so that the nitride particles found are partly oxidized. The oxidation could only be decreased but not avoided. The wear rate amounts to 3 µm/h, which is significantly higher than under atmosphere. 3.3 Wear under water lubrication Fig. 3 XPS depth profile of the worn-out CrN coating In Fig. 2 a, a SEM image of a worn-out CrN surface is given, which shows long grooves as a result of the sliding motion. The roughnesses of such surfaces are 0.2 µm for CrN and 0.5 µm for TiN. In Fig. 3 an XPS depth profile from a CrN coating is given. On the surface-near layers, carbon and oxygen are detected, which belong as well to hydrocarbon contamination as to chromium oxide. Nonetheless, nitride is detectable on the unsputtered surface, the thickness of the oxide layer must be of the same order as for the coatings prepared and stored in the laboratoy atmosphere. TiN shows similar results. The debris particles were collected from the coatings by means of the impactor after running-in until the surfaces are

The encapsulated test equipment can also be used for testing the durability of the coatings under water lubrication. Only the CrN coating was tested to get more insight into its wear process. Figure 5 shows the Auger lines of Cr LMM, O KLL and N KLL of the unsputtered surface of a debris particle (curve a) and after removal of an about 4 nm thick layer (curve b), which has mainly been atrributed to contamination. This finding clearly demonstrates that tribooxidation does not take place under water lubrication. The wear rate amounts to 5 µm/h, which again is also significantly higher than under atmosphere.

280

Fig. 5 Auger spectra of two regions taken from the debris particle of the CrN coating tested under water lubrication, (a) original surface, (b) after sputtering for 1 min

off. The interface of tempered CrN with a surface layer consisting of Cr2O3 showed no intermediate compounds like CrOxNy, which should improve the adhesion between CrN and Cr2O3[4]. In contrast to our measurements, Chiba et al. [5] measured the surface temperature by means of an optical pyrometer in a test equipment which also simulates sliding conditions. They observed a mean temperature of less than 100° C during the test and conclude that oxidation does not play an important role. But they did not analyze the wear particles. Ye et al. [6] analyzed the debris particles worn out from a (Ti, Nb) N coating after sliding motion. They observed oxidation of the TiN coating by using TEM. Because of our examinations we conclude that the temperatures of the hot spots produced under our test conditions should exceed 100° C, by far. When changing the experimental conditions by applying water lubrication or running under inert gas atmosphere, tribooxidation is reduced or completely stopped. Unexpectedly, this leads to a higher wear rate. The tribooxidation of CrN seems to be a process which influences the wear rate positively. A possible explanation for these findings is that the heat generated at prominent spots on the surface (e.g. droplets) leads to oxidation. These oxidized parts have no adhesion and flake off, so that the number of the hot spots is diminished. Thus, the tribooxidation helps to produce a smoother surface than the original one and to reduce the wear rate due to the decreased thermal loading. Another possible interpretation is that the generated oxidized surface has gained lubricating properties, which protect the surface against the sliding motion. It is known for example that the oxidized surface of TiN acts as a solid lubricant [7]. In general, the only critical aspect of the PVD technique is the low thickness of the coatings which is in the range of the manufacturing tolerances. First results of our actual experiments indicate that thicker CrN coatings (30 µm) deposited as CrN multilayer with different stoichiometries, can be achieved, which have lower wear rates.

References 4 Discussion In the tests CrN is found to be superior to TiN with respect to durability (wear rate in atmosphere 0.5 µm/h and 5 µm/h, resp.). The CrN wear process can be described as tribooxidation because of the formation of fully oxidized debris particles. TiN is worn out chemically unchanged. Oxidative wear occurs, when a sliding surface is oxidized in atmosphere by frictional heating of contacting parts. The oxidized parts on the surface have insufficient adhesion and flake

1. Holmberg K, Mathews A (1994) Thin Solid Films 253 : 173 2. Niebuhr T, Bubert H, Dämgen U, Kauder K (1996) Mikrochim Acta (in press) 3. Seah MP, Dench WA (1979) Surf Interf Anal 1 : 2 4. Bubert H, Niebuhr T (1996) JMSA, J Microsc Soc Amer 2 : x 5. Chiba Y, Omura T, Ichimura H (1993) J Mater Res 8 : 1109 6. Ye X, Celis JP, De Bonte M, Roos JR (1994) Wear 171 : 91 7. Schneider T, Meier zu Köcker G, Santner E (1996) Surf Interf Anal 24 : 7