Wear resistance of a high-nitrogen austenitic stainless steel coated with nitrogenated amorphous carbon films

Wear resistance of a high-nitrogen austenitic stainless steel coated with nitrogenated amorphous carbon films

Surface and Coatings Technology 161 (2002) 224–231 Wear resistance of a high-nitrogen austenitic stainless steel coated with nitrogenated amorphous c...

1MB Sizes 0 Downloads 44 Views

Surface and Coatings Technology 161 (2002) 224–231

Wear resistance of a high-nitrogen austenitic stainless steel coated with nitrogenated amorphous carbon films A. Di Schinoa, L. Valentinia, J.M. Kennya,*, Y. Gerbigb, I. Ahmedb, H. Haefkeb a Materials Engineering Centre, University of Perugia, 05100 Terni, Italy CSEM Centre Suisse d’Electronique et de Microtechnique, Rue Jaquet Droz 1, CH-2007 Neuchatel, Switzerland

b

Received 8 April 2002; accepted in revised form 26 July 2002

Abstract In this paper, the dependence of wear resistance on grain size of nitrogen-alloyed austenitic stainless steel is investigated and compared to measurements for the same samples coated with nitrogenated amorphous carbon wa-C:H(N)x film, deposited by means of plasma-enhanced chemical vapour deposition. To this aim, ball-on-disk (BoD) tests were performed to investigate the wear stability of coated and uncoated substrates. Long-time test results on the uncoated material show that no improvement in the low-friction performance (LFP) duration was found with decreasing grain size of the substrate. On the other hand, an improvement in the low-friction performance duration with substrate microstructure refinement was found in the case of nitrogendoped overcoats. An effect of the film structure on the LFP duration was also detected in the case of the finest-grained samples, indicating a synergic effect between the ultrafine grain structure and the nitrogen-doped carbon coating. 䊚 2002 Elsevier Science B.V. All rights reserved. Keywords: High-nitrogen austenitic stainless steel; Grain size; Carbon thin films; Wear

1. Introduction Nitrogen-alloyed austenitic stainless steels exhibit attractive properties, such as high strength and ductility, good corrosion resistance and a reduced tendency of grain boundary sensitisation w1x. The high austenitic potential of nitrogen allows the nickel content in steel to be reduced, offering additional advantages, such as cost savings. The production of these low-nickel steels is made possible by the addition of manganese, which allows an increase in N solubility in the melt and decreases the tendency of Cr2N formation w2x. Although there have been many studies on finely grained ferritic steels (e.g. w3x), only a few research reports are available on refined austenitic stainless steels. The grain size of ferritic steels can easily be refined by phase transformation, but in austenitic alloys, due to the absence of phase transformation, the grain diameter is *Corresponding author. Tel.: q39-0744-492939; fax: q39-0744492925. E-mail address: [email protected] (J.M. Kenny).

usually controlled by recrystallisation after cold working w4x. This method is mainly affected by the working temperature, working ratio and recrystallisation temperature. Recrystallisation after hot rolling is reported to have the effect of grain refining w5x, but this method seems to be limited. In previous papers w6,7x, we examined the effect of subzero working on the grain refining of austenitic stainless steels. In particular, ultrafine-grained AISI 304 stainless steel with an average grain size of approximately 0.8 mm was obtained by applying the reverse transformation of martensite to austenite on subzeroworked steel annealed at low temperatures. Furthermore, a strong increase was found in both the mechanical w6x and localised corrosion w8x resistance. In order to increase the strength further, it is possible to combine the effect of nitrogen addition and grain refining. Due to the absence of high-temperature phase transformation and to the very low aptitude to martensitic transformation of this class of stainless steels w9x, the only way to promote grain refining seems to be by

0257-8972/02/$ - see front matter 䊚 2002 Elsevier Science B.V. All rights reserved. PII: S 0 2 5 7 - 8 9 7 2 Ž 0 2 . 0 0 5 5 7 - 1

A. Di Schino et al. / Surface and Coatings Technology 161 (2002) 224–231 Table 1 Chemical composition of the high-nitrogen austenitic stainless steel studied Element

Composition (mass %)

C Si S P Mn Cr Ni Mo Cu N

0.026 0.04 0.003 0.026 10.5 18.6 0.42 0.07 0.07 0.37

225

the N2 gas flow rate to total precursor gas flow rate, rswN2x y(wCH4 xqwN2 x), was fixed to 0% and 60%, while the total gas flow rate was kept constant at 28 sccm. The films were deposited onto HN substrates that were mounted on the water-cooled cathode of an RF (13.56 MHz) plasma-enhanced chemical vapour depo-

dynamic or static recrystallisation after cold rolling, promoted thereafter by elevated annealing temperatures. Due to the above-mentioned properties, many efforts w10x have recently been devoted to the production of high-nitrogen austenitic stainless steels. Recently, metallurgists have been very active in developing basic scientific research concerning this class of steels; in particular, great attention has been devoted to the solidification mechanisms w11x and recrystallisation processes w12x of these materials. However, to the best of our knowledge, no results have yet been reported concerning the effect of grain size on the wear resistance of highnitrogen steels. In this paper, the dependence of the wear resistance on the grain size (down to approx. 2.5 mm) in a nitrogen-alloyed austenitic stainless steel is investigated and compared to measurements for the same material coated with carbon thin film, deposited by means of plasma-enhanced chemical vapour deposition. Furthermore, the effects of nitrogen doping in carbon overcoats are also explored in terms of longer wear durability as a function of the substrate microstructure. 2. Experimental details The chemical composition of the high-nitrogen stainless steel (hereinafter HN) investigated in this paper is reported in Table 1. Samples in the cold-rolled state were annealed at 1100 8C for 20 s (sample A1), 3 min (sample A2) or 10 min (sample A3) to obtain the three different austenitic microstructures reported in Fig. 1. The sample surfaces were polished using increasingly finer abrasive papers down to a 1-mm grit. The dependence of yield strength and micro-hardness on the grain size of the steel were measured by means of tensile and Vickers tests, respectively. Results, reported in Fig. 2, show a strong effect on the mechanical properties when the grain size changes in the range from 2.5 (sample A1) to 20 mm (sample A3), in accordance with the Hall–Petch relationship w13x. Nitrogen-containing a-C:H film was produced by plasma processing of a CH4 –N2 mixture. The ratio of

Fig. 1. Microstructures of the high-nitrogen austenitic steel samples annealed at 1100 8C for: (A1) 20 s; (A2) 3 min; and (A3) 10 min. Grain sizes: 2.5, 10 and 20 mm.

226

A. Di Schino et al. / Surface and Coatings Technology 161 (2002) 224–231

sition (PECVD) system. Film deposition was carried out at room temperature, with the pressure fixed at 9 Pa, at a bias voltage fixed at y400 V with a power of 30 W supplied by the plasma. The thickness of the films deposited, measured by X-ray reflectivity, was approximately 0.5 mm. The X-ray reflectivity was performed using a Siemens D5000 X-ray diffractometer with a copper target (Cu-Ka X-ray source with ls 0.15604 nm). Quantitative analysis of the carbonaceous coatings obtained from r values ranging from 0% to 60% were carried out using a least-squares fit of the XRR patterns (Fig. 3) w14x. The friction coefficient was measured using a tribometer in a ball-on-disk (BoD) configuration, whereby a sapphire ball (⭋s6 mm) slides over the coated HN steel with a constant linear speed of 10 cm sy1 and under a load of 5 N (max. Hertzian contact pressure of 1.3 GPa for sapphire on steel). During the test, the temperature and relative humidity of the air atmosphere were kept constant at 21"1 8C and 50"5%, respectively. In order to investigate the low-friction performance (LFP) duration of the coated samples, long-time tests were carried out. The LFP duration is defined as the distance or time passed during the test, until the coefficient of friction (COF) between the sapphire ball and the sample reached the sapphireysteel COF value (f0.6). The adhesion behaviour was assessed by scratch tests. Scratch testing consists of introducing stresses at the interface between the coating and substrate by pressing a diamond stylus on the sample surface with a normal load. As the sample is displaced at a defined speed, the resulting stresses at the interface cause damage to the

Fig. 3. Chemical composition obtained by XRR investigation of aC:H and a-C:H:N thin films.

coating, e.g. flaking or chipping. The smallest load at which a specific failure event is recorded is called the critical load. In this case, scratch tests were performed with a Rockwell C-type diamond stylus (with a cone angle of 1208 and a tip radius of 200 mm) under a progressive normal load increasing from 0.5 to 30 N with a constant loading rate of 30 N miny1. The coated sample was displaced at a constant speed of 10 mm miny1. Measurements of hardness and elastic modulus were performed using a nanoindenter and a continuous stiffness measurement (CSM) technique. The resolution of the loading and displacement systems was 1 nN and 0.2 nm, respectively. With this technique, each indent gives hardness and elastic modulus as a continuous function of the indenter displacement into the samples. Multiple indentations were made at different locations on the film

Fig. 2. Yield strength and micro-hardness grain-size dependence in the HN steel.

A. Di Schino et al. / Surface and Coatings Technology 161 (2002) 224–231 Table 2 Duration of the low-friction performance (LFP duration) of the uncoated steel substrates Sample

LFP duration (s)

A1 A2 A3

3.9"1.1 3.8"1.4 5.4"2.1

surface with different loads. Loading was controlled such that the loading rate divided by the load was held constant at 0.05 sy1. Experiments were performed up to a depth of approximately 10% of the film thickness. For each indentation, the load vs. displacement curve was recorded, and the film hardness was determined from these curves according to the method proposed by Oliver and Pharr w15x. In addition, the permanent plastic deformation was determined at the end of each loading cycle. In order to investigate the degree of graphitisation in our carbonaceous coatings, Raman scattering spectra were recorded with a Jobin Yvon micro-Raman LabRam system in backscattering geometry. A 632.8-nm (1.96 eV) He–Ne laser was used as the light source and the power of the laser was adjusted by optical filters. Using a 100= objective lens, the illuminated spot on the sample surface was focused to approximately 2 mm in diameter. The resolution of the Raman spectra was better than 1 cmy1. Typical acquisition time for the spectra was 30 s. 3. Results and discussion The time dependence of the friction coefficients for the uncoated steel substrates is shown in Table 2. No clear changes in the duration of the stable friction, LFP, with the grain size are observed; in particular an increase in friction coefficient was detected after 3.9, 3.8 and 5.4 s for samples A1, A2 and A3, respectively. On the contrary, a similar effect was not detected in the AISI

227

304 steel, in which an improvement in the LFP duration with the grain size reduction was detected w16x. As shown in Table 2, HN steel samples can only perform low-level friction for a very short duration (10 s). After that, damage mechanism occurs due to adhesion and abrasion between the sapphire and the steel surface. To confirm these experimental findings, material transfer from the uncoated steel to sapphire caused by adhesion was clearly detected by optical observation (Fig. 4). The area of the wear tracks of samples after the BoD test as a function of the steel grain size is shown in Fig. 5. The results obtained show that the wear track area diminishes with the steel grain size, thus demonstrating the beneficial effect of grain refinement on the wear resistance. These results are in accordance with those reported in Fig. 2, where an increase in both the yield strength and hardness with grain refinement is shown. The LFP values for the long-time test performed on the coated substrates are reported in Fig. 6. From the results shown in Fig. 6, no effects of the grain size are evident in the case of the steel coated with a-C:H film. On the other hand, clear evidence of the influence of the substrate microstructure is observed if nitrogendoped carbon films are considered. In particular, the finer the microstructure, the longer the LFP duration, thus underlining the beneficial effect of grain size refining on the wear resistance of coated high-nitrogen austenitic stainless steel. Moreover, a significant increase in the LFP duration was observed for the finest-grained substrate when nitrogen-doped amorphous carbon films were used as protective coatings. On the other hand, no effects of the film composition on the LFP duration were detected for larger grain structures, showing a synergic effect between the fine grain structure and the nitrogen-doped carbon coating. The same effect was detected in the standard AISI 304 stainless steel as shown in Fig. 7 w16x. A possible explanation of the positive effects of nitrogen is based on recent experimental findings w17x,

Fig. 4. Optical micrograph of a sapphire ball after the BoD test on the substrate steel with the finest grains (sample A1).

228

A. Di Schino et al. / Surface and Coatings Technology 161 (2002) 224–231

Fig. 5. Wear track area as a function of grain sizes for the uncoated steel.

showing how nitrogen doping enhances graphitisation of the amorphous carbon matrix. To understand the microstructural evolution due to nitrogen doping, Raman spectra from films grown with 0–60% fraction of N2 in the plasma atmosphere are shown in Fig. 8. The spectra consists of two features, a G peak at approximately 1550 cmy1 due to optical zone centre vibrations (E2g mode) of the aromatic rings in graphite, and a D peak or shoulder at approximately 1360 cmy1 due to the disorder-activated optical zone edge modes of microcrystalline graphitic sheets w18x. When the nitrogen content in the film increases, the position of the G band shifts to higher frequencies, accompanied by a slight reduction in width, while the

intensity of the D peak increases. According to previous results w19x, this dependence on the plasma composition suggests stronger graphitic character of films deposited from a N2-rich plasma with respect to films obtained from a pure CH4 plasma. The decrease in line width (observed for the G band) as the nitrogen fraction is increased is consistent with the removal of bond-angle disorder and the increasing dominance of crystallites w20x. Fig. 9 shows the results of micro-Raman spectroscopy recorded inside and outside the wear tracks of the films after the ball-on-disk tests. When the nitrogen content in the film increases (Fig. 9), the position of the G band shifts to higher frequencies, accompanied by a slight

Fig. 6. Coefficient of friction (LFP) for the long time test performed on the coated HN steel substrates.

A. Di Schino et al. / Surface and Coatings Technology 161 (2002) 224–231

reduction in width, while the intensity of the D peak increases. Thus, the experimental findings show that the increase in G band position is higher for films obtained at a fixed N2 fraction and exposed to the ball-on-disk test. Moreover, lower width of the G band and an increase in the ID yIG ratio immediately after tests was observed for all samples. These experimental findings indicate that, in our case, the use of a-C:H:N film produces a graphitic transfer layer on the counter body, inducing low friction in the system, although the real carbon coating has already been partially or totally destroyed. The changes in the different wear behaviour of the steel substrate observed as a function of the overcoat composition are certainly a result of micro-structural changes in the films. In the case of nitrogen incorporation, the increase observed in the LFP duration for the coated sample with the finest microstructure is a consequence of changes in both chemical bonds and hybridisation states w17x. In particular, the effect of nitrogen incorporation on the Young’s modulus and film hardness can be observed in Fig. 10. Progressive nitrogen incorporation causes the Young’s modulus and film hardness to sharply decrease to less than 90 and 8 GPa, respectively. In the scratch tests the same failure event was detected for all tested films. Beginning from an applied load value of between 10 and 20 N, depending on the sample, the formation of Hertzian-type circular cracks was observed. With further increase in the load, only a growing network of cracks occurred, but no spallation or delamination of the films appeared (Fig. 11) in the load range tested. Concerning the critical load values reported in Table 3, the differences between both test

229

Fig. 8. Comparison of Raman spectra as a function of N2 fraction (0 and 60%) in the feed gases.

series of the films are negligible for N2 ratio in the gas mixture lower than 60%. However, for N2 ratio of 60% the film shows two-fold higher values for the critical load. As discussed in previous works w21,22x, structural changes that result in an increase in network connectivity, such as carbon sp3 atoms, contribute to network rigidity and higher internal stresses. So, for films grown without nitrogen addition, the co-actions of the high internal and external stresses (induced by the indentery ball) caused cracking of the film in the sp3 carbon bonding area. Cracking of the a-C:H coating before the

Fig. 7. Coefficient of friction (LFP) for the long time test performed on the coated AISI 304 steel substrates.

230

A. Di Schino et al. / Surface and Coatings Technology 161 (2002) 224–231

Fig. 10. Hardness and Young’s modulus of a-C:H and a-C:H:N thin films.

Fig. 11. Sequence of photographs of the scratch test on the film deposited with different nitrogen fraction in the feed gases. Table 3 Critical load of the carbon films as a function of nitrogen fraction (average values on five measurements on each sample

Fig. 9. G peak position, G band width and IDyIG ratio for a-C:H:N films as a function of the N2 fraction in the feed gases before and after wear testing.

occurrence of any spallation is typical for brittle coatings and indicates (Fig. 11) that insufficient film cohesion due to the high internal stress limits the coating behaviour to a greater extent than insufficient adhesion between the film and substrate. The increase in the carbon sp2 fraction, together with the network termination introduced by nitrogen-containing groups previously observed in other works w23x, contributes to a reduction in the bond angle and bond length distortion, thus reducing the internal stress. As discussed above, cracking of the coating is indicative of

N2y(N2qCH4) (%)

Critical load (mN)

0 25 50 60

15.0 13.6 16.0 25.0

poor film cohesion. Thus, N–sp2 C in films deposited with nitrogen could relax the cross-linked structure and contribute greatly to the material elastic properties, increasing the tribological properties. 4. Conclusions The effects of both the substrate microstructure and coating composition on the tribological properties of a high-nitrogen austenitic stainless steel have been investigated. Long-time test results on the uncoated material

A. Di Schino et al. / Surface and Coatings Technology 161 (2002) 224–231

have shown no improvement in the low-friction performance (LFP) duration with decreasing substrate grain size. On the other hand, an improvement in the lowfriction performance duration with substrate microstructure refinement was found in the case of nitrogen-doped overcoats. A strong effect of film structural changes on the LFP duration was also detected in the case of the finest-grained sample, indicating a synergic effect between the ultrafine grain structure and the nitrogendoped carbon coating. References w1x A. Rechsteiner, M. Speidel, Proceedings of the 1st European Stainless Steel Conference, vol. 2, 1997, p. 107. w2x A. Di Schino, J.M. Kenny, M.G. Mecozzi, M. Barteri, J. Mater. Sci. 35 (2000) 4803. w3x K. Kiyoshige, S. Pyusuke, Jpn. Inst. Met. 44 (1980) 1037. w4x R.E. Reed-Hill, Physical Metallurgy Principles, PWS Publishing Company, 1994. w5x I. Salvatori, T. Hayashy, K. Nagai, Proceedings of the International Workshop on Innovative Structural Materials, Tsukuba, 2000, p. 24. w6x A. Di Schino, J.M. Kenny, I. Salvatori, J. Mater. Sci., in press. w7x A. Di Schino, J.M. Kenny, M. Barteri, J. Mater. Sci. Lett. 21 (2002) 753.

231

w8x A. Di Schino, J.M. Kenny, J. Mater. Sci. Lett., in press. w9x M. Speidel, Proceedings of the 5th International Conference on High-Nitrogen Steels, Stockholm, 1998, p. 241. w10x A. Bailitski, M. Speidel, Proceedings of the 4th International Conference on High-Nitrogen Steel, 1998, p. 401. w11x A. Di Schino, M.G. Mecozzi, M. Barteri, J.M. Kenny, J. Mater. Sci. 35 (2000) 375. w12x A. Di Schino, J.M. Kenny, G. Abbruzzese, I. Salvatori, J. Mater. Sci. 36 (2001) 593. w13x J.N. Petch, J. Iron Steel Inst. (1953) 174. w14x S. Santucci, L. Valentini, A. Menelle, L. Lozzi, J.M. Kenny, Diamond Relat. Mater. 11 (2002) 1188. w15x W.C. Oliver, G.M. Pharr, J. Mater. Res. 10 (1992) 1564. w16x L. Valentini, A. Di Schino, Y. Gerbig, H. Haefke, J.M. Kenny, Wear, in press. w17x L. Valentini, S. La Rosa, L. Lozzi, J.M. Kenny, S. Santucci, Thin Solid Films, in press. w18x W.C. Chan, B. Zhou, Y.W. Chung, C.S. Lee, S.T. Lee, J. Vac. Sci. Technol. A 16 (1998) 1907. w19x A.C. Ferrari, J. Robertson, Phys. Rev. B 61 (2000) 14095. w20x N. Hellegren, M.P. Johansson, E. Broitman, L. Hultman, J.E. Sundgren, Phys. Rev. B 59 (1999) 5162. w21x J. Robertson, Phys. Rev. Lett. 68 (1999) 220. w22x D.F. Franceschini, C.A. Achete, F.L. Freire, Appl. Phys. Lett. 60 (1992) 3229. w23x S.E. Rodil, A.C. Ferrari, J. Robertson, W.I. Milne, J. Appl. Phys. 89 (2001) 5425.