Sliding wear behaviour of Ni–P–W composition-modulated coatings

Sliding wear behaviour of Ni–P–W composition-modulated coatings

Surface and Coatings Technology 105 (1998) 224–231 Sliding wear behaviour of Ni–P–W composition-modulated coatings V.D. Papachristos a,*, C.N. Panago...

953KB Sizes 0 Downloads 26 Views

Surface and Coatings Technology 105 (1998) 224–231

Sliding wear behaviour of Ni–P–W composition-modulated coatings V.D. Papachristos a,*, C.N. Panagopoulos a, P. Leisner b,c, M.B. Olsen b, U. Wahlstrom c a Laboratory of Physical Metallurgy, National Technical University of Athens, 157 80 Zografos, Athens, Greece b Technical University of Denmark, 2800 Lyngby, Denmark c Industrial Microelectronics Center, 581 83 Linkoping, Sweden Received 21 July 1997; accepted 27 January 1998

Abstract We studied the unlubricated sliding wear behaviour of Ni–P–W compositionally modulated coatings on mild steel pins, sliding against hardened steel discs in a pin-on-disc set-up. The multilayered coatings consisted of high and low W-content layers. The linear velocity was constant for all tests, and the normal loads used were 200, 500, 800 and 1100 g. Coatings of various layer periodicities were studied. Friction coefficients were recorded during the tests, and it was found that after a running-in period of about 30 m, friction coefficients reached a steady state. An increase in the normal load caused a decrease in friction coefficients regardless of the layer periodicity of the coating. Coatings with a smaller layer periodicity showed a better wear resistance than those with larger layer periodicities. Two major wear mechanisms were identified: brittle fracture of the individual layers accompanied by crack deflection at the interfaces between the layers and delamination of parts of the coating. The first mechanism was active for the whole range of loads, whereas the second was operative only for loads of 800 and 1100 g. © 1998 Elsevier Science S.A. Keywords: Composition-modulated coatings; Ni–P–W; Sliding wear behaviour

1. Introduction During the last 20 years, compositionally modulated materials (or multilayers) have attracted a lot of attention. The reason is that new or improved techniques have been developed, which, in most cases, offer excellent control of the structure, composition and layer thickness of the multilayered material. The unique control of microstructure has led to the production of materials with enhanced or novel behaviour concerning their magnetic [1], electrical [2,3], optical [4–6 ] and mechanical [7,8] properties. These multilayered materials are usually deposited as coatings. Two of the major methods for the production of multilayered coatings are: (1) deposition in vacuum, and (2) electrodeposition from an aqueous solution. Although vacuum techniques are most often used for the production of multilayered coatings, electrodeposition (either a dual bath or a single bath technique with the aid of pulsed current) can offer a less expensive and a more industrialised production method with certain * Corresponding author. 0257-8972/98/$19.00 © 1998 Elsevier Science S.A. All rights reserved. PII S 02 5 7 -8 9 7 2 ( 9 8 ) 0 0 45 9 - 9

advantages over the vacuum deposition methods in some applications. Since multilayered materials are mainly used as coatings and one of the most important applications of coatings is antiwear protection, it is of great interest to examine the behaviour of these materials in this regard. The enhanced mechanical properties of multilayers (for example, high hardness and strength for certain layer periodicities) are expected to lead to improved tribological properties, that is increased resistance to wear. The majority of studies dealing with the behaviour of multilayers in wear are concerned with ceramic or ceramic/metallic multilayers produced by vacuum techniques. Vancoille et al. [9] studied the wear and friction of PVD TiC/Ti(C,N )/TiN multilayer in a ball-on-disc test and found that wear rates in the multilayer were three to four times smaller compared to the reference TiN. Minevich [10] investigated the tribological properties of multilayered coatings consisting of TiC and TiN layers deposited by cathodic arc plasma in longitudinal turning of stainless steel. Zhang et al. [11] studied the wear behaviour of CVD TiN–TiC multilayers in a ballon-disc test with better results for the multilayer coating compared to the monolayers.

V.D. Papachristos et al. / Surface and Coatings Technology 105 (1998) 224–231

In the field of metallic multilayers, the number of publications is much smaller. Zhang et al. [12] used ion beam mixing (IBM ) to produce Ni–Mo multilayers. The wear behaviour of the multilayers in unlubricated sliding was examined and compared to the wear behaviour of monolayers. Wear rates were found to be lower for the multilayers, and the friction coefficient was more stable during the tests. The wear mechanism of the multilayers was identified to be fatigue and delamination. Ruff and Myshkin [13] investigated the lubricated sliding wear behaviour of composition modulated Ni–Cu coatings on steel prepared by electrodeposition. These coatings were found to have improved wear resistance sliding against AISI type 52100 steel, when compared with the wear of the two pure metals nickel and copper under the same conditions. Ruff and Wang [14] studied the unlubricated sliding wear behaviour of Ni–Cu multilayered coatings on steel prepared by electrodeposition. They used a crossed cylinder geometry, and the counter material was AISI type 52100 steel. The composition-modulated coatings showed less wear than coatings of pure copper and nickel. The wear was found to be dependent on the layer spacing for the composition modulated coatings. This paper presents the results of a study of the unlubricated sliding wear behaviour of a composition modulated coating consisting of alternate layers of ternary Ni–P–W alloys with different phosphorus and tungsten contents, produced by electrodeposition from a single bath using pulsed current. Measurements of the wear volume and friction coefficient were made, and the wear scar and wear debris were analysed using optical, electron microscopy and EDAX techniques to reveal the wear mechanism.

2. Experimental The Ni–P–W multilayered coatings were deposited on mild steel cylindrical pins with a hemispherical tip of 5 mm radius. Prior to the deposition, the pins were cleaned for 3 min in an anodic degreasing bath, rinsed with water and activated for 1 min in a bath containing 0.1 M NaHF and 0.1 M NaHSO . The plating solution 2 4 had the following composition: NiSO .6H O: 17 g l−1 4 2 Na WO .2H O: 66 g l−1 2 3 2 H PO : 25 g l−1 3 3 H Cit.H O: 63 g l−1 3 2 pH: 5.5 temperature: 70 °C. From this solution, alternate Ni–P–W layers of low and high tungsten content were deposited. The average composition of the layers was Ni–10%P–10%W and Ni–5%P–45%W in weight for the low and high tungsten layers, respectively, as measured by EDAX on cross-

225

sections of the coatings. The low tungsten content layer was deposited during the low current density step of the pulse, which was 20 mA cm−2, whereas the high tungsten content layer was deposited during the high current density step of the pulse, which was 200 mA cm−2. The thickness of the alternate layers comprising the multilayer was always equal and the layer periodicities (bilayer thickness) of the multilayers studied were 6, 20, 60, 200, 600 and 2000 nm. The total thickness of the coating was approximately 25 mm. The top layer was always the layer with the high tungsten content. Wear testing of the coatings was conducted using a pin-on-disc apparatus. The coating was deposited on the standstill pin, which was sliding against the rotating disc of the counter material. The counter material was a CALMAX hardened steel provided by Uddeholm. The hardness of the steel disc after hardening was 57.5 HRC. All discs were polished to an average roughness of 0.05 mm using SiC polishing paper. The average roughness of the coatings was always less than 0.08 mm. Both pins and discs were cleaned immediately before testing with ethanol followed by acetone in an ultrasonic bath, to remove traces of abrasive particles and surface contaminants introduced by polishing or handling. The applied loads were 200, 500, 800 and 1100 g. The sliding speed was 14 cm s−1, and this was kept constant for all tests by adjusting the diameter of the wear track and the rotational speed of the disc. Tests were performed in air at 25 °C with a relative humidity of about 35–45%. Preliminary tests were carried out to ensure that there was no penetration of the coating for the selected sliding distances, which ranged from 150 to 350 m with the smaller values for the higher loads. The frictional force was recorded continuously during the test using a transducer linked to a personal computer. In this way, the friction coefficient was continuously monitored during the test. The wear volume was calculated by the diameter of the wear scar on the pin (average of two or three diameters) according to ASTM G99-95 instructions. The diameter of the wear scar was measured using an image analysis programme. Two tests were carried out for each load condition. The wear scar on the pin was examined using an optical and scanning electron microscope equipped with EDAX before and after the removal of loose debris particles from the surface by acetone in an ultrasonic bath. The wear debris was examined using a scanning electron microscope and EDAX.

3. Results and discussion Cross-sections of the multilayered coatings that had been produced under identical conditions and from the same solution were examined in the transmission electron microscope ( TEM ). In Fig. 1, the cross-section

226

V.D. Papachristos et al. / Surface and Coatings Technology 105 (1998) 224–231

equation [15]: (1)

F=As,

Fig. 1. TEM electron micrograph of the cross-section of a multilayered coating. The dark layers are those with a high W content, and the bright layers are those with a low W content.

where F is the frictional force, A is the true area of contact, and s is the shear strength at the contact. In Fig. 3, the friction coefficients for the different layer periodicities versus normal load are shown. It can be seen that the values of the friction coefficients decrease from an average of about 0.9 for the load of 200 g to an average of about 0.7 for the load of 1100 g. The decrease in coefficients of friction with the increase in normal load can be explained taking into account the theory of Hertz for an elastic contact between two bodies. A contact between metals is most commonly a plastic one, but under certain conditions, the contact between two metals can be elastic [16 ]. In this study, it can be suggested that the contact between the multilayered pin and the hardened steel disc is predominantly elastic since both bodies have a very low roughness (<0.08 and 0.05 mm for the pins and the discs, respectively), the normal loads used have low to medium values, and the coating is a brittle amorphous or nanocrystalline material possessing no ductility as the authors of the paper have observed. Therefore, the Hertzian theory for an elastic contact can be applied. In this case, the contact area between the pin and the disc is given by the following equation [16 ]:

A B

3wr 2/3

A=p of a multilayered coating is shown. The layer thickness is about 4 nm for this coating. The dark layers are the W-rich layers, whereas the light layers are the W-poor layers. Generally, the layers were regularly spaced and uniform in thickness. The interfaces were relatively sharp, and both layers were amorphous or nanocrystalline with a grain size smaller than the minimum layer thickness (3 nm). In Fig. 2, recorded values of the friction coefficients versus sliding distance for three different tests are shown. The shape of the curves shows the typical behaviour of the friction coefficients. As can be seen, there is a running-in period for about the first 30 m of sliding, after which the friction coefficient enters a steady-state period. During the running-in period, the wear rate is high, and the friction coefficient changes very rapidly. The high rate of wear is due to the formation of a flat at the tip of the hemispherical pin where contact takes place. After this flat—which is going to be the wear scar—has been formed, it grows at a reduced rate. The slight increase in friction coefficients with sliding distance when they are in the steady-state may be attributed to the gradual increase of the size of the wear scar, leading to an increase in contact area and hence to an increase in frictional force which is given by the following

4E

,

(2)

where w is the normal load, r is the radius of the hemispherical tip of the pin, and E is the combined Young’s modulus of the two materials at the pin and at the disc. The frictional force is given by Eq. (1), and the coefficient of friction is given by the well-known equation: m=

F w

,

(3)

where m is the friction coefficient, F is the frictional force and w is the normal load. Combining Eqs. (1)–(3), we get: m=p

A B

3r 2/3

4E

s(w)−1/3.

(4)

From the above result, it is shown that for an elastic contact, an increase in the normal load will lead to a decrease in the friction coefficient as is the case in this study. The above explanation is a formalistic one, based on basic friction theory and the elastic and surface properties of the moving materials consisting the system under

V.D. Papachristos et al. / Surface and Coatings Technology 105 (1998) 224–231

227

Fig. 2. Friction coefficients versus sliding distance for three different tests.

Fig. 3. Friction coefficients versus normal load for the studied layer periodicities.

friction. A factor that might also play a role in the decrease of friction coefficient with increasing load is that under higher loads, the brittle wear of the coating (as will be discussed below), leads to detached particles that become wrapped between the moving surfaces and act as ‘‘inefficient castors’’ that facilitate relative movement, thus leading to a lower friction coefficient. These

particles could also partly account for the higher wear observed at higher loads since the local pressures at the contact points of the particles with the coating will be much higher than those acting on the coating as a whole, leading to cracking along the tracks taken by the loose particles. The wear volume of the multilayered coatings on the

228

V.D. Papachristos et al. / Surface and Coatings Technology 105 (1998) 224–231

pins, normalized per meter of sliding distance versus normal load, is given in Fig. 4 for the layer periodicities studied. The first point that can be observed from the above figure is that the wear volume values are relatively low, indicating that the multilayered coating offers significant antiwear protection for the studied conditions and range of loads. The existence of a critical load is also apparent. Up to a load of 500 g, the wear volume values are very low and close for all coatings, regardless of the layer periodicity. For higher loads, the wear volume increases more rapidly, and the difference between the coatings with a small layer periodicity (up to 60 nm) and those with a higher layer periodicity is evident, with the small layer periodicity coatings being more resistant to wear. As will be discussed later in this paper, the change in the wear rate above the critical load is probably associated with a change in the mechanisms that act during wear. In an effort to reveal the wear mechanisms acting during sliding of the multilayered coated pins against the steel discs, the wear scars at the pins and the wear debris were examined using optical and electron microscopy. In Fig. 5, the top micrograph is an electron micrograph of a wear scar at the pin, typical for loads of 200 and 500 g, and the bottom micrograph is an optical micrograph of a wear scar at the pin, typical for loads of 800 and 1100 g. By examining the wear scars under the optical and electron microscope, it can be seen that there are certain features on them that are common for all scars, regardless of the normal load used for the test,

and some features that appear only at higher loads (800 and 1100 g). The features that are common for all scars have been designated by letters in the top micrograph of Fig. 5. At the leading edge and the sides of the wear scar, there is fine adhered debris ( letter a) that has a brown colour when viewed under the optical microscope. EDAX analysis of this debris showed the presence of iron and oxygen, and, as its colour suggests, the debris is probably iron oxide that has been transferred to the pin from the disc material worn just ahead the tip of the pin and the edges of the wear track on the disc. The compacted and cracked debris at the leading edge of the wear scar ( letter b Fig. 5) also exists in patches inside the wear scar, as can be seen in Fig. 6. Its colour is grey–brown when viewed under the optical microscope, and EDAX analysis showed that this compacted debris consists of coating material, iron and oxygen. There are probably nickel and iron oxides in the debris and perhaps metallic coating particles since it is not possible to distinguish metallic nickel from nickel in an oxide form in EDAX, and the amount of debris was too little for X-ray analysis. The compacted debris is believed to be formed by the swept of the debris produced in the wear scar at the pin and the debris produced in the wear track at the disc, as the pin slides on the steel. The third common feature for all scars is the presence of the film designated by the letter c in Fig. 5. For wear scars that have been formed under the action of a lower load (200 and 500 g), this film usually covers continuous areas in the wear scar, whereas for higher loads, patches

Fig. 4. Wear volume of the multilayered coatings on the pins, normalized per meter of sliding distance versus normal load.

V.D. Papachristos et al. / Surface and Coatings Technology 105 (1998) 224–231

(a)

(a)

(b)

(b)

Fig. 5. The top micrograph is a scanning electron micrograph of a wear scar at the pin, after 280 m of sliding under load of 500 g for a coating of a 6-nm layer periodicity: (a) debris transferred around the wear scar; (b) compacted debris at the leading edge of the wear scar; and (c) adherent oxide film. The bottom micrograph is an optical micrograph of a wear scar at the pin, after 200 m of sliding under a load of 800 g for a 20-nm coating layer periodicity. Arrows show the direction of sliding.

of this film coexist with patches of clean metal in the wear scar. This film is very adherent since it is not removed from the surface by ultrasonic cleaning in acetone. EDAX analysis of the film showed the presence of oxygen and a small amount of iron along with the coating elements, so this film might consist of nickel oxides, as well as iron oxides formed by the transferred material from the disc. When viewed under the optical microscope, this film seems to be very thin since the underlying material can be seen and also has a wide range of coloured patches, probably indicating the participation of various oxides in the formation of the film. Those different oxides are believed to be a result of the heat generated by friction. The temperature rise in the wearing surfaces can be high enough locally for oxides to be formed, although the average temperature can be much lower. However, the local increase in temperature leads to different temperatures from spot to spot on the

229

Fig. 6. Areas of compacted debris on the surface of the wear scar. Arrows show the direction of sliding. Top: 200 m of sliding, load 800 g, layer periodicity 6 nm. Bottom: 150 m of sliding, load 1100 g, layer periodicity 600 nm.

wear scar, and the result is the formation of different oxides. From the above analysis, a first conclusion is that oxidation of the pin surface is involved in the wear mechanisms. The other two described features on the wear scar (adherent debris at the leading edge and the sides of the scar and compacted debris at the leading edge of the scar) originate from geometric factors related to the set-up of the test and are not significantly involved in the wear mechanisms. More important conclusions concerning the wear mechanisms that cause the loss of pin material can be gathered by examining the wear scar on the pin in more detail. In Fig. 7, the electron micrograph of a clean metallic area in the wear scar is seen without an overlayer of oxide film, after 280 m of sliding and under 500 g normal load. The fractured individual layers of the multilayered coating and terraces with a large width compared to the thickness of the individual layers can be seen. The wear mechanism seems to be a brittle fracture of the layers owing to microcracks probably originating from within the layers. When these microcracks reach the interface between the layers, they deflect, and so wide terraces

230

V.D. Papachristos et al. / Surface and Coatings Technology 105 (1998) 224–231

Fig. 7. Scanning electron micrograph of a metallic area in the wear scar without an overlayer of oxide film, after 280 m of sliding and under 500 g normal load ( layer periodicity 200 nm). The arrow shows the direction of sliding.

are formed. A wear mechanism based on the brittle fracture of the individual layers of the multilayer is in agreement with the properties of the layers, including the structure and the properties of the material as a whole, since both Ni–P–W layers are brittle, and as a result, the coating itself is a brittle material. Deflection of cracks at the interfaces between the different layers of multilayered materials has been observed for other systems that behave either in a brittle [17] or in a more ductile way [18]. In Fig. 8, an electron micrograph of an area in the wear scar that is covered by the thin oxide film is shown, after 350 m of sliding and under 200 g normal load. The fractured individual layers and the terraces can be seen in the underlying coating. The oxide film has copied the worn surface of the coating and imparts a smoothing effect. Up to 500 g normal load, the above mechanism is the main wear mechanism leading to the loss of material. For higher loads (800 and 1100 g), this mechanism

Fig. 8. Scanning electron micrograph of an area in the wear scar that is covered by the thin oxide film, after 350 m of sliding and under 200 g normal load ( layer periodicity 600 nm). The arrow shows the direction of sliding.

continues to operate (terraces and fractured individual layers can be seen in Fig. 9, which is an electron micrograph of an area in a wear scar after 150 m of sliding distance and under 1100 g normal load ), but it is not the only acting mechanism any more as will be discussed below. At higher loads, the wear scar contains a network of cracks, as can be seen in the bottom image of Fig. 5. These cracks may extend to a certain depth in the coating, leading to the delamination of large parts of the coating (as shown in Fig. 9), or may penetrate the coating reaching to the substrate. In this case, pieces of the coating become loose (this is shown in Fig. 10, pieces a and b, for a wear scar after 200 m sliding distance and under 800 g normal load) and may delaminate at a next step, revealing the substrate. The extended cracking of the coating at higher loads leading to delamination may be attributed to the higher contact pressures induced by the higher loads. An examination of the wear debris produced during the tests did not shed any more light on the nature of the acting wear mechanisms on the pins. The reason for

Fig. 9. Scanning electron micrograph of an area in the wear scar after 150 m of sliding distance and under 1100 g normal load ( layer periodicity 600 nm).

Fig. 10. Pieces of the coating that are loose and ready to spall ( letters a and b), for a wear scar after 200 m sliding distance and under 800 g normal load ( layer periodicity 20 nm).

V.D. Papachristos et al. / Surface and Coatings Technology 105 (1998) 224–231

this is that the debris was either compacted (this means that debris particles are not valid for examination) or transferred from the steel disc containing only a few particles from the coating and thus little information on the wear mechanisms on the pin. The analysis of the wear scars on the pins leads to the conclusion that two wear mechanisms are active. Up to 500 g, loss of material takes place mainly by brittle fracture of the individual layers, whereas for higher loads, the above wear mechanism co-operates with delamination of the coating, owing to extensive cracking. The calculation of the wear volume loss versus load, showed that multilayered coatings with small layer periodicities (up to 60 nm) showed a higher wear resistance than multilayered coatings with greater layer periodicities. Wear mechanisms in this study do not involve any plastic deformation or dislocation movement since the coating is brittle. Koheler’s theory [7] or Hall–Petch strengthening concepts are not applicable in this case. The better behaviour of the small layer periodicity coatings might be attributed to the increased number of interfaces included in their structure. The microcracks responsible for the fracture of the individual layers or the cracks propagating to the substrate deflect at the interfaces absorbing energy. In this way, more energy is needed for the propagation of a crack, and the toughness of the coating increases [17]. The greater the number of the interfaces, the greater energy needed, and this explains the better behaviour of the multilayered coatings with the smaller layer periodicities.

4. Conclusions From the study of the unlubricated sliding wear behaviour of pins coated with Ni–P–W multilayered coatings of various layer periodicities against steel, the following conclusions have been reached: (1) For all coatings, the friction coefficient reaches a steady state after a running-in period of about 30 m. (2) The friction coefficient decreases as the normal load increases. (3) An adherent oxide film consisting of nickel and iron oxides forms on the pin wear scar and covers parts of it. (4) Transferred debris in the form of iron oxide adheres round the leading edge and the sides of the wear scar on the pin, and compacted debris consisting of nickel oxides, iron oxides and perhaps metal par-

231

ticles from the coating accumulates at the leading edge of the wear scar. (5) For a normal load up to 500 g, the predominant wear mechanism is brittle fracture of the individual layers. For higher loads (800 and 1100 g), the above mechanism continues to operate, but delamination of the coating acts at the same time. (6) The increased wear resistance of the small layer periodicity coatings (up to 60 nm) is attributed to the increased number of interfaces in their structure compared to the larger layer periodicity coatings, which make the cracks deflect, absorbing more energy during their propagation and thus increasing the toughness of the coating.

Acknowledgement This research was totally funded by European Union under the BRITE/EURAM research programme BE 8002 with contract number BRE2-CT94-0608.

References [1] K. Terakura, in: T. Shinjo, T. Takada ( Eds.), Metallic Superlattices, Elsevier, Amsterdam, 1987, pp. 223–232. [2] V. Matijasevic, M.R. Beasley, in: T. Shinjo, T. Takada ( Eds.), Metallic Superlattices, Elsevier, Amsterdam, 1987, p. 187. [3] M. Falco, I.K. Schuller, in: L.L. Chang, B.C. Giessen ( Eds.), Synthetic Modulated Structures, Academic Press, New York, 1985, p. 339. [4] N. Holonyak, Jr., K. Hess, in: L.L. Chang, B.C. Giessen ( Eds.), Synthetic Modulated Structures, Academic Press, New York, 1985, p. 257. [5] E. Spiller, A. Segmuller, J. Rife, R.P. Haelbich, Appl. Phys. Lett. 37 (1980) 1048. [6 ] H. Underwood, T.W. Barbee, Nature 194 (1981) 492. [7] S. Koehler, Phys. Rev. B 2 (1970) 547–551. [8] T. Foecke, D.S. Lashmore, Scripta Metall. Mater. 27 (1992) 651–656. [9] E. Vancoille, J.P. Celis, J.R. Roos, Tribol. Int. 26 (1993) 115–119. [10] A.A. Minevich, Surf. Coat. Technol. 53 (1992) 161–170. [11] R. Zhang, Z. Lin, Z. Cui, Q. Song, Wear 147 (1991) 227–251. [12] W. Zhang, X. Zhang, D. Yang, Q. Xue, Wear 197 (1996) 228–232. [13] W. Ruff, N.K. Myshkin, J. Tribol. 111 (1989) 156–160. [14] W. Ruff, Z.X. Wang, Wear 131 (1989) 259–272. [15] P. Bowden, D. Tabor, The Friction and Lubrication of Solids, Oxford University Press, Oxford, 1950. [16 ] M. Hutchings, Tribology: Friction and Wear of Engineering Materials, Hodder&Stoughton, London, 1992. [17] O. Jorgensen, A. Horsewell, B.F. Sorensen, P. Leisner, Acta Metall. 43 (1995) 3991–4000. [18] E.M. Benzon, P. Leisner, C.N. Panagopoulos, V.D. Papachristos, Surf. Coat. Technol. to be published.