Wear, 40 (1976) 7 5 - 84 0 Elsevier Sequoia S.A., Lausanne - Printed in the Netherlands
SLIDING WEAR RESISTANCE OF METALLIC COATED SURFACES
E. P. ABRAHAMSON II and N. P. SUH
Department of Mechanical Engineering, Massachusetts Cambridge, Massachusetts 02139 (U.S.A.)
(Received March 9, 1976)
The role of metallic coatings in sliding wear is examined experimentally. The results indicate that the tribological behavior of soft coatings is consistent with the delamination theory of wear, especially the critical nature of the plating thickness. It is shown that a reduction in wear rate of three orders of magnitude is possible when the coating material is softer than the substrate and thinner than a critical thickness. The optimum plate thickness is found to be of the order of 0.1 p for cadmium, silver, gold or nickel plated on various types of steel. Cadmium, silver and nickel reduce wear only in non-oxidizing environments, whereas gold reduces wear both in air and in inert atmospheres. The roughness of the substrate surface prior to plating and the nature of the coating/substrate bond have significant effects on the life of these coatings. The life of the coatings is increased by polishing the substrate to 0.1 fl (c.1.a.) prior to plating, and also by diffusion of the plated material into the substrate, which increases the coating/substrate bond strength.
Introduction Earlier work [ 1 -31 has shown that the delamination theory of wear may provide a theoretical basis for reducing wear by the application of soft metallic coatings to the metal surfaces. The hypothesis was that if a thin layer of a softer metal is deposited on a harder substrate, large plastic deformation and wear of the substrate by delamination may be prevented. It was found that the plate thickness is an important parameter in major wear reduction. The critical nature of the plate thickness was explained in terms of dislocation instability very near the surface, although this was not shown explicitly*. *The existence of a soft layer due to the escape of dislocations is and has been a controversial subject. Definitive work has not been done to date because of experimental difficulties.
Soft metallic coatings have been used quite extensively [ 4 - 151 in the past to reduce friction and wear of such components as gears, bearings and electrical contacts.However, the friction and wear reduction mechanism and the effects of important parameters such as the plate thickness and the coating/substrate bond strength were not fully explained. Previous attempts have been made [15,16] to determine the effect of the coating thickness of soft metallic plates on wear, but the results were inadequate. The purpose of this paper is to show that a large wear reduction is possible by plating a soft metal on a harder substrate if the deposited material is not thicker than a critical thickness* (much less than 1 m). It is shown that, as the plate thickness is increased, wear by delamination within the plate occurs until the thickness is reduced to a stable value. Cadmium, silver, gold and nickel platings on steel are wear tested and the influence of the substrate surface condition prior to plating is investigated. The role of soft metallic coatings in wear reduction is discussed in terms of both dislocation mechanics and the mechanics of delamination.
Review of earlier work on coated metal surfaces Electroplating with precious metals to increase the life of ball bearings [ 10 - 131 and gears [ 13,141 operated in vacuum has been reported in the literature. However, the components still exhibited a high degree of wear. The coatings were 30 - 75 pm thick, well above the probable optimum values, which may explain the observed large wear rates. Other investigators [ 5 - 71 have noted a reduction in the sliding wear rate of steel when plated with soft materials such as lead, gold, gold alloys, silver, copper and nickel. For the most part wear was reduced only by a factor of 2 - 3. These results were obtained in tests which were run with an unplated slider rubbing against a plated specimen, leading to the possibility of abrasive wear caused by hard asperities and loose wear particles. Kuczkowski and Buckley  coated nickel and AISI 440 C stainless steel with 25 pm of various binary and ternary alloys of gallium, tin and iridium. A ternary alloy of gallium, indium and tin reduced the wear rate of the stainless steel by four orders of magnitude when tested in a vacuum of lo-l1 mmHg. Similar results were found for nickel with a coating of a binary alloy of gallium and indium. No mention was made, however, of the possible influence of plating thickness on the wear rate. In previous papers by the authors [2, 31 it was shown that a thin coating of cadmium (0.1 c) on AISI 1020 steel can reduce the wear rate of steel under an inert atmosphere by a factor of 5000. The thickness of the coating was found to be critical, since delamination within the cadmium plate occurred for thicknesses greater than 0.1 m. It was also shown that
the subsurface deformation of the parent metal below the coated wear track was much less than the corresponding deformation beneath an uncoated track. The reduced deformation, which retarded substrate delamination, was due to the lower friction coefficient of the thin cadmium-plated steel. The critical thickness of the soft plating could be predicted by utilizing the delamination theory of wear. This prediction was based on the fact that dislocations generated near the surface during sliding are not stable and some of them can be pulled out of the free surface by the image forces acting on these dislocations. Consequently, the thickness of the “low dislocation density” layer is controlled by the magnitude of the image forces which depend on such material properties as shear modulus and friction stress. Therefore, if the plated layer is made so thin that most of the dislocations in it can be eliminated, the plate deforms continuously without appreciable strain-hardening and subsequent fracture. This thin plated layer also reduces the friction coefficient because of its low flow stress and non-strainhardening properties, thus resulting in less substrate deformation and wear. However, if the plate exceeds this critical thickness, dislocation accumulation and strain-hardening can take place in the plated layer, leading to wear by delamination within the plate. Due to the significant wear reduction possible with the thin cadmium plating, the investigation was continued to include other plating materials. The result of wear tests on silver, gold and nickel platings on steel is reported in this paper. Important parameters such as the plate thickness, the relative hardness between the plating material and the substrate, and the coating/ substrate bond strength are also considered.
Experimental procedure Wear tests were carried out with a cylinder-on-cylinder arrangement. The specimens were 0.63 cm in diameter and 7.6 cm long with a 0.4 m (c.1.a.) ground finish prior to plating. In one case the surface was metallographically polished with a 0.25 fun diamond paste. The specimens were rotated at a surface speed of 1.8 m min-l and the stationary pins were pushed against the specimens by a normal load of 2.25 kg. The inert atmosphere tests were carried out dry in a chamber surrounding the mating surfaces under argon flowing at a rate of 10 1 min-l. The substrate materials used had a variety of hardnesses: AISI 1018 (84 kg mms2), AISI 1095 (170 kg mme2) and AISI 4140 (270,370,460 kg mm- 2 - obtained by different heat treatments). In all cases the sliders were made of the same material as the specimen and were heat-treated and plated inthe same manner. The plating thickness varied from 0.05 to 10 p. The platings tested were gold, gold over a nickel flash, cadmium, nickel and silver. Some of the gold-plated specimens were plated with a flash of gold first and then heated at 500 “C for 2 h in vacuum to obtain a diffused bonding between the plate and the substrate. The required thickness of gold was then plated on the samples.
78 TABLE 1 Experimental results on the wear resistance of 1 Mm plates on steel, wear-tested in argon, normal load 2.25 kg, sliding distance 108 m Coating
Substra?gand hardness (kg mm )
Wear zjte (X 10 mg cm-l)
Coefficient of friction
Cd Cd Cd Cd
AISI AISI AISI AISI
(84) (170) (270) (370)
3.6 1.8 3.6 3.6
0.35 0.25 0.35 0.25
AISI 1095 (170)
AISI 1095 (170)
1018 1095 4140 4140
Effused) AISI 1095 (170)
$ri underlayer) Ni Ni Ni Ni Ni
AISI AISI AISI AISI AISI
1018 1095 4140 4140 4140
(84) (170) (270) (370) (460)
immediate immediate immediate immediate 1.8
failure failure failure failure
*These specimens gained weight.
(b) Fig. 1. Comparison of wear tracks of AISI 4140 steel tested in argon, with a normal load of 2.25 kg: (a) unplated against unplated (after 30 min of testing); (b) 1.0 pm Ni plated against 1.0 pm Ni plated (after 2 h of testing). Fig. 2. The effect of the initial Ni plate thickness on the wear rate of AISI 4140 steel in argon, for 30 min tests.
Experimental results The wear data for 1 m plates of cadmium, silver, gold and nickel on steel substrates with various hardnesses for 60 min tests (108 m of sliding) in argon are summarized in Table 1. The wear rates of all coatings, with the exception of Ni on steels with a hardness less than 460 kg mm-‘, were lower than the wear rates of unplated materials by at least three orders of magnitude. It should be mentioned that greater wear reductions are possible since the tests were discontinued after 60 min while the plates were still effective in reducing wear. To determine the maximum possible wear reduction the tests should have been continued up to the final failure of the plates. The effect of plate thickness The effectiveness of plating in wear reduction is dramatically shown in Fig. 1 for a nickel plate of initial thickness 1 m on AISI 4140 steel. The large difference in the size of the wear tracks between the unplated (Fig. l(a)) and the plated sample (Fig. l(b)) should be noted. The unplated sample was tested for 30 min, whereas the plated sample was tested for 2 h. Since no further wear could be detected on the plated specimen after 2 h (216 m of sliding) the test was terminated. The important effect of the nickel plate thickness on the wear rate of AISI 4140 steel is shown in Fig. 2. Wear reduction by three orders of magnitude is observed for an initial nickel plate thickness of 1 fl. This minimum thickness of the plate also decreased the coefficient of friction from 0.63 to 0.45, probably because of less plowing when the plated layer is thin. Figure 2 shows that the wear rate increases with the nickel plate thickness for plates thicker than 1 c. This increase in wear rate is caused by dislocation accumulation leading to strain-hardening and delamination within the plate. Cracking and delamination within a thick (25 pm) plate of gold on AISI 1018 steel is shown in Fig. 3. The thick layer wears by delamination until the thickness of the layer is reduced to the optimum thickness in which the dislocations are presumably not stable. Afterwards, the wear rate is very low, as in the case of thin platings. The transient behavior of thick coatings is shown in Fig. 4 for an initial 10 I.tmnickel plate on AISI 4140 steel. The steady state wear rate of thick platings may be larger under some sliding situations where the wear particles cannot be removed from the contact and act as abrasive particles. The effect of surface roughness and bond strength The steel samples which were plated with a thin layer of gold without any special treatments failed immediately at the beginning of sliding. This failure was found to have been caused by the weakness of the bond between the plated layer and the substrate, which was accentuated by the roughness of the substrate surface at the time of plating. In these tests, since the optimum plate thickness is very small, the substrate surface roughness may
Fig. 3. Crack formation in a thick Au plate (25 pm initial thickness) after wear testing; AISI 1020 steel substrate. Fig. 4. Wear of a thick Ni plate (initial thickness 10 pm) vs. time.
play a major role in the life of the coatings. The bond strength between electrodeposited gold and steel has been reported to be very low and to become even weaker with increasing substrate surface roughness [ 171. Under sliding conditions, the roughness can cause further deterioration of the bond strength by the deformation and final fracture (Fig. 5) of the original substrate asperities. Therefore, the gold-plated steel samples were specially treated to achieve a good bond strength. By diffusion of a very thin layer of gold or by plating a flash of nickel over the substrate before gold plating, it was possible to increase the life of the thin gold plates from immediate failure to more than 108 m of sliding (60 min). The influence of the substrate surface finish on cadmium-plated specimens was checked with a 0.05 fun cadmium coating on AISI 1018 steel with both a fine ground finish and a metallographically polished finish. The surface roughness had only a moderate effect on the life of the cadmium plates, and polishing only increased the life of the coatings from 25 m to 31 m of sliding. Since the steel-cadmium adhesion strength is much greater than the adhesion strength of steel-gold [ 171, this result suggests that moderate changes of surface roughness may only influence plating-substrate combinations which have a low adhesive strength. Hardness matching between the coating and the substrate According to the delamination theory the plated material must be softer than the substrate to minimize the wear rate. The results on the wear rate of cadmium- and nickel-plated steels in Table 1 support this hypothesis. The wear rate data indicate that nickel plate on steel specimens which are softer than 460 kg mmB2 fails immediately at the start of sliding; However, the wear rate of a 1 lun nickel-plated 4140 steel specimen with a hardness of 460 kg mmm2 is very low. Table 1 also indicates that a 1 cun cadmium plate with a hardness of 30 - 50 kg mmF2  was successful on all substrates with hardnesses ranging from 84 - 374 kg mmV2.
Fig. 5. The deformation of substrate asperities due to the sliding action on a thick plate of Au on AISI 1020 steei substrate.
The effect of environment The wear tests conducted in air on 1 m plate of various materials on steel indicated that only gold platings (with special treatments) were effective in wear reduction. Other platings such as cadmium, silver and nickel were not successful in air, because they oxidized and were removed early in the tests. The oxide particles of the plating materials also caused abrasive wear. These findings are consistent with previous results  on cadmium platings tested in air and under corrosive lubricants. In that study it was found that, under lubricated conditions, cadmium plating is effective if the lubricant does not cause any corrosion of the plating material. Discussion The results presented in this paper further support the predictions made by the del~ination theory of wear on the wear resistance of soft metal coatings. For best wear resistance the plated material must be softer than the substrate, thinner than a critical thickness and bonded strongly to the substrate. The condition of the surface of the substrate is an important factor since it was shown that smoother plated surfaces last longer. It has been shown recently  that subsurface damage caused by machining operation greatly influences initial wear behavior. Therefore, components for sliding applications must be prepared carefully so as to minimize the damage to the substrate during machining. Otherwise, the damaged substrate may cause premature failure and offset the beneficial effects of plating. If the delaminated particles are not removed from the contact, they may oxidize and serve as abrasive particles, thus degrading the coating and enhancing delamination. Soft metals have been used previously for wear resistance [5 - 151. Some of these investigators assumed that the metal layer softens during sliding
and acts as a lubricant. This explanation can only be correct when the sliding speeds and the normal loads are high enough to cause a large flash temperature at the contact. The present work has shown that even a high melting point metal such as nickel can be effective in wear reduction under low speed sliding wear, and that the thickness of the plate has a pronounced influence. These findings are contradictory to the previous assumptions on the mechanism of wear reduction of soft metal layers. It has been shown in this paper that it is not the absolute softness which is important; it is the ratio of flow stresses between the substrate and the coating which is the determining factor. The relative lattice spacing, elastic modulus and crystal structure of the coating and the substrate may also influence the effectiveness of the coating material. When the shear modulus of the coating is less than that of the substrate, the dislocations generated in the coating during sliding will be repelled by the interface, while the dislocations generated in the substrate will be attracted towards the interface. If the coating is sufficiently thin, some of the mobile dislocations in the plated layer may be eliminated by the image force, and also by the stress field established by other dislocations, when the surface is unloaded after the slider asperity has passed. In this case the coating will remain soft and function as a protective layer. Another important influence of these relative material properties is the dislocation mobility across the plate-substrate interface. Ideally, for the substrate to become harder and to resist plastic deformation, the dislocations in the substrate should not cross the interface and move into the coating, but rather they should tangle up near the interface. This phenomenon will arise if the lattice spacing and the crystal structure and orientation of the coating and the substrate are sufficiently different to prevent the dislocations generated in the substrate from penetrating across the interface. In this respect the choice of f.c.c. metals (such as nickel) or h.c.p. metals (such as cadmium) for the coating material and b.c.c. metals (such as steel) for the substrate is quite appropriate. The foregoing discussion presumed that there is a strong bond between the substrate and the coating material. It is well-known that those metals which readily form an alloy also form a strong adhesive bond. However, the requirements for solubility of one metal in another are precisely those qualities which tend to promote penetration of dislocations across the interface. Therefore, the material choice is rather limited. What may help the situation, however, is the texturing of the surface layer during deformation. The surface layer orients during deformation in such a manner that the primary slip planes become nearly parallel to the surface [ 191. If the substrate metal has a different crystallographic structure to the plating material, the orientation of crystallographic planes in the coating will be different to that in the substrate 2nd the interface will retard the transport of dislocations across the interface. The experimental results on the role of coating in sliding wear presented here are not direct and explicit proof of the assumed dislocation
instability caused by the image force, but they are in accordance with the original predictions made by the del~ination theory of wear. It appears, however, that the original postulates predicting the effects of thickness and hardness of the metallic coatings are justified by the indirect evidence provided by the wear tests. In the preceding paragraphs the role of thin soft coatings was explained in terms of dislocation mechanics. However, other plausible explanations were also attempt without much success. One may attempt to explain the observed phenomenon in terms of the state of stress during sliding wear and the stress relaxation of the material. It has been shown that there is a region following the slider asperities where the state of stress is tensile . In materials without coating, a correlation exists between the location of the subsurface cracks and the location of the maximum tensile stress perpendicular to the surface. The magnitude of this stress is very low close to the surface. Therefore, one may argue that in thin coatings (~0.1 pm) the tensile stress is not sufficient to propagate subsurface cracks in the coating. In order to explain the non-hardening nature of the surface layer, one may then speculate that stress relaxation occurs to counter the work-hardening effects. However, these arguments fail since they cannot explain why work-hardening occurs when the plated layer is thick. Therefore, it may be concluded that the dislocation behavior very near the surface is different from the bulk and that the difference is responsible for the observed wear resistance of thin metallic coatings.
Conclusions 1. The delamination theory of wear provides a theoretical basis for reducing wear through the development of a coated metal surface. 2. For major wear reduction the coating material must be softer than the substrate material. 3. There exists an optimum thickness of soft coating for wear resistance. This thickness is, in general, less than 1 p for steel plated with Cd, Ag, Au or Ni. 4. Cd, Ag and Ni plates are effective only in an inert atmosphere, but Au is effective in air or in an inert atmosphere. 5. The surface roughne~ of the substrate and the ~oat~g/sub~rate bond strength are two important factors for the wear resistance of soft metallic coatings.
Acknowledgments The work reported in this paper was sponsored by the Defense Advanced Research Projects Agency through the Office of Naval Research under contract N00014-67-A0204-0080. We are grateful to Dr. Edward van
Reuth and to Lt. Richard S. Miller for their personal support and guidance of our work.
References 1 N. P. Suh, S. Jahanmir and E. P. Abrahamson, The Delamination Theory of Wear, Progress Report to ARPA, Contract No. N00014-67-A-0204-0030, Sept. 1974. 2 E. P. Abrahamson, S. Jahanmir, N. P. Suh and D. A. Colling, Application of the delamination theory of wear to a composite metal surface, Proc. Int. Conf. on Production, Tokyo, Japan, 1974, pp. 408 - 413. 3 S. Jahanmir, N. P. Suh and E. P. Abrahamson, The delamination theory of wear and wear of a composite metal surface, Wear, 32 (1975) 33 - 49. 4 Y. Tsuya and R. Takagi, Lubricating properties of lead films on copper, Wear, 7 (1964) 131 - 143. 5 R. Takagi and T. Liu, The lubrication of steel by electroplated gold, ASLE Trans., 10 (1967) 115 - 123. 6 R. Takagi and T. Liu, Lubrication of bearing steels with electroplated gold under heavy loads, ASLE Trans., 11 (1968) 64 - 71. 7 A. J. Solomon and M. Antler, Wear mechanisms of gold electrodeposits, Plating, East Orange, N.J., Aug. (1970) 812 - 816. 8 T. J. Kuczkowski and D. H. Buckley, Friction and wear of low melting binary and ternary gallium alloy films in argon and in vacuum, NASA Tech. Note D-2721,1965. 9 H. E. Evans and T. W. Flatley, Bearings for vacuum operations, retainer material and design, NASA Tech. Note D-1339, 1962. 10 H. E. Evans and T. W. Flatley, High speed vacuum performance of gold-plated miniature ball bearings with various retainer materials and configurations, NASA Tech. Note D-2101,1963. 11 T. W. Flatley, High speed vacuum performance of miniature ball bearings lubricated with a combination of barium, gold and silver films, NASA Tech. Note D-2304, 1964. 12 P. E. Brown, Bearing retainer material for modern jet engines, ASLE Trans., 13 (1970) 225 - 239. 13 T. L. Ridings, Operational evaluation of dry thin film lubricated bearings and gears for use in aerospace environmental chambers, AEDC-TR-65-1,1965. 14 R. D. Lee, Jr., Lubrication of heavily loaded, low velocity bearings and gears operating in aerospace environmental facilities, AEDC-TR-65-19, 1965. 15 E. Rabinowicz, Variation of friction and wear of solid lubricant films with the film thickness, ASLE Trans., 10 (1967) 1 - 9. 16 F. P. Bowden and D. Tabor, Friction and Lubrication of Solids, Clarendon Press, Oxford, 1954. 17 W. H. Safranck, The Properties of Electrodeposited Metals and Alloys, Elsevier, New York, 1974. 18 S. Jahanmir and N. P. Suh, Surface roughness and integrity effects on sliding wear, Wear, to be submitted for publication. 19 D. R. Wheeler and D. H. Buckley, Texturing in metals as a result of sliding, Wear, 33 (1975) 65 - 74. 20 N. P. Suh, Microstructural effects in sliding wear of metals, Proc. Battle Materials Science Colloquium, Fundamental Aspects of Structural Alloy Design, Plenum Press, to be published, 1976.