Sliding and abrasive wear of composite solgel alumina ... - Ewp.rpi.edu

Surface and Coatings Technology 133᎐134 Ž2000. 389᎐396

Sliding and abrasive wear of composite sol᎐gel alumina coated Al alloys S. Wilsona,U , H.M. Hawthorne a , Q. Yang b, T. Troczynski b a

b

Surface Technologyr Tribology Group, NRC Inno¨ ation Centre, 3250 East Mall Vancou¨ er, BC, Canada V6T 1W5 Metals and Materials Engineering Dept., 309-6350 Stores Rd., Uni¨ ersity of British Columbia, Vancou¨ er, BC, Canada V6T 1Z4

Abstract Alumina coatings, 60-␮m thick ŽHV10g 647 " 144 kg mmy2 ., were deposited using a novel sol᎐gel technique on 6061 Al. The samples were subjected to dry sliding wear tests against hard bearing steel balls ŽSAE 52100 steel. and softer mild steel pins ŽAISI 1018 steel. at different sliding speeds and contact loads. Wear at low contact loads and sliding speeds was characterized by steel oxide ŽFe 2 O 3 . transfer to the coating. Coating rupture arose at a critical transition boundary of sliding contact loads and speeds which were approximately twice as high for mild steel as against the harder bearing steel. The coated alloy was also subjected to abrasive wear against SiC abrasives of increasing coarseness. Coating wear rates approached those of the substrate alloy as the indentation depth increased. A similar reduction in coating hardness was observed for Vickers diamond pyramid indentations at greater depths. When abrasion indentation depths were greater than ; 20% of the coating thickness, a rapid degradation in wear resistance arose and the same effect was seen for scratch test indentation depths. The use of normalizing parameters for sliding wear, hardness, abrasive wear and indentation depth were found to be useful in interpreting coating wear performance. 䊚 2000 Elsevier Science B.V. All rights reserved. Keywords: Composite sol᎐gel; Abrasion; Pin-on-disc; Aluminum

1. Introduction Sol᎐gel ŽSG. technology has enjoyed success in the last 20 years, for deposition of thin films of hard, corrosion resistant ceramics Že.g. SiO 2 , ZrO 2 , Al 2 O 3 . or functional ceramics for sensors, membranes, etc. w1,2x. Nonetheless, SG technology often fails if the film is to be thicker than approximately 1 ␮m because of the damaging shrinkage strains during drying and densification w2x. However, well established sol᎐gel technology ŽSG., powder processing ŽPP., and chemical bonding ŽCB. techniques can be used to overcome this principal weakness to produce chemically bonded composite sol᎐gels ŽCB-CSG.. When calcined ceramic powder particles Ž‘filler’. are dispersed in an SG matrix U

Corresponding author. Tel.: q1-604-221-3124; fax: q1-604-2213088. E-mail address: [email protected] ŽS. Wilson..

Ž‘binder’. to produce a composite sol᎐gel ŽCSG., a dramatic decrease of densification strain results w3᎐5x, allowing processing of ceramic coatings and bulk shapes of unlimited size and shape. However, CSG ceramics must be further heat treated at relatively high temperatures, i.e. 1300᎐1400⬚C for Al 2 O 3rAl 2 O 3 CSG in order to form ceramic bonds and eliminate porosity w6᎐8x. Recently, chemical bonding ŽCB. of porous composite sol᎐gel Žor CSG. alumina films has been shown to produced dense, hard, wear- and corrosion-resistant ceramic coatings, after heat treatment at temperatures as low as 300᎐500⬚C w9x. The lower heat treatment temperatures allow this technology to be used to produce low cost, wear resistant coatings on a variety of substrates that can survive ; 300⬚C heat treatment, including low melting point metal alloys, e.g. Al and Mg alloys. Whilst the main function of hard coatings is to provide protection to softer substrates from indenta-

0257-8972r00r$ - see front matter 䊚 2000 Elsevier Science B.V. All rights reserved. PII: S 0 2 5 7 - 8 9 7 2 Ž 0 0 . 0 0 9 6 4 - 6

390

S. Wilson et al. r Surface and Coatings Technology 133᎐134 (2000) 389᎐396

tion modes of wear, it should be noted that they can be subject to a wide range of surface strains, depending on the tribological system. A coated system’s hardness is dependent on the extent to which an indenter penetrates the coating thickness, t. When the depth of indentation, d, is well below t, measured hardness is coating dominated. Generally, when d is less than 5% of the thickness for hard coatings, substrate effects are eliminated w10x. However, for d) t, hardness response becomes increasingly substrate dominated. This variation in hardness is now well understood, with a model to predict the composite coating hardness having successfully been developed by Korsunsky et al. w11x, based on dimensionless plasticity, elastic and fracture parameters scaled with respect to coating thickness, t. However, given the complexity of most tribo-systems, determining the influence of the coating composite properties on tribological performance in different wear environments has presented significant challenges. The development of ‘wear maps’ has led to greater understanding of these systems w12᎐14x. In the present study, a tribo-systems approach is adopted to characterize the sliding and abrasive wear performance of a CB-CSG ŽAl 2 O 3Žbinder.rAl 2 O 3Žfiller. . coating, spray deposited onto a commercial aluminum ŽAl6061. alloy. The CB-CSG

coating was subjected to a range of imposed strains by sliding wear tests against hard and soft steels at different contact loads and sliding speeds. Similarly, abrasion experiments were conducted against the coating using abrasives of increasing coarseness and accompanying harshness. In a similar approach to that of Korsunsky et al. w11x, better understanding of coating abrasion and sliding wear properties is sought through use of dimensionless parameters such as normalized hardness Žrelative to substrate or another bulk material., indenter penetration depth Ž dr t . and abrasion resistance Žrelative to substrate abrasion resistance ..

2. Experimental 2.1. Materials Alumina sol᎐gel ŽCB-CSG. was deposited on gritblasted 4-mm-thick coupons of a commercial Al alloy ŽAl 6061.. The CB-CSG coating thickness was measured as 60 ␮m using the ball cratering method w15x on a Plint 䊛 TE-66 micro-scale abrasion tester. The lowviscosity water-based hydrated alumina sol contained finely dispersed calcined alumina powder ŽA-16 grade

Fig. 1. Ža. Wear rates of the CB-CSG coating is represented in map form on load vs. sliding speed axes for tests against AISI 1018 mild steel. The coating wear rates Ž10y7 g my1 . are given numerically for tests at different load and speed combinations and the map has been demarcated into mild wear and rupture regimes. Žb. SEM micrograph ŽBSE. showing typical wear track morphology in the mild wear regime of Ža.. Mild steel pin material has transferred to the worn surface and has been oxidized as is indicated by the EDX spectrum in Žc.. Žd. SEM micrographs of worn CB-CSG coating surface taken from test in the mild region but performed close to the rupture boundary. Coating surface has been removed to produce a deep groove. Že. SEM micrograph showing subsurface taper cross-section Ž20 = vertical magnification. of wear track groove.


by Alcoa, ; 0.3 ␮m average particle size.. The coating was spray-deposited using an air-pressure driven nozzle, dried and heat treated at 300⬚C to decompose the hydrated alumina to a super-fine Ž10᎐50 nm. gamma᎐alumina phase. A proprietary chemical bonding process w9x followed to eliminate porosity in the coating and to provide high interfacial bond strength Ž30᎐50 MPa, as determined in tensile pull adhesion test w16x. of the coating to the aluminum alloy substrate. The coatings microstructure is shown in Figs. 1 and 3, together with the effects of wear tests. 2.2. Hardness Hardness measurements were made on the coating using a Vickers diamond pyramid indenter at different indentation loads ranging between 10 gf and 10 kgf. The depth of maximum penetration for each indent Ž d H . was calculated using the measured indent impression width and diamond indenter geometry and normalized by dividing indentation depth by coating thickness Ž t s 60 ␮m.. These hardness values were also normalized by dividing by the Al6061 substrate hardness ŽHV 36.4 kgf mmy2 .. 2.3. Dry sliding wear Flat square coupons Ž25 = 25 = 4 mm. of the CBCSG coated 6061 Al alloy were subjected to dry sliding wear tests against hard bearing steel balls ŽSAE 52100 steel, HV 858 kgf mmy2 , 6.35 mm diameter. and softer mild steel pins ŽAISI 1018 steel, HV 297 kgf mmy2 , 5 mm diameter. at different sliding speeds and contact loads. The mild steel pin ends were ground on 1000-grit SiC paper to give an approximately hemispherical end. Sliding tests were performed on the CB-CSG specimens using a pin-on-disc sliding wear apparatus. The apparatus is comprised of a variable speed rotating shaft arrangement to which the pin Žor ball. specimen is attached. The rotating pin is then lowered onto the stationary SG coupons to produce a circular wear track 14.0 mm in average diameter. Each of the experiments was performed at a set load and sliding speed and run up to a constant sliding distance of 1800 m Ž; 41 000 rev... Both pin Žor ball. and SG discs were ultrasonically cleaned in ethanol, air dried and weighed to "0.1 mg prior to wear testing. After each test, specimens were cleaned of loose debris using compressed air and weighed to determine mass loss Žor gain.. Wear rates of coupons were obtained by dividing mass lossrgain by the total sliding distance, 1800 m. All wear experiments were conducted under atmospheric humidity conditions ranging between 25 and 30% RH. Scanning electron microscopy ŽSEM. using backscattered electron ŽBSE. imaging and energy dispersion X-ray ŽEDX. analysis techniques were used to characterize worn surfaces.

391

2.4. Scratch tests Incremental loading scratching was performed on the CB-CSG coatings with a CSEM Revetest 䊛 scratch test instrument using diamond stylus indenters ŽRockwell ‘C’, 200 ␮m radius.. Acoustic emission signals and transverse friction force measurements acting on the stylus were also recorded as a function of indentation load during each scratch experiment. The depth of the scratch track was measured using a Wyko 䊛 optical interference surface profilometer. Scratch track depth readings Ž d S . were normalized relative to the coating thickness Ž t s 60 ␮m.. 2.5. Abrasi¨ e wear tests The CB-CSG coatings were subjected to two-body dry abrasive wear tests against SiC abrasive bonded papers with grit sizes of differing coarseness, i.e. 1000 grit Žfinest., 600, 400, 280, 220 and 120 grades. Experiments were performed using a modified dimple grinder apparatus with an abrasive paper covered cylindrical brass wheel Ž15 mm diameter = 5 mm width. that is vertically loaded and rotated against the stationary specimen surface. A load of 1 N was applied for all tests using a wheel velocity of 100 rev. miny1 Ž0.08 m sy1 circumference velocity. and sliding distance of 4.7 m. Fresh abrasive paper strips Ž5 mm wide. were applied to the wheel’s outer surface before each experiment. Wear loss from the abraded surface was determined by measuring the wear scar volume, using a Wyko NT-2000 䊛 optical surface profilometer Žvertical resolution to 3 nm. which incorporates image analysis and metrology software. Abrasion tests were also performed on the uncoated Al6061 substrate alloy using the same SiC abrasive sizes and procedures used for the coated specimens. The wear volumes Žaverage of three tests. obtained for abrasion of the coating against different mesh SiC abrasives were then normalized by dividing them by those obtained for the uncoated substrate Al6061 alloy. The depth to which abrasive particles penetrated the worn coating was estimated from peak-to-valley heights Ž dABR . of the optical profiles within the wear scar. These were then normalized by dividing by the coating thickness, t.

3. Results 3.1. Wear maps Wear rates of the alumina coated 6061 Al alloy coupons are represented in map form on load vs. sliding speed axes for tests against mild steel pins in Fig. 1a, and bearing steel balls in Fig. 2. The coating wear rates are given numerically for tests at different

392


Fig. 2. Wear map of CB-CSG coating for tests against SAE 52100 bearing steel. The coating wear rates Ž10y7 g my1 . are given numerically for tests at different load and speed combinations and the map has been demarcated into mild wear and rupture regimes.

load and speed combinations and maps have been demarcated into regimes based on wear rate and surface damage data. 3.1.1. CB-CSG ¨ s. mild steel The wear map for mild steel in sliding contact against the CB-CSG coating ŽFig. 1a. can be demarcated into mild wear and coating rupture regimes which are separated by the critical boundary for rupture, R MS . In the mild wear regime, wear rates of the order of 10y6 to 10y7 g my1 prevail for contact loads ranging below limits of approximately 40 N at 0.04 m sy1 and 20 N at 0.51 m sy1 . Some coating wear rates within this regime are negative, which is attributable to mass gains by transfer of the pin material to the coating wear track. The SEM micrograph in Fig. 1b shows the typical morphology of the worn coating surface where pin material has transferred to the wear track to form a surface layer. The formation of such transfer layers is characteristic of many dry sliding systems w17,18x and common to metal vs. ceramic w19x or coating wear w13,20,21x. The transferred material had a characteristic reddish-brown appearance and EDS analysis ŽFig. 1c. shows the presence of Fe and O species, indicating oxidation of the transferred steel. The formation of transfer layers on coating surfaces tends to mask actual coating wear rates, so it is important to assess damage using SEM or other techniques w13x. SEM examination of a worn coating surface from a test conducted at 30 N and 0.17 m sy1 ŽFig. 1d. shows increased damage by formation of a groove in the wear track and there is comparatively little pin material transfer. The increased damage is manifested by a coating wear rate Ž50 = 10y7 g my1 . that is high in comparison to experiments at lower speeds and contact loads in the mild regime. A sub-surface taper section ŽFig. 1e. through the coating wear groove edge, shows that the alumina has remained intact beneath the wear

track and appears to have been removed by a microattrition process, as opposed to any large scale spallation. The cracks present in the coating arise mostly from moderate processing shrinkage strains but tend to close again during subsequent processing Žchemical bonding.. It is interesting to note that they have not acted as paths for bulk coating removal, even where the coating is at its thinnest. Some porosity is also visible at the interface where the spray deposited CB-CSG material did not fully penetrate the sand blasted aluminum substrate. Wear tests conducted at loads and sliding speeds above the mild wear regime resulted in rapid rupture and removal of the coating, exposing the substrate Al alloy to wear by the mild steel pin. Coating wear rates exhibited in this regime were at least an order of magnitude above those in the mild region. Coating rupture at sliding speeds above 0.2 m sy1 arises at contact loads above approximately 20 N, while this limit is raised to between 30 and 40 N at lower sliding speeds. 3.1.2. CB-CSG ¨ s. SAE 52100 steel CB-CSG coating wear rates against the harder SAE 52100 bearing steel balls are summarized on load vs. sliding speed axes in Fig. 2. The transition boundary, R 52100 , between mild wear and coating rupture occurs at a significantly lower range of contact loads Žapprox. 15 N at 0.04 m sy1 to 10 N at 0.34 m sy1 . compared to that seen for wear against SAE 1018 steel ŽFig. 1a.. 3.2. Abrasion tests Damage to the surface of the CB-CSG coatings arising from abrasive wear against fine Ž1000 grit. and coarse Ž120 grit. SiC is shown in the optical interference micrographs in Fig. 3a,b, respectively. Wear against the fine abrasive has produced a polishing effect on the coating surface. A surface profile of a polished region ŽFig. 3c. indicates the extent to which abrasive particles penetrate the coating, where the average peak-to-valley Ž dABR . value was measured at 5.2 ␮m. In contrast, wear against the coarsest SiC abrasive resulted in exposure of the Al6061 substrate ŽFig. 3b. and formation of deep grooves in the coating wear scar ŽFig. 3d, average dABR of 60.4 ␮m.. Abrasion of the coating by the 120 mesh SiC was the only set of tests where the Al alloy substrate became exposed. Coating wear scar dABR values after abrasion against 600, 400, 280 and 220 mesh SiC were measured as: 4.4, 6.7, 17.6 and 33.4 ␮m, respectively. Normalized wear scar volumes ŽWCB-CSG rWAl6061 . for abrasion against each grade of SiC have been plotted as a function of their normalized Žabrasion. indentation depth readings Ž dABR r t . in Fig. 4b. Similarly, normalized hardness ŽHVCB-CSG rHVAl6061 . measurements ob-


393

Fig. 3. Optical interference micrograph of CB-CSG coating abrasive wear scar abraded against 1000 grit SiC Ža. and associated wear track surface roughness profile Žb.. CB-CSG coating wear scar after abrasion against the coarsest SiC Ž120 grit. Žc. and wear track surface roughness profile Žd..

tained at different indentation loads have also been plotted on the same figure against d H rt, the normalized indentation depth. A scratch test indentation volume ‘wear’ term, derived from d S ŽFig. 4a. as Ž d S . 3 is also plotted on Fig. 4b against d S r t, the normalized scratch indentation depth. Wear rates of the coating are lowest when dABR r t falls below 0.2, or one-fifth the penetration of the coating thickness. Coating wear rates range between approximately 0.3 and 0.4 of the uncoated matrix alloy for dABR r t - 0.2. In a similar manner, coating hardness is greatest where d H r t - 0.2 and falls between ; 18 and 15 times that of the substrate. The peak coating hardness, measured using an indentation load of 10 gf, was HV10g 647 " 144 kg mmy2 . For abrasion against the coarser SiC abrasives where dABR r t ) 0.2, the coating wear rate increases significantly and approaches that of the matrix alloy. Hardness measured on the coated surface also decrease for d H r t ) 0.2 and reach those of the matrix alloy for d H r t f 8᎐10. In addition, the scratch test ‘wear’ term Ž d S . 3 shows a marked increase for d s r t ) 0.2 which correlates with the onset of cracking in the scratch wear track ŽFig. 4a. and spallation at a critical indentation load of 17.6 N.

4. Discussion In the pin-on-disc experiments, the CB-CSG coated Al alloy specimens can withstand contact loads against the soft mild steel ŽHV 297 kgf mmy2 . pin approximately twice those when sliding against the harder SAE 52100 steel pin ŽFigs. 2 and 3.. For the softer pin case, the sliding contact load may be normalized by dividing it by the hardness of the softer of the two materials and . y1 Ž the nominal contact area w22x, i.e. Fn s FAy1 Fn o H is normalized load, A o is nominal contact area and H

is the hardness.. Whilst this can be done for the mild steel, which is softer than the coated Al alloy’s highest hardness reading Ži.e. HV10g 647 " 144 kg mmy2 ., it is not feasible for contact with the harder SAE 52100 steel as most damage is expected to be concentrated in the softer coating microstructure. A much-simplified approach can be adopted using the bulk hardness ratio of mild steel to that of SAE 52100 Ži.e. 0.35:1. as a simple normalizing parameter for plastic accommodation of the pin material in the contact zone. This would bring the range of critical coating transition loads for wear against mild steel down by a factor of 3. However, extensive work hardening of the worn mild steel pin surface occurs. Hardness measurements taken from a subsurface section of the work hardened zone of worn mild steel pin, gave a maximum hardness value of HV10g 467 kg mmy2 , making its hardness ratio with SAE 52100 steel to 0.56:1. Normalizing contact loads with this ratio reduces the range of critical coating transition loads for wear against mild steel by approximately half and brings it close to what is seen for wear against the SAE 52100 bearing steel Ži.e. R⬘ms in Fig. 5.. The close proximity between R⬘ms and R⬘52100 suggests that a single Žnormalized. rupture limit can now be identified for the coating and used to predict the critical transition boundaries for wear against pin materials of different hardness, using hardness ratio measurements. Conceptually, the onset of coating rupture at the R⬘ms and R⬘52100 boundaries for sliding contact against the two steels, is similar to critical failure or spallation loads seen in scratch tests on brittle coatings from which coating-interface bonding energies can be determined w23x. The normalized critical boundaries for CB-CSG wear against both steels, R⬘ms and R⬘52100 ŽFig. 5., cover a 10᎐22-N range within which also falls L c for the diamond indenter scratch test at 17.6 N ŽFig.

394


Fig. 4. Ža. SEM micrograph of scratch test wear track surface morphology showing critical load for rupture at 17.6 N and scratch indentation depth Ž d S . profile. Žb. Summary of normalized coating hardness Žrelative to substrate ., abrasion resistance ŽWcoatrWsubstrate . and a ‘wear’ term d S3 derived from the scratch test surface profile, shown as a function of normalized indentation depth Ž dAB R r t, d H r t and d S r t ..

4a.. Nonetheless, this apparent correlation does not imply that coating rupture in sliding wear arises by the same large-scale spallation and buckling failure mecha-

Fig. 5. Coating rupture boundaries for wear against mild steel Ž R ms . and SAE 52100 bearing steel Ž R 52100 . and their normalized values Ž R⬘ms and R⬘52100 ..

nisms observed at L c in scratch testing ŽFig. 4a.. Indeed, the removal of the coating by attrition ŽFig. 1d,e. suggests that micro-fracturerabrasion is the dominant wear mechanism, possibly occurring between heavily work hardened steel particles and the coating surface. It should be emphasized that coating failure evolving from cumulative deformation w12,24x and localized friction heating effects w13x may also be contributing factors. Increasing the scale of indentation during abrasive wear of the coated surface produces wear rates that approach those of the Al alloy substrate and reduces the coating effectiveness ŽFig. 4b.. The same effect is observed for hardness indentations at different levels of coating penetration ŽFig. 4b.. The normalized coating hardness of ; 15᎐18 for d H r t - 0.2 imply that similar coating abrasive wear rates relative to the matrix Al6061, should also be seen at equivalent indentation depths, if a direct hardness᎐abrasion resistance


ŽArchard. type relationship w25x is to be obeyed. However, the WCB-CSG rWAl6061 ratio falls into the range 0.2᎐0.4 Ži.e. coating wear rates ; 2.5᎐5 times below those of the matrix alloy. for dABR r t - 0.2, showing a deviation from hardness-based predictions. The deviation is similar to that associated with initiation of indentation microfracture phenomena in brittle materials Ži.e. median, radial or lateral cracking w26,27x., and plasma sprayed coatings Že.g. w28x. which are accompanied by high wear rates. The good correlation between the scratch test volume ‘wear’ term d S3 and the onset of both hardness and abrasion resistance degradation for d S r t ) 0.2, shows that some measure of success can be attained in using the scratch test as a predictive tool for coating abrasion performance. Resistance to indentation modes of wear damage on coated surfaces are scale-dependent, where the substrate influence on coating hardness increases with the indentation depth. By normalizing the scale of indentation depths with the coating thickness for hardness, abrasion and scratch test experiments, indentation depth parameters d S r t, dABR r t and d H r t can be compared and used to identify critical indentation depths at which transitions in wear behavior arise. Further work would be beneficial in establishing whether similar relationships exist with other coating systems.

5. Conclusions Sliding wear rates and wear mechanisms of a 60␮m-thick alumina CB-CSG coated Al6061 alloy in contact against a soft mild steel and hard bearing steel were summarized on wear maps with contact load vs. sliding speed axes. Abrasive wear experiments were also performed against the coated Al alloy against SiC abrasives of varying coarseness. The following conclusions can be drawn from the present set of experiments. 5.1. Sliding wear Coating rupture arose at a critical transition boundary of sliding contact load and speed when sliding against both steels. The transition loads for wear against mild steel were approximately twice as high as those against the harder bearing steel. Critical boundary loads, normalized using a simplified steel pinrcoating hardness ratio, were found to be in reasonable agreement with the critical rupture loads found in scratch tests on the same coating.

395

5.2. Abrasi¨ e wear When CB-CSG coatings were subjected to abrasive wear against SiC abrasives of increasing coarseness, coating wear rates approached those of the substrate alloy as the scale of indentation increased. A similar reduction in coating hardness was observed for Vickers diamond pyramid indentations at greater depths. Abrasion indentation depths greater than ; 20% of the coating thickness produced a rapid degradation in wear resistance. The same effect was seen for scratch test indentation depths. The use of normalizing parameters for hardness, abrasive wear and indentation depth was found useful for interpreting coating wear performance.

Acknowledgements The authors are grateful to the Canadian Lightweight Materials Research Initiative ŽCLiMRI. for support of the research. References w1x C.J. Brinker, G.W. Sherer, Sol Gel Science, Academic Press, San Diego, 1990. w2x C.J. Brinker et al., Sol᎐gel derived ceramic films ᎏ fundamentals and applications, in: K. Stern ŽEd.., Metallurgical and Ceramic Protective Coatings, Chapman & Hall, London, 1996, pp. 112᎐151. w3x D.A. Barrow, T.E. Petroff, M. Sayer, Method for Producing Thick Ceramic Films by a Sol-Gel Coating Process, US Patent No. 5 585 136, 1996. w4x D.A. Barrow, T.E. Petroff, M. Sayer, Surf. Coat. Technol. 77 Ž1995. 113᎐118. w5x T.E. Petroff, M. Sayer, S.A. Hesp, 78 Ž1993. 235᎐243. w6x Q. Yang, T. Troczynski, J. Am. Ceram. Soc. 83 Ž2000. 958᎐960. w7x Q. Yang, T. Troczynski, Composite Alumina Sol᎐Gel Ceramics, Proceeding of the Symposium on Sol᎐Gel Processing, at 100th annual convention of American Ceramic Society, Cincinnati, May 1998. w8x Q. Yang, T. Troczynski, J. Am. Ceram. Soc. 82 Ž1999. 1928᎐1958. w9x T. Troczynski, Q. Yang, Process for Making Chemically Bonded Sol Gel Ceramic, US Patent pending, UBC-UILO, 1999. w10x S. Veprek, J. Vac. Sci. Technol. A. 17 Ž1999. 2401᎐2420. w11x A.M. Korsunsky, M.R. McGurk, S.J. Bull, T.F. Page, Surf. Coat. Technol. 99 Ž1998. 171᎐183. w12x K. Kato, Surf. Coat. Technol. 76-77 Ž1995. 469᎐474. w13x S. Wilson, A.T. Alpas, Surf. Coat. Technol. 120-121 Ž1999. 519᎐527. w14x S.J. Bull, Tribol. Int. 30 Ž1997. 491᎐498. w15x J. Valli, J. Palojarvi, U. Makela, Measurement of Coating Thickness by Using a Ball Cratering Method, Technical Research Centre of Finland, report no. 435, 1985. w16x E. Breslauer, T. Troczynski, J. Adhesion Sci. Tech. 12 Ž1998. 367᎐382. w17x W. Hirst, J.K. Lancaster, J. Appl. Phys 27 Ž1956. 1057᎐1065. w18x D.A. Rigney, Annu. Rev. Mater. Sci. 18 Ž1988. 141᎐163.

396


w19x N. Chen, K. Adachi, K. Kato, Trans. Jpn. Soc. Mech. Eng. 61 Ž1995. 1605᎐1612. w20x S.J. Bull, R. Kingswell, K.T. Scott, Surf. Coat. Technol. 82 Ž1996. 218᎐225. w21x Z.P. Huang, Y. Sun, T. Bell, Wear 173 Ž1994. 13᎐20. w22x J.F. Archard, W. Hirst, Proc. R. Soc. Lond. 236A Ž1956. 392᎐410. w23x S.J. Bull, D.S. Rickerby, A. Matthews, A. Leyland, A.R. Pace, J. Valli, Surf. Coat. Technol. 36 Ž1988. 503᎐517.

w24x w25x w26x w27x

S.J. Bull, D.S. Rickerby, Thin Solid Films 181 Ž1989. 17᎐21. J.F. Archard, J. Appl. Phys. 24 Ž1953. 981᎐988. B.R. Lawn, A.G. Evans, J. Mater. Sci. 12 Ž1977. 2195᎐2199. K.-H. Zum Gahr, ‘Microstructure and wear of materials’, Tribology Series, Elsevier, Amsterdam, 1987, vol. 10. w28x L.C. Erickson, R. Westergard, U. Wiklund, N. Axen, H.M. Hawthorne, S. Hogmark, Wear 212 Ž1998. 30᎐37.