Wear 217 (1998) 175-181
Wear-life diagram of TiN-coated steels Young-Ze
L e e ~'*, K i - H u n J e o n g b
J St'hool of Mechanical Engineering, Sung Kyun Kwan Universi~.. Suwon. KynngGi.l~J. 440-746. Sonth Korea h Graduate Sclu~d of Mechunical Engineering. Sung Kyun gwan Unirer.~ity. Suwnn. K~'nngGi-Do. 440-746, South Korea
Received27 November 1997:accepted 12 March 1998
Abstract This paper will present the results of indentation, one pass .,¢ratch and the repeated sliding tests to detennine the critical load and the wearlife of TiN coated specimens. The sliding tests of cone(diamond)-on-di~ type were performed under dry sliding conditions as well as indentation and .~ratch. The test specimens were 0.2% plain carbon steels. The coefficient of friction and the wear cycles to failures were measured with different coating thickness, substrate hardness and roughness. From the ~ratch te.'.¢,~it w ~ confirmed thai the critical increased with the hardness of the substrate, furthermore the critical loads ate inerea.~d with the coating thickness ~ ~ with inereasing surface roughness. Using the percent contact load, the ratio of sliding normal load to the critical scratch load, the cycles to failure are measured to predict the wear-life of TiN film. On the wear-life diagram, the percent contact loads and the cycles to failure ~ the good linear relation on log coordinate. With a decreasing load percentage, the diagram shows that the wear-limiLs,at which the coated.qeels survive more than 4000 cycie.~,are about 6-10% of the critical ~ratch Ioa&s. © 1998 Elsevier Science $.A. All rights re~rved. Keywords: TiN film:Wear-lifediagram:Criticalscralchload: Wcar-fimit:Coatings:Surface roughness:Titaniumnitride:Tribo]ogy
l, Introduction With good resistance to high temperature, wear and corrosion, TiN coatings have found their uses in applications such as cutting tools, bearing spindles and many mechanical components [ 1-31. Various materials are u ~ d as coatings on all kinds of suhstrates and in different coating proces~s [ 4 I. This remarkable step forward is a result of the improvement in the coating process and in the control of the coated substances. Whatever its function, the expected performance of the coated products can he only achieved if the adhesion and the intrinsic cohesion of the coating are sufficient. The bonding strength of a coating with a substrate is of vital importance lbr coated componcnts. In the past, many methods [ 1,5,61 based on indentation have been developed for assessment of the bonding strength of coatings. Those tests were done on the base of indentation and .scratch, in which the load was applied just once on the specimen. Test results tell us just wear mode or wear mechanism of each coated specimen. However, there are only few guidelines on the design, and more importantly the evaluation methods on wear-life of the coated components under repeated sliding 17 I. *Corresponding author. Fax: +82-331-290-5276: e-mail: yzlec(~ yudm.skku.ac.kr 0043-16~8/98/$19.00 © 1998EL,;x~vicrScienceS.A. All rights rc.~rvcd. PII S0043-1648 ( 98 ) 00185-9
This paper wesents the results of repeated sliding tests performed on TiN coatings obtained by catbodic arc ion plating on steels, as well as the results of.scratch and indentation tests. For a practical application, the cycles to failures with various loads are u ~ d to construct the wear-life diagram as the S--N curve in the failure analysis. The main objects of the paper are: I. To compare the results of convcntioual coating tests such as indentation and one pass scratch with those o f r e ~ sliding test. 2. To determine the effects of coating thickness and the substrate conditions like hardness and roughne~ on wear-life of TiN coatings. 3. To make a plot of the wear-life diagram and find the endurance limits, called wear-limiL~, of coatings under repeated sliding.
2. Experimental details
In the present work, indentation. ~ratch and repeated sliding tests were carried out to measure the failure modes, the critical loads and the failure times of TiN coated specimens.
Y.Lee, K. Jetmg/ Wcar217 (1998) 175-181
2. I. Specimen
Normal loading rate 100 Nlmln
The substrates of the test specimens were made of 0.2% plain carbon steels with different hardness, roughness and coating thickness as shown in Table I. The specimens were heat treated to produce three different hardness and polished to get three different surface roughness, such as 0.05, 0.1 and 0.25 p.m in center-line average roughness. Ra. On the steel substrates, TiN coatings were applied using the cathodic arc ion plating, which process allowed mechanically resistant materials to he deposited at relatively low temperatures and makes them very attractive with increasing bias voltage and high ionization ratio [ 8,9 I. Depending on actual deposition times the coating thickness were changed to 1.0, 2.5 and 4.0 p.m. The carbon steel surfaces were precleaned through oil degreasing followed by alkaline cleaning; and then argon sputter cleaning was carded out for about 2 min prior to coating. 2.2. Indentation test
In order to examine the possible fracture pattern of TiN coating in indentation and to evaluate the critical indentation load for coating removal, static indentation tests were carded out using a micro-Vickers hardness tester with the pyramidical diamond indenter. The tests were repeated several times to get the averaged values of critical loads, at which the part of the coating was spalled out. Experimental range of the normal load was 5 N-50 N and the indentation time was 30 s. After indentation, the contact surfaces were observed by the optical microscope. 2.3. Scratch test
The bonding strength was evaluated through the scratch test methods. The scratch test apparatus was designed and made in the laboratory. A ,schematic diagram of the apparatus is shown in Fig. I. The frictional forces were measured using the load cell. A microphone was attached near the diamond stylus to detect the acoustic signals emitted from the film fracture. The diamond stylus used in the unit was in the form Table I Test conditionsfi)rthe TiN ctrltcd Mccls
Tesls ¢,n substratc hardness Tc~I~on ~.uh~tr-.~tc roughnc~.~. TCM~(m fi¢)ating thickness
TiN coaling thickness
6 Hu, 38 H~, 58 Hm6 Hx~
o. i / ~ m ().')5 /ZlIi n.l/~m
l .()/J.m 2.5 p.m 4.0/zm
Ftic fional force 0 . 2 R D i a m o n rt cone ~ . _
Bending type Ioadcell for fflctlonalforce 100 Nlmin
I ~ L J
Scra fching speed 10 ram/ram
Fig. I. Schematic diagram o f scratch test.
of a cone having a spherical tip with a 0.2 mm radius. The scratch speed was maintained at 10 mm/min. Specimens were scratched while increasing the normal force at a rate of 100 N/min. The applied load was increased continuously until the coating could not follow the deformation of the substrata. This resulted in failure of the coating. The smallest load, called the critical scratch load, at which the coating was damaged, was determined by on-line measurement of friction and acoustic signals, and confirmed by optical microscope. The indicated critical scratch loads were obtained by averaging the values of more than three different scratches. 2.4. Sliding test
The sliding tests were carried out using a specially designed cone-on-disc tribometer for repeated sliding, which was capable of measuring the frictional forces. To keep the stable rotation, a servo-motor was used to rotate the shaft. The test rig used is shown schematically in Fig. 2. The top specimen of a diamond cone with a 0.2 mm tip radius was located in a holder which was clamped to a fixed arm with transducers for friction force. The lower specimen, a coated disc with 50 mm diameter, was mounted on a rotating shaft. After ultrasonic cleaning with acetone, the specimen were mounted on the holders and brought precisely into contact without initial indenting. Contact was achieved by pressing the cone against the flat surface of the coated specimen under a normal load applied by dead weights. The advantage of using diamond cone of Rockwell hardness tester as one counter body in all sliding is that they are manufactured for commercial purpo~s, so that little variation occu~ in the size Nonnalload
. Loadce I Jndmatot
Riction force ~
_. 02R L~amono cone
dir~.ction Fig. 2. Sl:hcmati¢ diagram of rCwalcd sliding test,
Y. Lee, K, Jeon~ I Wear 217 t 199,";) 17.5-181
and the mechanical properties. Also. the diamond slid on coated steel avoided the formation of transfer film, .so called "lubricious oxides' [ 10,11 ], which could be formed on surfaces in case of sliding with steel or ceramic hall. The sliding tests were carried out under ambient conditions. New diamonds were used on each specimen to reduce the effect of tip radius changes due to wear. A slow rotating speed of 0.055 m / s ( 35 rpm) to avoid the frictional heating was used in all tests. To construct the wear-life diagram, we defined the percent contact load, which was the ratio of the sliding load to the critical scratch load. At each percent contact load the cycle to failure was measured with an indication of sudden changes in the coefficients of friction and also confirmed by optical microscope. The tests were repeated several time to get an average value of cycle to failure.
Nomml l.ud, N
3. E x p e r i m e n t a l results a n d d i ~ . n d o n s
Fig. 5. The ~:hangcof m:oustlc emission signals and frictional Forceswilh scralch normal I~rads.
3. I. hldentation test spalling had occurred. For He(- = B. the inner-rectangle cracks were generated at a load o f 5 N a.s shown in Fig. 4a. and the outer-rectangle cracks were generated at a load o f i 0 N. A t
Fig. 3 shows the fracture patterns of the TiN coating after indentation with a pyramidical diamond indenter within 30 s after indentation. The contact surfaces were observed by the optical microscope, it can he seen that inner-rectangle cracks [ 121 are generated at a light normal load. At higher loads. the formation and the growth of outer-rectangle cracks spalled out part of the coatings. Fig. 4 shows the effects of substrate hardness, and the coating thickness on the fracture patterns after indentation. The critical load was defined as the load at which the partial
NoCmck Inner"Rlclgngu~Ir Ouler4~Iclan0ular Crack Crack
30 N load. the part of the coating outside the contact area was spalled out due to the propagation of outer-rectangle cracks. Fig. 4a shows the effect of hardness on the fracture patterns and the critical loads, which were greatesl at Hac = 6. It seems that the deformation of.soft substrate reduced the crack propagation. In Fig. 4b. the smooth substrate shows the lower critical loads for partial spalling. Fig. 4(: shows that thecritical normal loads for the transition of fracture patterns are .strongly affected by the coating thickness. Of the special impormnee is the fact that the critical load for the generation of partial spalling shows a m:aimum at a certain coating thickness similar to the test results of Diao et al. I 13 I. 3.2. Scratch test
As shown in Fig. 5. the TiN coating failed with increasing loads. The coefficients of friction also transited from stable
Fig. 3. Propagationand notationof crack initialionm(xlcin indcnlalion, 6O
l t, ;
• JA II S A E:] L$ o
o o o
Rockweg Hardnesl Number, Hac (a) Tests on hardn¢l~
Rou0hne~ Ra(prn) (b) Tests on rn~ghncm;
Coaling Alickne~lLpm (c) Tests o n coating thickness
Fig. 4. Varialion of crack nll~]¢5and crilical loads from indcnlation lcsls.
)~ l.ee, K. ,h'otIg / Wear 217 ~19981 175-181
3O i ~ughne=R=) o1.111 i co=,~ ~,ckness25vm
Hafane= 6 RougnneSl(Ra) 01;,m
Hardne=l. 6H=¢ Coaling "~lCkneSs 25;ml Z
o, -.. =
ol O2 Roughne~ Ra(pm)
o~ . . . . . . . . 2 Goaung ~,cknes
(c) Tests on coating thickness
(b) Tests on Substrate roughness
(a) Tests On substrate hardness
Fig, 6. The resultsof scratchtest. Previous studies [ 14,15 ] indicated that the critical scratch load increases linearly with the hardness oftbe substrate. This is conlirmed by this work. as seen in Fig, Oa. The present work also shows that the critical load increases with the coating thickness and decreases with increasing surface roughness. All tests were done more than three times to get the averaged values of critical loads. The critical loads obtained on 2.5 /.tm coating on steel of HR(-= 58 are nevertheless superior to those obtained on 4 /xm coating on steel of HR(, = 6. It must be pointed out, h o w e v e r , that the critical loads ( Fig.
Fig. 7, Formationof adsorl~d filmon TiN coatingusing aluminaball.
6) measured using scratch test showed different trends from the results of the indentation test in Fig. 4. It seems that the indentation test using just normal load would be inadequate to determine the bonding strength of the coating, but the scratch test using both normal load and frictional force should be used to determine the critical loads of the coatings.
low values to unstable high values as increased loads. The smallest load. at which the coating was damaged, was determined by the transition of the acoustic signals and conlirmed by frictional forces and optical microscope, because the acoustic signals were more sensitive than the frictional forces. which showed the transition a little bit later. For the present work, there was an excellent correlation between the onset of the acoustic emission signal and the microscopic observation of the first occurring TiN coating damage.
3.3. Sliding test a n d u'ear-I[fe diagrtml With the slider of the diamond cone, the repeated sliding tests w e r e carried out to lind the cycles to failures of coated
I. Before failure
i o 0
Number of Cycle. N Fig. 8. "l'xpk-ul IriL-liuual tran~iti.n ~d TiN I In) o ~ d ng
III. After failure
Y. Lee. K. Je,ng / Wear 217 ¢199H~175-181
!~0 L1 1
Fig. 9. Partial failure of coating on transition peri(KI.
"-o."--~' " ~ r~ .~ RDughnel RB
" ' - ~'4¢~ ~'~.j
,,,,10,3N o O.25jl, 1O.IN ~0J
steels, which also had three different kinds of hardness and roughness of substrate, and coating thickness like the tests of indentation and scratch. Because the diamond cone reduced the formation of adsorbed films on sliding surfaces, the life cycle of the coating could be evaluated on each specimen without the effect of the adsorbed film 1121. Fig. 7 shows adsorbed film on coating when using the alumina ball as a slider. In this case. the friction and wear occu~ between the alumina and adsorbed film. not TiN coating, As shown in Fig. g. the coefficients of friction transit from low values to alternating high values with repeated sliding cycles. The friction cycle was determined to be 850 cycles at the transition between wear-life and transition period. The friction diamond and TiN coated steel was about half of that between diamond and substrate, on which the coating was completely removed alter 3500 cycles. Also. it can be seen that transition period lasts for more than ~veral thousands cycles before the total failure of coating. Fig, 9 shows the partial coating failure on wear track during the transition period. The repeated contacts with abrasive asperities resulted in fatigue wear. An inclusion or void in the interlace .serves as a local stress concentration for crack nucleation, When subsurface cracks grow to a critical size. or link up to produce a critical size. a tiny spall will he ejected from the hearing
1 lO IOO lO0O 1000o Cyclesto failure Fig. I I. Wear-lil~: diagram (ffTiN ¢o',,ings with difl'crcnl coaling subslrelc roughne~.~.
surfaces. Eventually, spalled areas may grow large enough to he easily ob~rvable.
3.3. I. Wear-life diagrams with different £'~mting thickne,,s The wear-life diagram on specimens with different coating thickness were constructed to .see the failure cycles at each sliding load. As ~ e n in Fig. IOa. the sliding load and wearlife shows a good linear relation on log coordinate. It can be .seen that there are wear-limits similar to the endurance limits of S-N curve in fatigue tests, for each specimen to survive more than 4000 cycles. Once the surface slide over 4(100 cycles, the coated surface survive more than I0 h (21.000 cycles) with the low values of sliding loads. To reduce the test times the wear limits are defined at the cycle of 4000. For t = 4 . 0 ~ m of coating thickness, the cycles to failures were the longest among the other coating thickness at each load. The thin coatings show very poor lives of wear. Eventhough the specimens with intermediate coating thickness had only 63% of the critical ~ m t c h load of thick coating, the wearlives of thin coating was almost comparable to t h o ~ of thick
"~ ," -¢'.',~
"'--''~' Not failed
O25 ,~ O 4 ,~ 10 100 t00O Cyclesto failure
Coating ~icknes. C ntical efatch load
"~'~..-" "'J Nol faint
o 2.5~,. 10.3N 0 4 ~ 16.3N 10 100 I000 Cyclesto failure
tb) 06zgrem using the percent ~ ta) Otag~ram~ the sliding ~ load Fig. It). Wear-lift: diagram o r T i N ¢(~atings with ditt;Jrcnl coating thickness.
Y. Lee. K. Jeung l Wear 217 (199bl) 175-181
Harass. ~c zx6 D3B O58
10o 1000 cycles to failure
Hardness, I-~ r Critical. . . . . . . ~ 6,10.3N 038.14.1N o 5B. 203N
-~,-,,, ",,, ,.~, ",,, ,..,
(a) Diagram uling the sliding normal load
...... ~ / N , fa';';ecl
10 100 1000 10000 Cycles to failure (b) Olagcam using the percent contact load
Fig. 12. Wear-lil~:diagramof TiN coatingswith differentcoatingsubstratehardness. coating. Fig. lOb was reconstructed using the percent contact load, the ratio of sliding load to the critical scratch load. The critical loads on wear-limits were about 6-9% that of the critical scratch load of one-pass sliding. 3.3.2. Wear-life diagrams with different substrate roughness The effect of substrate roughness on wear-life is seen in Fig. I I. Wear-limits are about 6-9c~ of the critical scratch loads depending on surface roughness. Surfaces with smooth and intermediate roughness shows higher wear lives than rough surfaces at each percent contact load. The substrate with the intermediate surface roughness i R a = 0 . 1 /,tm) shows highest wear-limit to survive very well more than 4000 cycles, at which the wear limit was about 9% at the critical scratch load. The coating on rough substrate shows lowest wear resistance on repeated sliding similar to the scratch tests ( F; 2. 6b ). 3.3.3. Wear-life diagrams with different substrate hardness Using coated specimens with different sub::trate hardness. the results are shown in Fig. 12a, which represents good linear relation between sliding load and cycles to failures. Eventhough the critical scratch load of Hkc = 6 specimens was half that of HRc-= 58 specimens, the surfaces with HRt. = 6 and H R c = 5 8 shows very similar wear lives at each sliding load. but the harder one has a higher wear-limit. While the critical scratch loads increased with increased hardness as .seen in Fig. 6c. the repeated sliding tests show that the surfaces with intermediate hardness have lower wear lives and higher wear-limits than soft surfaces. Fig. 12b was reconstructed using the percent contact load. The wear-limits existed in the range of 6--10% of the critical scratch loads.
4. Conclusions To predict the wear-life of TiN coating, indentation. scratch and repeated sliding tests were carried out using diamond indenter. The following results were obtained.
( I ) On the wear-life diagram the percent contact loads and the cycles to failure show a good linear relation on log coordinate. With decreasing load percentage, the diagram shows the wear-limit at which the coated discs survive more than 4000 cycles. From the wear-life diagram it is possible to predict the failure cycles of repeated sliding components. (2) The wear-limits of TiN coated specimens are about 6 10% of the critical scratch loads. Using the wear-life diagram, it is possible to predict the safe sliding conditions of the coated specimens. (3) From the scratch tests it is confirmed that the critical loads are increased with the hardness of the substrate, but it is seen that the critical loads are increased with the coating thickness and decreased with increasing surface roughness. (4) The indentation test a using normal load would be inadequate to determine the bonding strength of the coating, but the scratch test using both normal load and frictional force should be used to determine the critical loads of the coatings under sliding.
Acknowledgements The authors are grateful f o r the support provided by a grant from the Korea Science & Engineering Foundation (KOSEF) and the Safety & Structural Integrity Research Center at the Sung Kyun Kwan University.
References I i I E. Posti, I. Nieminen. lniluencc of coaling thickness on the lifeoi" TiN-coated high speed cutting ttmls.Wear i29 (1989) 273-283. [ 21 F.E. Kennedy, L. Tang, Factors afl~cling the sliding peril)finance of titanium nitridc coalings, Mechanics of Coatings. Elsevier i 1989) pp. 409-415. 131 Y.G. Yeong. F.L. Jen. C.F. AI, The Irihologicalcharacteristicsof lilaniumnhridecoatings.Wear 194 (1996) 12-21.
14] K. Holml~rg. A. Mallhews.Coaling tribology: a concept, critical aspectsand filturedirections,Thin Solid Films253 11994) 173-178.
Y. Lee, K. Jeong/Weur217(19987 175-181
15 [ HJ. Boving. H.E. Hintermann, TiC-coalnd ball bearing for spin-axis gym application. ASLE Trans. 37 ( 1981 ) 534-537. 16[ P.A. Steimann. H.E. Hin:ermann. Adhesion of TiC and TH C.N ) coatings on steeh J. Vac. Sci. Technol. A 3 ( 1991) 2394-2400. 171 K.C Ludema. Friction, Wear. Lubricalion. CRC Press (1996) pp. 238-239. I g I S.G. Oh. J.G. Han. The effects of precoating and bias voltage on the adhesion of reactive arc evaporated TiN and ZrN coatings. J. V~. Sci. Technoh A 9 ( 1991) 2557-2562. 191 B. Bhushau. B.K. Gupta. Handbook of Tribology. McGraw-Hifi ( 1991 ) pp. 9.43-9.45. ] 10l I.L. Singer. S. Fayeulle. P.D. Ehni, Friction and wear hehavior of TiN in air: the chemistry of transfer films and debris formation. Wear 149 ( 1991 ) 375-38,;, I I I J E. Vancoille. J.P. Cefis, J.R. Roos. Dry sliding wear of TiN based ternary PVD coatings. Wear 165 (1993) 41-49. I 121 K. Jeongo Y. Lee. A study on friction and wear of TiN film for the wear-life prcdicfion. J. KSTLE 13 (1997) 28-32. 113l D.F. Diao. K. Kato. K. Hokkirigawa. Fr~ture mcchanisms of ceramic coatings in indentalion. J. Tribe1.116 (1994) 860-869. 1141 AJ. Perry. Scratch adhesion testing: a critique. Su~. Eng. 2 (1986) [151 R. Rezakhanlon. J. Vonstebut. Damage mechanisms of hard coafing~ on hard substrate: a critical analysis of failure in scratch and wear
testings. Proc. of the 16th Leeds-Lyon Syrup. on Tribology (1989) pp. 183-192..
Y o n n g - Z e Lee: graduated in mechanical engineering from Seon! National University in 1983 and w a s a w a r d e d a P h D degree from the U n i v e ~ i t y o f Michigan in 1990 under the direction o f Dr. L n d e m a . At presenU he is an associate prof e s ~ r in the ~ h o o l o f mechanical engineering. Sung Kyun K w a n University. His research interest include the coating tribology, the scuffing m e c h a n i s m , and the d e v e l o p m e n t o f sintered metal bearings and air bearings. Ki-Hun Jeong: graduated in mechanical engineering from Sung Kyun University in 1995 and just finished a m a s t e r ' s degree from the ~ m e university. Currently he is a researcher at L G Industrial Electronics in Korea.