Adhesion and abrasive wear resistance of TiN deposited on electrical discharge machined WC–Co cemented carbides

Adhesion and abrasive wear resistance of TiN deposited on electrical discharge machined WC–Co cemented carbides

Available online at Wear 265 (2008) 490–496 Adhesion and abrasive wear resistance of TiN deposited on electrical discharge mac...

1MB Sizes 8 Downloads 49 Views

Available online at

Wear 265 (2008) 490–496

Adhesion and abrasive wear resistance of TiN deposited on electrical discharge machined WC–Co cemented carbides B. Casas a , U. Wiklund b , S. Hogmark b , L. Llanes a,∗ a

Departament de Ci`encia dels Materials i Enginyer´ıa Metall´urgica, ETSEIB, Universitat Polit`ecnica de Catalunya, Avda Diagonal 647, 08028 Barcelona, Spain b Uppsala University, The Angstr¨ ˚ om Laboratory, Tribomaterials Group, SE-751 21 Uppsala, Sweden Received 20 October 2006; received in revised form 15 October 2007; accepted 30 November 2007 Available online 25 January 2008

Abstract Electrical discharge machining (EDM) is a non-traditional machining method extensively used to manufacture complex geometries of hard and brittle materials such as WC–Co cemented carbides (CC). Although the thermal action of the EDM process is known to yield a relatively poor surface integrity in these materials, it may be minimized through the implementation of multi-step sequential EDM and post-EDM surface treatments. Particularly, hard coating application has been demonstrated to be effective for decreasing the EDM-induced mechanical degradation. However, additional studies are required on such coating–EDMed substrate systems to determine other crucial properties in terms of applications, e.g. adhesion and micro-scale wear behaviour. In this work the adhesion strength and the microabrasive wear resistance of TiN deposited on EDMed substrates have been evaluated by means of scratch and crater grinder testing, respectively. The results indicate that both critical load for decohesion of the coating from the substrate and coating specific wear rate increase with finer-executed EDM, reaching values close to those measured for a TiN coating deposited on a ground and polished substrate. © 2007 Elsevier B.V. All rights reserved. Keywords: Adhesion strength; Microabrasive wear resistance; Electrical discharge machining; Cemented carbides; TiN coating

1. Introduction High-dimensional accuracy combined with complex geometries are frequent requirements in applications involving WC–Co cemented carbides (CC), materials also commonly referred to as hard metals. The elevated hardness, and even more important, the intrinsic brittleness that CC exhibit results in significant inconvenience and high cost for ensuring close tolerances and surface finish under traditional grinding techniques. Electrical discharge machining (EDM) turns out to be a very good alternative for shaping CC because it is a thermal machining process, i.e. it does not involve any physical contact between the processed material and the electrode. EDM of CC has proven to be a satisfactory technique in terms of machining performance indexes such as material removal rate or surface roughness [1–5]. On the other hand, EDM is also known to yield a thermally affected

Corresponding author. Tel.: +34 934011083; fax: +34 934016706. E-mail address: [email protected] (L. Llanes).

0043-1648/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.wear.2007.11.016

zone just beneath the shaped surface, with poor surface properties such as unfavourable residual stresses, cracks and craters [1–10]. In recent years, this shortcoming has been drastically reduced by the implementation of multi-step sequential EDM and post-EDM surface treatments [8,10–12]. Particularly, the implementation of surface modification technologies, such as physical vapour deposition (PVD) of hard coatings, have been demonstrated to markedly decrease the negative effect of EDM on the mechanical strength of CC [12]. However, information regarding other critical contact-related properties, such as adhesion strength and wear resistance, of PVD coatings over EDMed CC does not exist in open literature. Considering that these material systems are widely used in applications involving extreme mechanical contacts, such knowledge is decisive for their effective implementation as tools, wear resistant parts and structural components. It is then the aim of this work to evaluate the adhesion strength and the microabrasive wear resistance, by means of scratch and crater grinder testing, respectively, of PVD TiN coatings deposited on CC substrates shaped by both EDM and conventional machining.

B. Casas et al. / Wear 265 (2008) 490–496

2. Experimental procedure The base material studied was a commercial fine-grained WC–10 wt% Co cemented carbide grade produced by DURIT Metalurgia Portuguesa do Tungst´enio. Longitudinal sections of rectangular bars (3 mm × 4 mm × 45 mm dimensions) were shaped by EDM. Sequential EDM yielded four different surface variants for each grade. A wire cut EDM machine equipped with an advanced pulse-type generator (Model ROBOFIL 2020SI, Charmilles Technologies) was employed. The wire chemical composition was Cu–37 wt% Zn and its diameter was 0.25 mm. The control unit was used either to keep constant the working voltage between the CC sample and the wire or to limit maximum cutting rate. Control based on imposed voltage was applied for rough and finish shaping operations (conditions A and B). Under this set-out, working voltage values ranged from 15 to 70 V. Aiming to optimize surface condition rather than precision or machining time, a limiting maximum cutting rate was used as control parameter for surface finish sequences (conditions C and D). In this regime, maximum cutting rates were between 1.2 and 2 mm/min and multi-step machining was performed for attaining the corresponding surface conditions. For comparison purposes, one reference surface finish condition was produced by conventional mechanical abrasion. This variant, referred to as P, resulted from a two-step grinding through 200 grit (68 ␮m) and 600 grit (30 ␮m) diamond-containing discs, followed by polishing, ultimately with 3 ␮m diamond paste, up to optical finish. Titanium nitride (TiN) films were deposited on all substrates following Balzers’s PVD arc ion plating process. Microstructure, roughness and the amount of surface deterioration were determined by means of scanning electron microscopy (SEM) on the deposited films, coating/substrate interfaces and substrate subsurfaces. The nomenclature used in this work, the operations used for substrate surface preparation and the resulting surface roughness for each specimen are presented in Table 1. The adhesion of a coating to the substrate is an important property of the film/substrate system. It was evaluated by scratch testing and is given in terms of the normal force required to cause any exposure of the substrate, i.e. adhesive failure. Scratch tests were performed using a commercial equipment (Model Revetest, CSEM Instruments) with a detector for acoustic emission (AE). During scratch testing a 200 ␮m radius Rockwell C diamond stylus was drawn across the coated surfaces under a normal load continuously increased up to 100 N. Five experiments were carried out for each coating/substrate system, using a


loading rate of 10 N/mm. The normal load and AE were continuously recorded during scratching. Critical loads for adhesive failures of the coatings were determined by the change in the AE–normal load curve as well as by postoptical and scanning electron microscope examination. The wear assessment was focused on evaluating the influence of EDM-induced surface defects on the intrinsic abrasion wear rate of the subsequently deposited TiN film. Thus, the crater grinder method, a technique widely used to assess the wear properties of thin (about 1–10 ␮m) hard coatings since the early 1990s [13–22] and under continuous test methodology refinement [23–29], was selected. In this study this technique was implemented by using a dimple grinder (Model 656 Precision Dimple Grinder, Gatan), commonly employed for conventional transmission electron microscopy specimen preparation. The equipment consists of a stainless steel grinding wheel with a diameter of 20 mm and a thickness of 2 mm, which rotates about a horizontal axis. During test, the wheel grinds the specimen, which in turn rotates about a vertical axis. This generates wear craters in the surfaces of the coating–substrate composites (Fig. 1). The load and the rotational speed of the grinding wheel were set to 20 g and 375 rpm, respectively. Approximately 0.1 ml of Hyprez Liquid Diamond (Type K), a commercial diamond slurry of an unstated standard concentration with 2.5 ␮m particles, was applied directly on the specimen surface. The grinding was interrupted at regular intervals of 100 revolutions (a sliding length of about 12 m) and the crater diameters were measured using an optical microscope with an accuracy of ±5 ␮m. For all EDM conditions, and particularly for those corresponding to rough cutting and fine shaping, inner crater boundaries (i.e. at the interface between substrate and coating) were not as well defined as for the P case. Hence, crater size throughout the study was assessed as the mean value of at least eight measurements taken all around the craters. These mean values were then used for calculation of the coating and substrate wear volumes. In doing so, and attempting to account for equipment specific deviations from ideal test conditions (particularly misalignment of axes), the depth of worn craters was measured using a white light optical profilometer and the corresponding value then introduced as experimental input for evaluation of the real crater shape (with

Table 1 Specific shaping operations involved, nomenclature and resulting surface roughness for each surface finish conditions investigated Condition


Machining operations

Rough wire-cut EDM As A + fine shaping As B + surface finishing As C + surface microfinishing Grinding + polishing with diamond

Roughness (␮m) Ra


3.75 1.33 0.41 0.11 0.01

20.17 8.3 2.84 0.88 0.07

Fig. 1. Schematic drawing of the crater grinder test, as implemented in this study.


B. Casas et al. / Wear 265 (2008) 490–496

a smaller depth than the one apparent from the measured crater diameter) [23]. At least three experiments were carried out for each coating/substrate system. Finally, an inspection of the test craters in the SEM was conducted in order to reveal any coating or substrate defects as well as prevalent abrasion mechanisms. 3. Results and discussion 3.1. Surface quality of coated and uncoated EDMed substrates Surface defects associated with WC–Co substrates shaped by EDM have been reported in detail elsewhere [8,10]. EDM machining of CC results in local melting of the workpiece surface followed by rapid solidification. Typically a quite homogeneous network of microcracks is formed, extending both parallel and perpendicular to the surface [3,6]. In this study the penetration depth of the cracks varied between 5 and 15 ␮m depending on the EDM parameters. In particular, the rough wire EDM (A) generated a 10–15 ␮m thick, homogeneously deposited, brass layer on the shaped surface from the wire used as tool electrode (Fig. 2). Deposition of electrode material was not observed for the other EDMed surfaces. An extensive microstructural and mechanical characterization of TiN coatings deposited on the CC substrates shaped by EDM has been reported earlier [12]. It shows the formation of a uniform, about 3 ␮m thick, TiN layer with dense and fine-grained columnar structure. No significant morphological difference was associated with the different substrate surface preparations, as shown in Fig. 3. Cross-sections were examined to investigate EDM effects on surface roughness, a key parameter for the adhesion strength of the coating/substrate interface. From Fig. 3, it is clear that the EDMed surface consists of uniformly distributed microcraters whose depth and size is reduced as EDM is more finely executed. This is in agreement with the roughness values reported in Table 1. As a consequence, the interface for all EDM-related Fig. 3. Polished cross-sections of TiN coatings deposited onto WC–Co substrates machined with different EDM sequences (A, B, C and D surface variants).

surface conditions exhibits a variable and discrete tortuousness, different from the flat profile observed for the control condition P. 3.2. Adhesion strength

Fig. 2. Polished cross-section of surface with the A condition. It exhibits a 10–15 ␮m thick brass layer, homogeneously deposited from the wire used as electrode during the EDM process.

Critical normal loads for adhesive failure of the coating during scratch tests are shown in Table 2. These loads were determined by the change in the AE versus normal load curve as well as by postoptical microscope examination. Experimental results indicate that coating adhesion increases with finer-executed EDM, reaching high values similar to that exhibited by the TiN coating deposited on the ground and polished condition. Considering the general rule of thumb that a critical load of 30 N in scratch testing with a Rockwell C diamond tip is generally sufficient for tooling applications [30], the adhesion strength

B. Casas et al. / Wear 265 (2008) 490–496 Table 2 Critical normal loads for adhesive failures of TiN coatings deposited onto different surface variants, determined using the acoustic emission vs. normal load in curves from scratch tests as well as from postoptical and scanning electron microscope examination Condition

Pcrit (N) TiN


4 29 50 78 74

± ± ± ± ±

1 5 1 3 4

exhibited by surface conditions B, C and D should be described as satisfactory towards excellent as sequential and multi-step EDM is further implemented. Regarding dominant failure mechanisms (Fig. 4), relative changes were discerned depending on the surface quality of the substrates. On one hand, systems with surface microcracked CC tended to exhibit early substrate exposure in terms of small chipped or spalled regions along the track edges (e.g. Fig. 4, A and C conditions). On the other one, on substrates whose surface is almost free of microcracks or without damage (i.e. D and P conditions, respectively) adhesive failure is only evidenced as a complete detachment of the coating at relatively high loads, usually initiated at through-thickness cracks (cohesive failure) developed inside the track. Based on these experimental findings and taking into account the “discretely uneven” character evidenced at the interface, the relatively different scratch resistance assessed for the EDM surface variants is speculated to be a consequence of the effective role played by the substrate microcraters on providing mechanical interlocking at the


coating–substrate interface, as opposite to the detrimental influence that they may invoke in terms of existence of intrinsic interfacial flaws and/or stress concentrators at geometrical discontinuities [31–33] and effective stiffness of the load-bearing substrates. For the particular case of the A condition, the fact that coating delamination was already observed at loads as low as 4 N is rather due to the soft brass layer between the coating and the CC substrate. Qu et al. [9] recently characterized the mechanical properties of surface layers of EDMed CC by nanoindentation. They concluded that Cu and Zn from the brass wire electrode, oxides of Co and Cu as a result of the high temperature involved, and thermal cracks and bubbles all reduce the hardness and modulus of elasticity of the recast layer. Thus, removal of any brass-containing recast layer resulting from rough EDM, for instance by subsequent finer EDM operations, is expected to yield better surface quality and adhesion strength. This is also the observation of this study. 3.3. Microabrasion wear resistance The theoretical basis of the crater grinder test is Archard’s classical wear equation for homogeneous materials [13,17]: V = kL S


where V is the wear volume, S is the sliding distance, k is the (load) specific wear rate and L is the applied load. At the beginning of the microabrasion test, only the coating is worn and Archard’s equation can be used directly. However, as soon as the coating is worn through, the coating and substrate are worn simultaneously and the equation has to be modified to give the intrinsic wear rate of the individual materials. In this case, the

Fig. 4. Coating failure scenario during scratch testing (scratching direction: from left to right) for surface conditions A, C, D and P. Images are focused on showing adhesive failure modes although some cohesive ones are also evidenced.


B. Casas et al. / Wear 265 (2008) 490–496

total load (L) and total wear volume (V) may be distributed on the two materials assuming that the wear of one material does not influence the wear of the other, except through load distribution, as follows: L = L c + Ls


V = Vc + Vs


The subscripts s and c refer to the substrate and coating materials, respectively. For constant loads, the wear equation for the two materials can thus be written as Vc = kc Lc S


Vs = ks Ls S


Combining Eqs. (2)–(5), a general formulation of the wear equation for two materials, is obtained: Vc Vs + = SL kc ks or   Vs Vc = kc SL − ks



The specific wear rate of the substrates (ks ) with different surface finish conditions was determined in a specific microabrasion test on uncoated specimens. Wear volumes of the substrate and the coating (Vs and Vc ) were calculated assuming a central flat bottom crater geometry and a constant crater diameter deviation (misfit) from the ideal spherical shape, the latter being evaluated from the measured values of real crater depth and diameter [23]. The wear constants of the coatings were then derived as the slope of the straight-line plot of Vc versus (SL–Vs /ks ). Finally,

Table 3 Calculated specific microabrasion wear rate for uncoated substrates (ks ) and coatings (kc ) Surface condition

ks (␮m3 /(N mm)) Uncoated substrates


468 425 460 459 464

± ± ± ± ±

14 31 16 38 14

kc (␮m3 /(N mm)) TiN 606 258 214 208 175

± ± ± ± ±

21 14 16 19 11

coating failure mechanisms in the microabrasion wear test were determined using SEM. Experimentally determined wear coefficient values of coatings, deposited onto substrates previously shaped by different routes, as well as the wear rate of uncoated substrates, are given in Table 3. Scatter for the attained Kc and Ks values was always less than ±10%, and it was mainly associated with uncertainties on the measurement of inner crater diameters due to the unclear definition of the corresponding boundaries (i.e. at the interface between substrate and coating). This was particularly true for the EDM conditions corresponding to rough cutting and fine shaping (i.e. A and B), as shown in Fig. 5; thus, in these cases additional tests were conducted and a larger number of measurements of the inner crater diameter (up to eight) were taken. Two main observations can be done from the results presented in Table 3. First, no significant differences were obtained between uncoated substrates. In contrast to microscratch testing of uncoated EDMed CC [8], the microabrasion wear test does not seem to be sensitive to EDM damage. This difference is expected since the material is stressed very differently in the two tests. In the microabrasion test the full contact load is distributed through

Fig. 5. Craters ground using the dimple grinder tests in TiN coatings deposited on A and B substrates obtained by EDM. The enlargements showing the central parts of the craters with exposed substrate reveal the crack networks induced by EDM.

B. Casas et al. / Wear 265 (2008) 490–496

the abrasive particles to a vast number of local contact points. Consequently, the deformation caused by an individual contact is also extremely local and the wear rate thus essentially insensitive to defects. In comparison, the scratch test is more like a macroscopic contact, locally and instantaneously deforming a relatively larger volume, and thus much more sensitive to defects. Second, the TiN intrinsic wear coefficients measured on coated substrates with conditions B, C and D are relatively close to that determined for the coating deposited on the reference substrate P. The value for the coating on the rough EDM surface is significantly higher. A ranking of the coatings based on wear coefficient seems to follow a similar trend as a ranking based on adhesion strength. However, differences among intrinsic wear coefficients are much smaller than those discerned in terms of critical loads for adhesive failure of the coating. This suggests that: (1) the influence of different EDM parameters on nucleation and growth of a TiN coating is not significant (with the exception


of such parameters resulting in condition similar to A) and (2) given that the adhesion strength is satisfactory enough, possible EDM-related coating defects play just a secondary role (if any at all) on the determined intrinsic wear resistance, as expected from the analytical considerations behind the crater grinder test. These statements are further supported by the similar failure mechanisms observed for the coating in the microabrasion wear test for all the conditions except A. As it is shown in Fig. 6, in these cases wear occurs by microcutting induced by abrasive particles either rolling or sliding between the two surfaces (grinding wheel and specimen) but always without any sign of chipping or spallation of coating fragments. Within this context, the above considerations are quite relevant because they allow one to emphasize the complementary character of the scratch and crater grinder testing techniques for assessing how EDM pre-treatment affects two different key properties for the coated system, i.e. adhesion strength and intrinsic wear resistance of the coating. Additionally, from the perspective of evaluating EDM as a feasible machining route for CC surfaces to be coated using PVD, the above results point out sequential and increasingly finely executed EDM as a method as effective as conventional diamond grinding and polishing. 4. Conclusions From a consideration of the experimental results obtained in this investigation, the following conclusions can be drawn: (1) In terms of providing a suitable substrate surface, sequential and multi-step EDM can be just as effective as conventional diamond grinding and polishing when forming CC substrates that are to be PVD coated. (2) The influence of different EDM parameters on nucleation and growth of a TiN coating is insignificant. (3) Scratch testing and microabrasion testing complement each other and can be used to assess two different critical parameters: adhesion strength and intrinsic wear resistance, for coatings deposited on EDMed CC substrates. Acknowledgements This work was supported by the Spanish Ministerio de Ciencia y Tecnolog´ıa (Grants Nos. MAT03-01685 and MAT2006-13480-C02-02). Research work was conducted within a cooperative effort among DURIT Metalurgia Portuguesa do Tugst´enio, Balzers-Elay, Mecanizados Gin´es, AMP-Tyco Electronics, Uppsala University and Universitat Polit`ecnica de Catalunya. The authors sincerely thank M. Marsal (UPC) for experimental assistance in SEM examination. Finally, B.C. acknowledges the scholarship received from the Spanish MCYT. References

Fig. 6. Parts of the wear crater rims after the microabrasion test for surface variants C, D and P.

[1] C.J. Heuvelman, Summary report on the CIRP cooperative research on spark-erosion machining of cemented carbides (die-sinking), Ann. CIRP 29 (1980) 541–544.


B. Casas et al. / Wear 265 (2008) 490–496

[2] A.M. Gadalla, W. Tsai, Machining of WC–Co composites, Mater. Manuf. Proc. 4 (1989) 411–423. [3] A.M. Gadalla, W. Tsai, Electrical discharge machining of tungsten carbide–cobalt composites, J. Am. Ceram. Soc. 72 (1989) 1396–1401. [4] J. Qu, A.J. Shih, R. Scattergood, Development of the cylindrical wire electrical discharge machining process. Part 1. Concept, design, and material removal rate, J. Manuf. Sci. Eng. 124 (2002) 702–707. [5] J. Qu, A.J. Shih, R. Scattergood, Development of the cylindrical wire electrical discharge machining process. Part 2. Surface integrity and roundness, J. Manuf. Sci. Eng. 124 (2002) 708–714. [6] E. Lenz, E. Katz, W. K¨onig, R. Wertheim, Cracking behaviour of sintered carbides during EDM, Ann. CIRP 24 (1975) 109–114. [7] S.M. Pandit, K.P. Rajurkar, Analysis of electro discharge machining of cemented carbides, Ann. CIRP 30 (1981) 111–116. [8] L. Llanes, E. Ida˜nez, E. Mart´ınez, B. Casas, J. Esteve, Influence of electrical discharge machining on the sliding contact response of cemented carbides, Int. J. Refract. Met. Hard Mater. 19 (2001) 35–40. [9] J. Qu, L. Riester, A.J. Shih, R.O. Scattergood, E. Lara-Cuirzo, T.R. Watkins, Nanoindentation characterization of surface layers of electrical discharge machined WC–Co, Mater. Sci. Eng. A344 (2003) 125–131. [10] L. Llanes, B. Casas, E. Ida˜nez, M. Marsal, M. Anglada, Surface integrity effects on the fracture resistance of electrical-discharge-machined WC–Co cemented carbides, J. Am. Ceram. Soc. 87 (2004) 1687–1693. [11] E. Mart´ınez, B. Casas, A. Lousa, J. Esteve, L. Llanes, Diamond coatings on electrical-discharge machined hard metals, Diamond Relat. Mater. 12 (2003) 762–767. [12] B. Casas, A. Lousa, J. Calder´on, M. Anglada, J. Esteve, L. Llanes, Mechanical strength improvement of electrical discharge machined cemented carbides through PVD (TiN, TiAlN) coatings, Thin Solid Films 447/448 (2004) 258–263. ˚ Kassman, S. Jacobson, L. Erickson, P. Hedenqvist, M. Olsson, A new [13] A. test method for the intrinsic abrasion resistance of thin coatings, Surf. Coat. Technol. 50 (1991) 75–84. [14] P. Hedenqvist, M. Bromark, M. Olsson, S. Hogmark, Mechanical and tribological characterization of low-temperature deposited PVD TiN coatings, Surf. Coat. Technol. 63 (1994) 115–122. [15] K.L. Rutherford, I.M. Hutchings, A microabrasive wear test, with particular application to coated systems, Surf. Coat. Technol. 79 (1996) 231–239. [16] K.L. Rutherford, S.J. Bull, E.D. Doyle, I.M. Hutchings, Laboratory characterisation of the wear behaviour of PVD-coated tool steels and correlation with cutting tool performance, Surf. Coat. Technol. 80 (1996) 176–180. [17] K.L. Rutherford, I.M. Hutchings, Theory and application of a microscale abrasive wear test, J. Test. Eval. 25 (1997) 250–260.

[18] J.C.A. Batista, C. Godoy, A. Matthews, Micro-scale abrasive wear testing of duplex and non-duplex (single layered) PVD (Ti,Al)N, TiN and Cr–N coatings, Tribol. Int. 35 (2002) 363–372. [19] J.C.A. Batista, M.C. Joseph, C. Godoy, A. Matthews, Micro-abrasion wear testing of PVD TiN coatings on untreated and plasma nitrided AISI H13 steel, Wear 249 (2002) 971–979. [20] J. Richter, Micro-scale abrasion testing of PVD TiN coatings on conventional and nondeleburitic high-speed steels, Wear 257 (2004) 304– 310. [21] A. Ramalho, Micro-scale abrasive wear test of thin coated cylindrical surfaces, Tribol. Lett. 16 (2004) 133–141. [22] C.K.N. Oliveira, R.M.M. Riofano, L.C. Castelleti, Micro-abrasive wear test of niobium carbide layers produced on AISI H13 and M2 steels, Surf. Coat. Technol. 200 (2006) 5140–5144. [23] R. G˚ahlin, M. Larsson, P. Hedenqvist, S. Jacobson, S. Hogmark, The crater grinder method as a means for coating wear evaluation—an update, Surf. Coat. Technol. 90 (1997) 107–114. [24] R.I. Trezona, D.N. Allsopp, I.M. Hutchings, Transitions between two-body and three-body abrasive wear: influence of test conditions in the microscale abrasive wear test, Wear 225/229 (1999) 205–214. [25] M.G. Gee, A. Gant, I. Hutchings, R. Bethke, K. Schiffman, K. Van Acker, S. Poulat, Y. Gachon, J. von Stebut, Progress towards standardisation of ball cratering, Wear 225 (2003) 1–13. [26] Y. Kusano, K. Van Acker, I.M. Hutchings, Methods of data analysis for the micro-scale abrasion test on coated substrates, Surf. Coat. Technol. 183 (2004) 312–327. [27] Y. Kusano, I.M. Hutchings, Sources of variability in the free-ball microscale abrasion test, Wear 258 (2005) 313–317. [28] M.G. Gee, A.J. Gant, I.M. Hutchings, Y. Kusano, K. Schiffman, K. Van Acker, S. Poulat, Y. Gachon, J. von Stebut, P. Hatto, G. Plint, Results from an interlaboratory exercise to validate the micro-scale abrasion test, Wear 259 (2005) 27–35. [29] K.I. Schiffman, R. Bethke, N. Kristen, Analysis of perforating and nonperforating micro-scale abrasion tests on coated substrates, Surf. Coat. Technol. 200 (2005) 313–317. [30] S. Hogmark, S. Jacobson, M. Larsson, Design and evaluation of tribological coatings, Wear 246 (2000) 20–33. [31] A.G. Evans, G.B. Crumley, R.E. Demaray, On the mechanical behavior of brittle coatings and layers, Oxid. Met. 20 (1983) 193–216. [32] A. Strawbridge, H.E. Evans, Mechanical failure of thin brittle coatings, Eng. Fail. Anal. 2 (1995) 85–103. [33] U. Wiklund, J. Gunnars, S. Hogmark, Influence of residual stresses on fracture and delamination of thin hard coatings, Wear 232 (1999) 262–269.