CONTACT FATIGUE DAMAGE OF MONOLAYER AND BILAYER CERAMIC COATINGS DEPOSITED ON CEMENTED CARBIDES

L. Llanes1,2, E. Tarres1,3, G. Ramírez1,4, E. Jimenez-Piqué1,2, N. Salán1 and A. Mateo1

1CIEFMA- Departamento de Ciencia de los Materiales e Ingeniería Metalúrgica, Universitat Politècnica de Catalunya, ETSEIB, Barcelona 08028, Spain

2CRnE, Campus Diagonal Sud, Universitat Politècnica de Catalunya, Barcelona 08028, Spain

3Sandvik Hard Materials, Coventry CV4 0XG, UK

4CTM Centre Tecnològic, Manresa 08242, Spain

RESUMEN

En este trabajo se estudia el comportamiento a fatiga bajo solicitaciones de contacto de un carburo cementado WC-Co recubierto con dos películas cerámicas distintas: TiN y WC/C, ya sea como monocapas o bicapas, mediante técnicas de indentación. Los resultados obtenidos en ensayos de indentación esférica indican que la nucleación de fisuras circulares en las capas es un criterio más apropiado que la delaminación interfacial para definir las condiciones de daño crítico en los sistemas investigados. Desde esta perspectiva, se encuentra que los carburos cementados recubiertos son susceptibles a ser degradados mecánicamente bajo solicitaciones de contacto cíclicas, aunque la sensibilidad a fatiga y el escenario de daño efectivos difieren en cada caso, en función de la naturaleza cerámica (cristalina/amorfa) y la arquitectura (mono- o bicapa) del recubrimiento.

ABSTRACT

The contact fatigue behavior of a WC-Co cemented carbide coated with two distinct ceramic films: TiN and WC/C, either as monolayers or bilayers, is studied by means of indentation techniques. Spherical indentation tests indicate that circular cracking at the coating is a more appropriate choice than interfacial delamination for defining critical damage in the coated hardmetals studied. From this perspective, coated cemented carbides are found to be fatigue susceptible under contact loading, although effective fatigue sensitivity and failure scenario are discerned to be dependent on ceramic nature (crystalline/amorphous) and coating assemblage (mono- or bilayer).

KEYWORDS: contact fatigue; ceramic film; coated cemented carbide; mono- and bilayer assemblage

1. INTRODUCTION

Cemented carbides belong to a class of composite materials, in which hard particles, tungsten carbide (WC), are bound together by a soft and ductile metal binder, cobalt (Co). Such a particular microstructure assemblage, usually referred to as hardmetals, yields an extraordinary combination of mechanical properties, allowing the use of these materials in a wide range of applications. From this perspective, the role of cemented carbides in metal cutting and forming tools is remarkable, the corresponding components being here subjected to a complex state of loading (abrasive and adhesive wear, impact, repetitive mechanical contact, etc.) that determines their lifetime (e.g. Ref. [1]). On the other hand, the need to extend the lifetime of parts and/or, for example, minimize the use of lubricants in order to reduce time, maintenance and manufacture costs has led to coat base hard materials with different

thin films or coatings. In this regard, the development of vacuum deposition technologies, physical vapor deposition (PVD) and chemical vapor deposition (CVD), has been of major impact, since they make it possible to deposit a thin layer of only a few micrometers on the surfaces of most engineering materials The geometrical change is minimal and the surface layers may have properties covering an extremely wide range of values [2]. In the metal cutting and forming areas, research efforts have mainly ranged from productions of coatings designed for high hardness for cutting applications to low-friction films for sheet forming ones, although optimal choices usually involve combined approaches [3].

There exists extensive literature on the tribomechanical response of coated systems to be used as metal cutting or forming tools, though the use of film deposition technology in the former is far more established than in

the latter [4,5]. In general, it has centered primarily on investigating hardness, scratch resistance, friction coefficient and wear behavior as a function of either film nature and architecture - single versus multilayer - (e.g. Refs. [5-8]) or surface roughness (e.g. Refs. [7,9,10]). However, these experimental approaches are rather deficient in evaluating issues directly pertinent to mechanical performance of coated tools or components subjected to repetitive contact loading. On the other hand, experimental and analytical approaches using spherical indenters to deliver stresses over a small area of specimen surface, i.e. testing protocols based on the Hertzian theory, have proven to be successful on the assessment of contact damage in bulk polycrystalline ceramics as well as layered structures (e.g. Refs. [11-13]). Within this framework, it is the aim of this work to assess the contact mechanical response and the corresponding damage mechanisms, under both monotonic and cyclic loading conditions, on two different PVD coatings deposited, as either individual monolayers or a combined bilayer, on a fine-grained hardmetal. In doing so, experimental protocols based on spherical indentation testing techniques are implemented.

2. EXPERIMENTAL PROCEDURE

A commercial fine-grained WC-10%wt Co cemented carbide grade was used as substrate. Elastic modulus and hardness for such base material, as determined in a previous study [14], were 540 GPa and 14.5 GPa, respectively. Two coatings were investigated, both deposited following Oerlikon-Balzers’s PVD processes: (1) a single-layer TiN film by arc ion plating, and (2) a multilayer diamond-like carbon (DLC) one, in this study referred to as WC/C, by magnetron sputtering. They were deposited as 3.7 m (TiN) and 3.3 m (WC/C) individual monolayers as well as a combined bilayer of 2.7 m WC/C on 3.5 m TiN (Figure 1). In all the cases dense coatings were attained. TiN layers exhibited a crystalline fine-grained structure whereas WC/C films

Figure 1. SEM micrographs of: focused ion beam milled cross-sections of a) TiN and b) WC/C monolayers; and fractured cross-section of

WC/C//TiN bilayer.consisted of a nanostratified arrangement of amorphous WC and C lamellas. Young’s modulus and intrinsic

hardness for both coatings were determined by nanoindentation (MTS Nanoindenter XP) equipped with a continuous stiffness modulus. As expected, the crystalline coating exhibited significantly higher stiffness (430 GPa as compared to 135 GPa) and hardness (29 GPa with respect to 12 GPa) than the alloyed amorphous one [15]. For the bilayer system the hardness – penetration depth showed an early initial plateau of 12 GPa (for depths about 250 nm) associated with the intrinsic hardness of the top WC/C layer. As load was increased, hardness also rose up to values close to 20 GPa, where it described a second plateau behavior starting at 2 m and lasting up to final penetration depth tested, i.e. 6 m. Rising hardness values are intimately related with the influence ascribed to harder base layer and substrate, although the intrinsic hardness of TiN film was never attained. One possible reason for such finding is the fact that local chipping was discerned as penetration depth got levels of 3 m, i.e. close to the internal bilayer interface.

Similar trend was also discerned regarding adhesion, as given by values of 98 N and 38 N for TiN and WC/C respectively, for the critical normal load related to initial coating detachment under scratch testing. Such critical load rose to 54 N for the bilayer system and it was very close to the one needed (51 N) for inducing decohesion at the internal interface between WC/C and TiN layers. Although these findings are in agreement with previous literature reports (e.g. Ref. [15]), it should be highlighted that critical loads in all the cases were higher than 30 N, a level generally described as sufficient in scratch testing with a Rockwell C diamond tip for tooling applications [4].

The mechanical contact response of the coated systems was investigated by means of spherical indentation. Hertzian tests were conducted in a servohydraulic testing machine (Instron 8511) by using a hardmetal spherical indenter with a curvature radius of 1.25 mm. Definition of the critical failure event was done on the basis of “crack prevention”, an approach different from the coating delamination criteria usually invoked in previous works involving cyclic impact testing (e.g. Refs. [16,17]). As it will be seen later, this definition is sustained on the fact that through-thickness circular cracks at the coating lead to hardmetal substrate cracking, independent of the loading conditions, before any interfacial failure (at the coating/substrate level) takes place. Accordingly, the first stage on the experimental protocol followed was the determination of the critical load for circular crack emergence at the coating surface under monotonic loading, Pc. Once it was assessed, fatigue testing was carried out by applying fractions of such critical load as the maximum cyclic load (Pmax). The cyclic loading was imposed by means of a sinusoidal waveform at a frequency of 10 Hz and corresponding load ratio of 0.1. Main outcome of such cyclic tests was critical load for discerning similar

a)

b) c)

circular cracks at the films after a very high number of cycles, namely 106 cycles, Pf.

Evolution of subsurface indentation damage, with increasing load or number of cycles, was investigated by means of specific monotonic and cyclic Hertzian indentation tests conducted on “clamped-interface” specimens. In doing so, a procedure similar to that commonly employed in ceramics by Lawn’s group (e.g. Ref. [18]) was followed; although here extended to coated surfaces. Briefly, these may be described as coated samples consisting of two previously ground and polished cross-sections clamped together by artificial means. Indentation across the surface trace of the existing interface and subsequent mechanical separation of the referred halves permits then examination of damage features under the surface, through optical and scanning electron microscopy (SEM).

3. RESULTS AND DISCUSSION

3.1 Monotonic spherical indentation

As applied load increases irreversible deformation of the coated systems was evidenced through residual surface traces. At relatively high level loads, first signs of damage were discerned at the edge of the corresponding residual imprints in terms of circumferential cracks. The critical load for the emergence of these cracks, Pc, for the three studied systems is listed in Table 1. Nevertheless, plastic yielding of the coated sets was required for the subsequent cracking in the film, in agreement with previous studies by other authors (e.g. Ref. [19]). Regarding experimental data analysis, it should be pointed out that data from Hertzian tests are usually presented in terms of contact pressure or indentation stress (p) as well as the resulting indentation strain (e.g. Ref. [13]). Such approach was also implemented in this study. Accordingly, the corresponding critical mean contact pressures, given by pc = Pc /ac

2 where ac is the contact radius measured on the residual impression after applying Pc, are also listed in Table 1. The results attained clearly indicate that load and pressure levels for discerning film rupture is higher for the DLC-coated hardmetals, either as monolayer or as bilayer, than for the one simply coated with TiN. Taking into consideration the intrinsic properties of both coatings, such a finding should be associated with the lower stiffness and hardness exhibited by the WC/C multilayer, which then render more elasticity and tolerance to follow the deforming substrate before experiencing brittle rupture. The above monotonic critical loads and indentation stresses were then used as baseline reference for comparison purposes with the ones determined under cyclic loading conditions.

Table 1. Critical load and indentation stress under monotonic contact loading for the appearance of circular cracks in the mono- and bilayer ceramic coatings deposited on the hardmetal substrate.

Coating

Critical monotonic load for coating

cracking, Pc (N)

Critical monotonicindentation stress

for coating cracking, pc (GPa)

TiN 800 10.2

WC/C 1300 11.3

WC/C//TiN 1600 12.5

3.2 Cyclic spherical indentation

Cyclic indentation tests were conducted attempting to evaluate the susceptibility to damage appearance for the coated systems as either applied cyclic stress (for a given number of cycles) or number of cycles (for a given applied stress) increases. In agreement with previous studies on other hard-coated systems [16,20,21], they show that the critical damage emergence is sensitive to cycle loading. However, two interesting remarks should be done. First, different from the trend observed under monotonic loading, maximum cyclic loads for which time-differed damage is still evidenced after a very large number of cycles, here referred to as Pf, are lower for the DLC-coated systems, either as monolayer or as bilayer, than for the TiN coated one (Table 2). Second, damage mechanisms distinct from the cohesive one observed under monotonic loading (i.e. adhesive failure, for instance) are not discerned, even after 106 cycles. Following the analysis described in the previous section, critical cyclic indentation stresses for coating cracking, pf, were calculated from the determined Pf and corresponding residual spherical imprint diameters. They are also included in Table 2. It should be noticed that although relative difference on contact fatigue strength between DLC- and just TiN-coated systems decreases, it is still higher for the latter.

From the critical parameters listed in Tables 1 and 2, the fatigue susceptibility of the coated systems here studied may be analyzed by means of the ratio [1 – (pf / pc)] a parameter usually referred to in the fatigue literature as fatigue sensitivity. In this sense, the closer to 0 the ratio is, the less sensitive to contact fatigue, as related to the failure mechanism defined as critical event, the material should be, and vice versa. For the TiN coated system the fatigue sensitivity is found to be 0.13, i.e. a relatively low value, even when compared to similar low values usually reported for other brittle-

Table 2. Critical load and indentation stress under cyclic contact loading for the appearance of circular

cracks in the mono- and bilayer ceramic coatings deposited on the hardmetal substrate.

Coating

Critical cyclic load for coating

cracking, Pf (N)

Critical cyclicindentation stress

for coating cracking, pf (GPa)

TiN 400 8.9

WC/C 200 7.0

WC/C//TiN 200 7.1

like materials (e.g. see Refs. [22,23]). Considering that the failure event chosen here as critical involves cracking of the TiN film, it is speculated that such low fatigue sensitivity should be intimately related to the ceramic nature of the coating, especially if the high compressive residual stresses exhibited by it are considered [24]. Such a hypothesis may be somehow supported by a recent work, by Cairney and coworkers [25] on degradation of TiN coatings under cyclic loading, where it is shown that differences between intergranular shear stresses required for promoting sliding along intercolumnar cracks, postulated by the authors as the principal deformation mechanism in the coating, under cyclic and monotonic (nanoindentation) loading are minimal. However, the ceramic nature of the coating does not seem to be a sufficient condition for expecting low fatigue sensitivity. This is clearly evidenced by the higher [1 – (pf / pc)] ratios: 0.38 and 0.43, determined for the monolayer and bilayer DLC-coated hardmetals, respectively. Such high fatigue sensitivity level is close to those determined for PVD-coated steels under the consideration of interfacial decohesion as critical damage event, i.e. one where a metallic substrate is directly involved in the failure mechanism [26]. In this regard, it is speculated that the amorphous nature of the WC/C film may be playing a key critical role on effectively determining the intrinsic cyclic degradation susceptibility of this ceramic coating. Indeed, the existence of alike microstructural features (e.g. equiaxed/columnar assemblage or crystalline/ amorphous nature) has been pointed out by Jayaram et al. [27] as plausible reason for the differences reported by two distinct research groups [25,28] on the indenter displacement response under cyclic loading of TiN films.

3.3 Subsurface indentation damage

Spherical indentation tests are specially attractive because, different from more conventional techniques involving sharp-like indenters, they allow to monitor damage evolution within an otherwise uncracked microstructure as a function of increasing either applied load (monotonic tests) or number of cycles (cyclic tests). This is particularly interesting in the case of coated systems due to the fact that here it is important to evaluate not only the intrinsic competition between deformation and fracture mechanisms but also where they are first developing, i.e. within the coating or the substrate, or even at the interface. Within this context, the damage resulting from contact loading, under monotonic and cyclic conditions, was inspected by means of optical and scanning electron microscopy on both the contact surface and below it, the latter by using “clamped-interface” specimens. In doing so, the number of testing conditions studied was limited and focused on identifying damage micromechanims.

The evolution of Hertzian-induced damage with increasing applied load or number of cycles was qualitative similar. Optical inspections of imprint half-surfaces and cross-sections in the bonded-interface specimens revealed a significant plastic damage in the substrate under the indentation contact area (Figure 2), as expected from the range of contact stress imposed. More accurate inspection by SEM showed that damage within the coating evolves, beyond the assumed stretching conjugated with the referred plastic deformation of the substrate, into circular crack nucleation at the periphery of the contact and at the

Figure 2. Induced damage under spherical indentation for the WC/C-coated hardmetal: a) 1400 N (monotonic)

and b) 600 N – 105 cycles. Upper (darker) and lower (brighter) optical micrographs in each image correspond to half-surface and cross section views respectively of

clamped-interface specimens.coating surface, and its subsequent advance through the thin film up to the interface. The cohesive failure through the whole coating thickness is quite straight for monolayer films (Figure 3). However, similar finding was not always evidenced for the bilayer-coated hardmetal. Rather, in several imprints cracks were observed to grow, from the surface into the film, following inclined paths, and exhibiting deflection-like features, particularly at the internal bilayer interface. Considering that critical cyclic load for cracking is higher for the inner TiN layer (Table 2), the role of the referred interface as crack stopper for applied maximum loads lower than 400 N was an interesting query to answer in this study. However, validation of such hypothesis was not possible in this investigation. Nevertheless, it should be highlighted that crack - layer assemblage interactions as the one detailed in Figure 3c points out an improved contact response for the DLC-coated system if it is supported by a harder layer underneath. Indeed, this could be one of the “unknown reasons” for explaining the existing empirical knowledge that additional mechanical support from hard coatings results in enhanced tribomechanical response

Figure 3. Typical contact damage scenario for the coated systems studied, as assessed from spherical

indentation of clamped-interface specimens: a) TiN monolayer; b) WC/C monolayer;

and c) WC/C//TiN bilayer.of low- friction films (not only DLCs but also MoS2) (e.g. Ref. [6]).

Finally, in all the cases such through-thickness fissures lead to substrate cracking along the metallic binder surrounding the ceramic particles. Here, it should be noted (1) the absence of any intermediate interfacial delamination stage (at substrate/coating level) before substrate failure is observed, and (2) the role of coating circular fissures as precursors of crack nucleation in the underlying brittle-like hardmetal. Both experimental facts are quite significant, within the context of the investigation conducted here, because they sustain the choice of the early circular cracking at the coating as the critical damage event, mainly because it induces a loss of mechanical integrity of the substrate before anyadhesive failure takes place. Moreover, they also point out the relevant influence of the substrate nature on the contact response of the coated system, on the basis that such final damage evolution stage (substrate cracking) is usually not evidenced when considering tougher substrates, e.g. steels [26].

4. CONCLUSIONS

a)

b)

Micron bar

a) b)

c)

Based on the experimental study of the mechanical contact behavior, under monotonic and cyclic spherical indentation, of a WC-Co cemented carbide coated with two distinct ceramic films: TiN and WC/C, either as monolayers or bilayers, the following conclusions can be drawn:

(1) Circular cracking at the coating is a more appropriate choice than interfacial delamination for defining critical damage under spherical indentation for the coated hardmetal here studied. Such a statement is based on the fact that it not only emerges as the first failure feature observed with increasing load or number of cycles but also leads to substrate cracking without any intermediate failure at the substrate-coating interface.

(2) Coated cemented carbides are susceptible to “real” contact fatigue. However, the effective fatigue sensitivity is dependent on ceramic nature, i.e. crystalline versus amorphous. Additionally, the consequent failure scenario associated with the induced damage may also be dependent on the coating assemblage regarding mono- or bilayers, the latter in terms of a hard coating offering mechanical support to a softer low-friction top-layer film.

ACKNOWLEDGEMENTS

This investigation was partly funded by the Spanish Ministerio de Ciencia e Innovación (MAT2009-14461). The support received from Direcció General de Recerca del Comissionat per a Universitats i Recerca de la Generalitat de Catalunya through recognition of CIEFMA as Grup de Recerca Consolidat 2009SGR 1285 is also acknowledged. The authors would also like to acknowledge the scholarship received from the Consejo Nacional de Ciencia y Tecnología de México (G.R.).

REFERENCES

[1] Upadhyaya GS, Cemented Tungsten Carbides: Production, Properties and Testing. Noyes Publications, New Jersey, 1998.[2] Holmberg K, Matthews A, Coating Tribology – Properties, Techniques and Applications in Surface Engineering. Tribology Series 28, Elsevier, Amsterdam, 1994.[3] Bewilogua K, Bräuer G, Dietz A, Gäbler J, Goch G, Karpuschewski B, Szyszka B. Surface technology for

automotive engineering. CIRP Ann. – Manuf. Technol. 2009; 58: 608-627.[4] Hogmark S, Jacobson S, Larsson M. Design and evaluation of tribological coatings. Wear 2000; 246: 20-33.[5] Mitterer C, Holler F, Reitberger D, Badisch E, Stoiber M, Lugmair C, Nöbauer R, Müller Th, Kullmer R. Industrial applications of PACVD hard coatings. Surf. Coat. Technol. 2003; 163-164: 716-722.[6] Fox V, Jones A, Renevier NM, Teer DG. Hard lubricating coatings for cutting and forming tools and mechanical components. Surf. Coat. Technol. 2000; 125: 347-353.[7] Podgornik B, Hogmark S, Sandberg O. Influence of surface roughness and coating type on the galling properties of coated forming tool steel. Surf. Coat. Technol. 2004; 184: 338-348.[8] Carlsson P, Olsson M. PVD coatings for sheet metal forming processe: a tribological evaluation. Surf. Coat. Technol. 2006; 200: 4654-4663.[9] Harlin P, Carlsson P, Bexell U, Olsson O. Influence of surface roughness of PVD coatings on tribological performance in sliding contact. Surf. Coat. Technol. 2006; 201: 4253-4259.[10] Heinrichs J, Jacobson S. Laboratory test simulation of aluminium cold forming influence from PVD tool coatings on the tendency to galling. Surf. Coat. Technol. 2010; 204: 3606-3613.[11] Guiberteau F, Padture NP, Cai H, Lawn BR. Indentation fatigue: a simple cyclic Hertzian test for measuring damage accumulation in polycrystalline ceramics. Phil. Mag. A 1993; 68: 1003-1016.[12] Lawn BR. Indentation of ceramics with spheres: a century after Hertz. J. Am. Ceram. Soc. 1998; 81: 1977-1994.[13] Lawn BR, Deng Y, Miranda P, Pajares A, Chai H, Kim DK. Overview: damage in brittle layer structures from concentrated loads. J. Mater. Res. 2002; 17: 3019-3036.[14] Torres Y. Comportamiento a fractura y fatiga de carburos cementados WC-Co. PhD Thesis, Universitat Politècnica de Catalunya, Barcelona, 2002.[15] Vercammen K, Haefke H, Gerbig Y, Van Hulsel A, Pflüger E, Meneve J. A comparative study of state-of-the-art diamond-like carbon films. Surf. Coat. Technol. 2000; 133-134: 466-72.[16] Knotek O, Bosserhoff B, Schrey A, Leyendecker T, Lemmer O, Esser S. A new technique for testing the impact load of thin films: the coating impact test. Surf. Coat. Technol. 1992; 54-55: 102-107.[17] Stoiber M, Panzenböck M, Mitterer C, Lugmair C. Fatigue properties of Ti-based hard coatings deposited onto tool steels. Surf. Coat. Technol. 2001; 142-144: 117-124.[18] Guiberteau F, Padture NP, Lawn BR. Effect of grain size on hertzian contact damage in alumina. J. Am. Ceram. Soc. 1994; 77: 1825-1831.[19] Bantle R, Mathews A. Investigation into the impact wear behaviour of ceramic coatings. Surf. Coat. Technol. 1995; 74-75: 857-68.

[20] Lugscheider E, Knotek O, Wolff C, Bärwulf S. Structure and properties of PVD-coatings by means of impact tester. Surf. Coat. Technol. 1999; 141-146: 116-119.[21] Bouzakis K-D, Vidakis N, David K. The concept of an advanced impact tester supported by evaluation software for the fatigue strength characterization of hard layered media. Thin Solid Films 1999; 355-356: 322-329.[22] Fleck NA, Kang KJ, Ashby MF. The cyclic properties of engineered materials. Acta Metall. Mater. 1994; 42: 365-81.[23] Ritchie RO. Mechanisms of fatigue-crack propagation in ductile and brittle solids. Int. J. Fracture 1999; 100: 55-83.[24] Bermejo R, Torres Y, Anglada M, Llanes L. Fatigue behavior of alumina-zirconia multilayered ceramics. J. Am. Ceram. Soc. 2008; 91: 1618-25.[25] Cairney JM, Tsukano R, Hoffman MJ, Yang M. Degradation of TiN coatings under cyclic loading. Acta Mater. 2004; 52: 3229-3237.[26] Ramírez G, Mestra A, Casas B, Valls I, Martínez R, Bueno R, Góez A, Mateo A, Llanes L. Influence of substrate microstructure on the contact fatigue strength of coated cold-work tool steels. Surf. Coat. Technol. 2011; doi:10.1016/j.surfcoat.2011.12.012[27] Jayaram V, Bhowmick S, Xie Z-H, Math S, Hoffman M, Biswas SK. Contact deformation of TiN coatings on metallic substrates. Mater. Sci. Eng. A 2006; 423: 8-13.[28] Dharma Raju T, Nakasa K , Kato M. Relation between delamination of thin films and backward deviation of load-displacement curves under repeating nanoindentation. Acta Mater. 2003; 51: 457-467.