ICMCTF2007 Session E2-1: Mechanical Properties and Adhesion

The deposition of thin films and coatings has in recent years prompted significant efforts by ISO and ASTM to standardize tests which can adequately characterize coating adhesion and scratch resistance.

Since 1991 when the VDI 3198 standard appeared, many companies started to use this test method which has since become known as the "Daimler", "Mercedes" or "Rockwell" test method. It consists of making a Rockwell C indentation with an applied load of 1500 N on the coating and evaluating the resultant cracking on a 6-level scale. This test has proved to provide only qualitative information and does not necessarily correlate to coating adhesion. This makes it very limiting when trying to assess the mechanical properties of a coating in a quantitative and reproducible way.

The progressive load scratch test has since proved to be a far better method (ASTM C1624 & ISO 20502) for evaluating coating adhesion in hard ceramic coatings. This paper presents a direct comparison between the Mercedes test and the scratch test on a range of common coatings including TiN, DLC and nanocomposites on Steel and carbide substrate.

In performing tests to compare the Rockwell Test with the Scratch Test it was noticed that the results were significantly influenced by the indenter tip geometry and wear. When using the progressive load scratch test to characterize the mechanical properties of a substrate-coating system, the geometry of the indenter as well as its condition, are very important in obtaining quantitative information and in being able to compare results between different labs.

In order to better understand variations in the measured critical load as a typical indenter becomes worn, we have compared various 200 µm Rockwell indenters, both of natural and synthetic diamond material, from various manufacturers.

This study was carried out using the standard operating procedures as outlined in the ISO 20502 and ASTM C1624 standards, including up to 100 scratches per indenter on a standard DLC-coating of thickness 2 µm. The results to be presented show very clearly the importance of regular quality control of a scratch test indenter and the influence of its wear on the measured critical load. Some recommendations will also be made on the best ways of achieving this in the modern industrial environment.

For the case of flexible printed circuits, peel tests have been commonly used to determine the strength of adhesion. However the estimated strength usually depends on the thickness of adhered films due to the plastic dissipation energy. It is expected that extrapolation of the results toward zero thickness may exclude the influence of plastic deformation, which resulted in 250 J/m² for a Cu/Polyimide sample studied in this paper. Kinloch¹ tried to estimate the effect of plastic deformation on the basis of beam bending theory. His theory yielded the adhesion energy 100 J/m² when applied to the same sample. However, there still remains plastic deformation on the substrate side and at the interface crack tip which was not taken into these considerations.

In this study, instead of peel test, an experimental technique recently developed by Kamiya et al.² was applied, where interface crack was effectively extended by a local load with a minimum amount of deformation. Elastic-plastic finite element model was calculated to simulate the interface crack extension process and to obtain the amount of plastic deformation. The sum of the increments of elastic strain energy and plastic dissipation energy was found not to agree with the work done by the external load. The difference was consumed to separate the interface and thus defined as the interface fracture energy. The same amount was obtained by calculating the work done by the nodal force in the vicinity of interface crack tip. This concept concluded the interface fracture energy of the system mentioned above to be 35 J/m² and is expected as a new reliable scheme to determine the strength of adhesion independent of plastic deformation in film-substrate systems.

¹A.J.Kinloch, C.C.Lau, J.G.Williams, Int. J. Frac., 66(1994), 45-70.

²S. Kamiya, H. Nagasawa, K. Yamanobe, M. Saka, Thin Solid Films, 473(2005), 123-131.

Buckling and wrinkling of thin objects, such as sheets or rods, are well known phenomena, studied from wide to fine-scale in micro-electronic or bio-mechanic for instance. Using continuum mechanics, buckling of thin sheets has been studied by Foppl and Von-Karman at a macroscopic scale as far back as the beginning of 1900s. Since the 1980s, the reduction to the microscopic scale of experimental observations has allowed us to study thin film buckling phenomena in the same framework. Taking advantage of the high resolution offered by scanning probe microscopy, the fine structure of buckling patterns is now intensively investigated at a nanometre scale. It has been demonstrated that a thin film sustaining a stress greater than a critical value may buckle into different patterns, such as circular blisters, straight-sided wrinkles or telephone cords. The influence of substrate morphology and elasticity has been also investigated. However, the effect of the crystalline substrate plasticity has not yet been considered on the buckling instability phenomenon.

It is the purpose of the present paper to report experiments where dislocations coming from the substrate and emerging at the interface modify the further buckling structures of the thin film. Formation of straight-sided blisters just above step structures resulting from the dislocation emergence process is experimentally shown and explained in the frame of the Foppl/Von-Karman theory of thin plates. A new critical stress above which the buckling may occur has been determined and asymmetry of the resulting blisters has been also explained. Finally, the role of the crystallographic angle of the emerged steps with respect to the interface has been quantified and appears to be a relevant parameter in the study of the mechanical behaviour of the film.

Previous analysis on the Rockwell C indentation of systems with titanium nitride (TiN) films and high-strength tool steel substrates indicated the formation of circular and radial cracks at the film surface. Further analysis indicated that the radial cracks propagated after the unloading step of the indentation process, since they were only observed in the outer region of indentation.

In this work, a Finite Element Analysis was carried out to simulate the penetration of Rockwell conical and spherical indenters in coated systems and to study the intensity of the peaks in radial and tangential stresses at the indentation edge, during the loading and unloading steps of the indentation. The objective was to determine the conditions where the peak in tangential stresses overcomes the peak in radial stresses near the edge of indentation mark, favoring the formation of radial cracks. In the simulations, the behavior of the substrate and the film were considered to be elastic-plastic and different values of substrate yield stress were considered.

The results indicated that, during loading, the peak in radial stress at the indentation border is higher than the peak of tangential stress. During unloading the peak of tangential stresses increases, while the peak of radial stresses decreases, favoring a change in the crack pattern. Results have indicated that the lower intensity of the peak in radial stresses is due to a change of film curvature when the system is unloaded.

The rate at which the peaks in radial and tangential stresses increase was lower for substrates having higher yield stresses. Substrates with higher yield stresses also showed a larger decrease in the peak of radial stresses during the unloading step. The results were similar for both conical and spherical indenters.