ICMCTF2004 Session F1/E3-2: Mechanical Properties and Adhesion

Surfaces can be protected against mechanical damages by coatings that have to be adapted to the substrate and the conditions in the application. To reach the optimum protection the covering coating should not be a single layer but an optimised layer stack or a layer with graded mechanical properties. Recently [1] it was shown that B-C-N films can be used as model system for the production of graded coatings and for the investigation of the mechanical properties due to the large variation range of their properties. For the purpose of the following investigation two single layers on the opposite ends of the variation range were deposited with a thickness of about 0.8 µm on silicon. Additionally a 1.55 Â realised in 31 deposition steps on the same substrate as model for a graded coating. The mechanical properties of all coatings were measured by nanoindentation with a Berkovich indenter and two different spherical indenters of 3 µm and 12 µm radius. The Young's moduli of the homogeneous layers were derived from wholly elastic measurements with the spherical indenters. Elastic force-displacement curves were also measured on the graded coating. A new mathematical model was used to calculate the force-displacement curve for a graded coating on a substrate and to fit it to the measurement data. Using the data from two different indenters it was possible to derive the correct moduli at both ends of the graded coating. The agreement of both curves over the whole depth range was better than 2% of the maximum depth. The results confirm the usefulness of the mathematical model to predict the mechanical behavior of graded coatings and to analyze mechanical stresses in such applications.

[1] F. Richter et al., Depth-depending Young's modulus of thin films for increased load carrying capacity, ICMCTF 2002.

Strength, friction, and wear are dominant factors in the performance and reliability of materials and devices fabricated using microsystem technologies. While adequate for some applications, as-fabricated strength and wear properties severely restrict use of these devices in many dynamic applications. Applying coatings and films is one method to enhance device performance and reliability. Atomic Layer Deposition (ALD) is a method ideally suited for applying these films as as it creates conformal, stress free, and well-adhered films. However, measurement of mechanical and tribological properties of ALD films presents a significant challenge due to their extremely small scale. We have therefore begun a study of ALD tungsten and aluminum oxide films and nanolaminate properties. The films were deposited to thicknesses ranging from several monolayers to 200 nm on mirror-polished silicon. The range of thicknesses served to highlight the evolution of film and substrate contributions to strength and adhesion. Nanolaminates were then created from sequential deposition of tungsten and aluminum oxide films to a thickness of 200 nm. Following deposition, nanoindentation and nanoscratch techniques were employed to evaluate the strength, friction and adhesion of each film system at small loads characteristic of microsystem operation. These tests showed that the modulus and hardness varied with film thickness and composition. Of most interest were results showing that the hard tungsten films were very susceptible to spallation while aluminum oxide films were not. The properties and spalling resistance of the nanolaminates reflect the properties of both systems. These results will be discussed in terms of film composition and structure and how their use in nanolaminates can be used to tailor performance.

This work supported by U.S. DOE Contract DE-AC04-94AL85000.

Mechanical strength of thin films on substrates is an issue of critical importance. However, making quantitative comparison of strength among different films and substrates is not easy. This is because the results evaluated by conventional techniques are strongly influenced by the factors other than strength such as film thickness etc. In order to eliminate these ambiguities, a series of systematic measurement techniques has been established by the authors. A number of hard coatings, such as TiN, TiAlN, CrN and also CVD diamond, were subjected to successful evaluation by applying the new techniques. The results were quantitatively compared with each other, as well as with the results obtained by scratch tests.

Similar to the other conventional techniques, scratch test measures the strength of coatings basically in terms of just one parameter that is the critical load to damage the coatings. Therefore evaluated strength is also influenced by e.g. the residual stress in the coating, thickness of the coating, and even the deformation characteristics of substrates. Although the critical load is commonly mentioned as a measure of "adhesion", it is obviously influenced also by the strength of coating itself. In contrast, the proposed techniques independently measure the toughness of the coating, the toughness of interface between coatings and substrates, residual stress in the coating, Young's modulus of the coating. Overall information given by these four parameters dramatically improves the foresight onto the mechanical strength of coatings. It was newly found that the toughness of interface could be significantly different even though the same level of critical load was obtained by scratch test. CVD Diamond coatings, whose weak adhesion has always been an issue, was also confirmed to have the same level of adhesion compared with PVD coatings. The background of these interesting findings will be extensively discussed in the presentation.

The impact test, supported by its finite elements method (FEM) simulation, has been successfully used to characterize the fatigue performance of coatings. In this test the specimen is loaded perpendicularly to its coated surface by a cemented carbide ball. In the inclined impact test the successive impacts are acted onto two symmetrically located surfaces. Both surfaces belong to the same appropriately prepared specimen inclined to the impact direction. In this way, the coated surfaces are loaded vertically and tangentially simultaneously. The coating fatigue failure modes were classified by means of scanning electron microscopy observations and energy dispersive X-ray spectroscopy microanalyses. Critical values for stress components, responsible for distinctive failure modes of the coating substrate compounds were determined and the coating fatigue performance was investigated.

The inclined impact test implies a new reference to the prediction of the coatings failure, managing to approach loading directions for a variety of coated surfaces in different applications. The experimental method is supported by a developed FEM simulation, which considers the mechanical elastoplastic properties of the coating and the substrate, as well as of the carbide ball during the impact test.

In this work, the surface layers of steel samples have been modified using the tribological process of scuffing in order to improve the resistance of the surface to subsequent scuffing failure. In general, scuffing is known to involve large plastic strains and high local temperatures. The temperatures of the scuffed surface layers cool rapidly as heat is transported into the bulk material. This deformation and thermal treatment have been observed to form highly work-hardened surface layers with small grain size. Resistance to scuffing failure typically improves with increasing hardness; for this reason, we have assessed the scuffing resistance of previously scuffed material as follows.

We have modified a High-Frequency Reciprocating Rig tribological tester to allow movement of the flat, in a direction perpendicular to the direction of reciprocation, during a ball-on-flat scuffing test. After initiation of scuffing, each flat was moved at a constant velocity, propagating the scuffing effect sideways from the reciprocating line in order to affect an area. After cleaning, the resistance of such areas to scuffing was measured in a second test. Flats of hardened 440C stainless steel were used.