CuNi films are interesting for resistive and thermoelectrical applications. A heat treatment (e.g. under Ar atmosphere) is used for microstructural stabilization. During annealing (i) a surfacial oxidation occurs and (ii) a high tensile film stress develops being critical for stability and reliability. In this paper, the correlation between stress evolution and oxidation is considered.

The stress evolution was investigated on a 400 nm thick Cu_0.57Ni_0.42Mn_0.01 film on oxidized silicon during a thermal cycle to 550 °C using a laser-optical substrate-curvature technique. Cycle-stop-prepared samples were used to clarify the correlation between stress development, microstructural evolution, and oxidation. The microstructure was studied by transmission electron microscopy. Oxidation was investigated by Auger electron spectroscopy and resistance measurements.

The grain morphology is columnar with a medium diameter of 20 nm and is nearly unaffected by annealing. The degree of oxidation depends on the ambient atmosphere. Whereas in Ar the oxide layer grows with increasing temperature, only a thin oxide is formed in a N₂/H₂ atmosphere. By comparing the results, the oxidation-induced stress contribution under Ar could be separated. In the stress curve, the striking feature associated with oxidation is a tensile stress component of about 500 MPa around 350 °C. At this temperature, a surfacial NiO layer of some nanometers thickness is formed. The stress component is semiquantitatively explained by grain-boundary diffusion of Ni to the surface and the induced contraction within CuNi due to Ni lost. With progressive oxidation (1) a CuO-NiO double layer forms, growing without essential contributions of diffusion within CuNi and, therefore, without oxidation-induced stress development and (2) the tensile stress relaxes presumably by plastic deformation (diffusional flow, mechanical twinning) as already observed in unoxidized CuNi films.

Ultra hard films deposited on steel substrates are usually compressed. Therefore, they tend to swell and surge. If this happens, the coating breaks. We study the conditions for the splitting of a compressed thin film off a substrate. The force that tears down the film from the substrate is due to the compression (negative tension) in the film. The origin of the compression is twofold: (i) the deposited films are formed compressed and (ii) the thermal expansion causes additional compression when the films cool down. The force that affixes the film to the substrate is the surface tension. We found the critical tension for the splitting instability. The compression stronger than critical leads to the film rising.

We studied the following problem: the substrate is semi-infinite, and it fills the space with z<0. A solid film with the thickness h and with the surface tension coefficient between the film and substrate γ, is affixed to the substrate. We found that the film splits off when absolute value of the negative the in-plane tension exceeds 6γ/h. Note that the tension depends on temperature because of the difference in thermal expansion of the film and substrate.

Actually, hard coatings consist of many layers of materials with different elastic properties. We found a condition for the splitting instability of a layered structure on a substrate surface, which is a generalization of the above criterion. In this case, splitting can occur between any pair of layers or between the substrate and the first layer. We developed software, which computes the critical tension for a layered coating. This allowed us to optimize the coating composition with respect to the applied tension.

Boundary element modelling is a numerical tool especially well suited for the analysis of stress states in media. In a series of preceding papers we have shown this technique to be especially economic and powerful for bi-layer modelling under elastic contact loads.

In the present paper we extend the 2D mechanical analysis of very thin coatings to the more complex situation of a 3D ball on flat contact.

We show that the main results of the 2D case still hold ; an important stress gradient in the coat-ing and a sharp stress discontinuity at the interface.

Results are given for coatings roughly twice as stiff as the substrate as well as for the inverse situation.

The effect of frictional surface stress and various "model" internal stress fields is also analysed. These findings are of practical interest for general elastic contact as occurring in many lubricated friction applications of very thin films. However they also apply to the initial phase of scratch-testing where, for a given, absolute coating thickness, t, the relative thickness t/a_H (a_Hbeing the hertzian contact radius) keeps decreasing as the contact load increases.