A comparison of models for predicting the true hardness of thin films

Instrumented indentation is widely used to characterize and compare the mechanical properties of coatings. However, the interpretation of such measurements is not trivial for very thin films because the hardness value recorded is influenced by both the deformation of the film and that of the substrate. An approach to extract the mechanical properties of films or coatings as an alternative to the experimental hardness measurement versus the indentation depth involves the use of composite hardness models. However, there are always uncertainties and difficulties in correctly deconvoluting the film hardness in experiments on composite materials.To justify their approach, some authors argue that their model is correct if the predicted hardness obtained for the coating provides a good fit to the experimental data. This condition is, of course, necessary, but it is not sufficient. A good fit to the experimental curve does not guarantee that a realistic value of the film hardness is deduced from the model. In this paper, different models to describe the composite hardness were tested by indenting a Ni–P coating. Its thickness was chosen to be sufficiently large such that its mechanical properties were perfectly known. We show that some models extensively used in the literature are inadequate to extract the film-only hardness without the effects of the substrate when the indentation range is too limited, although they predict the composite hardness very well.

► Models with only 2 parameters are more robust than models with 3 parameters.
► The Jönsson Hogmark model works well even so the indentation depth is low.
► The reverse analysis may give erroneous coating hardness if data are pertubated.