In contact to the substrates of mating parts reactive adhesives can develop interphases. Influenced by the substrate these interphases show altered properties and deviate from the polymer bulk far from the surface. Depending on the kind and the size of this volume the bondline thickness may influence the macroscopic characteristics of the joint. Mechanically it is possible to find effects like “thinner is less stiff” or “thinner is stiffer”. Such size-effects foil simple calculus of joints with varying bondline thickness, because bulk characteristics can no longer be attributed to thin bonds. Such a size-effect of the kind “thinner is less stiff” was examined by extensive shear-testing of anodized aluminum joints bonded with polyurethane in varying thicknesses. Together with measured bulk-properties the obtained data is used to test a newly developed model of continuum mechanics, which is able to account for such size effects by interpolating between the interphase and the bulk. Additionally similar and other bonds with different polymers and different substrates are examined for interphase effects by DSC and Nanoindentation.

In this study we examine polyurethane bonds of varying thickness between anodised aluminium substrates. The performed shear tests showed an intriguing size effect of the kind “thinner equals softer”. This size effect occurs not only in the basic elasticity (relaxed state), but also in the viscoelastic behaviour of the tested material. The cause of such size effects is supposed to be found in the existence of so-called interphases or boundary layers, which may differ considerably from the bulk in terms of mechanical behaviour, thus having an enormous impact on thin bonds. In thick bonds, however, these interphases or boundary layers have a minor effect on the overall mechanical behaviour. To account for these experimental results in bond modelling, an extended phenomenological continuum mechanics-based model, which explicitly includes such size effects in its calculation, is developed and presented. For this purpose, an abstract structure parameter with its corresponding balance equation is established describing the formation of the interphases by means of a phase transition. This makes it possible to define the bond stiffness at a macroscopic level, without entering into the microstructure. The extended model brings up a set of model parameters, which are determined efficiently by an ES (evolution strategy). The study concludes with a summary and an outlook on our further research work.

The non-linear viscoelastic behaviour of a polyurethane (PUR) network is determined by continuous uniaxial tension tests and stepwise relaxation tests. Following the concept of internal variables on the modelling side, the finite Neo-Hookean material model combined with linear evolution equations for the internal variables is applied to include the time-dependence caused by viscoelasticity. The parameters are identified by an evolution strategy combined with a non-linear finite element analysis which solves the boundary value problem given by the specimen geometry and testing conditions. In conclusion, the combination of the described experiments and modelling provides a full description of the mechanical behaviour of the given PUR.

It is known from applications that the mechanical behaviour of polymer bonds does not only depend on the properties of the polymer itself but also on the substrate. Therefore, the mechanical behaviour, i.e. the stiffness, of a polymer joint becomes thickness dependent. In the present work we describe experiments performed on polymer joints and we develop a continuum-based model which is able to describe the experimentally observed size effects without suggesting the microstructure in detail. The continuum mechanical model is enhanced by a scalar-valued structure parameter which describes all the effects taking place in the boundary layer which arises near the substrate. It is shown that the model parameters can be determined on the basis of simple shear experiments performed on polymer layers of different thickness.