Reinforcement of rubber by nanoscopic fillers induces strong nonlinear mechanical effects such as stress softening and hysteresis. The proposed model aims to describe these effects on a micromechanical level in order to predict the stress-strain behaviour of a rubber compound. The material parameters can be obtained by fitting stress-strain tests. These quantities have a clear defined physical meaning. The previously introduced "dynamic flocculation model" was extended for general deformation histories. Stress softening is modelled by hydrodynamic reinforcement of rubber elasticity due to strain amplification by stiff filler clusters. Under stress these clusters can break and become softer, leading to a decreasing strain amplification factor. Hysteresis is attributed to cyclic breakdown and re-aggregation of damaged clusters. When stressstrain cycles are not closed, not all of these clusters are broken at the turning points. For the resulting "inner cycles" additional elastic stress contributions of clusters are taken into account. The uniaxial model has been generalized for threedimensional stress states using the concept of representative directions. The resulting 3D-model was implemented into a Finite Element code, and an example simulation is shown. Good agreement between measurement and simulation is obtained for uniaxial inner cycles, while the 3D-generalization simulates the behaviour closer to the experiment than the original model.
IntroductionReinforcement of elastomers is essential for their application. Without nanoscopic filler additives such as carbon black or silica, rubber products as e.g. tires or seals would rupture or wear off very rapidly. Besides making the elastomer stiffer and tougher, the incorporation of fillers leads to a non-linear dynamic-mechanical response. An example is the Payne effect: the amplitude dependence of the dynamic moduli [1-3]. With increasing strain amplitude an increasing loss angle and a decreasing storage modulus are seen. The drop will be even more pronounced at higher amount of filler in the elastomer. We point out that the uniaxial stress-strain cycles and hysteresis of filler-reinforced rubbers are indeed strongly non-linear effects also at small strain amplitudes. The apparent linear behavior that is observed during standard dynamic mechanical testing even up to 100% strain amplitude and more is only a special case found in simple shear experiments. This is also called the "harmonic paradox" of the Payne effect [4,5] since an almost linear viscoelastic stress-strain behavior is observed also at large strain amplitudes. Because of the pronounced decrease of the storage modulus with increasing amplitudes, one would rather expect non-linear viscoelastic stress-strain behavior and no neat sinusoidal response as at low amplitude. One reason being that an ideal Gaussian statistics of polymer networks with constant volume (Neo-Hook model), which approximately applies up to moderate strain, leads to a linear relation between shear stress and strain, τ = Gγ, though the ...