The nonlinear viscoelastic behavior of filled elastomers is examined in detail using a variety of samples including carbon-black filled natural rubbers and fumed silica filled silicone elastomers. New insights into the Payne effect are provided by examining the generic results of sinusoidal dynamic and constant strain rate tests conducted in true simple shear both with and without static strain offsets. The effect of deformation history is explored by probing the low amplitude modulus recovery kinetics resulting from a perturbation by a large strain deformation such as a sinusoidal pulse or the application or removal of a static strain. It is found that a static strain has no effect on either the fully equilibrated dynamic (storage and loss) moduli or the incremental stress-strain curves taken at constant strain rate. The reduction in low amplitude dynamic modulus and subsequent recovery kinetics due to a perturbation is found to be independent of the type of perturbation. Modulus recovery is complete but requires thousands of seconds, and is independent of the static strain. The results suggest that deformation sequence is as critical as strain amplitude in determining the properties, and that currently available theories are inadequate to describe these phenomena. The distinction between fully equilibrated dynamic response and transitory response is critical and must be considered in the formulation of any constitutive equation to be used for design purposes with filled elastomers.

A series of in situ synchrotron X-ray diffraction experiments are performed during the stretching of weakly and highly vulcanized carbon black (CB), silica and grafted silica filled natural rubber sample (NR). Conversely to literature, Mullins effect observed after one stretching cycle modifies the strain induced crystallization (SIC) behaviour of the sample. The onset of crystallization is ruled by the strain amplification induced by the filler presence. Moreover, fillers (CB and silica) behave as additional crosslinks into NR network, through fillererubber interactions that either accelerate or slow down the crystallization rate depending on NR matrix chemical crosslink density. This is consistent with the assumption that effective network density, which is due to chemical crosslinks, entanglements, and fillererubber interactions, controls the crystallization rate.

No abstract