In the literature, the loss factor of an elastomer, expressed as tan d, has been reported to decrease with strain. This has been interpreted as a lowering of the internal viscosity due to chain orientation under strain. This contrasts with experimental findings showing that up to large strains, loss modulus does not change with strain. These experiments show that, as measured, tan d does decrease with strain. However, a simple analysis shows that this effect is due to geometric changes alone and that the essential viscoelastic behavior, expressed as the loss modulus, is constant with strain up to an extension of about 2 for the simple unfilled compounds investigated here. This article also shows that the prestrain does not induce any significant anisotropy in the loss modulus, at least up to an extension ratio of 2. This was measured by the stretching of a rubber sample by simple extension and then the subjection of the sample to free vibrations in either torsion or tension. By the measurement of the damping in this way, the anisotropy induced by the preextension was measured. The experiments in this work were all conducted on unfilled natural rubber or styrenebutadiene rubber compounds, with 2 different natural rubber compounds being used to investigate the effect of the crosslink density.

Rivlin derived an elastic solution for the problem of a rubber cylinder subjected to combinations of tension and torsion. The theory correctly predicts the experimentally observed behaviour where the length of the cylinder increases as a twist is applied and conversely there is a reduction in the moment resulting from the application of an additional tensile loading. This problem is of interest for two reasons. First, the reduction in axial force with twist is second order by virtue of the change in the axial force being a symmetric function of the square of the twist angle. Therefore, the problem makes a demanding test of the applicability of the finite element method at solving this large strain elasticity problem. Second, the Rivlin elasticity solution could in principle be applied to any form of stored energy function. As a result, it is a useful test of the validity of different stored energy functions. Also, this test geometry has been examined previously by the authors to investigate any anisotropic changes in the viscoelastic behaviour in rubber with prestrain and it is important to validate the elastic solutions for this work.

Suphadon et al. (J App Polym Sci 2009, 113, 693) showed using small oscillations of less than 1% strain superimposed on a larger prestrain that the loss modulus, referred to the test piece dimensions after the application of the prestrain, did not vary with prestrain for unfilled rubber materials for a wide range of prestrains up to 100%. Also for unfilled rubbers it was observed that up to 100% prestrain that the loss modulus behavior was isotropic. This paper extends this previous work to larger prestrains for styrene butadiene rubber (SBR) compounds and natural rubber (NR) compounds some of which incorporate carbon black fillers. Both the storage modulus and the loss modulus are again calculated relative to the dimensions of the test piece after the application of the prestrain. These results show that for materials with 25 phr of carbon black filler, the loss modulus was still independent of the prestrain for normal engineering strains but at filler contents of 50 phr the loss modulus increases with prestrain at extension ratios less than 2. Even so over the typical engineering strains of below 50%, the loss modulus was still independent of strain. This increase in loss modulus at large prestrains can in part be explained by considering the molecular orientation of the polymer in combination with a consideration of the molecular slippage that takes place at the polymer filler interface.

No abstract

No abstract