Strain-stiffening elastomers fail from the edge

Abstract: The accurate measurement of fracture resistance in elastomers is essential for predicting the mechanical limits of soft devices. Usually, this is achieved by performing tearing or peeling experiments on thin-sheet samples. Here, we show that these tests can be surprisingly thickness-dependent, with thicker samples being significantly stronger than thinner ones. Even for a simple geometry, direct imaging of the fracture surface shows that the fracture process actually involves three distinct cracks: an inner crack, and two edge cracks. Ultimately, samples fail when two edge cracks meet at the sample's mid-plane. The opening angle of edge crack, $2 \theta$, determines how far the sample has to be stretched before the edge cracks meet. Conveniently, $\theta$ is a material property that can be inferred from the elastomer's non-linear elastic response. To yield thickness-independent fracture-test results, sample thickness should be much smaller than the smallest lateral sample dimension divided by $\tan \theta$. Our results have direct implications for characterizing, understanding, and modelling fracture in soft elastomers.