A molecular picture of elastomer fracture

In threshold conditions the molecular model of Lake and Thomas predicts that when a network breaks, the minimum amount of energy dissipation is due to the breakage of chemical bonds crossing the fracture plane over a thickness of the order of the mesh size, and each broken strand frees an energy which evolves linearly with the strand’s length. This model predicts the right scaling and order of magnitude for fracture energy as a function of strand areal density and length even when viscoelastic dissipation isn’t negligible. Up to this day, why this model works so well remains an open question. The stochastic nature of the polymer network inevitably leads to early breakage of shorter strands which encourages us to believe that bonds have to break over a thickness much larger than the mesh size when a crack propagates. The existence of significant bond scission in the bulk was confirmed experimentally by Slootman et al. for a simple acrylate networks.

In this work, we synthesized soft networks of poly ethyl acrylate (Tg = -20°C) at various crosslink densities, labeled with mechanosensitive crosslinkers that become fluorescent when they break. Crack propagation tests were then carried out on notched samples and we extracted the energy release rate during crack propagation and mapped the concentration of broken crosslinkers from the activation of mechanophores. For a given energy release rate, the areal density of broken crosslinkers was less dependent than expected on crosslink density but each broken crosslink appeared to dissipate energy proportional to its average length to the power 2.4. In our experimental system the two errors compensate each other. This experimental observation brings light to why the molecular model of Lake and Thomas predicts so well the scaling of the fracture energy with network structure even though the mechanism is physically more complex.

[1] Lake, G.J. and A.G. Thomas, Proc. Roy. Soc. A, 1967. A300: p. 108-119.

[2] Slootman, J., et al., Physical Review X, 2020. 10(4): p. 041045.