Abstract:A dynamic crack will travel in a straight path up to a material-dependent critical speed beyond which its path becomes erratic. Predicting this critical speed and discovering the origin of this instability are two outstanding problems in fracture mechanics. We recently discovered a simple scaling model based on an effective elastic modulus that gives successful predictions for this critical speed. We now show that a simple atomic picture based on broken-bond relaxation at the dynamic crack tip provides an explanation for the origin of the effective elastic modulus.
PACS: 62.20.MKIn 1951, Yoffe [1] made the physically intuitive suggestion that mode I crack growth occurs in the direction of maximum asymptotic hoop stress and found the crack speed for the onset for branching to be about 70% of the Rayleigh wave speed c R [2,3]. However, this high speed is rarely observed in experiment [4,5]. An obvious shortcoming in Yoffe's analysis is the assumption of a constant linear elastic response for all deformations. In our recent study of brittle fracture [6], we showed that hyperelasticity, the elasticity at large strain, plays a governing role in the onset of the crack instability from unidirectional motion. We discovered a simple, yet remarkable, scaling based on an effective elastic modulus for our modelled solid (the secant modulus at the stability limit of the bulk solid), which led to successful predictions for the onset speed of the crack instability. We have also applied this scaling to the same-modelled solid with the exception that the crack is constrained to travel unidirectional irrespective of its speed [7]. This allowed the crack to achieve a unique steady-state speed that has a dependence on hyperelasticity. Using our scaling law, we found that the steady-state crack speed scales to a constant value equal to a crack speed of a linear solid with our effective elastic modulus. In this paper, we demonstrate that atomic relaxation of breaking bonds at the crack tip governs these dynamic features of the travelling brittle crack.We summarize our earlier findings. Our simulation model is based on a generalized bilinear force law composed of two spring constants, one associated with small deformations (k 1, r < r on ) and the other associated with large deformations (k 2 , r > r on) . This is shown in Fig. (1a). This model allowed us to investigate the generic effects of hyperelasticity by changing the relative magnitude of the spring constants = k 2 /k 1 and transition distance r on of the potential [in terms of 0 = (r on /r 0 ) -1]. We considered the propagation of a crack in two-dimensional hexagonal lattice geometry. The slab is loaded in mode I with a constant strain rate. The dynamic crack instabilities for the various = k 2 /k 1 are associated with the precipitous drops in crack speed (see Fig. 1b), as indicated by the arrows, and *Address correspondence to this author at the Lawrence Livermore National Laboratory, Livermore, CA 94550, USA; E-mail: abraham4@llnl.gov are a consequence of the crack ...