6 Summary and conclusion

We have presented a review of some of the recent results on large scale MD modelling of dynamic fracture. We have reported progresses in three areas of focus: (i) crack limiting speed, (ii) crack tip instabilities, and (iii) crack dynamics at interfaces.

The main conclusion is that hyperelasticity, the elasticity of large deformation, is crucial in order to form a clear picture of the failure process near a rapidly moving crack tip: Both the maximum crack speed and the dynamic instability are strongly influenced by the large-strain elastic properties. Phenomena such as intersonic mode I fracture (Figure 9) and supersonic mode II fracture (Figure 10) can not at all be addressed by classical, linear elastic theories. The observation of stable cracks in homogeneous materials at speeds beyond the Rayleigh-wave speed, as shown in Figure 23, can only be understood from a hyperelastic point of view. Further, we have shown that hyperelastic softening can significantly reduce the critical instability speed in cracks, as shown for instance in Figure 19. Classical linear elastic theories such as the Yoffe-model [78] fail when significant softening is present at the crack tip. Our new concept of a characteristic energy length scale % helps to understand the relative significance of hyperelasticity in dynamic fracture (see Figure 13), both for the limiting speed of cracks (Figure 12) and for the instability dynamics (Figure 20).

Further, we find that interfaces and geometric confinement can play an important role in the dynamics of cracks.