A review was presented of the mechanisms by which cracks nucleated in polycrystalline samples. In the absence of pre-existing cracks, crack nucleation was the first step towards mechanical failure. Various mechanisms were considered, mainly involving the development of micro-cracks from their initial sizes, below the level of easy detection, to easily observable cracks. It was noted that true crack nucleation, without pre-existing precursors, required that the stresses be locally concentrated to levels that attained the theoretical cleavage or cohesive strength. Possible mechanisms for this stress concentration included dislocation glide (leading to pile-ups of dislocations on particular slip planes), grain-boundary-sliding (leading to stress concentrations at the triple junctions), thermal expansion of inclusions (such as those produced by freezing of brine pockets) and elastic anisotropy of the crystals. Dislocation glide and boundary sliding were processes during which stress redistribution occurred at a finite temperature-dependent rate. They also produced a temperature-dependent internal friction and anelasticity and their operation could be independently measured. Elastic anisotropy led to stress concentrations in an athermal manner and became more important at very low temperatures and high loading rates. In the absence of stresses due to brine pockets, the inherent elastic anisotropy was insufficient to nucleate cracks in purely 2-dimensional models: such as perfectly columnar grains in plane strain. The estimated nucleation energies far exceeded any available thermal activation, and some stress-concentrating effect was required. Experiments on fresh-water ice had provided evidence that crack nucleation was consistent with grain-boundary sliding and associated stress-concentration fields.

Mechanisms of Crack Nucleation in Ice. H.J.Frost: Engineering Fracture Mechanics, 2001, 68[17-18], 1823-37