Fabric development and macroscopic deformation are studied here by examining the effects ofnearest-neighbour interaction (NNI) among crystals and dynamic recrystallization (Thorsteinsson, 2002).
The 'strength' of grain interaction can vary from no interaction (no-NNI), with homogeneous stress (Sachs, 1928), to 'strong' interaction (full-NNI), with significant stress redistribution. Increasing the NNI leads to a more homogeneous strain of all the crystals.
There are three dynamic recrystallization regimes for ice: normal grain growth (which has no effect on fabric; Gow, 1971), polygonization (small effect on fabric) and migration recrystal-lization (significant effect on fabric). As the grains strain, sub-boundaries (dislocation walls) may form due to heterogeneous deformation within grains that relieves stress concentrations. The formation of sub-boundaries can lead to the division of the
parent grain into two, or more, new grains, as the misorientation of the subgrains increases. This is called polygonization. The effect on the fabric development is small, because the orientation of the new grains usually deviates by less than 5° from the parent grain, but the effect on grain size is clearly visible in thin-sections (Thorsteinsson et al., 1997).
The formation of sub-boundaries can create small sections of grains that are in a strain shadow. Being strain-energy-free, these small grains can act as seeds for migration recrystallization. The idea adopted in the model is that within the crystal aggregate there are many such 'seeds' that provide potential nucleation sites for new grains. When the temperature is high enough for grain-boundary migration to be efficient (Duval & Castelnau, 1995), and/or the strain energy is greater than the grain-boundary energy, these seeds can quickly consume highly strained crystals, thus reducing the free energy of the system (Montagnat & Duval, 2000).
In studies of high-temperature (—5°C to 0°C) creep of ice, Kamb (1972) found that after only about 0.04 shear strain there was already strong evidence of recrystallization. A totally random orientation for the new grain is not to be expected. In uniaxial compression of ice, for instance, a small-circle girdle fabric forms (Budd & Jacka, 1989). This indicates that new crystals that form in orientations with high resolved shear stress (soft orientations) are favoured to grow. Figure 61.1 shows the model results when polygonization and migration recrystallization are active in uniaxial compression. The results are encouraging because they show both the typical girdle type fabric (Fig. 61.1a), and the nearly constant strain rate beyond 10% strain (Fig. 61.1b).
Was this article helpful?