The control of crystallographic preferred orientation (CPO) development and grain size reduction on ice mechanical weakening (enhancement)

Abstract:

Ice creep experiments show a change in ice “strength” as initially isotropic polycrystalline ice is deformed. Mechanical weakening manifests as a strain rate “enhancement” after secondary creep (minimum strain rate) in constant load experiments or a stress decrease after peak stress in constant displacement rate experiments. This mechanical weakening correlates with development of a crystallographic preferred orientation (CPO) and also with other microstructural changes, including grain size reduction. We deformed isotropic polycrystalline pure water ice samples to successively higher strains under constant load at -4 °C or under constant displacement rates at -10 and -30 °C. Most of the samples were deformed under uniaxial compression. We also deformed samples at -30 °C under direct shear to strains higher than those reached in uniaxial compression experiments. We collected the CPO and grain size data from these samples using cryo-EBSD (electron backscatter diffraction). CPO and grain size data were used to model the potential effects that these might have on mechanical strength. CPO data were used to calculate the critical resolved shear stress (CRSS) on the ice basal plane for each pixel of the EBSD maps and from this an area integrated CRSS calculated for the sample. The ratio of each sample’s integrated CRSS with the integrated CRSS of an isotropic sample (the starting material) gives an estimate of the weakening effect of the CPO. Grain size data were used together with grain size sensitive ice flow laws to model the evolution of bulk strain rate for constant load experiments or the evolution of bulk stress for constant displacement rate experiments. The modelling results were compared with measured mechanical data.

The ratio of the measured peak and flow stresses (the normalised magnitude of weakening) is about the same in all constant displacement rate experiments, regardless of temperature and strain rate. Integrated CRSS values suggest that basal planes become better aligned to resolved shear stresses with increasing strain in all experiments; however, the increase in the integrated CRSS on basal planes is less significant at lower temperatures. This observation suggests that mechanical weakening cannot be attributed entirely to development of a CPO. Grain size reduces with strain in all experiments and application of a grain size sensitive flow law predicts weakening. The predicted stress reduction in constant rate experiments changes with temperature; it is lower than that measured at -10 °C, but higher than measured at -30 °C. Modelling grain size reduction underestimates the magnitude of mechanical weakening at a warm temperature and overestimates it at colder temperatures.

When projected onto a deformation mechanism map, stress – strain rate data, from our experiments and from other published experiments, lie close to the mechanism boundary where the strain rate contribution of grain size sensitive (dislocation creep) and grain size insensitive (grain boundary sliding: GBS) mechanisms are equal. Given that dislocation creep is likely to lead to CPO development and GBS is inherently grain size sensitive, this is consistent with both CPO development and grain size reduction contributing to weakening in laboratory experiments. The different balance of their contributions at different temperatures may relate to different balances in recrystallisation mechanisms.