Abstract
This study investigates the effectiveness and limitations of grain-size piezometers in ice undergoing either grain-size growth or grain-size reduction during deformation. Microstructural analyses were conducted on laboratory deformed ice samples with controlled starting grain-sizes and initially random crystallographic preferred orientations (CPOs). The resulting ‘steady state’ microstructures were analysed using high resolution cryo-EBSD to quantify final grain-sizes as well as CPO and misorientation data.
We identified a piezometric relationship of grain size in polycrystalline ice (ice 1h) samples that have undergone grain-size reduction during deformation. We use these samples to calibrate a grain-size, subgrain-size, and recrystallised grain-size piezometer for ice. Samples that underwent grain-size growth were excluded from the piezometer due to their deviation from expected piezometric relationships; it is uncertain if they obtained a steady state. This thesis explores a new barycentre method for segregation of subgrains, to define a subgrain size piezometer. The relative paucity of subgrain structures across the range of stresses makes it difficult to define a useful subgrain piezometer. Segregation of recrystallised grains is difficult as the dominant recrystallisation process is grain boundary migration and existing segregation methods, established in materials at lower homologous temperatures and based on intragranular distortion, do not work well. A segregation method based on grain boundary irregularity has some success and a recrystallised grain size piezometer based on this approach has some potential.
Measurements of intragranular distortion are used to calculate geometrically necessary dislocation densities for specific basal, prismatic and pyramidal dislocation systems. Highest dislocation densities are for basal dislocations (c) and <a>. Data from samples, including those that underwent grain size reduction and grain growth, have increased dislocation densities with increased flow stress during deformation. A log-log plot of dislocation density versus stress for the basal dislocations gives a good linear slope suggesting that a geometrically necessary dislocation density piezometer might be developed for ice.
We conducted static grain growth experiments with undeformed samples over the same time period at -2 °C as samples undergoing deformation. These experiments show that grains grow more in the deforming samples than in the static samples. We infer that there must be an additional driving force to promote this faster growth, and the likely candidate is strain energy. When we examine the growth data relative to a deformation mechanism map, we see that there is an identified rate controlling factor of grain-size growth relative to grain-size reduction. This controlling factor is suggested to be due to the distance from the mechanism boundary. Samples with starting grain-sizes closer to the boundary experience significantly greater rates of growth.
Sample microstructures have complex correspondence to grain size and stress. To help understand these better we defined simple thresholds of qualitative microstructural parameters such as the sphericity parameter (boundary shape irregularity), M-index (CPO strength), subgrain proportions, and dislocation density. Sample points on a deformation mechanism map were coloured based on being above or below thresholds and this analysis used to infer transitions between deformation (grain boundary sliding, dislocation creep) and recrystallisation mechanisms (grain boundary migration, subgrain rotation).
We applied the grain-size piezometer to natural ice from the Priestley Glacier to obtain stress estimates. A recent experimentally calibrated easy slip flow law for the Priestley Glacier allowed a stress estimate independent of the piezometer. Stress values of 0.12 and 0.18 MPa were calculated for the piezometer and flow law methods respectively, which is within error of ± 0.05 MPa.