The dynamics of substructures, which encompass all structures present at the subgrain-scale, were investigated by static, in-situ annealing experiments. Deformed, single crystal halite was annealed inside a scanning electron microscope at temperatures between 280-470 ºC. Electron backscatter diffraction maps provided detailed information about crystallographic orientation changes. Three temperature dependent regimes were distinguished based on boundary misorientation changes. In regime I (280-300 ºC) some low angle boundaries (LABs), i.e. with 1º-15º misorientation, increase in misorientation angle, while others decrease. In regime II (~300 ºC) all LABs undergo a decrease in misorientation angle. Regime III (>300 ºC) is defined by enhancement of the subgrain structure as remaining LABs increase and some undergo a rotation axis change. Throughout regimes I and II, new LABs develop, subdividing subgrains. LABs could be divided into four categories based on annealing behaviour, orientation and morphology. We suggest that these observations can be directly related to the mobility and activation temperature of climb of two dislocation groups introduced during deformation. Therefore, with in-depth investigation of a substructure with known deformation geometry, we can infer ratios of dislocation types and their post-deformation and post-annealing location. These can potentially be used to estimate the post-deformational annealing temperature in crystalline materials.

In-situ 3D X-ray diffraction (3DXRD) annealing experiments were conducted at the ID-11 beamline at the European Synchrotron Radiation Facility in Grenoble. This allowed us to non-destructively document and subsequently analyse the development of substructures during heating, without the influence of surface effects. A sample of deformed single crystal halite was heated to between 260-400 ºC. Before and after heating a volume of 500 by 500 by 300 mm was mapped using a planar beam, which was translated over the sample volume at intervals of 5-10 µm in the vertical dimension. In the following we present partially reconstructed orientation maps over one layer before and after heating for 240min at 260 ºC. Additional small syn-heating “maps” over a constrained sample rotation of 12-30º. The purpose of this was to illuminate a few reflections from 1 or 2 subgrains and follow their evolution during heating.

Preliminary results show that significant changes occurred within the sample volume, for which, surface effects can be excluded. Results show a number of processes, including: i) change in subgrain boundary misorientation angle and ii) subgrain subdivision into areas of similar lattice orientation with new subgrain boundary formation. These results demonstrate that 3DXRD coupled with in-situ heating is a successful non-destructive technique for examining real-time post-deformational annealing in strongly deformed crystalline materials with complicated microstructures.

Most in-situ heating experiments where substructure is investigated have been restricted to 2D. We compare a 2D experiment to a 3D X-ray diffraction experiment to evaluate the validity of the 2D method. Until now 3D X-ray diffraction has been limited to well-recovered substructures. We conducted a 3D in-situ annealing experiment on a halite crystal with a significant orientation gradient. This is the first experiment of its kind on a geological material and shows that even complicated microstructures can be resolved. Comparison of 2D and 3D showed that, although general results were similar, subgrain boundary movement occurred with higher frequency in 3D. We suggest this discrepancy is due to enhanced drag force on subgrain boundaries by surface thermal grooving. Thus, while results from 2D experiments largely reflect what is happening in the volume, analysis of boundary movement with regard to absolute mobilities needs to be considered with some care.

A new, deterministic model for recovery integrated with the microstructural modelling platform Elle is presented here. Experimental data collected from 2D in-situ annealing experiments were used to develop and verify the simulation. The model is based on the change of strain energy related to misorientation when a virtual rotation is applied to a crystallite (i.e. physical data point). Boundary energies are calculated using the Read-Shockley relationship. The axes of rotation were selected based on the deformation geometry and potentially activated slip systems. Crystallographic rotation was applied in the case where largest reduction of energy was observed. The effect of parameters such as rotation mobility, neighbourhood size, critical misorientation and specific energy calculation method were systematically investigated. Simulations reproduced many aspects of the experiments and showed that processes were highly dependent on dislocation type and increase of long-range effects with temperature. The results suggest that dislocations remain independent entities for longer than expected, even in an organised subgrain boundary. The model could not, however, retain higher angle boundaries and always resulted in a general shift of boundary distributions towards lower angles. We suggest that the classic interpretation of boundary energies does not entirely work for misorientations that lie in the less defined part of the Read-Shockley relationship.