ABSTRACT
Modern X-ray diffraction techniques are now allowing researchers to collect long-desired experimental verification data sets that are in situ, three-dimensional, on the same length scales as critical microstructures, and using bulk samples. These techniques need to be adapted for advanced material systems that undergo combinations of phase transformation, twinning and plasticity. One particular challenge addressed in this article is direct analysis of martensite phases in far-field high-energy diffraction microscopy experiments. Specifically, an algorithmic forward model approach is presented to analyze phase transformation and twinning data sets of shape memory alloys. In the present implementation of the algorithm, the crystallographic theory of martensite (CTM) is used to predict possible martensite microstructures (i.e. martensite orientations, twin mode, habit plane, twin plane and twin phase fractions) that could form from the parent austenite structure. This approach is successfully demonstrated on three single- and near-single-crystal NiTi samples where the fundamental assumptions of the CTM are not upheld. That is, the samples have elastically strained lattices, inclusions, precipitates, subgrains, R-phase transformation and/or are not an infinite plate. The results indicate that the CTM still provides structural solutions that match the experiments. However, the widely accepted maximum work criterion for predicting which solution of the CTM should be preferred by the material does not work in these cases. Hence, a more accurate model that can simulate these additional structural complexities can be used within the algorithm in the future to improve its performance for non-ideal materials.
