- Ruhr-Universität Bochum

In this funding phase, the need for applying and validating data science approaches for the analysis and interpretation of microstructures and thermomechanical properties represents a new challenge. We use a Bridgman seed technique for single crystal solidification experiments, which allows to precisely control all parameters, which govern solidification. The figure shows the Bridgman furnace which is used in project B7.

A special effort is made to investigate the formation and evolution of crystal mosaicity in single crystal Ni-base superalloys. In general, mosaicity is a measure of the spread of crystal plane orientations. A superalloy single crystal exhibiting mosaicity is typically characterized by the presence of small sub grains with misorientations of up to a few degrees. We show that crystal mosaicity is associated with dendrite deformation processes. Therefore, we develop a new tomographic method, which is based on serial sectioning. In a first step, optical micrographs of as-cast superalloys are obtained by a quantitative metallographic procedure using precisely aligned montages of large numbers of optical micrographs. In a second step, neuronal network-based techniques are applied to detect and to evaluate the growth behavior of a large number of individual dendrites, to determine their growth trajectories and to investigate local dendrite spacings.

This approach allows to study the evolution of microstructures by monitoring the growth behavior of more than 2500 dendrites. Sudden changes in growth directions can be observed and documented. These types of orientation defects are related to dendrite bending which is based on a slight rotation of the crystal lattice due to local stress fields. Furthermore, using tomographic microstructure snapshots, we can document how dendrite branching and the interaction between slightly misaligned individual dendrites affect the microstructural features after solidification.