Pattern Selection in Model Alloys Under Additive Manufacturing Conditions
Brian Rodgers, Colorado School of Mines
Additive manufacturing has rapidly risen from a Star Trek-like concept to an industry worth tens of billions. The term Additive Manufacturing (AM) applies to a variety of technologies; this work focuses on laser based fusion methods. Laser melting leads to high interface velocities during solidification, with speeds approaching or exceeding the maximum solute flux in the liquid. Rapid solidification and cooling forms unusual non-equilibrium microstructures. These microstructures can have advantageous structural and functional properties. Utilizing the advantages afforded by rapid solidification microstructures requires understanding and predicting their formation. Current methods are routinely unable to predict which solidification pattern will be selected under a given set of conditions. This work conducts laser melts of model alloys under AM conditions; solidification is directly observed with two in-situ methods. The first is x-ray radiography at the Advanced Photon Source (APS) synchrotron at ANL (Argonne National Laboratory), which is more representative of bulk behaviors at length scales like an AM build. The second is Dynamic Transmission Electron Microscopy (DTEM) at Lawrence Livermore National Laboratory (LLNL), with finer temporal and spatial resolutions well suited to exploring fundamental solidification behaviors. Early results from these experiments confirmed the onset of absolute stability is not well predicted by conventional stability analysis. Updating values in the model with experimental data, rather than theoretical values, still yields incorrect predictions. Additionally, experimental analysis has revealed several oddities with the crystallographic orientations of the solidified microstructures. Dendritic growth directions are misaligned from the expected directions, indicating a dendrite orientation transition. Regions of concentrated solute develop perpendicular to the grain boundaries, and twin-like misorientations develop across pairs of these regions. Some of the concentrated solute regions precipitate out as a new phase, but with an inconsistent orientation relationship. There is an expected, minimal free energy orientation relationship, yet the precipitates will sometimes form a novel orientation relationship or simply do not follow one at all.