Modeling of Solidification and Microstructure Evolution in the Scanning Laser Epitaxy (SLE) Process for Additive Manufacturing With Nickel-Base Superalloy Powders

Abstract

This paper investigates effects of natural and Marangoni convection on the resultant solidification microstructure in the scanning laser epitaxy (SLE) process. SLE is a laser-based additive manufacturing process that is being developed at the Georgia Institute of Technology for the additive manufacturing of nickel-base superalloys components with equiaxed, directionally-solidified or single-crystal microstructures through the laser melting of alloy powders onto superalloy substrates. A combined thermal and fluid flow model of the system simulates a heat source moving over a powder bed and dynamically adjusts the thermophysical property values. The geometrical and thermal parameters of the simulated laser melt pool are used to predict the solidification behavior of the alloy. The effects of natural and Marangoni convection on the resultant microstructure are evaluated through comparison with a pure conduction model. Inclusion of Marangoni effect produces shallower melt pools compared to a pure conduction model. A detailed flow analysis provides insights into the flow characteristics of the powder, the structure of rotational vortices created in the melt pool, and the solidification phenomena in the melt pool. The modeling results are compared with measurements and observation through real-time thermal imaging and video microscopy to understand the flow phenomenon. In contrast to the single weld-bead approach, the raster scan in SLE allows every position in melt pool to be visited twice by the solid-liquid interface as the scan source progresses. To properly address this situation, time tracking is incorporated into the model to correctly couple the microstructure prediction model. An optimization study is carried out to evaluate the critical values of the transition parameters that govern the columnar-to-equiaxed transition (CET) and the oriented-to-misoriented (OMT) transition. This work is sponsored by the Office of Naval Research through grant N00014-11-1-0670.