Computational Modeling of Microstructure Formation During Laser Powder Bed Fusion of Bulk Thermoelectric Materials

Additive manufacturing has enabled the production of complex geometries with enhanced functionalities. The advancements in additive manufacturing, particularly in laser powder bed fusion (PBF-LB), underscore the need for computational simulation efforts to better understand grain formation in laser-processed thermoelectric (TE) materials. Accurately predicting the process-structure relationship is essential for extending the capabilities of PBF-LB in additively manufacturing multifunctional materials. PBF-LB works by scanning a laser upon powder layers, melting and fusing the material into hierarchically structured shapes. Processing parameters such as laser power, laser scan speed, laser spot size, hatch distance, scanning strategy as well as the setup parameters such as powder layer thickness influence the melt pool solidification behavior, which directly affects the resulting microstructure in bulk parts. This study focuses on two well-known thermoelectric materials: bismuth telluride, a low-temperature thermoelectric material, and silicon germanium, a high-temperature thermoelectric material. The microstructure formation of bulk bismuth telluride and silicon germanium parts are computationally predicted using the finite difference Monte Carlo (FDMC) method used in the Stochastic Parallel PARticle Kinetic Simulator (SPPARKS) code developed by Sandia National Laboratories. The model achieves experimental bulk dimensions on the millimeter scale. Simulations of uni-directional and bi-directional laser scan strategies, including the 90-degree rotation between layers, are compared to experimental results. This work represents the first computational study of grain structure formation in thermoelectric bulk parts produced via PBF-LB. The findings indicate that microstructure formation can be controlled by adjusting scan strategies, which influence grain orientation, and by modifying laser processing parameters such as laser scan speed and laser power, which affect grain size. This approach offers a pathway to optimize the process-microstructure-property relationship in additively manufactured multifunctional materials.