Influence of feedstock properties on the injection molding of aluminum nitride

Powder injection-molding (PIM) simulations are useful to identify appropriate combinations of material, process, and geometry variables required for successful manufacturing outcomes. PIM simulations can identify optimized processing parameters without the requirement of expensive trials when changes are made to feedstock composition or geometry design. However, PIM simulations require physical, thermal, and rheological feedstock properties as input data that require additional time, expertise, and expense. Whereas injection-molding simulation platforms typically offer over 5000 listings of property datasets for polymers, there are presently fewer than 5 such listings for ceramics and metals. The present work compares experimentally measured physical, thermal, and rheological properties for an aluminum nitride (AlN) feedstock to estimated values based on known filler properties and semi-empirical models. Injection-molding experimental studies carried out on a simple test geometry showed reasonable correspondence to PIM simulations using the two datasets. Further, mold-filling simulations were performed on complex heat-sink substrate geometries to compare the output of PIM simulations using experimentally measured and estimated feedstock property datasets. The present study indicates the merit of using estimated feedstock properties as input parameters in mold-filling simulations that could be extended for a variety of material systems and geometries early in the PIM design stage.