Metal injection molding is a process enabling net-shape production of intricate parts with adequate dimensional accuracy and excellent mechanical properties. This process consists in four steps. First, the powder is mixed with a molten binder to produce the feedstock; second, the feedstock is injected into a mold cavity, cooled down, and then ejected from the mold; third, a debinding process is applied to the part to remove the binder; and fourth, sintering consolidates the part by solid-state diffusion which finally generates a dense metallic part. All these steps are important, but the injection is particularly critical since it can generate partimpairing defects such as cracks, voids and distortions. Its success depends on adequate control of the feedstock's fluidity, where injection parameters, powders characteristics and binder formulation are all factors at play. Although the general influence of each of these factors is well-known, their correlation with moldability is still too complex to allow efficient prediction of the feedstock performance during injection other than by real-scale injections or by melt rheology which are resource-intensive methods. In other words, development of the metal injection molding process is still impaired today by the lack of knowledge about the fundamental mechanisms of the rheological behavior of feedstocks. The objective of this work is to study some of these mechanisms by the exclusive observation of the impacts of dry powder characteristics on the injection step, which means that the injection parameters and the binder formulation remain unchanged, in order to isolate its influence. Four stainless-steel powders of different sizes and shapes were first characterized by the FT4 powder rheometer which generated fourteen metrics describing their flow properties. Following, each powder was mixed with the same molten binder to produce four feedstocks that were measured with a rotational rheometer to obtain their viscosity profile. Finally, the performance of the feedstock during the injection (i.e., the moldability) was evaluated by real-scale injections. The results demonstrated a strong correlation between moldability and melt viscosity, which confirms that moldability can be efficiently predicted by melt rheology although this method remains resource-intensive. Yet, strong correlations of the moldability were also observed with many metrics obtained by the dry powder rheology, especially the conditioned bulk density (CBD) and the normalized basic flow energy (NBFE). Considering that dry powder rheology requires much less resources than melt rheology, this method shows good promise to enhance the development and optimization of feedstocks for the injection step.
Langlais, D. (Author),
Demers (Supervisor) &
Brailovski (Co-supervisor),
19 Nov 2021Student thesis: Master's thesis › Master in Engineering: Mechanical Engineering