Thumbnail Image

Producing near-net shape titanium alloy parts by metal injection moulding

Metal injection moulding (MIM), is a metalworking process for producing complex parts from metal powders. It is often more cost effective to use MIM to make small, precise componentry from expensive materials such as titanium than traditional metal processes such as casting, forming, extrusion, investment casting and machining. The four processing steps in MIM are: formulating homogenous feedstock of metal power and binders, injection moulding, removing the binder (debinding) and sintering. The binder affects feedstock rheology, allowing the required shape to be moulded. It then needs to be removed so the object can be sintered to consolidate into a near-net shaped part. As MIM has many advantages, including low wastage of material and the ability to produce intricate shapes, it is used in sophisticated applications such as biomedical instruments, in the aerospace industry, and in niche applications that use expensive materials such as titanium. The objective of this research was to identify binders and processing methods that produced easily moulded titanium-based feedstock and final products with minimal defects. This involves understanding how factors such as binder composition, powder loading, and processing method affect feedstock rheology, homogeneity, debinding and sintering, and to develop a model for investigating feedstock rheology. Feedstocks were manufactured by mixing different amounts of hydride-dehydride titanium alloy powder (HDH Ti-6Al-4V) with a three-component binder containing different amounts of polyethylene glycol (PEG), polyvinyl butyral (PVB) and stearic acid (SA). Most research on titanium powder metallurgy uses spherical, fine powders with mean diameters (d50) between 0.3 and 19 μm. This research used HDH Ti-6Al-4V, an irregular, coarse powder with a d50 of 52 μm and a d90 of 117 μm. The effect of using different types of mixers for various mixing times was evaluated by measuring mixing torque and thermal analysis. Feedstock homogeneity was evaluated by density measurements and DTA/TGA analysis, and mouldability was evaluated by rheological properties. After mixing and extrusion, binders were removed from the formed (green) part in a two-stage process involving aqueous debinding followed by thermal debinding. The part was then sintered to remove any remaining binder and to create the final product. Morphological characteristics of feedstocks, moulded, debound and sintered parts were assessed using optical and scanning electron microscopes. Polyethylene glycol, PVB and SA were used as binder components because they are inexpensive and readily available. PEG is environmentally safe, easily extractable in water and reduces feedstock viscosity. The PVB is water-insoluble and provides strength to the formed (green) parts and debound (brown or grey) parts so used as a backbone component. The SA acts as a surfactant and helps wet the powder particles. Thermal characteristics were determined with DTA/TGA and rheological analyses with a capillary rheometer. The four-stage process developed to manufacture homogeneous material suitable for titanium-MIM (Ti-MIM) involved dry-mixing irregular HDH Ti-6Al-4V powder and binder components in a planetary mixer for 30 min at room temperature, followed by 16 h further mixing at room temperature in a roller mixer rotating at 250 rpm, then melt-mixing at 125 ˚C for 45 min in a roller compounder. The final step involved extruding the processed feedstock in a twin-screw extruder operating with a feed-to-nozzle temperature profile of 125, 140, 135, 135, 140 ˚C. Viscosity, density and mixing torque data for the PEG-PVB-SA binders indicated that the maximum (so-called critical) powder loading was 65 vol. %. Feedstocks made with 75:20:5 vol. % PEG:PVB:SA and 80:15:5 vol. % PEG:PVB:SA binders and 60 vol. % powder loading, were optimal for Ti64-MIM. This powder loading is higher than reported for spherical, fine powders. The flow behaviour index of the feedstock could be used to identify optimal powder loading and temperatures for injection moulding. A feedstock with low viscosity (<80 Pa-s), low flow activation energy (<30 kJ/mol), high mouldability index (15 – 22), high melt flow rate (up to 860 g/10 min), and high stability was selected for Ti-MIM. Solvent debinding green parts made of 60 vol. % Ti64 powders and PEG-PVB-SA binders at 35 °C was affected by binder composition. Between 4.67 – 5.20 h, PEG removal changed from dissolution-controlled to diffusion-controlled. Increasing debinding temperature to 55 °C decreased the time at which the transition occurred. Debinding time affected PEG removal, porosity and reaction depth to a similar extent, indicating these parameters are inter-related. The empirical relationship describing solvent debinding kinetics of Ti-MIM parts showed PEG removal (weight loss %) was linearly related to the nth root of debinding time (n√t). This model can be used to predict optimum solvent debinding temperature and time. The maximum relative density, tensile strength and the oxygen content of as-sintered parts were 95%, 455 MPa and 1.58 wt. % respectively. Fracture surfaces exhibited transgranular fracture. The poor mechanical properties are mainly due to porosity and high oxygen content. New rheological models to predict feedstock flow properties, relative viscosity and critical powder loading were developed and validated. An improved model for predicting relative viscosity and critical powder loading from powder loading, surface area, shear rate, and flow behaviour index is proposed. The Einstein relationship for relative viscosity of dilute suspensions was modified and used for MIM feedstocks, which have a much higher powder loading. Recommendations for further work are given.
Type of thesis
Gnanavinthan, T. (2016). Producing near-net shape titanium alloy parts by metal injection moulding (Thesis, Doctor of Philosophy (PhD)). University of Waikato, Hamilton, New Zealand. Retrieved from http://hdl.handle.net/10289/10354
University of Waikato
All items in Research Commons are provided for private study and research purposes and are protected by copyright with all rights reserved unless otherwise indicated.