Technical Paper: PIM International, Vol.2 No. 4 December 2008, pages 50-55, 2653 words

Authors: Martin Jenni [1], Lukas Schimmer [1], Rudolf Zauner [2], Juergen Stampfl [3] and Jeff Morris [4]

[1] Austrian Research Centers GmbH – ARC, A-2444 Seibersdorf, Austria
[2] Verbund – Austrian Renewable Power GmbH, A-1010 Wien, Austria
[3] Vienna University of Technology, Institute of Materials Science and Technology, A-1040 Vienna, Austria
[4] City College of New York, Levich Institute & Chemical Engineering, NY 10031 New York, USA

Abstract

The rheological properties of feedstocks and their flowability and tendency for powder binder separation have a strong influence on the successful manufacturing of PIM components. A systematic comparison of the filling behaviour of different tungsten and aluminium feedstocks has been carried out experimentally. Using specially designed moulds, designed experiments (DOE) with variations in the nozzle temperature, mould temperature an injection speed of feedstocks with extremely different physical and thermal properties (tungsten and aluminium) were carried out. The results were statistically analysed. The observed effects are explained and implications for optimised PIM processing are suggested. Furthermore parts of the samples are analysed for powder separation and compared to predictions made by the pseudo-continuum separations model named: Balance Model.

Introduction

Powder injection moulding (PIM) is a relatively new processing technology used in powder metallurgy and ceramic processing industries. This process is especially cost-effective and beneficial for manufacturing small and complex components in large quantities. Powder injection moulding is used in an increasing range of different fields, including automotive, medical and telecommunication industries. It includes four basic steps consisting of (1) mixing the powders and binders, (2) injection moulding, (3) debinding and finally (4) sintering [1]. Both injection moulding and sintering are the most important steps related to forming the green part and the final part, respectively. In particular, the injection step often requires expensive and time consuming trial and error methods to resolve design problems associated with raw material, product dimensions, tooling factors and process issues during manufacturing. A direct relationship between input and output parameters is often not obvious and a lot of testing is needed to find empirical relationships for the influence of feedstock properties, processing conditions and mould properties on the mould filling behaviour. Today’s computer aided engineering tools for plastic injection moulding and PIM have shown promising results in resolving problems of material, part and mould design.

Similar to injection moulding of thermoplastic feedstocks, defects such as jetting, air traps, dead zones, or welding lines can also occur in PIM. The available simulation software addresses these points already. However, the powder binder separation, also called phase segregation, is a phenomenon which is frequently observed in PIM and happens during the high speed and high pressure injection moulding process due to the different densities associated to powder and binders. These can induce inhomogeneities of green parts. After the debinding step, the binder is removed and the remaining component results in a porous brown part. In the consecutive sintering step, the debinded parts shrink substantially. The shrinkage between the green component and resulting net component is typically in the range of 10 to 20% and the final density in the range of 95 to 100% [2]. .....

Further sections of this article include:

- Experimental Investigation
- Two dimensional measurements
- Simulation models of suspension flows
- Balance model
- Conclusion
- References

Figures and Tables:

Fig. 1 Viscosity measured with capillary rheometry for different loadings. The material was tungsten with ARC binder
Fig. 2 Characterisation moulds; three different forms: spiral (white), square spiral (pink) and zigzag (blue)
Fig. 3 Position of the three samples for powder content measurements
Fig. 4 Influence of the nozzle, mould temperature and injection speed on the powder content; mean and median powder content over the experiment series
Fig. 5 Powder distribution along a corner; material: tungsten feedstock with 60% loading
Fig. 6 Measurement with computer tomography (top) and presentation of the greyscale intensity (bottom) along the marked line 60% loading (intensity in the middle)
Fig. 7 Measurement with radiography (greyscale)
Fig. 8 Typical geometry with computational grid
Fig. 9 A cell with nodes at the center and wall-mid points
Fig. 10 “Solver-Evolver” method
Fig. 11 Powder concentration in a curved channel with tungsten feedstock (60% loading)
Fig. 12 Corner with tungsten feedstock (60% loading); Top: “balancemodel”; powder content is shown here as a fraction of maximum flowable solid content, assumed as 68%; 0.6/0.68 = 0.88 Bottom: measured by DSC
Fig. 13 Powder distribution along a corner; material: tungsten feedstock with 60% loading. Top left: Balance model; Top right: measured by DSC; Bottom left: measured by computer tomography

Table 1 Data sheet for DOE measurements; material: ARC tungsten feedstock with 60% loading; mould: square spiral