Avoidance of surface defects through uniform particle distribution in MIM feedstocks

 (c) SIGMA Engineering GmbH

Figure 1 - Simple round part with distinct defects due to different particle concentrations, left side shows the nozzle side surface, right side the ejector side surface

Figure 1 - Simple round part with distinct defects due to different particle concentrations, left side shows the nozzle side surface, right side the ejector side surface (c) SIGMA Engineering GmbH
Figure 2 - Comparison of the visible defects (left) with the simulated local particle concentration (right) (c) SIGMA Engineering GmbH
Figure 3 - Main influences on the differences in local particle concentration: influence of the filling time (top left), influence of the gate position (top right) and influence of the runner length (bottom left) (c) SIGMA Engineering GmbH

During sampling of a supposedly simple MIM component, clear flow marks were visible on the surface of the sintered part. The part is a simple round component with a constant wall thickness made of a titanium alloy. Figure 1 shows the component and its defects on the surface on both the nozzle and ejector side.

The project aimed to explain the defect and to develop suitable improvements. The first step was to simulate the component taking the process settings used in production into account. In the simulation it was apparent that the surface defects were due to locally varying particle concentrations. In the area of high shear gradients, binder and particles of the normally homogeneous titanium feedstock separate. This segregation leads to different particle concentrations, which are then transported further through the component. At the surface of the component, differences in particle concentration lead to different shrinkage behavior at the latest after sintering. The surface is then no longer homogeneous, reflects the light slightly differently in all locations and the overall picture appears noisy. A lot of times, the progress of the flow front can be seen on the surface as in the current example (Fig. 1). In the simulation, the distribution of the particle concentration could be well reproduced (Fig. 2).

With the help of the simulation, not only the defect pattern was simulated, but it was also proven that the mold filling of the component and the segregation occurring in the gate lead to this defect pattern. Although initial segregation also occurs in the runner, it is still sufficiently homogenized in the runner system.

Once the component has been simulated and the problem understood, meaningful solutions can be tested with the help of a virtual DoE. In contrast to DoEs at the injection molding machine, the simulation helps to include changes inside the mold into the calculation. For this example, different process parameters of the filling phase (volume flows), as well as different gate geometries (position of the gate on the component and manifold length) were varied.

The influence of the process parameters on the segregation was evaluated based on the difference in particle concentration. A smaller value on the y-axis is equivalent to less separation and a smaller density difference at the surface of the component. The x-axis lists the various simulated parameters (Fig. 3). During the evaluation it becomes evident that the selected volume flow and thus the filling time of the component has a significant influence on the segregation. Longer filling times in this component lead to lower shear gradients in the gate and thus to reduced segregation. This in turn improves the surface quality of the finished component. The positioning of the gate also has an influence on segregation. However, there is no continuous correlation, but rather a minimum that characterizes an ideal position. A slightly off-center positioning of the connection results in the lowest segregation and thus the best surface quality. The length of the manifold has hardly any influence on the density and can therefore be selected freely.

The simulation was used to determine the best design for the gating of the component as well as good process parameters for the production. Thus, the surface quality of the component could be significantly improved.