A combination of scanning electron microscopy, energy dispersive x-ray analysis, and a mathematical approach quantify the extent of separation of highly filled polymer melts.
Injection molding of highly filled (about 60vol.%) polymer melts brings several quality and efficiency issues. One critical factor is powder and polymer separation during molding. This is especially significant for powder injection molding (PIM), where—in contrast to classical composite products—the polymer component is removed (debinded) from the molded part, so that it has a porous structure. The molded powder is then sintered (heated to a temperature below the melting point to fuse the particles together by diffusion) to the final density. The origin, mechanism, and cause of phase separation are still not fully understood, and similarly its onset and extent are not satisfactorily quantified. Separation causes non-uniform distribution of powder in a polymer binder, and thus results in cracks, voids, and distortions on the final (sintered) parts.Early (i.e., prior to debinding and sintering steps) recognition could save material, energy, and time. PIM combines the advantages of polymer processing and metallurgical tools, and thus PIM products offer the material flexibility of powders and design flexibility of plastics. There is a substantial, constant stream of new applications in the medical, IT, electronics, automotive, aircraft, and other industrial sectors.The testing molds (zigzag, spiral, or square spiral) usually employed to study the separation phenomenon only partly fulfill the purpose.1 With colleagues at the Fraunhofer Institute for Manufacturing Technology and Advanced Materials, Bremen, Germany, we developed a new testing mold design to examine performance at inner and outer corners, radical thickness changes, weldlines, and very thin sections.2 The mold is composed of four elements connected by gates: see Figure 1. The first three elements contain external and internal corners. The inner wall of the first three elements measures 10mm, and the side of the square cross section decreases for successive elements from 3 to 2.5 to 2mm.Figure 1.Design of testing mold with highlighted areas A–C prone to powder/binder separation. Figure 2.Development of the flow front obtained at the positions A–C (left to right, respectively) of the testing mold for stainless steel feedstock. Axes are as indicated in Figure 1. We found that separation was most evident close to the entrances to each element of the testing mold (positions A–C in Figure 1). The flow into the mold starts before position A and continues through B to C. That is, the mold is filled first at the last area around position C, and last at position A. Figure 2 shows clearly how the phase separation develops with the flow front at the indicated positions, for highly filled commercial PIM feedstock based on stainless steel powder 316L. Bright points reveal the powder particles and dark areas account for separated polymer binder.To quantitatively assess the severity of the separation, we used scanning electron microscopy (SEM), combined with energy-dispersive x-ray (EDX) analysis of the distribution of elements typical for powder and binder.3 SEM/EDX is a semi-empirical method that allows determination of major and minor element concentration, despite lacking the sensitivity for trace analysis. We created a quantitative EDX map of the distribution of iron, which we selected as a proxy measure of the distribution of powder particles within the binder: see Figure 3. At positions A–C of the tested stainless steel feedstock sample, our iron distribution map ranges from 0% (black) to 100% (white). Clearly, there is evidence of separation, which is more pronounced at the flow front.Figure 3.Energy-dispersive x-ray quantification maps of iron distribution at positions A–C (left to right, respectively) of the testing mold for stainless steel feedstock. Colors show the distribution of iron: (black) 0% and (white) 100%. Axes are as indicated in Figure 1. Finally, we have proposed a method to quantify the tendency of feedstocks to separate with a single characteristic parameter we call ‘variability.’ The rate of phase separation represents the non-uniformity of powder and binder distribution, i.e., the non-uniformity of bright and dark points on EDX maps. The images were converted into matrices, where completely bright areas were assigned as 100% and purely dark pictures as 0%. We then smoothed the data to eliminate local inhomogeneities. This was performed by averaging the value of the element concentration, where each pixel was considered as an average of (5×5) neighboring pixels. Finally, we computed variability coefficients as standard deviations of the variance in element content in the same pixel neighborhood.Low variability—uniform distribution—reveals an efficient PIM process without defects arising from the separation. We tested the method on two commercially available feedstocks. In the tested separated areas, the separation developed differently in the two materials, suggesting that a ‘variability coefficient’ might indeed serve as a single parameter to quantify separation and consequently reduction/elimination of the issue and production of defect-free items. The first feedstock had higher variability (about 20%) but the separation was still similar in each square element, whereas for the second feedstock we obtained rather high variability, from 10% (area C) to 40% (area A), depending on the element tested: see Figure 3.Future research will focus on linking the factors responsible for the separation to a relevant rheological model described elsewhere.4 Preliminary investigations suggest exploring wall slip as a rheological parameter that indicates the onset of separation.AuthorsBerenika HausnerovaTomas Bata University in Zlíin (TBU)Berenika Hausnerova is a full professor in technology of macromolecular compounds at TBU. Her area of expertise is rheology and powder injection molding. Her work has been recognized with the Werner von Siemens Excellence Award (1999) and For Women in Science (2006).Daniel SanetrnikTomas Bata University in Zlíin (TBU)Petr PonizilTomas Bata University in Zlíin (TBU)ReferencesM. Jenni, L. Schimmer, R. Zauner, J. Stampfl and J. Morris, Quantitative study of powder binder separation of feedstocks, PIM Int’l 2 (4), pp. 50-55, 2008. L. Jiranek, B. Hausnerova and T. Hartwig, Testing sample, Community Design 001704974, 2010. Submitted to the Office for Harmonization in the Internal Market, Alicante, on 6 May 2010.B. Hausnerova, D. Sanetrnik and P. Ponizil, Surface structure analysis of injection molded highly filled polymer melts, Polym. Compos., 2013. First published online: 13 July. doi:10.1002/pc.22572B. Hausnerova, L. Marcanikova, P. Filip and P. Saha, Optimization of powder injection molding of feedstock based on aluminum oxide and multicomponent water-soluble polymer binder, Polym. Eng. Sci. 51 (7), pp. 1376-1382, 2011. DOI: 10.2417/spepro.005097