It is a well-known fact that weldlines are unavoidable in most injection-molded products of even moderate complexity. While there are many situations where they are barely perceptible, weldlines represent a potential source of weakness in molded parts. In injection molding weldlines are generated when two separate melt streams join either in multigated molds or as a consequence of flow around obstacles. The development of many interesting materials has been hampered by poor weldline strength. Among such materials are plastics reinforced with fibers or platelets, liquid crystal polymers, and a number of multiphase polymer blends. Weldlines have ever been called the "Achilles' heel" of these multiphase materials. This article is a review of the literature published on weldlines in injected parts. It deals primarily with the aspects related to the mechanical behavior of weldline-containing parts. It begins with a brief description of the phenomena important for the part formation in the mold, including those leading to weldlines, in addition to the techniques used to characterize weldline-containing parts. The following three sections consider the structure and properties of weldlines in neat amorphous and semicrystalline polymers, filled and reinforced plastics, and finally in polymer blends and alloys. In the last section methods developed for increasing the weldline strength are discussed.

Presence of weldlines introduces an element of uncertainty to the performance of injection molded parts. Weldlines are particularly problematic in reinforced plastics because, unlike molecular orientation in neat polymers, the flow induced fiber orientation does not relax. This paper deals with the structure and mechanical behavior of weldlines in glass fiber reinforced nylon 66, a plastic known for excellent fiber‐matrix adhesion. Two molds were used to generate weldlines: a double gated tensile sample shaped cavity in which the weldline is formed by a head‐on collision of melt fronts flowing in opposite directions and a film gated rectangular plaque with a circular insert in which the weldline formation behind the insert is followed by additional flow. In both cases the weldline zone is several millimetres wide: in the plane where the melts fronts have met fibers are oriented parallel to this plane (random‐in‐plane in the double‐gated cavity and unidirectional in the cavity with insert). The transition zone between the weldline plane and the rest of the sample is characterized by an increased presence of microvoids. Weldline tensile depends little on the fiber concentration and on the sample shape or thickness: values close to the matrix strength are found: in samples without weldlines strength increases with the fiber content. However, in instrumented impact penetration test during which the material is subjected to multiaxial loading, the weldline effect appears negligible.

This article deals with the stress-strain behavior of two viscoelastic polymers, polypropylene and polyamide 6, filled with rigid particles in the range of axial strain of 0 to 8%. These materials, when subjected to a constant strain rate test lose stiffness via two mechanisms: filler-matrix debonding and the viscoelastic softening of the matrix. A model that combines the concepts of damage mechanics and the time dependence of the interfacial strength is described and compared to the experimental results of polypropylene and polyamide 6 filled with up to 50 vol % of untreated and silane-treated glass beads. The matrix behavior is described in terms of an empirical equation selected to fit the stress-strain behavior of neat polymers in the range of strain rates between 0.12 and 0.5% s 01 and strains between 0 and 8%. The stiffness of the damaged, partially debonded composite is calculated using the Kerner-Lewis equation assuming that debonded particles do not bear any load. The model is able to generate stress-strain curves that are in good agreement with the experimental data. The void volume attributable to debonding calculated using the model is much smaller than the experimental total determined void volume (which is a sum of several deformation mechanisms).

ABSTRACT:The mechanical behavior of multiphase materials is closely related to the interfacial adhesion between their various components. There is considerable interest in the development of simple experimental techniques for characterization of interfacial debonding during mechanical loading. Probably the best known method is tensile dilatometry, in which the onset and progression of debonding are related to the volume of microvoids generated in the material as it undergoes mechanical loading. Several authors have suggested that equivalent information can also be extracted from stressstrain data generated during a simple constant strain rate test. In practice, however, the transition between the initially well-bonded and the debonded state is obscured by the strain-induced softening of the matrix, which is usually observed in the same strain range as the debonding. In this work the filler/matrix debonding in polypropylene and polyamide 6 filled with up to 50 vol % of glass beads is examined using both tensile dilatometry and an analysis of tensile stress-strain curves. It was found that, depending on the level of adhesion, either a complete or partial debonding occurs in the strain range studied (0-8%). It appears that the volume change due to debonding is a small part of the total volume strain recorded. Therefore, the accuracy of the tensile dilatometry is not sufficient to detect the onset of debonding. The loss of stiffness of the composite, particularly when compared to the loss of stiffness of the matrix offers a more promising way to follow the debonding process.

A 58% (by weight) long glass fiber reinforced (LGF)‐HDPE master batch was blended with a typical blow molding HDPE grade. HDPE composites having between 5% and 20% (by weight) long fiber content were extruded at different processing conditions (extrusion speed, die gap, hang time). The parison swell (diameter and thickness) decreased with increasing fiber content. Although the HDPE exhibited significant shear rate dependence, the LGF/HDPE composites were shear rate insensitive. Both the diameter and weight swell results also indicated very different sagging behavior. The LGF/HDPE parisons did sag as a solid‐body (equal speed at different axial locations) governed by the orientation caused by the flow in the die. Samples taken from blown bottles showed that fiber lengths decreased to 1‐3 mm, from the original 11 mm fiber length fed to the extruder. No significant difference in fiber length distribution was found when samples for different regions of the bottle were analyzed. SEM micrographs corroborate the absence of fiber segregation and clustering or the occurrence of fiber bundles (homogeneous spatial fiber distribution) as well as a preferential fiber orientation with the direction of flow. The blowing step did not change the orientation of the fibers. Five‐percent (5%) and 10% LGF/HDPE composites could be blown with very slight variations to the neat HDPE inflation conditions. However, 20% LGF/HDPE composites could not be consistently inflated. Problems related to blowouts and incomplete weldlines were the major source of problems.