Weld lines in plastic components are usually mechanical weak points; they have reduced strength compared to the base material. There are many scientific studies on weld line strengths. However, there is a lack of structured processing in such a way that the practitioner can at least make a plausible estimate of the weld line strength with the material data usually available. This article addresses this problem and presents a procedure that makes this possible. First, the origin of weld lines is discussed. Then different types of weld lines are characterized. Furthermore, the plastic classes are grouped according to their sensitivity to weld lines. Finally, a practical procedure for estimating weld line strength is presented on the basis of a literature review and a statistical analysis derived from this.
How do weld lines develop?
Weld lines are formed when separate melt flows merge. This can be caused by flow obstacles in the flow channel such as holes (Fig. 1, top right, bottom left) or by several injection gates (Fig. 1, top left, bottom right). Different wall thicknesses can also lead to weld lines, as smaller wall thicknesses lead to a flow delay compared to areas of greater wall thickness.
The resulting weld lines can be characterized on the one hand as actual weld lines, which are also referred to as stagnant welds, and on the other hand as so-called meld lines (or also flow lines). This indicates that in the case of weld lines, no further movement takes place after the melt fronts merge, whereas in the case of meld lines, the weld is transported through the component with the flow. The formation of the two weld forms can be quantified via the flow angle. Figure 2, left, shows a weld line, here the flow angle is 0°, the melt fronts meet butted. Figure 2, right, shows the formation of a meld line. While a stagnant weld line can be assumed directly behind the flow obstacle (small flow angle), the flow angle increases with increasing distance from the flow obstacle. Up to a flow angle of 135°, it is still referred to as a weld line, whereas larger angles are referred to as meld lines. The greater the flow angle, the lower the strength-reducing influence of the weld line. At angles in the order of 140°-150° (weld line vanishing angle [2]), there is no longer any noticeable influence of the seam on the strength. This therefore represents an initial orientation point, as injection molding simulation programs usually provide the flow angle, weld positions with angles ≥ 135° can be neglected as a first approximation. This only relates to the influence on strength, although weld lines with larger angles can lead to optical defects. The stagnant weld line with a flow angle of 0° is the most critical weld in terms of loss of strength.
Factors influencing the weld line strength
In order to be able to make a meaningful estimate of the weld line strength, it is helpful to understand the physical effects that take place on a micromechanical and molecular scale during the formation of the weld line. Based on this, factors influencing the weld line strength can then be defined. Initially, only unfilled and unreinforced materials are discussed here; fiber orientation is described below.
Essentially, two physical effects are decisive: molecular interdiffusion and molecular orientation (Fig. 3). Molecular diffusion refers to the mutual exchange of macromolecules across the contact area of the melt fronts. With ideal diffusion, the weld line would be completely healed and no strength-reducing influence would be noticeable. Diffusion is favored by high molecular mobility, which in turn is favored by lower molecular weights or chain lengths of the macromolecules. Equally favorable are higher temperatures, which lead to greater micro-Brownian motion and thus free volumes between the molecules. A macroscopically accessible measure of molecular mobility is viscosity. The lower it is, the better interdiffusion can take place. A greater distance between the melt front temperature and the glass transition temperature Tg of the material also leads to longer periods of time for the healing process, which can only be considered complete when the temperature falls below Tg, as the molecular mobility is then almost frozen. As a result of the expansion flows in the melt front (see Fig. 7, top right), molecular orientations parallel to the weld plane are formed which thus have an unfavorable effect on the weld strength; the higher the viscosity, the more pronounced these orientations are. It is also argued that these molecular orientations are responsible for a V-shaped notch on the weld surface, as the orientation creates stresses that pull material into the weld. The notch has a depth in the order of a few micrometers up to 50 µm (see Figure 5). A quantifiable influence on the strength is not known. The orientations can lead to residual stresses in the weld line, which then also remain in the component. In the case of semi-crystalline plastics, crystallization processes play a role in addition to the influences described above, but these will not be discussed further here.
Various models for mathematically modeling the weld line strength are known from the literature (see review e.g. in [2]). These include molecular dynamic models, which attempt to model the molecular mobility of the macromolecules, and empirical models, which establish a relationship between parameters and weld line strength on the basis of experiments. From the correlations shown, it can be deduced that the following measurable parameters probably have an influence on the weld line strength: Viscosity, glass transition temperature, melt temperature.
Estimation of weld line strength
In the following, a procedure is derived that allows a practicable estimation of weld line strengths to be carried out using simple means. Only material properties that are usually available are used; process parameters from an injection molding simulation are not required.
Study design
The study was based on a literature review on the topic of weld line strength and molecular dynamic diffusion processes, in which more than 60 publications from the publication period from 1950 to 2023 were considered. The usability of the sources was limited by the fact that some of the test conditions were not reproducible, results were incomplete or no consistent material properties could be determined for the plastic grades used. In this respect, the study results are based on a subset of the sources, which can nevertheless be considered representative. A total of 42 data points for a wide variety of plastic grades were included in the quantitative analysis. We did not carry out any experimental studies of our own. Only data for stagnant weld lines are considered (see Figure 1), so a conservative estimate is made.
Sensitivity of plastic classes to weld lines
As a starting point for the analysis, it has proved useful to categorize the plastics according to material classes with regard to their weld line sensitivity as follows:
- semi-crystalline unreinforced plastics
- amorphous unreinforced plastics
- short-fiber-reinforced plastics (amorphous and semi-crystalline)
Materials filled with fillers such as glass beads, talcum, mica etc. are not discussed further here, as these must be considered separately.
As an example, Figure 4 shows weld line factors as a ratio of weld line strength to base material strength by material class. The box indicates the range from the first to the third quartile of the values, the horizontal line in the box indicates the median value. There are clearly different distributions of weld line factors by material class.
Semi-crystalline unreinforced plastics
It is noteworthy that the semi-crystalline unreinforced plastics (green box) consistently exhibit high weld line factors (median 0.96) and in some cases even achieve the strength of the base material with comparatively low scattering. Figure 5 shows an example of a micrograph of the weld line of an unreinforced PA66. A weld line area is not morphologically recognizable, apart from the structural V-notch. The crystallites have grown homogeneously over the former contact area of the melt fronts. The weld line has healed completely.
Amorphous unreinforced plastics
The weld line factors of the unreinforced amorphous plastics are significantly lower on average (median 0.55) than for the semi-crystalline unreinforced plastics. However, the scatter is higher, which is largely due to the outlier for polycarbonate (PC). Figure 6 provides a physically based explanation for this. For non-reinforced amorphous and semi-crystalline plastics, the weld line factor is shown on the left versus the temperature difference between the injection molding processing temperature (Tmelt) and the glass transition temperature (Tg) and on the right versus the viscosity. It is obvious here that both the temperature difference and the viscosity clearly discriminate between the two material classes, in the opposite direction in each case. The semi-crystalline plastics consistently exhibit a higher temperature difference and lower viscosity (except PC!) than the amorphous plastics. The very low viscosity of polycarbonate compared to all other amorphous plastics and even to most semi-crystalline plastics is the reason for the high weld line strength. In line with the physical effects described above, viscosity is a measure of molecular mobility. Low viscosities favor interdiffusion and thus higher bond strengths. Similarly, high temperature differences favor interdiffusion. Until the glass transition temperature is reached, below which there is almost no molecular mobility, more time is available for interdiffusion. The temperature difference is generally higher for semi-crystalline plastics than for amorphous plastics (Fig. 6, left).
Short-fiber-reinforced plastics (amorphous and semi-crystalline)
According to Figure 4, the short-fiber-reinforced plastics show a wide spread of weld line factors in the range of approx. 0.15 to 0.9 (orange box). These plastics require more detailed consideration. Figure 7 shows the flow processes at the melt front at the top right and the resulting fiber orientations at the bottom right. As a result of the biaxial elongation flows, the fibers align themselves parallel to the weld line plane in a similar way to the molecular orientations discussed above (see Fig. 3). This is very unfavorable for the strength of the seam - there is no reinforcement effect by the fibers across the weld plane.
Figure 7, left, shows a micrograph of the fracture surface of the weld line of a short-fiber-reinforced plastic in a perpendicular direction. It is easy to see that the majority of the fibers are oriented parallel to the plane of the weld line. A molecular healing process can also only take place in the matrix material. In terms of surface area, the matrix cross-sectional area in the weld line is reduced by the proportion occupied by the fibers. It can be deduced from this that fiber or matrix volume fraction and matrix strength play a role as decisive influencing variables in fiber-reinforced plastics. This also explains the large scatter of the weld line factors in Figure 4, as plastics with very different fiber contents were examined.
Figure 8 shows the weld line factor for short-fiber-reinforced plastics as a function of the fiber volume fraction (not fiber weight fraction!). In a semi-logarithmic plot, the relationship can be approximated well in a first approximation using a linear equation (blue in the diagram). The fiber content on its own therefore explains the development of the weld line strength well.
The accuracy of the model can now be increased by including further factors; the decisive physical effects provide indications of this. A multiple linear statistical regression model is determined in which only factors that are statistically significant with a statistical probability of error of α = 5% are included. The significance of the model was tested using a global F-test of the regression model. The coefficient of determination R2 was determined to assess the goodness of fit of the model, i.e. how much variance in the data is explained by the model.
The following factors are included in the model:
- f (fiber volume fraction)
- Matrix viscosity
- Matrix-strength/composite-strength ratio
Since there is a non-linear relationship between the weld line factor and the fiber volume fraction, as shown above, the fiber volume fraction is first subjected to a non-linear transformation using a suitable function. The results provided by the model and the model equation can be seen in Figure 9.
Predicted over measured weld line factors are plotted; points on the 45° straight line correspond to an exact match. An adjusted coefficient of determination R2 = 91.1 % is achieved, the individual factors as well as the overall model are statistically highly significant, although only a comparatively small sample of 25 data points was used. This is an astonishingly good result, especially because only basic material properties and no process data were taken into account and the model covers a broad range of matrix base polymers. It should be mentioned at this point that the model is valid for stagnant weld lines (mostly conservative estimation); process parameters may have a greater influence on meld line (flow lines).
The transformation function and the numerical values of the model coefficients are not disclosed here. The model is implemented in this or a further optimized version in MatScape and S-Life Plastics.
Takeaways
Plausible estimates of the weld line strengths for the various plastic classes can already be made with on-board tools, without a more precise multiple regression model. The following rules of thumb can be used:
- weld lines with flow angles ≥ 135° can be neglected (all plastic classes)
- a general weld line factor of 0.85 can be used for semi-crystalline unreinforced plastics (see Figure 4)
- general weld line factor of 0.45 can be used for amorphous unreinforced plastics (except easy-flowing plastics) (see Figure 4)
- amorphous unreinforced easy-flowing plastics with viscosities comparable to polycarbonate can be treated like semi-crystalline unreinforced plastics (see Figure 4)
- for short-fiber-reinforced plastics (amorphous and semi-crystalline), the regression equation in Figure 8 or an individual function f(fiber volume fraction) can be used
The general factors and equations mentioned apply to short-term static stresses. Other values may apply for long-term and cyclic stresses. The correlations are also based on the assumption of “normal” injection molding process conditions; in highly oriented thin-walled component areas (e.g. film hinges), other correlations may apply.
Conclusion
Weld lines regularly pose a problem with regard to strength assessment. At least some of the uncertainty can be eliminated with the simple rules of thumb mentioned above, and the estimates can be regarded as conservative. The question of which failure hypothesis should be applied in weld lines (ductile, brittle) and the alignment of the weld line to the main stress direction was not addressed here. The position and orientation of the weld line in the component must be known. These aspects will be addressed with the help of the weld line mapping functionality in Converse, the strength assessment with S-Life Plastics and the integrated material modeling module MatScape in future versions of the software.
Converse and S-Life Plastics can be obtained either directly from PART Engineering or via the Altair Partner Alliance.
[1] C. Hopmann, „Integrative Festigkeitsanalyse kurzfaserverstärkter Thermoplastbauteile unter Berücksichtigung des Einflusses von Bindenähten“, Institut für Kunststoffverarbeitung an der RWTH Aachen, Aachen, IGF 16311N, Juni 2012.
[2] S. Fellahi, A. Meddad, B. Fisa, und B. D. Favis, „Weld lines in injection-molded parts: A review“, Advances in Polymer Technology, Bd. 14, Nr. 3, S. 169–195, 1995, doi: doi.org/10.1002/adv.1995.060140302.
[3] G. Jadhav u. a., „Weld-lines and its strength evaluation in injection molded parts: A review“, Polymer Engineering & Science, Bd. 63, Nr. 11, S. 3523–3536, 2023, doi: doi.org/10.1002/pen.26470.
[4] I. S. Dairanieh, A. Haufe, H. J. Wolf, und G. Mennig, „Computer simulation of weld lines in injection molded poly(methyl methacrylate)“, Polymer Engineering & Science, Bd. 36, Nr. 15, S. 2050–2057, 1996, doi: doi.org/10.1002/pen.10600.
[5] I. Kuehnert, Y. Spoerer, und M. Zimmermann, „Weld lines in injection molded parts: strength, morphology and improvement“, in Proceedings of SPE Antec, Indianapolis, Indiana, USA, 2016.
[6] J. Kim, J. Song, S. Chung, und T. Kwon, „Morphology and mechanical properties of injection molded articles with weld-lines“, Polymer Engineering & Science, Bd. 37, S. 228–241, Jan. 1997, doi: doi.org/10.1002/pen.11665.
Author: Dr. Wolfgang Korte is Managing Director at PART Engineering GmbH, Bergisch Gladbach, Germany
