Effect of processing on weld line strength in five thermoplastics.
Weld lines are formed during melt processing of polymers when two separated melt streams recombine. In injection molding, weld lines will form during mold filling because of (1-7): 1) multiple gating, 2) flow-obstacles in the mold, 3) variable part thickness, and 4) jetting.
Weld lines can be seen on the finished part as thin hair-like cracks on the surface. Important for the design is that the local mechanical strength can be much lower in the weld line than in the rest of the part.
Two main types of weld lines are usually distinguished. Cold or "butt" weld lines are formed when two melt streams meet head to head at approx. 180 [degrees] to each other. No additional flow occurs after the recombination. This type gives the lowest strength. Hot or "streaming" weld lines form when two more or less parallel melt streams recombine and there is an additional flow after the recombination.
The low mechanical strength in weld lines is usually explained by (8-10): 1) unfavorable molecular orientation, 2) insufficient bonding (due to insufficient entanglements), and 3) formation of a V-notch. For unfilled materials, the lowest weld line strengths are obtained for brittle, amorphous polymers. Typical values are 50% for cold weld lines and 75% for hot, compared with the values of the bulk (2, 7, 11). Higher values are usually obtained for ductile amorphous or crystalline polymers. The low weld line strength in certain filled or reinforced materials is well known. A high aspect ratio of the filler reduces strength more than a low aspect ratio.
Computer-aided mold filling simulations can nowadays be used to predict where weld lines will form in an injection molded part. However, it is still difficult to predict the strength of the weld lines. Factors like melt temperature, injection velocity, holding pressure, holding time, and mold temperature will affect weld line strength to various degrees for different materials.
In this study, weld line strength in five injection molded engineering thermoplastics has been measured. The effect of injection molding parameters was studied by using an experimental design. Four parameters were varied in two levels. For each parameter setting, the weld line strength was measured and related to the bulk strength via a weld line factor.
Five engineering polymers often used in load bearing applications were chosen:
1. PA 6, with 35% glass fiber (GF) (Ultramid A3WG7), from BASF.
2. PPS, with 40% GF, extra tough (Ryton R-4XT), from General Electric.
3. PP with 40% talc (Hostacom M4 U02), from Hoechst.
4. PPO (Noryl 110) from General Electric.
5. ABS, temperature resistant (Novodur GMT), from Bayer.
A Demag D150-42 microprocessor-controlled injection molding machine was used. The molding tool, of dimensions 250 x 140 x 2 mm, was gated in such a way that both cold and hot weld lines could be formed simultaneously. This was accomplished by having two gates at one short end of the tool and one gate at the other end; see Fig. 1.
The effect of injection molding parameters on weld line strength is usually studied by changing one parameter at a time. This could lead to wrong conclusions, since important comparisons, e.g., on interaction effects, cannot be made. Another way of studying the effect of molding parameters, which has been used here, is to use experimental design (12). Each parameter is varied in two levels. The number of parameters decides the number of tests. In this work only four parameters were studied to reduce experimental effort. The following were chosen: injection velocity, holding pressure, mold temperature, and melt temperature. This means [2.sup.4] = 16 various machine settings for each material. For each setting, weld line and bulk strengths were measured and the weldline factor calculated.
For mold and melt temperature, the two levels were chosen according to the maximum and minimum temperatures recommended by the material suppliers. The two levels for injection velocity and holding pressure were chosen by trial and error. Several settings were tried for each material, and those that consistently gave a product that was visually acceptable, i.e., without too much flash or over-large sinkmarks, were finally chosen. The various parameter settings are shown in Tables 1 through 5, together with the sequence of testing.
Flexural test specimens, 60 x 12 x 2 mm, were cut from the plates at four different positions from the gate, numbered from 1 to 4 as shown in Fig. 2a. Position 4 is across the weld line. Flexural testing was performed in accordance with ASTM-D790, with a support span of 40 mm and a crosshead rate of 20 mm/min. To study the weld line strength it is necessary that the central load be applied over the welded zone. This adjustment was made visually. Three specimens were tested at each position, and mean values and standard deviations were calculated.
In order to study the strength of hot weld lines, specimens 60 x 12 x 2 mm were cut perpendicular to flow at eight positions, numbered from 1 to 8, with their centers over the weld line, as shown in Fig. 2b. Also, unwelded specimens were cut perpendicular to flow at eight positions, as shown in Fig. 2b.
Weld line properties are often presented in the form of a weld line factor (WL-factor, [A.sub.wl]), here defined as:
WL-factor = Strength with weld line/Strength without weld line (1)
Flexural strength without weld line (bulk strength) was measured at position 3 [ILLUSTRATION FOR FIGURE 2A OMITTED].
For hot weld lines [ILLUSTRATION FOR FIGURE 2B OMITTED], strength values at each position were measured both over the weld line and perpendicular to flow in unwelded specimens and WL-factors calculated according to Eq 1.
For PPO the flexural testing was supplemented with an experimental design using tensile specimens. This was accomplished with a dog-bone shaped mold, as in Fig. 1. Provided with one gate on each side, a cold weld line could be produced almost at the center of the specimens. By shutting one of the gates, specimens without weld lines could be produced. We used a fractional factorial, containing eight parameter settings. With a fractional factorial, only main effects are studied and not interaction effects (12). The parameter settings, shown in Table 6, are not exactly identical to those used for the plate. The tensile specimens were tested in an Instron tester at a crosshead rate of 5 mm/min. Five specimens were used to determine tensile strength, and mean values and standard deviations were calculated.
The flexural tests on cold weld lines were in some cases supplemented by impact testing. Square plates, 45 x 45 x 2 mm, were cut from the injection molded plates in four positions, numbered 1 to 4, as in Fig. 2c, where position 4 is over the weld line.
[TABULAR DATA FOR TABLE 1 OMITTED]
Impact testing was performed in an instrumented falling weight impact tester, using a hemispherical head with a diameter of 10 mm and a drop height of 0.5 m. The square plates were laid flat, without clamping, on a circular support with a central hole of 20 mm diameter. The force that the plates exert on the head during impact was measured with a piezo-electric crystal. Impact strength was calculated by integrating the force vs. time trace. Absorbed energy after full penetration (i.e., when load returned to zero) was taken as impact strength. Weld line factors were calculated as above, by comparing strength values from position 3 and 4. In impact tests on weld lines, it is necessary that the weight penetrate the specimen at the weld line. This adjustment was made visually.
Some flexural specimens with weld lines were also examined in a scanning electron microscope, type Jeol JSM-T100. Both untested and fractured specimens were cut into smaller pieces, and gold was sputtered onto the surfaces, followed by inspection in the SEM.
RESULTS AND DISCUSSION
Tables 1 through 5 summarize results from flexural strength measurements along the flow direction and at the cold weld line for the five materials, using 16 different parameter settings. At the bottom of each Table are also shown mean, maximum, and minimum values for each position. When moving from position 1 to 3, i.e., from the gate to the center, both increases and decreases in flexural strength can be observed, depending on the material. At the weld line there is a distinct drop in strength, especially marked for PPS(GF), and PA(GF)/PP(Talc). For PPS(GF) the weld line strength is on average only 0.33 = 33% of the bulk strength, varying from 0.25 to 0.39. For PA(GF) and PP(Talc) the average weld line factor is 0.67, varying from 0.58/0.60 to 0.77. Low weld line factors in certain reinforced and filled thermoplastics have also been reported (13-19). For unfilled PPO and ABS, higher values are obtained, between 0.84 and 0.98.
The variation in WL-factors resulting from different parameter settings is generally not very large, and can be summarized as shown below:
Variation in WL-factor Mean value PA(GF) 0.60-0.77 0.67 PPS(GF) 0.25-0.39 0.33 PP(Talc) 0.58-0.77 0.67 PPO 0.84-0.98 0.90 ABS 0.85-0.96 0.92
Limited variations in WL-factors have been reported elsewhere. Cloud et al. (14) tested tensile specimens of PA 6.6 with 30% GF. When melt temperature increased from 250 [degrees] C to 315 [degrees] C, [A.sub.wl] decreased from 0.63 [TABULAR DATA FOR TABLE 2 OMITTED] to 0.59. Decreasing injection time from 2.0 to 0.5 s increased [A.sub.wl] from 0.59 to 0.61. Increasing mold temperature decreased [A.sub.wl] from 0.61 to 0.59. Holding pressure had little effect. More examples can be found in the literature (14, 20-23). Data for WL-factors shown above are close to what has been published elsewhere for PA(GF) (14, 18), PPS(GF) (14), PP(Talc) (19), PPO (24), and ABS (7).
A simple way to analyze data from experimental designs is to plot Pareto- or effect diagrams. Here, main effects and effects of two or three (or four) parameter interactions are displayed in staple diagrams. The height of the staples corresponds to the magnitude of the effects, which are sorted in descending order. Positive effects (increases) are marked with a (+) sign and negative effects (decreases) marked with a (-) sign. Effect diagrams for weld line factors, weld line strength, and bulk strength for the five materials are shown in Figs. 3 through 7. Statistically not significant effects are marked with a parenthesis ().
For PA(GF) in Fig. 3a it can be seen that the largest effect on the WL-factor is obtained from injection velocity and holding pressure. Increased injection velocity decreases [A.sub.wl] by almost 6%, while increased holding pressure increases [A.sub.wl] by almost 4%. The negative effect of injection velocity is caused by a decrease in WL-strength [ILLUSTRATION FOR FIGURE 3B OMITTED] in combination with an increase in bulk strength [ILLUSTRATION FOR FIGURE 3C OMITTED]. Other effects, e.g., the effect of melt temperature, are smaller and not significant.
For PPS(GF) in Fig. 4a a negative effect on the WL-factor is obtained when mold temperature and injection velocity are increased. There is a positive effect of increased holding pressure. The negative effect of mold temperature was not expected, but was obtained for the other materials as well, and has been reported elsewhere (7, 14). Since the formation of a weld line is a complicated process (20) and factors like orientation and residual stresses might affect properties in the weld line as well as in the bulk, unexpected results can sometimes be obtained.
For PP(Talc), increased holding pressure had a positive effect on the WL-factor [ILLUSTRATION FOR FIGURE 5A OMITTED] and weld line strength [ILLUSTRATION FOR FIGURE 5B OMITTED]. There were also interactions between holding pressure and injection velocity/melt temperature. High velocity and temperature decrease the effect of increased holding pressure.
For PPO in Fig. 6 there is a positive effect of increased injection velocity on the WL-factor. This is due to a negative effect on bulk strength [ILLUSTRATION FOR FIGURE 6C OMITTED] while the effect on WL-strength is very small [ILLUSTRATION FOR FIGURE 6B OMITTED]. Other effects are not significant.
For ABS there is a positive effect of injection velocity and melt temperature on the WL-factor, and a negative effect of mold temperature; see Fig. 7.
In order to determine the optimum parameter setting for obtaining a high weld line factor, the following simple calculation was made: For each material the 16 different settings were ranked with respect to WL-factor, [TABULAR DATA FOR TABLE 3 OMITTED] [TABULAR DATA FOR TABLE 4 OMITTED] [TABULAR DATA FOR TABLE 5 OMITTED] so that a low WL-factor was given a low rank and a high WL-factor a high rank; see Tables 1 through 5. [For PA(GF), for example, this means giving parameter setting No. 7 a rank number of 1, and parameter setting No. 10 a rank number of 16.]
The various rank numbers were then summed up and mean values for each setting were calculated. The following results were obtained: On average, parameter settings No. 12 and 10 were given the highest rank, i.e., the highest WL-factor. Both correspond to high holding pressure, high melt temperature, and low mold temperature. Similarity, parameter settings No. 5 and 7 were given the lowest rank. Both correspond to low holding pressure, low melt temperature, and high mold temperature. These results are not unexpected, with the exception of mold temperature, which in this investigation did not contribute to a high WL-factor.
Injection velocity also affects the WL-factors. However, this effect can either positive or negative, depending on the material.
The effect of melt temperature on [A.sub.wl] was less than expected for most materials. Since this temperature varies only in two levels, there is no information about possible variations between these factor levels. To investigate this, a complementary investigation on PPO was performed, in which the melt temperature was varied in three additional steps. Other parameters were kept constant. Results are shown in Fig. 8. An almost linear relation between the WL-factor and temperature is obtained. Again the effect of melt temperature on [A.sub.wl] is small.
Weld line factors were also measured for hot weld lines for certain parameter settings. Results where WL-factors are plotted vs. distance from the gate are shown for the five materials in Fig. 9. Results are taken from parameter settings that gave low WL-factors for the cold weld [since these settings tend to give low WL-factors for the hot welds as well (7)].
From Fig. 9 it can be seen that most results for [A.sub.wl] are located between 0.9 to 1.0. This strength reduction in hot weld lines is less than in cold, similar to what has been obtained elsewhere (10, 25-27). Only PP(Talc) gives significantly lower values; 0.67 is obtained close to the gate, approximating what was obtained for the cold weld (0.59). A low WL-factor for hot welds in talc-filled PP near the gate has also been reported by Hashemi et al. (19).
For PPS the weld line factor becomes larger than 1.0. This can partly be explained by strong anisotropic properties, resulting in low bulk strength perpendicular to the flow.
Anisotropy was also measured for the five materials for the same parameter settings as in Fig. 9, and the results are shown in Fig. 10. Here strength values measured across flow in positions 2, 4, and 6 are compared with strengths measured along flow in positions 1, 2, and 3. As expected, PP(Talc) displays almost isotropic behavior and the strength ratio becomes close to 1.0. PPS and PA are highly anisotropic, [TABULAR DATA FOR TABLE 6 OMITTED] while ABS and PPO display some anisotropy. For PA(GF) and PPS(GF), strength values along the flow direction are almost 2.5 to 3 times higher than across. Here, fiber orientation is superimposed on the molecular orientation, resulting in this strong anisotropy. For PA(GF), the across-flow strength is even lower than the weld line strength.
Table 6 shows results from tensile tests on PPO with and without weld lines. Without weld lines the average strength becomes [approximately equal to] 41 MPa, and with weld line [approximately equal to] 36 MPa. This corresponds to an average WL-factor of 0.88, varying from 0.83 to 0.90. An analysis showed the largest effect (negative) of melt temperature and mold temperature. The average value of 0.88 is very similar to what was obtained in flexure (0.90) despite differences in test geometry and strength properties in tension and flexure. Menges has shown that mold geometry has little effect on [A.sub.wl] (25). Assuming fracture in tension and flexure occurs when local tensile strength at some point is exceeded, the insensitivity to test geometry can also be explained.
A common way to make a quality control test of weld lines is to perform impact tests. Weld line strength is compared with the bulk strength. Results from impact tests at different positions from the gate are shown in Figs. 11 through 15 for certain parameter settings. For PA, PPS, and PP(Talc), Figs. 11 through 13, brittle failures are obtained in all cases, giving low impact strength. For PA(GF) the weld line strength in some cases becomes significantly higher than the bulk strength. Meddad and Fisa obtained similar results for 40% GF-reinforced PA 6.6 (18). A visual inspection of the impacted plates showed, similarly to Meddad and Fisa, a larger damaged zone in the welded specimens. Forming a larger damaged zone requires a higher energy absorption.
Table 7 summarizes values of weld line factors from impact tests calculated according to Eq 1, together with corresponding values from the flexural tests.
PA(GF), PPS, and PP(Talc) all show high WL-factors around 1.0 in impact testing, despite the critically low values obtained in some flexural tests.
Both PPO and ABS show unexpectedly low WL-factors in impact testing but high values in flexural tests. The correlation between results from impact and flexural tests is thus quite poor. However, impact tests can sometimes be used to determine optimum parameter settings. For PPO, e.g., the impact tests showed WL-factors of 0.22, 0.45, and 0.58 for parameter settings No. 1, 4, and 11, respectively. In the flexural tests the same ranking was obtained, and the results become 0.84, 0.91, and 0.98. Applying the same reasoning for PP(Talc) gives poorer results since the ranking is converted.
Scanning Electron Microscopy (SEM)
Photographs were taken of the weld regions of PPO and ABS. A typical result is shown in Fig. 16. A weld line can be seen on the surface as a thin V-shaped notch. The opening of the notch is only 1-2 [[micro]meter]. (For materials other than PPO and ABS, the V-notches were either absent or hard to distinguish from other surface defects like scratches.) Fracture surfaces from welded specimens, broken in flexure, are shown for PPO and ABS in Fig. 17. Close to the specimen surfaces can be seen smooth regions with little or no plastic deformation. These should be the V-notches where no molecular entanglements have been formed. The depth of the notches can be measured from these pictures. For PPO the depth is approximately 25 [[micro]meter] and for ABS 12 [[micro]meter]. The question arises if these V-notches are large enough to cause a true notch effect, i.e., a strength reduction due to stress concentration associated with the notches. Comparison can be made with the size of the inherent flaw (28, 29) present in the materials. The inherent flaw size can be estimated [TABULAR DATA FOR TABLE 7 OMITTED] from a fracture mechanics analysis (28), and its magnitude is then estimated from the following relation (28):
[a.sub.c] = 1/[Pi][([K.sub.c]/(Y[Sigma]y)).sup.2] (2)
where [a.sub.c] is the inherent flaw size, [K.sub.c] the fracture toughness, [Sigma]y the yield stress, and Y a geometrical factor. (For small surface cracks in tension or flexure, Y [approximately equal to] 1.) Notches smaller than [a.sub.c] will not give strength reduction (28, 29). Inserting typical values of [K.sub.c] from quasi-static tests [[approximately equal to] 2 MPa[square root of m] for ABS and [approximately equal to] 3 MPa[square root of m] for PPO/HIPS (30, 31)] and yield stress (45/80 MPa, depending on loading mode, tension or flexure, Refs. 32, 33, for both materials) gives [a.sub.c] [approximately equal to] 200-1500 [[micro]meter]. The notches in Fig. 17 are significantly smaller than these values and should therefore not cause strength reduction during a slow-rate tensile or flexural test.
The following conclusions apply to the results presented above:
The lowest weld line factors, in the range 0.25-0.77, were obtained for cold weld lines in glass fiber-reinforced and talc-filled materials. Higher values, 0.840.98, were obtained for unfilled PPO and ABS. Changing injection molding parameters did not produce any large variations in weld line factors. From these results it can be argued that weld line factors might be given in materials data banks in a way similar to tensile or flexural strengths, etc., as a mean value together with some estimate of scatter.
By summing results from five materials, it can be concluded that high holding pressure, high melt temperature, and low mold temperature produced high WL-factors. The effect of injection velocity on WL-factors could be either positive or negative depending on material.
For PPO there was a linear relation between WL-factor and melt temperature. Equal WL-factors in flexure and tension were also obtained for PPO, despite differences in mold geometry, test geometry, and strength values.
Hot weld lines in all cases (except PP/Talc) produced high WL-factors.
Impact testing of cold weld lines sometimes showed WL-factors higher than 1.0. The high weld line strength could be explained from a larger fractured area in specimens with weld lines.
There is a low correlation between WL-factors obtained from flexural tests and similar data from impact tests. Using impact tests as quality control methods for weld lines thus might give misleading results. Impact tests might, however, in some cases be used to establish optimum parameter settings. In other cases this is not true.
Scanning electron micrographs taken at the weld regions of ABS and PPO showed V-shaped notches on the surfaces. The depth of the notches was relatively small compared with the estimated inherent flaw sizes, and therefore the associated stress concentrations should not affect strength properties.
1. G. Mennig, Mat. Wiss u. Werkstofftechn., 19, 383 (1988).
2. H. G. Mosle, R. M. Criens, and H. Dick, SPE ANTEC Tech. Papers, 30, 772 (1984).
3. K. Tomari, S. Tonogai, T. Harada, H. Harada, K. Lee, T. Morii, and T. Maekawa, Polym. Eng. Sci., 30, 931 (1990).
4. S. C. Malguarnera and A. Manisali, SPE ANTEC Tech. Papers, 28, 124 (1980).
5. S. C. Malguarnera and A. Manisali, Polym. Eng. Sci., 21, 586 (1981).
6. S-G. Kim and N. P. Suh, SPE ANTEC Tech. Papers, 30, 777 (1984).
7. R. M. Criens and H-G. Mosle, in Failure of Plastics, W. Brostow and R. D. Corneliussen, eds., Hanser Verlag, Munich (1986).
8. R. M. Criens, Mater. Chem. Phys., 14, 69 (1986).
9. R. Malloy, G. Gardner, and E. Grossman, SPE ANTEC Tech. Papers, 39, 521 (1993).
10. G. Mennig, Angew. Makromol. Chem., 185/188, 179 (1991).
11. J. L. Williams, and K. J. Cleereman, in Styrene: Its Polymers, Copolymers and Derivates, R. H. Boundy and R. F. Boyer, eds., McGraw-Hill, London (1979).
12. G. E. P. Box, W. G. Hunter, and J. S. Hunter, Statistics for Experimenters, John Wiley & Sons, New York (1978).
13. M. Christie, Plast. Eng., July 1986, p. 37.
14. P. J. Cloud, F. McDowell, and S. Gerakis, Plast. Technol., Aug. 1976, p. 48.
15. B. Fisa, J. Dufour, and T. Vu-Khanh, Polym. Compos., 8, 408 (1987).
16. B. Fisa and M. Rahmani, Polym. Eng. Sci., 31, 1330 (1991).
17. V. M. Nadkarni and S. R. Ayodhya, Polym. Eng. Sci., 33, 358 (1993).
18. A. Meddad and B. Fisa, Polym. Eng. Sci., 35, 893 (1995).
19. S. Hashemi, G. Gara, and B. Stanworth, Plast. Rubb. Compos. Process. Appl., 22, 105 (1994).
20. G. Mennig, Kunststoffe, 82, 235 (1992).
21. S. Brambhatt and R. Malloy, SPE ANTEC Tech. Papers, 38, 2563 (1992).
22. E. Worden and S. Kushion, SPE ANTEC Tech. Papers, 37, 2653 (1991).
23. S. L. Janicki and R. B. Peters, SPE ANTEC Tech. Papers, 37, 391 (1991).
24. A. Vaxman, M. Narkis, A. Siegmann, and S. Kenig, Polym. Compos., 12, 161 (1991).
25. G. Menges, Th. Schacht, and St. Ott, Kunstoffber., 4, 54 (1988).
26. B. Sanschagrin, R. Gauvin, and B. Fisa, Plast. Compounding, 37, (May/June 1987).
27. K. Engberg, A. Knutsson, P.-E. Werner, and U. W. Gedde, Polym. Eng. Sci., 30, 1620 (1990).
28. S. Hashemi, and J. G. Williams, J. Mater. Sci., 20(11), 4202 (1985).
29. K. L. DeVries, and R. J. Nuismer, Appl. Polym. Sci. 2nd Ed., ACS Syrup. Series 285, p. 277 Washington, D.C. (1985).
30. J. G. Williams, Fracture Mechanics of Polymers, John Wiley & Sons, New York (1984).
31. R. W. Herzberg, and J. A. Manson, Fatigue of Engineering Plastics, Academic Press, New York (1980).
32. Bayer Corp., Novodur ABS polymer (1981).
33. GE Plastics, Noryl modified PPO resins (1992).
|Printer friendly Cite/link Email Feedback|
|Publication:||Polymer Engineering and Science|
|Date:||Jan 1, 1997|
|Previous Article:||Interaction characterization of PA1010/PP blends using PP-g-AA as compatibilizer.|
|Next Article:||Incorporation of density relaxation in the analysis of residual stresses in molded parts.|