Long fiber reinforcement of polypropylene/polystyrene blends.
Polymer waste usually consists of a mixture of polymers, which cannot always be separated. As most of these polymers are immiscible (incompatible), recycling results in blends with poor mechanical properties. The use of compatibilizers can solve this problem, but compatibilizers have to be added, and the mixing can be critical. A new approach is adding glass fibers to the mixture of polymers. The mechanical properties of polymers improve significantly when glass fibers are added (1-5), particularly when the fibers are long (3). Adding glass fiber to a homopolymer like PP reduces the fracture strain but at the same time increases the fracture stress and modulus. The notched impact strength of PP increases with fiber concentration, and with increasing fiber length a further increase in tensile and impact properties is obtained (5, 6).
The main energy absorption during impact of a fiber-reinforced system is the fiber pull-out. The fiber pull-out depends both on fiber length and the critical fiber length. For optimal impact behavior long fibers are preferred with a long critical fiber length. The fibers can be kept long by mild compounding.
Short glass fibers, when added to mixtures of polymers, increase the modulus and the yield strength and also the impact strength (7, 8). The impact strength with short fibers is, however, lower than if the mixture had been compatibilized (8). Long fibers added to a mixture of polymers increase the fracture toughness (9).
We studied the incompatible blend PP/PS as a model compound for a polymer waste mixture. For this PP/PS (70/30) blend we studied the effect of glass reinforcement, with particular attention to long fiber reinforcement.
Of all the different methods of adding the fibers to the polymer, three methods were evaluated to investigate the influence of compounding. Previous reports show the influence of compounding on fiber-reinforced homopolymers (6, 10-14). Avalos et al. (15) investigated a short fiber-reinforced polyolefin blend. The resulting fiber length decreased with increasing mixing intensity during compounding. Also, the fiber content seemed to influence the fiber length negatively, as did the matrix type (11, 13), although this was not always clear (14, 16). The main energy dissipating mechanism is said to be fiber pull-out. Since this dissipating mechanism depends strongly on the pull-out length, the total fiber length should be as long as possible (17, 18).
The aim of this research is to investigate the long fiber reinforcement of the incompatible blend. Special attention was given to the influence of the compounding method and fiber length on the mechanical properties of the eventual injection molded product.
Three different series of injection molded specimens were made: The first was compounded with the twin screw extruder, the second with the single screw extruder, and the third was dry blended. All three series were made with 0 wt%, 10, 20, and 30 wt% glass fiber (PPG 3242, 4.5 mm long, with filament diameter 13 ([[micro]meter]) and matrix material PP/PS (70 wt% polypropylene DSM Stamylan P 17M10, 30 wt% polystyrene DOW Styron 678E).
The dry blend series was made directly with the specified fiber percentage, which means material with 0, 10, 20, and 30 wt% fiber was injection molded. The extrusion compounded material was made in a different way: Material was extruded with 30 wt% glass fiber, which was then mixed (just before injection molding) with virgin PP and PS to obtain the 10 and 20 wt% series.
The single screw compounding (Brabender 30/25D, 3 mm die, screw diameter 30 mm, with constant screw channel depth) conditions were 230 [degrees] C and a screw speed of 30 rpm. The twin screw compounding (Berstorff, co-rotating, screw diameters 25 mm, L/D = 33, 3 mm die) conditions were 230 [degrees] C and 120 rpm. The extruded material was chopped into granules of 9 mm. The conditions for the subsequent injection molding (Arburg Allrounder 221-55-250, screw diameter 30 mm) were: temperature 230 [degrees] C, mold temperature 40 [degrees] C, screw speed 100 rpm, [P.sub.inj] = 50 bar, [P.sub.hold] = 40 bar. Tensile test bars were made (10 by 3 mm, length about 14 cm). For the impact tests the middle part of these bars was used.
Tensile tests were conducted at 6 mm/min at room temperature. Izod impact tests were conducted. Both tests were carried with notched and unnotched specimens in tenfold. Glass fiber lengths were determined by burning off the polymer; the remaining glass fibers were suspended in a cellulose solution. Of this glass/cellulose suspension one drop was put on a slide. After drying the slide was projected and the fiber lengths measured. Both number ([l.sub.n]) and weight average ([l.sub.w]) fiber lengths were determined [the method was developed by Bijsterbosch (19)].
RESULTS AND DISCUSSION
After injection molding the dry blend, fiber bundles could be seen in the test specimens. It appeared that the injection molding alone did not result in a well-dispersed glass fiber reinforcement. The injection moldings of the single screw and twin screw extruder compound looked well dispersed.
Fiber Length Measurements
The fiber length in the samples was measured in order to study the influence of the fiber length on the properties of the reinforced PP/PS. The fiber length was studied both before and after injection molding.
Before injection molding the length of the fibers depended on the compounding method. For the dry blend the fiber length before injection molding was as received, 4.5 mm. The results are shown in Table 1. The 30 wt% glass fiber single screw extrusion compound had a number average fiber length of 1.33 mm. The fiber length of the twin screw extruded compound was 0.35 mm. Considerable fiber attrition was taking place with twin screw compounding. In our single screw compounds, the fiber length was much longer.
The injection molding of the dry blends and the compounds can cause further fiber attrition (Table 1, [ILLUSTRATION FOR FIGURE 1 OMITTED]). The dry blend showed increasing fiber length reduction with increasing fiber concentration. A similar fiber size reduction can be seen with homopolymers (7).
In the injection molding of the 10% and 20% single screw and twin screw compounds, the 30% compounds were diluted in the injection molding machine. In the single screw compounds the fiber size reduction with fiber concentration was smaller. In the twin screw compound no further size reduction on injection molding was observed. The fiber length reduction in the twin screw compounding step appeared to be the determining step. This was probably due to the higher shear forces in this compounding step.
The impact behavior was studied with the Izod method both on notched and unnotched samples (Table 2, [ILLUSTRATION FOR FIGURES 2 AND 3 OMITTED]). The results show the influence of fiber content on impact resistance. The notched tests give an indication of the notch sensitivity of the material.
The unreinforced PP/PS blend shows very poor impact properties. The unnotched Izod value of the [TABULAR DATA FOR TABLE 1 OMITTED] blend is 13.5-17 kJ/[m.sup.2]. This is much lower than the value for PP ([less than] 70 kJ/[m.sub.2]). The twin screw compound has a lower value than the single screw and dry blend compounds.
By the addition of glass fibers, the unnotched Izod values rise. For the twin screw compound this effect is very small. The values show a slight decrease before increasing. The values for 0 wt% and 30 wt% glass are nearly the same. This is probably caused by the extensive fiber attrition in the twin screw compounding step. The single screw compound shows a considerable increase in unnotched values. For the dry blended material the unnotched Izod value increases even more. By adding 30 wt% glass the value is raised by a factor of two.
The notched Izod results show that the values for the unreinforced blend (1.1-1.7 kJ/[m.sup.2]) are comparable to the value of unreinforced PP (about 1.6 kJ/[m.sup.2]). When glass is added, the Izod values rise and reach the high level of reinforced homopolymers.
The twin screw compound shows a slight increase. The single screw values increase strongly with increasing fiber content. The notched Izod value increases about nine times when 30 wt% glass is added. For the dry blend the increase is even larger. The value increases almost ten times.
In Table 2 the deviation in the test results is shown. Here the problem with the dry blended material becomes clear. The fibers are not as well mixed through the matrix as with extrusion compounding. In the injection molding process the bundles are not always broken into individual fibers. This results in a large variation in product quality and test results. This effect is greatly reduced with the single screw extrusion and even more with twin screw extrusion.
For all the compounds the notched Izod values of the dry blended material are higher than those of the extrusion compounded material. The twin screw compound especially shows lower values. The extensive fiber attrition causes these low Izod values. The limited fiber attrition in the single screw compounding results in a much smaller drop in Izod values.
Both notched and unnotched Izod results show that the poor impact properties of the blend can be improved by adding glass fibers. The dry blended material gives the best results. The notched Izod value becomes ten times higher by adding 30 wt% glass. Long fibers give better impact properties. The properties improve further by adding more fibers. More fibers mean more fiber attrition and thus shorter fibers [ILLUSTRATION FOR FIGURE 1 OMITTED]. The Izod values still increase with increasing fiber content [ILLUSTRATION FOR FIGURES 2 AND 3 OMITTED] and have not yet reached a constant level. This means the optimal fiber content and length have not been reached.
Like the impact testing, the tensile behavior was studied on both unnotched and notched samples. The results are shown in Figs. 4 and 5.
The unreinforced PP/PS blend has poor tensile properties, as in the impact tests. Especially the twin screw extruded compound has poor tensile properties.
Figure 4a shows the modulus results of the unnotched tensile tests. The values for the PP/PS blends are somewhat higher than the modulus of PP (about 1800 MPa).
With increasing fiber content the modulus increases considerably. For the single screw compound the modulus is raised by a factor of four by adding 30 wt% glass. The difference between the three different compounding methods is relatively small. The influence of the compounding route and the fiber length on the modulus is small.
The fracture stress increases with fiber content [ILLUSTRATION FOR FIGURE 4B OMITTED]. The unreinforced blends have values slightly higher than PP (about 25 MPa). The values for the twin screw compound are considerably lower than for the single screw and dry blend. Apparently the fiber length has a large influence on the fracture stress. The values increase with an increasing fiber content. A constant [TABULAR DATA FOR TABLE 2 OMITTED] value has not been reached. This indicates that the optimal fiber content has not been reached yet and the fiber length is still below the critical fiber length.
Figure 4c shows the fracture strain test results. Clearly the fracture strain decreases with an increasing fiber content. The strain values for the unreinforced blends are much lower than for PP (about 300%). The fracture strain values are lowest for the twin screw compound. This poor result is remarkable, since the twin screw extruder is widely used for polymer mixing to give the most homogeneous blend with the finest particle size.
Notched tensile tests were used to investigate the notch sensitivity of the material. These tests were conducted at the same speed as the unnotched tensile tests (6mm/min). The notched tensile tests gave about the same results [ILLUSTRATION FOR FIGURE 5 OMITTED]. In Fig. 5a the apparent modulus test results are shown. Here, too, the apparent modulus increases with fiber content. The difference between the three methods is small. The values are slightly lower than the unnotched modulus measurements. Figure 5b shows the fracture stress results. With an increasing fiber content the apparent fracture stress increases. The twin screw compound values are lower than the dry blend and the single screw compound values. Finally, in Fig. 5c the fracture strain results are shown. Again, with increasing fiber content the fracture strain decreases. Comparing Fig. 4c and Fig. 5c shows that the difference between the three compounding methods is more obvious in the notched than in the unnotched test results. The difference in the fracture strain between the unreinforced blends is also larger. It is not clear whether the influence of the fiber length is more apparent in notched testing than in unnotched testing. The influence of fiber length as measured by the notched and unnotched tensile tests seems to be comparable, with the notched tensile test values for fracture strain and fracture stress at a 40-50% lower level.
Like the impact test results, the tensile tests show a smaller variation in results for the twin screw compound. The dry blend test results have a much larger variation here, too. Both notched and unnotched test results show the same trends. The notched values are somewhat lower.
The results show that the unreinforced PP/PS blend has poor impact properties compared with a PP homopolymer. By adding glass fibers these properties can be improved significantly.
Of the three evaluated compounding methods, the twin screw extrusion resulted in the shortest fibers. The initial 4.5 mm fibers were broken to about a 0.3 mm length. The fiber content had no influence on this length. The single screw compound fibers were about 1.1-0.7 mm long. A higher fiber content increased attrition. The dry blending resulted in the longest fibers: 2.2-1.1 min.
The impact properties improved with an increasing fiber content. This effect was small for the twin screw compound. The single screw compound notched Izod values increased nine times when 30 wt% glass was added. The dry blend gave the best results. The notched Izod value increased ten times when 30 wt% glass was added. However, because of the poor fiber distribution the dry blend results showed a variation much larger than the single screw compound results. This means that the dry blend product quality is not constant.
Tensile properties improved also with an increasing fiber content. The twin screw compound again showed the lowest values. The single screw compound had somewhat lower results than the dry blend, but the variation was smaller.
The results show that for the fiber-reinforced PP/PS blend, long fibers result in strongly improved mechanical properties. The optimal fiber concentration and fiber length have not yet been reached in the experiments.
The mild compounding with a single screw extruder gave compounds with long fibers, which have good mechanical properties and a fairly constant injection molding product quality.
An incompatible blend with originally very poor mechanical properties can, with the use of glass fiber, reach the level of reinforced homopolymers.
1. J. L. Thomason and M. A. Vlug, Composites: Part A, 27A, 477 (1996).
2. D. E. Spahr, K. Friedrich, J. M. Schultz, and R. S. Bailey, J. Mater. Sci., 25, 4427 (1990).
3. V. B. Gupta, R. K. Mittal, and P. K. Sharma, Polym. Compos., 10, 8 (1989).
4. F. Truckenmuller and H.-G.Fritz, Polym. Eng. Sci., 31, 1317 (1991).
5. T. Moriwaki, Composites: Part A. 27A, 379 (1996).
6. M. Joshi, S. N. Maiti. A. Misra, and R. K. Mittal, Polym. Compos., 15, 349 (1994).
7. M. Arroyo Ramos and F. Avalos Belmontes, Polym. Compos., 12, 1 (1991).
8. A. Adewole, K. Dackson, and M. Wolkowicz. J. Thermopl. Compos. Mater., 8, 272 (1995).
9. T. Harmia and K. Friedrich, Compos. Sci. Technol., 53, 423 (1995).
10. D. M. Bigg, Polym. Cornpos., 6, 20 (1985).
11. R. Bailey and H. Kraft, Int. Polym. Processing, 14, 94 (1987).
12. H. J. Wolf, Polym. Compos., 15, 375 (1994).
13. B. Fisa, Polym. Compos., 6, 232 (1985).
14. B. Franzen, C. Klason, J. Kubat, and T. Kitano, Composites, 20, 65 (1989).
15. F. Avalos, M. Arroyo, and J.P. Vigo, J. Polym. Eng., 9, 157 (1990).
16. R. von Turkovich and L. Erwin, Polym. Eng. Sci., 23, 743 (1983).
17. J. K. Wells and P. W. R. Beaumont, J. Mater. Sci., 20, 1275 (1985).
18. A. Kelly, Proc. Roy. Soc., A282, 63 (1964); A319 95 (1970).
19. H. Bijsterbosch and R. J. Gaymans, Polym. Compos., 16, 363 (1995).
|Printer friendly Cite/link Email Feedback|
|Author:||Inberg, J.P.F.; Hunse, P.H.; Gaymans, R.J.|
|Publication:||Polymer Engineering and Science|
|Date:||Feb 1, 1999|
|Previous Article:||Extrusion characteristics of thermoplastic melts containing fluoropolymers.|
|Next Article:||In-Line residence time distribution of dicarboxylic acid oligomers/dioxazoline chain extension by reactive extrusion.|