Evaluating mechanical properties of reinforced PC structural foam.
Flexural properties of glass fiber reinforced polycarbonate (PC) structural foam were found to vary considerably with glass fiber content, density reduction, and sample orientation.
For the past twenty years, thermoplastic structural foam has been used to make cabinets and housings for electronic equipment. The major advantage of the material is its structural rigidity relative to weight: Because a structural foam system is composed of solid outer layers of a base thermoplastic material and a foamed inner core of the same material, it has the advantage of using the mechanical strength of the material very efficiently. This is particularly true in regard to flexure, in which the outer fibers of the material carry the bulk of the load. The resulting foam system uses less material than its solid counterpart yet has nearly equivalent stiffness. In addition, the large areas and thick sections that are typically present in electronic enclosures are more easily molded of structural foam than of solid material. During molding, the skin of a structural foam sandwich solidifies first. The flexible inner foamed core then acts as a shock absorber for differential cooling stresses, which occur during the molding of large, thick, flat parts and generally cause the sink and warp in solid parts. Thus, foamed parts are less sensitive to sink and warp flaws.
In the past, designers placed few demands on the structural behavior of foamed materials in electronic enclosures. Structural analysis of such sandwich materials was difficult and beyond the patience or time constraints of the average designer. When structural concerns were important, designers typically calculated section stiffness (EI) by using a composite flexural modulus for the material. To compensate for errors that could result from this type of approach, large safety factors were used in designing such materials.
However, electronics designs have become more sophisticated. Designers now prefer to lessen their reliance on safety factors by improving their ability to accurately calculate the mechanical behavior of the composites. In order to conduct a more exacting analysis of the material, the stiffness contributions of the foam core and outer skin of the composite must be determined independently, and the properties weighted.
In practice, the use of composite analysis for structural foam is difficult because the thickness of the skin and the properties of the individual core and skin layers must be known. The skin-foam transition region is not well defined for the low pressure foam process, and the foam density and properties are likely to vary through the cross section. A more complex analysis can be used, provided that the foam density-modulus relationship can be determined.
The prediction of part performance in service is extremely difficult because of the nonuniform discontinuous nature of thermoplastics structural foams. But the problem is further complicated when a polymer contains reinforcements such as glass fibers, which are likely to result in anisotropic behavior because of the influence of processing-related shear flow on fiber orientation. One study has shown that both the density of the composite and the orientation of glass fibers in the skins (outer layers) of the sandwich can affect the stiffness, strength, and creep of glass fiber reinforced foam. Fiber orientation and density of the material are determined by the molding process. Because of variations in density and fiber orientation in flexural strength calculations, designers do not typically consider the anisotropic nature of the sandwich.
This article investigates the mechanical properties of glass fiber reinforced polycarbonate (PC) structural foam relative to predicted values. Three grades of PC structural foam--5%, 10%, and 30% glass fiber reinforced--were investigated. The flexural properties of each material were determined at nominal density reductions of 0%, 10%, 20%, and 25%. Flexural properties of the solid, foam core, and composite "structural foam" materials were determined in both flow and cross-flow directions.
The results of the study show that the flexural properties vary considerably with glass fiber content, density reduction, and sample orientation; flow and cross flow values vary by as much as 35%. The individual moduli of the skin and foam core, obtained from machined samples, were used in predicting performance of the "composite" samples by means of the classical "equivalent width technique." For the various composite structural foams samples, agreement between the theoretical load-deflection predictions and the experimental load-deflection results was generally good. Although the experimental data were scattered, they show clear trends with respect to fiber orientation and density reduction.
Materials. The following glass fiber reinforced PCs from GE Plastics were evaluated: Grade FL-900, 10% glass fiber reinforced PC foam; Grade FL-910, 10% glass fiber reinforced PC foam; and Grade FL-930, 30% glass fiber reinforced PC foam. The "structural foam" grade PCs were of pellet form--the blowing agent had been precompounded into the formulation.
Each of the materials was dried according to manufacturers' recommendations, prior to molding.
Equipment. For processing, the injection molding machine was a Cincinnati Milacron T-100 reciprocating screw machine, featuring a 100-ton clamp and 8-oz injection capacity.
The injection mold was a single cavity, sprue-gated, rectangular plaque mold; its cavity dimensions were 6.35 x 127 x 230 mm. The mold was used to produce molded plates, from which test specimens were machined (see Fig. 1).
A band saw was used to cut flexural test samples (rectangular beams) from the rectangular plaques. The samples were then milled to finish dimensions by means of a high speed router, equipped with a six-flute carbide cutter and appropriate sample holding fixtures.
An Instron Corp. (Canton, Mass.) Universal Testing Machine, Model 1137, was used with an adjustable span three-point bending fixture to evaluate flexural properties of the machined test specimens.
Molding. To produce the solid (0% density reduction) specimens, the materials were dried and molded in accordance with conventional molding practices, on conventional injection molding equipment. The same equipment was used to produce the foam moldings, which had density reductions of 10%, 20%, and 25%, for each material. Foam samples were obtained by use of conventional short shot techniques, which reduced the shot size until an appropriate part weight (based on a percentage of the average "solid" part weight) was obtained. Low clamp forces (to assist venting), maximum injection speeds (which produced a fill time of approximately 1.0 sec), low back pressures, and tool temperatures of 175 [degrees] F were used for the foam molding trials. To ensure the correct nominal density reduction, the weight of each molded part was measured.
Specimen Preparation. The band saw was used to cut oversized test specimens from the molded plaques; Fig. 1 shows the locations from which the specimens were cut. Flow front progression, obtained from a series of short shot moldings, showed that test specimens cut from these locations would be representative of flow and cross-flow sample orientation.
The sides of the rough dimension samples were machined, by the high speed router apparatus, to a width of 12.7 mm. To ensure the sharpness of the cutting surface, the cutting tool was replaced periodically. Little evidence of softening (during machining) of the sample surfaces could be detected; after machining, foam cells could be observed. This procedure was used to prepare flow and cross-flow test samples for each material formulation, at density reductions ranging from 0% to 25%. A second set of samples was machined along the width and thickness (top and bottom) to remove the "solid skin layers," leaving only the foam core. An optical microscope was used to estimate the skin thicknesses (see Table 1).
Testing. The three-point bending apparatus was used to evaluate flexural properties of the solid, composite structural foam, and foam core specimens (see Fig. 2). ASTM D 790 was followed with one exception: Because of sample size limitations, the span-to-depth ratio was kept constant at 14:1, rather than the recommended 16:1. Initial modulus, flexural strength at 2% outer fiber strain, and outer fiber strain at break are reported.
The flexural properties of the solid and foam materials were then used to calculate the properties of the composite by means of classical elastic composite theory. The calculated values were then compared with those obtained experimentally.
Results and Discussion
Tables 2 through 4 show the flexural test results for the solid, structural foam, and foam core PC samples, all of which were cut from both the flow and cross flow directions. The results show that the flexural modulus of the structural foam composite is, as expected, between that of the molded solid and the foam core samples. Predictably, the stiffness (E) of the structural foam increases with glass content for the solid, foam, and composite samples. Figure 3 illustrates the linear variation of modulus with glass content for foam solids and composites that have been tested in the flow direction and have a 20% density reduction. The same trend holds for the other density reduction values.
The modulus of the composite material, and strength at 2% outer fiber strain, approached those of the solid moldings as the glass content was increased (see Tables 2 through 4). Table 1 shows that the skin thicknesses of the structural foam composites, too, increased with the glass content values (at a particular density reduction), causing the sandwich to increasingly resemble the solid in appearance and behavior.
The modulus and strength values presented in Tables 2 through 4 also show that the properties of the test specimens cut from the flow direction were, in some cases, significantly different from the properties of the specimens cut from the cross flow direction. Differences between flow and cross flow specimens actually increased with glass fiber content at all density reductions. It appears that such behavior can be attributed to fiber rather than molecular orientation, because of the strong effect that fiber content has on the mechanical properties of the system. For instance, the properties of the 5% glass fiber reinforced PC were nearly isotropic, although the cross flow modulus and strength values were somewhat higher in some cases. The flow direction property values for the 10% glass fiber grade are generally higher than those of the cross flow samples at each density, but the differences are generally less than 10% (see Fig. 4). However, the differences in the flow and cross flow direction properties for the 30% glass fiber reinforced structural foam are very significant, as Fig. 5 shows. Flow direction modulus values were 18% and 35% greater than cross flow values at density reductions of 10% and 20%, respectively.
Although the cross flow direction properties for the 30% glass fiber structural foam are significantly lower than those of the flow direction, it should be noted that even the cross flow modulus values are at least 30% higher than the flow direction values of the 10% glass fiber reinforced grade. However, the flow induced orientation is expected to be a significant concern in design, especially of parts having long flow lengths and thinner walls. To obtain the full structural benefit of these materials, extra care should be taken when locating gates and processing the more highly reinforced grades.
The data presented in Tables 1 through 4 relative to "foam only" samples were generated by machining the skins from the test specimens. The foam-only modulus data, together with the solid modulus data and estimated skin thicknesses, were used to predict the load-deflection behavior of the structural foam samples by means of the equivalent width technique. Consistent flow directions were used for all property sets. In practice, the use of such a method is limited because skin thickness is not easily quantified. However, the method appears to be very accurate (see Figs. 6 through 8) when the required data can be generated.
At low deflections, where material behavior is nearly linear and an elastic model is appropriate, agreement between experimental data and the predicted values obtained by means of the equivalent width technique was excellent. At the low deflections, agreement between the experimental and predicted load deflections curves ranged from 100% (as shown in Fig. 6) to 80% (as shown in Fig. 7). The correlation of predicted and calculated values appears to be strongest for the material of higher glass content, and at the lower density reductions.
The model that was used tended to overestimate the stiffnesses of the structural foam beams--particularly for materials of lower glass content. Such overestimation may be due to experimental errors, such as those relating to skin thickness measurement or surface melting during the machining operations. The model may also be too simplistic for analyzing a composite that may have compressive/tensile property differences,or differences between molded solid properties relative to skin properties, produced in the low pressure foam process. Figure 8 shows, however, that this model is a better tool for predicting the behavior of the composite than either the conservative "skin only" model (which underestimates the stiffness of the beam) or the "solid only" model (which overestimates the stiffness of the beam).
The use of cross flow properties to predict the behavior of a "flow" composite can introduce error as high as 10% at low glass contents, and 35% at high glass contents. This error, however, will be on the conservative side, for the best estimate of true structural foam beam behavior (especially for the more highly glass fiber reinforced grades) is obtained when the appropriate directional properties are used for the prediction. However, the availability of skin and core properties in the cross flow direction appears to allow the designer a reliable method of estimating the flexural properties of the structural foam composites--at least on a test sample. The applicability of this approach to molded parts will need to be verified in future work.
The authors would like to thank Apple Computer Inc., Product Design Department, Santa Clara, Calif., for providing funding for this project. Additional thanks go to Venkata Atluri and Saumil Brahmbhatt, of the University of Lowell, for assisting in the experimental portions of the study.
TABLE 1. Estimates of Average Skin Thickness Values at Test Sample Location (Polycarbonate With Glass Fiber Reinforcement).
% Glass Density Skin fiber reduction, thickness, % mm 5% 10 1.83 20 1.22 10% 10 1.88 20 1.27 25 0.81 30% 10 1.91 20 1.45 25 0.89
TABLE 2. Flexural Properties for 5% Glass Fiber Reinforced PC Structural Foam.
Flexural Outer fiber stress Outer fiber modulus, MPa at 2% strain, MPa strain at break, % Cross Cross Cross Sample Flow flow Flow flow Flow flow Solid 2490 2680 45.5 51.0 >5.0 >5.0 Foam core only 1880 2120 38.6 42.8 >5.0 >5.0
Structural foam 2210 2520 42.8 49.0 >5.0 >5.0
Foam core only 1480 1690 30.3 34.5 >5.0 >5.0
Structural foam 2040 1940 40.0 36.6 >5.0 >5.0
TABLE 3. Flexural Properties for 10% Glass Fiber Reinforced PC Structural Foam.
Flexural Outer fiber stress Outer fiber modulus, MPa at 2% strain, MPa strain at break, % Cross Cross Cross Sample Flow flow Flow flow Flow flow Solid 3110 2830 55.2 52.4 >5.0 >5.0 Foam core only 2640 2630 50.3 51.0 >5.0 >5.0
Structural foam 2860 2590 52.4 50.0 >5.0 >5.0
Foam core only 1900 1710 35.9 33.1 >5.0 >5.0
Structual foam 2210 2410 41.4 48.2 >5.0 >5.0
TABLE 4. Flexural Properties for 30% Glass Fiber Reinforced PC Structural Foam.
Flexural Outer fiber stress Outer fiber modulus, MPa at 2% strain, MPa strain at break, % Cross Cross Cross Sample Flow flow Flow flow Flow flow Solid 5300 3680 95.2 71.0 2.50 3.27 Foam core only 4140 3510 78.6 64.8 2.90 3.30
Structural foam 4660 3810 95.2 73.8 2.57 3.33
Foam core only 4010 2660 73.8 51.0 2.89 3.63
Structural foam 4820 3170 84.8 58.6 2.55 3.47
PHOTO : FIGURE 1. Flow and cross flow specimens were cut from molded test plaques at the locations shown.
PHOTO : FIGURE 2. A three-point bending apparatus and constant span-to-depth ratio were used to evaluate the test specimens.
PHOTO : FIGURE 3. Flow direction flexural modulus vs. glass fiber content for solid PC and 20% nominal density reduction PC.
PHOTO : FIGURE 4. Flexural modulus vs. density reduction for 10% glass fiber reinforced PC structural foam, cut from the flow direction.
PHOTO : FIGURE 5. Flexural modulus vs. density reduction for 30% glass fiber reinforced PC structural foam, cut from both the flow and cross flow directions.
PHOTO : FIGURE 6. Experimental and predicted load-deflection curve for the 30% glass fiber, 20% density reduction PC structural foam sample that was cut from the flow direction. Predicted behavior was based on skin and core moduli.
PHOTO : FIGURE 7. Experimental and predicted load-deflection curve for the 10% glass fiber, 20% density reduction PC structural foam sample that was cut from the flow direction. Predicted behavior was based on skin and core moduli.
PHOTO : FIGURE 8. Experimental and predicted load-deflection curves for the 30% glass fiber, 25% density reduction PC structural foam sample that was cut from the flow direction. Predictions were based on the skin and core moduli, the skin only, and a solid cross section.
Lee Hornberger Apple Computer, Inc. Santa Clara, California Robert Malloy and Prabodh Kadkol University of Lowell Lowell, Massachusetts
|Printer friendly Cite/link Email Feedback|
|Author:||Hornaberger, Lee; Malloy, Robert; Kadkol, Prabodh|
|Date:||Jun 1, 1991|
|Previous Article:||A novel nonreactive HALS boosts polyolefin stability.|
|Next Article:||Processing polyolefins on single-screw extruders.|