Particle Spacing Effects on the Strengthening of Discontinuously-Reinforced Polymer Matrix Composites

In metal matrix composites, the spacing between discontinuous reinforcements can affect strengthening by interfering with the motion of dislocations through the metal. This project looks for similar phenomena in polymer matrix composites (PMCs), since the molecular activity of the polymer chains should be altered in the vicinity of the reinforcements. Awareness of such a trend can improve the understanding of PMC mechanics, which in turn can improve PMC characterization and selection techniques. This project sought a relationship between particle spacing and overall strengthening in a discontinuously-reinforced PMC test case composed of alumina particles in a polyphenylene sulfide matrix. Tensile tests were run on hot-pressed composite samples with varying reinforcement volume fraction and particle size. Data showed that composite strength increases as particle spacing increases, except at high volume fractions where this trend reverses. These results provide preliminary data but demonstrate a need for more in-depth investigation.


Introduction
As designers and engineers implement polymer matrix composites (PMCs) into an increasingly wide variety of applica tions, the need for accurate and detailed information about the characteristics and properties of these materials grows. [1][2][3][4] En gineers commonly use strength data as a means for comparing and selecting materials for industrial application. The amount of reinforcement by volume and its effect on composite strength is the most frequently studied relationship in most discontinu ously-reinforced (DR) composites. In contrast, the effect of the reinforcement spacing, which accounts for volume fraction and size of reinforcements, on composite strength is still unclear in DR PMC systems, though there are well-understood behaviors observed in metal-matrix composites (MMCs). [5][6][7] To control the spacing in these systems as an independent variable in strength tests, the particle size can be varied for a constant reinforcement volume fraction (V R ). Producing samples to test this effect requires a processing method that can produce high-quality specimens with minimal variation between batches. From these specimens, tensile tests provide strength values that can be statistically analyzed to identify any relationships present between particle spacing and composite strengthening.

Theory
Current research contains little on the particle spacing effect on strength in PMCs. Since little research was found in PMCs, the work in MMCs was consulted for a basis in forming a hy pothesis. When considering the strength of a metal or MMC, the Orowan effect tells us that the presence of small, incoher ent phases or particles can increase the overall strength of the material. As dislocations move past these particles, they form loops that surround the particles and interfere with subsequent dislocation motion, thus raising the strength of the material. [8,9] This phenomenon implies that a small particle spacing will provide the best composite strengthening. It is also known that if the spacing between these particles is too small, then the dislocations will not pass between the particles; in contrast, widely spaced particles affected by these loops do not typically impact the strengthening significantly. An intermediate distance between the particles must therefore be determined to obtain the optimum strengthening.
Though this phenomenon accurately describes the behav ior of MMCs, it cannot be directly applied to PMCs, since dislocations are not present within polymers. With the basic understanding that polymer deformation is achieved primarily through the motion of polymer chains, particle spacing in a DR PMC should also affect the strength of that composite material. For two PMC samples with an equal V R but different spacing between the particles, implying varied particle sizes, the speci men with larger particles having more material between them should have more space available for the polymer chains to freely move. In contrast, smaller closely-packed particles should impede the motion of those chains, thereby strengthening the composite.

Experimental
To test for a relationship between particle spacing and com posite strengthening, techniques were developed to process a DR PMC and to obtain tensile test specimens from the pro cessed composite.

Material Selection and Characterization
The matrix needed to be a nontoxic, semicrystalline ther moplastic with a glass transition temperature (T g ) above room temperature; polyphenylene sulfide (PPS) was selected based on these criteria. Alumina (Al 2 O 3 ) polishing powder in 5.0-μm and 0.05-μm particle sizes was selected as the reinforcement phase. Tabulated densities of both materials and particle sizes of the reinforcement were verified through characterization prior to using these values in calculations.

Processing Options
Both V R and were varied to observe strengthening trends L e⋅e with respect to the particle spacing. Volume fractions of 1%, 3%, and 10% alumina were selected. Edge-to-edge spacing L e⋅e values for these conditions are reported in Table 1 accord ing to Equation 1 below: [10] 1  2p  2 2 (1)  Table 1, the spacing between particles decreases with increasing volume fraction and with decreasing particle size. These values illustrate the selected conditions (5.0 and 0.05 μm particles with 1%, 3%, or 10% V R ) should produce significantly differing spacings.

As shown in
Two techniques, injection molding and hot pressing, were considered for producing the DR PMCs for this project. Injec tion molding, using a Dynisco LMM injection molder, involved mixing the alumina with melted PPS; poor mixing and the settling of the alumina particles clogged the injection nozzle several times and did not produce any testable samples.
Hot pressing yielded better results, producing 7.5 cm x 12.5 cm composite plates. For this method, PPS and alumina powders were mixed by ball milling. The processing method developed for this experiment entailed hot pressing under the following parameters: These conditions allowed for a sufficient amount of time to ensure a fully melted sample while minimizing porosity and particle settling

Sample Preparation
Tensile test specimens were made from the pressed plates of composite through two steps: cutting blanks from the plates and punching dogbone specimens from each blank.
Using the punch on a material as brittle as the PMC in this lab causes substantial cracking in the surrounding part; to avoid excess material waste, small rectangles were cut from the large plate using a band saw. These blanks were cut slightly larger than the dogbones. Through this technique, a dozen dogbones can be quickly made from a single hot-pressed plate.

Tensile Testing
A Texture Technologies Corp. TA-XT2i Texture Analyser was used for the tensile testing. This machine proved to be suitable to our experiment by having a small test frame with a 50 g (~500 N) load cell; difficulties with this machine included a relatively fast strain rate (0.1 mm/s) and uncertainty in the calibration of the equipment. For this reason, exact values of the data presented herein are may not be accurate, though the observed trends should still apply. Flaws produced during the processing and sample preparation stages (porosity, notches, cracks) caused an extremely large scatter among the data, which is typical of failure testing of other brittle materials.

Tensile Test Results and Discussion
Preliminary tensile testing was performed to try and observe any general trends present in the strengthening of the PMC. Tables 2 and 3 present data for increasing volume fraction for particle sizes of 0.05 and 5 μm, respectively.
The strength values for both particle sizes show a decrease as V R is increased from 1% to 3%, indicating that strength de creases with particle spacing. With the 5.0 μm particles, though, strength increased from 3% to 10% V R , meaning that strength increases as particle spacing decreases. This discrepancy could occur because 3% may correspond to V min similar to that de scribed by Agarwal; this V would represent the V for which min R the composite strength is at a minimum, even compared to the strength of the unfilled matrix. [11] Tables 4 and 5 show the strength effects of increasing par ticle size for V R values of 1% and 3%, respectively.
For both V R values, increasing the particle size (thereby increasing the particle spacing) increases the strength of the composite. This matches the trend observed when increasing from 1% to 3% V R for a constant particle size but disagrees with the trend hypothesized in Section 2 (that strength should increase with decreasing particle spacing).

Conclusions
Hot pressing successfully produced DR PMCs. Using a punch to produce dogbones caused some imperfections that affected the tensile test results. It may be necessary to abandon punch ing dogbones and to cut samples (not necessarily dogbones) by some other means.
Results from tensile testing demonstrate that for lower par ticle volume fractions, increasing the V R (decreasing the particle spacing) decreased the composite strength, though at higher V R values this trend is reversed. For a constant V R , increasing the particle size (decreasing the particle spacing) increased the strength. Overall, these results show that composite strength Table 2. Data for 0.05 μm particles at 1%, 3%, and 10% V R increases with particle spacing, though the specific strengthen ing mechanism cannot yet be determined. These results are only preliminary; further research is necessary to draw any defi nite conclusions about the effect of particle spacing on PMC strengthening.

Future Work
To continue using a PPS matrix, more tensile tests must be performed, since PPS is a brittle materials and its failure can only be properly studied through in-depth statistical analysis. Using a more ductile material like polyethylene would allow the yield strength to be measured instead of fracture strength. In this situation, fewer samples would be necessary, since the yield phenomenon is less variable than fracture. This is because yield values generally show less variation than do fracture strength measurements.
While it appears that the processing method described above is adequate for producing DR PMC samples with good particle dispersion, additional characterization can be applied to verify this. Accurate microscopy could verify the spacing values pre dicted or scanning electron microscopy (SEM) could be used to look for particle agglomeration and flaws on fracture surfaces. Further characterization of the polymer matrix (i.e. crystallinity, molecular weight) could also help to explain the observed re sults. Finally, a more accurate and easily controlled tensile tester should produce more reliable data. In particular, a test frame with a high load capacity (~500 N), slower strain rates (0.01 mm/s or less), and finer resolution for measuring extension or strain (0.005 mm or less) would be ideal.

Acknowledgments
The investigators would like to thank the Materials Science and Engineering department at Virginia Tech for funding this senior design project. They are also extremely grateful to Dr. Jon Geibel of Chevron Phillips for donating the PPS used in this project and to Dr. Garth Wilkes of Virginia Tech for his ad vice and support. Dr. Brian Love's and Dave Berry's assistance Table 4. Particle size (spacing) effects on strength for 1% V R with the testing and processing equipment is also appreciated. Finally, thanks go to Dr. Marie Paretti and Dr. William Reynolds for supervising and advising the MSE senior design projects.