Effects of Flow Induced Orientation of Ferromagnetic Particles on Relative Magnetic Permeability of Injection Molded Composites

Transcription

1 Effects of Flow Induced Orientation of Ferromagnetic Particles on Relative Magnetic Permeability of Injection Molded Composites THOMAS FISKE,* HALIT S. GOKTURK, RAHMI YAZICI, and DILHAN M. KALYON** Highly Filled Materials Institute Chemical Sciences and Engineering Department Stevens Institute of Technology Hoboken, New Jersey Composite samples consisting of ferromagnetic asymmetric particles incorporated into a polyolefin binder were injection molded using custom designed molds which produced preferential fiber orientations. The relative magnetic permeability values of the composites were measured as a function of the filler volume fraction, injection rate, gate diameter, temperature, aspect ratio of the fibers, and fiber orientation. Fiber orientation was affected by the molding conditions and controlled the relative magnetic permeability of the composites. The degree of fiber orientation was significantly affected by the size of the opening (gate) to the mold, or by the mold geometry going from an edge-gated cylindrical to a center-gated disk cavity. Relative permeability values of the composites were observed to increase when the fiber orientation and the applied field were parallel to one another. For instance, highly aligned composite samples exhibited up to 30% greater relative permeability values compared to those samples which exhibit fiber orientation distributions approaching a random distribution. To our knowledge this is the first study that provides data linking the fiber orientation distribution functions of ferromagnetic asymmetric particles to the relative magnetic permeability values of injection molded composites. INTRODUCTION Polymer composites that exhibit relative magnetic permeability values greater than one can be prepared by mixing a magnetic filler material with a polymer binder. This is accomplished in a number of ways depending on the application of the composite. For instance, early researchers discovered that dust cores could be fabricated by compression molding of soft ferromagnetic powders combined with a polymeric binder (1 3). The binder serves to coat and insulate the individual particles in order to reduce power losses caused by eddy currents. Composites containing ferromagnetic particles and polymeric binders are also used extensively in magnetic recording and data storage media (4 6). For example, acicular -Fe 2 O 3 is a typical magnetic powder widely used in recording tape manufacture. Such powders typically exhibit aspect ratios in the range of * Currently with the Department of Energy. ** To whom correspondence should be addressed. two to eight and length of about 0.5 m. The magnetic particles are mixed with polymeric binders, solvents, plasticizers, lubricants and other additives to form a paint. The paint is applied to a polymer substrate (coated) and the particles are oriented by applying a magnetic field. The use of high aspect ratio particles in composites has various advantages over low aspect ratio particles. For example, high aspect ratio particles reduce the maximum packing fraction and the critical solid concentration levels where interparticle interactions (magnetic enhancement due to the close proximity of the particles) become important. The shape of high aspect ratio particles also contributes to magnetization. This means that it is possible to achieve the same magnetization with less filler or higher magnetization with the same amount of filler by using asymmetric particles. Another application of the ferromagnetic composites is in plastic magnets. Since their introduction (7) they have found widespread use in many industries. The ease with which plastic magnets can be manufac- 826 POLYMER ENGINEERING AND SCIENCE, MAY 1997, Vol. 37, No. 5

2 Effects of Flow Induced Orientation tured has contributed significantly to their rapid growth. For instance, injection molding offers high production rates at low cost and the ability to make complex shapes (8, 9). A novel application for soft ferromagnetic composites is for the attenuation of low frequency electromagnetic fields. In shielding applications it is often desirable to fabricate complex three dimensional shapes with high magnetic permeability values (10). Therefore, it would be advantageous to use injection molding to make high permeability shielding composites. The ultimate properties of injection molded articles are affected by various microstructural distributions including orientation, residual stress, crystallinity and density (11 19). Furthermore, the magnetic permeability of composites is influenced by the filler s magnetic quality, concentration, and particle shape and size (20 25). These microstructural variables are complex and do not allow the application of various available theoretical approaches (26 30), proposed on simplified systems to link magnetic permeability of a composite to the content and various characteristics of the magnetic powders. Thus, only experimental studies can provide the necessary guidance to proper selection of gating and mold design to tailor composites with targeted magnetic properties. Clearly, the orientation of ferromagnetic fibers upon the thermomechanical history of the injection molding process will play a key role in the development of the relative magnetic permeability value of the injection molded composite. A study similar in scope and aimed to link electrical properties of composites to the characteristics of conductive fillers was carried out by Weber and Kamal (31). The objective of this study was to experimentally examine the effects of processing on the orientation distributions of ferromagnetic fibers in injection molded composites and to elucidate the effects of the resultant fiber orientation distributions on the magnetic properties of the injection moldings. Furthermore, the effects of the concentration and the aspect ratio of the ferromagnetic particles were also investigated. EXPERIMENTAL Material The polymer used in this study was a commercial grade of high density polyethylene (HDPE), Quantum LR 734. This polymer has a melt density of kg/m 3 at 180 C under 1 atm. pressure and a melt flow index value of 0.35 g/10 min. The zero shear viscosity of this polyethylene is 150,000 Pa-s at 180 C. Several different ferromagnetic materials including a NiZn ferrite powder, flakes of an amorphous metal, a nickel powder with spherical particles, and two batches of nickel fibers were used as fillers. The ferrite powder (Steward s designation 28) was supplied by D. M. Steward Manufacturing Co., Chattanooga, Tenn. It is a fully reacted nickel-zinc spinel ferrite, which is spherical and exhibits a mean particle diameter of 60 POLYMER ENGINEERING AND SCIENCE, MAY 1997, Vol. 37, No. 5 m. The particle size distribution of this ferrite is narrow and the particle diameters range from 50 mto70 m. The nickel powder with spherical particles was supplied by Johnson Matthey of Ward Hill, Mass. The average particle size of the nickel powder was 110 m. Two grades of nickel fibers were also procured from National Standard of Mishawaka, Ind. The two batches of Ni fibers had a diameter of 20 m but differed in length. The shorter fibers had a length of 260 m (aspect ratio 13) and the longer fibers had a length of 1000 m (aspect ratio 50). The amorphous material, Metglas 2705M, was supplied by Allied-Signal, Parsippany, NJ. It has a chemical composition of 69% cobalt, 12% boron, 12% silicon, 4% iron, 2% molybdenum, and 1% nickel. Metglas was procured as a continuous ribbon of m (1 in.) width and 20 m (0.8 mil) thickness. The specification of the alloy are: Curie temperature of 365 C, crystallization temperature of 520 C, maximum dc permeability of 290,000 (as cast), saturation induction of 0.77 T, and saturation magnetostriction 1 part per million. Mixing Mixing of the high density polyethylene (HDPE) with the filler materials was accomplished by using a Haake Rheocord Torque Rheometer, UV5, with a Rheomix 600 mixing head attachment, at 180 C and 60 RPM. First, the HDPE pellets were introduced into the mixing chamber and allowed to melt and reach thermal equilibrium (approximately 2 minutes). Then, the filler was slowly added during the next 3 minutes. The mixing continued for an additional 5 minutes to ensure thorough dispersion of the filler. The degree of fill used in the mixing chamber was 70%. Injection Molding The specimens for magnetic property characterization were made by an injection ram molding apparatus which was converted from an Instron Floor Tester, Model FT and an Instron capillary rheometer. The attrition of fibers in this apparatus should be less than the typical attrition caused by the Archimedean screw driven injection molders. Two molds were designed and built; one to produce straight cylindrical specimens with a length over diameter ratio of 10 and the second to mold disk-shaped specimens with a diameter over the thickness ratio of 8. The disk mold consists of three sections as shown in Fig. 1. The bottom section of the mold is bolted to the top portion and holds the cavity section (middle) in place. The top portion of the mold attaches to a special adaptor which fits to the end of the barrel of Instron (Fig. 1). The suspension is forced through the die (located at the upper portion of the adapter) and the adaptor, and enters into the top section of the mold. It then enters into the disk shaped mold cavity and fills the mold. The mold has two thermocouple wells and the base plate has a pressure transducer cavity so 827

3 Thomas Fiske, et al. Fig. 1. Disk mold for injection molding: (a) the disk mold cavity (middle) is supported by the bottom plate (left), which attached to the top plate (right); and (b) the assembled injection mold. that both temperature and pressure in the mold can be monitored. The temperature was measured with two K-type thermocouples. The pressure was measured with a Dynisco force transducer, FT 44HM, and a microprocessor-based pressure indicator PR960. A Hewlett-Packard 3497A data acquisition box was used in conjunction with a HP 300 computer for collection and analysis of data. Both molds could be kept at a constant temperature and the injection temperature and speed could be systematically altered as parameters of the study. The dimensions of the gate could also be changed. Molding Procedure The Instron barrel was heated to 220 C. The cylindrical runner was placed in the adaptor and attached to the barrel. The molds were assembled and heated in mantles to the appropriate temperature. The barrel was loaded with the composite suspension and the ram was attached. The heated mold was attached to the barrel. The cross head speed of the plunger was set to correspond to the desired injection speed. The force acting on the plunger during injection molding was monitored. After the mold was cooled, it was disassembled and the injection molded sample was ejected. Sample Preparation The cylindrical injection moldings were m (0.25 in.) in diameter and m (2.5 in.) in length. The disk shaped injection molded samples were m (0.25 in.) in thickness and m (2.0 in.) in 828 POLYMER ENGINEERING AND SCIENCE, MAY 1997, Vol. 37, No. 5

4 Effects of Flow Induced Orientation diameter. Toroidal specimens with an inside diameter of m (1.64 in.) and an outside diameter of m were cut from the disk shaped injection molded samples. For comparison purposes aligned Metglas samples were prepared. The Metglas 2705M ribbon was cut into 16 strips with width of m (0.25 in.) and length of 0.22 m (8.66 in.). The strips were placed in a polyethylene toroid. The Metglas strips represented 5% of the toroid by volume. The Metglas ribbon was also cut into flakes of width and length of m (0.25 in.). The flakes were placed on a piece of doublesided tape with a small space between each flake. The tape was placed inside the toroid. This provides a mesh geometry where the Metglas flakes are insulated from one another. The Metglas again occupied 5% of the volume of the toroid. Characterization of Magnetic Properties The magnetic properties of cylindrical injection molded specimens were characterized with a Hewlett Packard 4284A LCR meter interfaced to a MacIntosh IIfx by an IEEE 488 card. Attached to this was an air coil in which the cylindrical samples could be slipped in and out without detachment. The magnetic property measurements of the toroidal samples were carried-out using a Hewlett Packard function generator model HP 200CDR, a Carver Power Amplifier model PT-1250, a Hitachi Oscilloscope model 425, and an EG&G Lock-In Amplifier model 5209 (22). The toroidal samples were uniformly wrapped with two sets of wire windings. One functioned as the primary coil (excited with a sine wave from the function generator) and the other served as the secondary or pick-up coil (the voltage induced in this coil was measured by the lock-in amplifier). Calibration of Two Magnetic Measurement Setups In order to ensure that the two magnetic measurement setups were consistent with each other, samples were prepared from HDPE/NiZn ferrite. The NiZn ferrite particles are spherical in shape and thus their orientation in different flow geometries is not a factor. Both cylindrical and toroidal samples were prepared at the 50 vol% loading level using a constant volumetric flow rate of m 3 /s at 155 C. There was no significant difference between the magnetic behavior of the NiZn ferrite samples molded into disk (relative permeability ) and cylindrical shapes ( ). Thus, the reported differences in magnetic permeability for high aspect ratio particles are associated with the differences in fiber orientation distributions caused by differences in mold geometry. Determination of Fiber Orientation Fiber orientation distributions of the injection molded samples were analyzed by applying X-ray microradiography on the transverse cross sections of the molded samples. The cross sections were selected POLYMER ENGINEERING AND SCIENCE, MAY 1997, Vol. 37, No. 5 from various locations in the samples at 200 m to 2000 m thicknesses. Microradiographs were taken with soft X-rays ( 10 4 ev) using a GE-GA5 unit and fine-grained X-ray films as described elsewhere (32, 33). The analyses were carried-out at 20 and 50 magnifications. The images of the microradiographs were digitized and further processed using Adobe Photoshop and Scan Maker software. Orientation distributions of individual fibers with respect to reference axes were characterized employing the digitized images using NIH Image/ppc processing software. Subsequent to the measurement of the orientation of the individual fibers, the orientation distribution and the orientation function (J) of the fibers were determined for each case. As defined by Stein (34), and utilized by Yaguchi et al. (35), the fiber orientation function (J) for two dimensional distribution is given by: /2 J 2 /2 Cos 2 q( ) d 1 (1) where is the fiber orientation angle and q( ) isthe distribution of the fiber orientation angles. The orientation function J varies such that: J 0, for random orientation; J 1, for unidirectional orientation in x direction ( 0); and J 1 for unidirectional orientation in y direction ( /2). In this study discrete values were obtained for each fiber in the microradiographs and the orientation function was determined by numerical integration. The accuracy of the NIH-Image based image analysis tools and the subsequent calculations in determining J values were tested using computer simulated fiber distributions prepared elsewhere (35). These control measurements were within 2% of the simulated J values. RESULTS AND DISCUSSION Fiber Orientation of the Toroidal and Cylindrical Profiles Typical micrographs of cylindrical and toroidal samples with the asymmetric Ni particles are shown at two different magnifications in Figs. 2a and 2b, respectively. It can be seen from the micrographs that the fibers in the cylindrical samples are highly aligned in the direction of flow (Fig. 2a). This is also the same direction of the applied magnetic field. The micrographs also show that the fibers in the toroidal samples are oriented in a random fashion (Fig. 2b). The typical orientation distributions of the fibers as obtained by image analysis in the cylindrical and toroidal samples are shown in Figs. 2c and 2d, respectively. The orientation function (J) of the toroidal samples varied from 0.16 to 0.05 depending on the region of analysis. For the toroidal samples a value of J 0.16 indicates a relatively low fiber alignment in the injection molding flow direction (y) which is perpendicular to the applied magnetic field direction (x) (see Fig. 2). 829

5 Thomas Fiske, et al. Fig. 2. Fiber orientation of 260 m fibers incorporated into HDPE at volume fraction 0.10: (a) x-ray microradiography of injection molded cylindrical specimens (magnifications 50 and 18 ), (b) x-ray microradiograph of the injection molded toroidal specimen (disk) (magnifications 50 and 18 ); (c) Fiber orientation distributions of the cylindrical; and (d) toroidal samples. The flow and applied magnetic field directions are as shown. 830 POLYMER ENGINEERING AND SCIENCE, MAY 1997, Vol. 37, No. 5

6 Effects of Flow Induced Orientation The cylindrical samples on the other hand exhibited relatively high degree of fiber orientation, i.e., alignment with respect to the flow direction, with the values of the orientation function ranging from 0.30 to The values for the cylindrical samples varied as functions of fiber content, fiber length or aspect ratio, gate size, and injection rate. Table 1 shows that for the HDPE/260 (fiber length 260 m, aspect ratio of 13) composites, the overall fiber orientation decreases as the fiber volume percent increases under injection molding conditions (temperature and injection rate), which were kept constant. This is probably due to increased interactions and friction between the fibers and the limited length of the gate used, which was not long enough to induce high alignment in the samples with high volume percent of fibers. The decrease in the orientation function due to high fiber volume percent in HDPE/260, however, was moderate. The J values, for example, dropped from 0.61 to 0.53, when the fiber content increased from 10 vol% to 20 vol%. However, in the 20 vol% sample with J 0.53, there are still more fibers aligned than in the 10 vol% sample with J Thus, the directional values of the physical properties such as permeability along the fiber alignment axis in the 20 vol% sample are still expected to be higher than the 10 vol% sample. The samples with the longer Ni fibers (HDPE/1000 and aspect ratio of 50) exhibit less fiber alignment, i.e., smaller J values in comparison to composites with a fiber aspect ratio of 13 under identical molding conditions. This effect should be primarily due to the bending of these longer fibers, especially in those cases where the fiber volume percentages were low. This finding indicates that the fiber stiffness needs to be considered when exploring relatively long fibers. The effect of the gate size was very significant on the fiber orientation distribution and the fiber orientation function, J, values. The J value for 20 vol% fiber samples decreased from 0.53 to 0.32 when the gate diameter decreased from 3.2 mm to 1.7 mm. This was most likely due to the different mold-filling mechanisms caused by the differences in gate design. The smaller gate might cause a jetting effect, which will produce misalignment of the fibers. The jetting effect may be eliminated with a larger gate, which should give rise to a higher degree of fiber orientation. The injection rate was another process parameter that noticeably affected the fiber orientation function. For the HDPE/260 composites at 20 vol% fill level, the J value increased from 0.53 to 0.63 when the injection rate was increased from the m 3 /s to m 3 /s. Implications of the Fiber Orientation with Respect to the Applied Magnetic Field The implication of the applied magnetic field being longitudinal or transverse to the fibers should become clear by examining the simple case of a single fiber Table 1. Summary of the Fiber Orientation Distribution Function (J) of the Injection Molded Cylindrical Samples as Functions of Processing Parameters. Property Orientation Function J Process Variables Fiber Content % Volume: all samples with m fiber length m 3 /s injection rate mm gate diameter Fiber Length m: all samples with % volume fiber content m 3 /s injection rate 3.2 mm gate diameter Gate Diameter: 3.2 mm 0.53 all samples with 1.7 mm % volume fiber content 260 m fiber length m 3 /s injection rate Volumetric Flow Rate, m 3 /s: all samples with % volume fiber content 260 m fiber length 3.2 mm gate diameter Shape of the Mold: cylindrical 0.61 all samples with disk ( ) % volume fiber content 260 m fiber length m 3 /s injection rate 3.2 mm gate diameter POLYMER ENGINEERING AND SCIENCE, MAY 1997, Vol. 37, No

7 Thomas Fiske, et al. subjected to a uniform magnetic field. The magnetic fiber is magnetized by the applied external magnetic field H, inducing a magnetic field which opposes the original field. This field is called the demagnetizing field, H d. Therefore, for a non-zero demagnetization field, the effective field, H eff, will always be less than the applied field because: H eff H H d (2) The apparent permeability of the fiber not only depends on the quality of the magnetic material but also on its shape. Bozorth (2) gives the relationship between the apparent permeability, the permeability, and the shape or demagnetization factor N as: 1 1 N 4 (3) For a fiber (or cylinder) N depends on the ratio of length to diameter, decreasing with increasing length to diameter ratio and also on the orientation of the fiber with respect to the applied field. Values for the demagnetization factor are given by Bozorth (2). To illustrate the effects of demagnetization, the apparent permeability,, is plotted in Fig. 3 for a value of 100. It can be seen from Fig. 3 that the apparent permeability decreases with decreasing effective aspect ratio. The effective aspect ratio is defined as the ratio of the dimension of the fiber parallel to the applied field to that of the dimension of the fiber transverse to the applied field. The effective aspect ratios of our fibers are and 0.02 for the fiber lengths of 260 m and 1000 m fibers, respectively, when the fibers are oriented in the transverse direction with respect to the applied magnetic field. At both of these effective aspect ratios the demagnetization factor, N, is greater than 0.3 and the apparent permeability value approaches to a value of one. Consequently, the fibers aligned transversely to the applied field contribute very little to the magnetization of the composite. Therefore, it can be concluded that composites with fibers oriented in the direction of the applied magnetic field will exhibit a higher apparent permeability than those aligned in the transverse direction. Figures 4 and 5 display the magnetic permeability values of injection molded cylinders and disks for the two composite systems with different lengths of fibers. The confidence intervals determined according to Student s t-distribution are included. In both cases it can be seen that the cylindrical samples indeed exhibit higher permeability values than the disk shaped samples suggesting a higher degree of fiber orientation for cylindrical samples as indeed found to be the case. The quantitative relationship between the relative permeability and fiber orientation distribution function will be elucidated later. Effects of Fiber Length Figures 4 and 5 can be used to compare the relative permeability of the two composite systems with differing fiber aspect ratios at volume fractions of 0.1, 0.2, and 0.3, respectively. For 10% and 20% loading levels, there is little difference in permeability values between the two aspect ratios. The demagnetization factor for a fiber with L/D 50 oriented in the direction of the magnetic field is N while that for a fiber with L/D 13 is N This is an order of magnitude difference. However, this does not translate to an or- Fig. 3. Apparent magnetic permeability vs. effective aspect ratio from Eq 3 (effective aspect ratio is ratio of dimension of fiber parallel to the applied field to the dimension of fiber found transverse to the field) Bozorth (2). Fig. 4. Relative permeability as affected by volume fraction of Ni fibers (mean length of 260 m) and mold shape. 832 POLYMER ENGINEERING AND SCIENCE, MAY 1997, Vol. 37, No. 5

8 Effects of Flow Induced Orientation der of magnitude difference in permeability. A closer look at Fig. 3 reveals that there is less than a factor of two difference in apparent permeability for these two effective aspect ratios. The same is true for the case when the fibers are oriented transverse to the applied field. Furthermore, it should be noted that the two fiber populations (mean lengths of 260 m and 1000 m) exhibit different size distributions, i.e., they are not unimodal. Such size distributions may also contribute to the observed similarities in permeability values. The cylindrical mold was used to evaluate the effect of mold temperature. Three fiber-volume concentrations were used for the HDPE/260 Ni fiber composite and one for the HDPE/1000 Ni fiber composite. Two mold temperatures were used: 75 C and 155 C. The results are plotted in Fig. 6. It is apparent that there is only a marginal increase in permeability with a decrease in mold temperature. Since the effect was marginal the fiber orientation distributions of these molded samples were not characterized. Fig. 5. Relative permeability as affected by volume fraction of Ni fibers (mean length of 1000 m) and mold shape. Effect of Injection Rate and Gate Diameter on Permeability Figures 7 and 8 show the effects of injection rate on the permeability of the composites. The mold and melt temperatures were held constant. Figure 7 summarizes the results of the HDPE/260 composite cylinders. It can be seen that an increase in injection volume flow rate improved the permeability values for the 20% and 30% filled composites. On the other hand, Fig. 8 shows that the increase in injection rate had no effect on the permeability values of the HDPE/1000 cylindrical composites. Figure 9 shows the results for the disk shaped composite specimens containing fibers with mean lengths of 260 m and 1000 m. Again, the permeability values are not affected by changes in injection rate. Gates with diameters of in. (1.7 mm) and in. (3.2 mm) were used to examine the effects they would have on orientation of fibers and ultimately the permeability of the composites. For these experiments the cylindrical mold was used in con- Fig. 6. Effect of mold temperature on relative permeability of injection molded cylindrical samples. POLYMER ENGINEERING AND SCIENCE, MAY 1997, Vol. 37, No

9 Thomas Fiske, et al. Fig. 7. Effect of injection rate on relative permeability of injection molded cylindrical composite samples for Ni fibers with mean length of 260 m at volume fraction, 0.1, 0.2, 0.3 and 0.4. Fig. 8. Effect of injection rate on relative permeability of injection molded cylindrical composite samples for Ni fibers with mean length of 1000 m at volume fraction, 0.1, 0.2 and 0.3. junction with the HDPE/260 composite at an injection rate of 0.55 in 3 /min ( m 3 /s) and a mold temperature of 155 C. The results are shown in Fig. 10. The difference in permeability values for the different gates is significant. The smaller gate produced samples which exhibit lower permeability values than those produced with the larger gate. The increased magnetic permeability is associated directly with the 834 POLYMER ENGINEERING AND SCIENCE, MAY 1997, Vol. 37, No. 5

10 Effects of Flow Induced Orientation Fig. 9. Effect of injection rate on relative permeability of injection molded toroidal composite samples at 155 C. greater degree of fiber orientation. The relationship between magnetic permeability and fiber orientation function is discussed next. The effects of the fiber orientation distribution function, J, on permeability were best revealed on those samples with constant fiber concentration but which were under different injection rate and gate diameter conditions. The data taken from such measurements on HDPE/260 samples with 260 m long fibers and fill level of 20 volume percent ( 0.20) are presented in Table 2. The orientation function, J, varies between zero for spherical Ni particles to 0.63 for fiber incorporated samples injection molded at the injection volumetric flow rate of m 3 /s. The data given in Table 2 are plotted in Fig. 11. The experimental relative magnetic permeability versus fiber orientation function (J) behavior obeys a power law relationship: o 4 J 2 (4) where o is the relative magnetic permeability of composite with spherical Ni particles. If the extrapolation to higher J values is permitted the maximum magnetic permeability value which is expected for fibers oriented uniaxially in the direction of the magnetic field is around 6. Figure 12 shows the difference between composites with spherical Ni particles and composites with 260 m fibers. Composites with high aspect ratio particles POLYMER ENGINEERING AND SCIENCE, MAY 1997, Vol. 37, No. 5 produce higher permeability values at all concentration levels. Thus, at the same loading level the composite with fibers will produce a higher permeability value than that of a composite with spherical particles. This finding is consistent with our earlier findings (22). It is possible to increase the relative magnetic permeability of composites over and above the permeability values reported. This can be accomplished by using fillers with greater relative magnetic permeability values, as will be demonstrated next. Orientation with Metglas So far magnetic materials with relatively low relative permeability were used to show how orientation of the fibers in the desired direction could enhance the permeability value of the composite. It is interesting to elucidate the orientation effect further with a magnetic filler with a greater relative magnetic value. For compression molded and randomly oriented Metglas flakes the relative permeability is 15.5 at A continuous 16 layers Metglas specimen exhibits a relative permeability of 240 at the 5 vol% concentration level. By cutting the Metglas strips into flakes (1/4 by 1/4 in.) and insulating them from one another a permeability value of 91 is obtained. This represents a substantial increase over the random oriented sample permeability of 15.5, upholding the fiber orientation and relative magnetic permeability relationship eluci- 835

11 Thomas Fiske, et al. Fig. 10. Effect of gate diameter on relative permeability of injection molded HDPE/260 composite cylindrical samples molded at an injection volumetric flow rate of m 3 /s using gate diameters of 1.7 and 3.2 mm. Table 2. Relative Permeability and Filler Orientation Function Values for Injection Molded Cylinders with 260 m long Ni Fibers at Process Variables J Orientation Function m Relative Permeability Injection Rate m 3 /s (Gate Size 3.2 mm) m 3 /s Gate Diameter 3.2 mm (Injection Rate 1.7 mm m 3 /s) Particulate Filler dated with nickel. This experiment also demonstrates that significantly higher relative magnetic permeability values can be obtained for composites with fillers exhibiting higher magnetic permeability. CONCLUSIONS Composites with relative magnetic permeability values greater than one were produced by injection molding of suspensions consisting of a high viscosity polymeric binder and various magnetic fillers in two custom designed molds, which produced differences in fiber orientation distributions. The magnetic properties of the composites were characterized in terms of their relative permeability values and related to their microstructure. It was determined that composites with a higher degree of orientation of fibrous magnetic particles generate higher permeability values. Quantitative data which link orientation distribution functions to magnetic permeability values of composites were provided for the first time. It was also shown that various processing conditions and die diameter influence the orientation distributions, hence the permeability. These findings can be utilized to tailor injection moldings with desired relative magnetic permeability values for various industrial applications. ACKNOWLEDGMENTS This research was supported by the Office of Naval Research through a student training grant N provided to Mr. Thomas Fiske. We are 836 POLYMER ENGINEERING AND SCIENCE, MAY 1997, Vol. 37, No. 5

12 Effects of Flow Induced Orientation Fig. 11. Relative permeability versus fiber orientation function (J) for injection molded HDPE/260 composite cylindrical samples with volume fraction of Ni fibers of 0.2. Fig. 12. Comparison of the relative permeability values of injection molded cylindrical samples with 260 m long Ni fibers (aspect ratio of 13) vs. those of injection moldings with spherical Ni particles. grateful for this support. The image analysis was carried-out by Ms. Esra Kucukpinar of HfMI. REFERENCES 1. G. Y. Chen and J. H. Wernick, in Ferromagnetic Materials, Vol. 2, 55, E. P. Wohlfart, ed., North-Holland, New York (1986). 2. R. Bozorth, Ferromagnetism, Van Nostrand, Princeton, N.J. (1968). 3. C. Goetzel, Treatise on Powder Metallurgy, Vol. 2, Interscience Publishers, New York (1950). 4. Y. Katoh, M. Nakayama, Y. Tanaka, and K. Takahashi, IEEE Trans. Mag., 17, 2742 (1981). 5. R. Matick, Computer Storage Systems and Technology, Wiley-Interscience Publishing, New York (1977). 6. B. Bhushan, D. Bogy, N. Eiss, and F. Talke, Tribology and Mechanics of Magnetic Storage Systems, American Soc. of Lubrication Eng., Park Ridge, Ill. (1984). 7. J. M. Dadek, French Patent No. 1,135,734 (1955). 8. T. Sakai, K. Nakamura, and A. Morii, Intern. Polym. Proc., 6, 26 (1991). 9. Z. Osawa, K. Kawauchi, M. Iwata, and H. Harada, J. Mater. Sci., 23, 2637 (1988). 10. S. B. Railkar, H. S. Gokturk, and D. M. Kalyon, J. Reinf. Plast. Comp., 12, 1212 (1993). 11. M. R. Kamal and R. Bakerdjian, SPE ANTEC Tech. Papers, 21, 297 (1975). 12. F. H. Moy and M. R. Kamal, Polym. Eng. Sci., 20, 957 (1980). 13. M. R. Kamal, D. M. Kalyon, and J. M. Dealy, Polym. Eng. Sci., 20, 1117 (1980). 14. A. H. Wagner, J. S. Yu, and D. M. Kalyon, Polym. Eng. Sci., 29, 1298 (1989). 15. J. S. Yu, M. Lim, and D. M. Kalyon, Polym. Eng. Sci., 31, 145 (1991). 16. A. H. Wagner, J. S. Yu, and D. M. Kalyon, Adv. Polym. Techn., 9, 12 (1989). 17. J. S. Yu and D. M. Kalyon, Polym. Eng. Sci., 31, 153 (1989). 18. J. S. Yu, A. H. Wagner, and D. M. Kalyon, J. Appl. Polym. Sci., 44, 477 (1992). 19. A. H. Wagner and D. M. Kalyon, Polym. Eng. Sci., 21, 1393 (1993). 20. T. J. Fiske, H. S. Gokturk, and D. M. Kalyon, SPE ANTEC Tech. Papers, 39, 614 (1993). 21. H. S. Gokturk, T. J. Fiske, and D. M. Kalyon, J. Appl. Phys., 73, 10, 5598 (1993). 22. H. S. Gokturk, T. J. Fiske, and D. M. Kalyon, IEEE Trans. Mag., 29, 6, 4170 (1993). 23. H. S. Gokturk, T. J. Fiske, and D. M. Kalyon, J. Appl. Polym. Sci., 50, 1891 (1993). 24. T. J. Fiske, H. Gokturk, and D. M. Kalyon, Enhancement of the Relative Magnetic Permeability of Polymer Composites with Hybrid Particulate Fillers, to appear in J. Appl. Polym. Sci. (1997). 25. T. J. Fiske, H. S. Gokturk, and D. M. Kalyon, Percolation in Magnetic Composites, to appear in J. Mater. Sci. (1997). 26. Z. Hashin and S. Shtrikman, J. Appl. Phys., 33, 10, 3125 (1962). 27. J. Lam, J. Appl. Phys., 60, 12, 4230 (1986). 28. J. Lam, J. Appl. Phys., 66, 8, 3741 (1989). 29. H. How and C. Vittoria, Phys. Rev. B, 44, 17, 44 (1991). 30. A. N. Lagarkov, A. K. Sarychev, Y. R. Smychkovich, and A. P. Vinogradov, J. Electromag. Waves and Appl., 6, 9, 1159 (1992). 31. M. E. Weber and M. R. Kamal, SPE ANTEC Tech. Papers, 38, 484 (1992). 32. R. Yazici, A. Wagner, D. M. Kalyon, and S. B. Han, SPE ANTEC Tech. Papers, 40, 1172 (1994). 33. A. Wagner, R. Yazici, and D. M. Kalyon, Polym. Compos., 17, 840 (1996). 34. R. S. Stein and S. N. Stidham, J. Appl. Phys., 35, 42 (1964). 35. H. Yaguchi, H. Hogo, D. G. Lee, and E. G. Kim, Intern. Polym. Proc., 10, 3, 262 (1995). Received April 2, 1996 Revised December 17, 1996 POLYMER ENGINEERING AND SCIENCE, MAY 1997, Vol. 37, No