Accuracy and mechanical properties of multiparts produced in one mold in microinjection molding.

INTRODUCTION

The development of microsystem technology on a vast scale is dependent on manufacturing systems that can reliably and economically produce microcomponents. The microinjection molding process, which permits cost-effective mass production of microstructures from a wide variety of high-performance plastics, is a key, enabling, and specialized technique with a unique set of challenges [1-3].

Kukla et al. [4] defined microinjection molded parts as (1) parts with microweight, (2) parts with microstructured regions, and (3) parts with microprecision dimensions. Parts with microweight are parts with masses of a few milligrams. but their dimensions are not necessarily on the micron scale. Parts with microstructured regions are characterized by local microfeatures on the micron order, such as microholes and -slots. Parts with microprecision are parts of any dimension that have tolerances in the micron range. This work focuses on the first and third types, that is, molded microparts with weights of less than 3 mg and precision on the order of a few microns.

A number of pure plastics, e.g., LCP (liquid crystal polymer), PC (polycarbonate), PS (polystyrene); PP (polypropylene), PMMA (polymethylmethacrylate), and POM (polyoxymethylene or acetal), have been successfully processed through micromolding [5-8]. Most parts, such as gears and fans, require high strength, wear resistance, and accuracy, but plastic materials cannot satisfy these requirements. Some researchers [9-13] have produced parts through micropowder injection molding. This technique includes blending, molding, debinding, and sintering processes. In addition to taking much time, this technique leads to cracks in and high shrinkage of parts although high strength and wear resistance can be obtained [14].

The inclusion of inorganic fillers in polymers for commercial applications is primarily aimed at reducing costs and improving stiffness [15]. It is worth noting that when micron-sized particulates are added to polymers, high filler content is generally required to produce the above-stated positive effects. For instance, the inclusion of fillers, such as glass particles that are 10 [micro]m in diameter, into POM is useful for reducing material costs and reinforcing strength.

Most researchers [16-18] who have studied the forming of microparts have focused on one mold with one part having very precise dimensions. However, this loses the value of batch production through microinjection molding. Using one mold to produce multiparts with high precision is the aim of this technique. Therefore, this study employed one mold to produce four microparts at one time to achieve batch production with highly precise dimensions. In addition, the critical properties, including shrinkage, wear resistance, and strength of polymer POM with added glass particles were systematically established in this study.

[FIGURE 1 OMITTED]

EXPERIMENTS

Materials and Specimens

High-performance POM is an engineering thermoplastic with excellent physical and processing properties. Products produced with POM compound include clock and watch parts, rollers, bearings, gearwheels, and housing parts. Thus, pure POM and POM with added glass beads, the weight content of which was 5 wt%, 20 wt%, or 25 wt%, were selected here as experimental materials. These materials were made by Formosa Plastic Inc., Taiwan. The polymer compound employed here did not include 10 wt% glass beads because these beads are not widely used or available on the market.

Four different microparts in one mold were used to study batch production with high accuracy through microinjection molding. The four microparts, which all had a thickness of 1.5 mm but differed in shape, are shown in Fig. 1. A microgear with 8 teeth and a 1.5 mm outer diameter was tested. Microannular plates, all having an outer diameter of 1.5 mm, were also tested. One was triangular in shape with a wall thickness of 0.2 mm; another had a square inner hole with an edge length of 0.85 mm; and a third one was circular with a donut hole having an inner diameter of 1.0 mm.

A round specimen with a diameter of 8 mm and a thickness of 6 mm was used in wear tests. A rectangular bar with a thickness of 3 mm, a width of 4 mm, and a length of 50 mm was used in flexural tests. In addition, a dog-bone with a thickness of 4 mm, a width of 10 mm, and a length of 60 mm was used in tension tests.

Processing Processes

The processing window is an effective tool for finding the range of processing parameters for the manufacturing of high quality parts. In this experiment, the injection pressure and the mold temperature were significant processing parameters. The melt temperature is a minor factor in the production of microparts when compared to the other two parameters [19-20]. Thus, we kept the melt temperature at a constant value of 210[degrees]C to find the processing window. The injection pressure and mold temperature were also varied point-by-point to find the processing window. Short-shot or flash in parts was located outside the boundary of the processing window. The permissible injection pressure, called the mechanical ability, could not exceed 2,300 bars in this window. All the tests were conducted in an Arburg injection machine (25 tons, 220S).

Quality Evaluation

To examine the shrinkage and accuracy of a polymer with added fillers, one mold with four different microparts was tested in this study. They were measured with a coordinate measurement machine (CMM; Poly, Italy). The precision level of the measurements was 1 micron. The shrinkage ([eta]) could be calculated in terms of the average mold cavity diameter d and average part diameter d' as follows:

d = ([d.sub.1] + [d.sub.2] + [d.sub.3] + [d.sub.4])/4,

d' = ([d'.sub.1] + [d'.sub.2] + [d'.sub.3] + [d'.sub.4])/4,

[eta] = [[d - d']/d] X 100%, (1)

where the outer diameters of the measured cavities, [d.sub.1], [d.sub.2], [d.sub.3], and [d.sub.4], were 1.509, 1.510, 1.509, and 1.509 mm, respectively. The outer diameters of the measured parts were denoted as [d'.sub.1], [d'.sub.2], [d'.sub.3], and [d'.sub.4], respectively. The average shrinkage of the microparts was determined based on 10 specimens. The reproducibility of the microparts was evaluated as a standard of accuracy. Ten measured microparts were selected from the 11th to the 20th mold. That is, the first 10 pieces were discarded because their quality was unstable. The allowable deviation in the precision of microparts on the market is normally 20 [micro]m. Thus, the level of accuracy achieved here was considered acceptable when the deviation in the precision of the microparts was less than 20 [micro]m.

Wear tests were carried out by using a pin-on-disk tester (Micro phonics, USA). All the testing conditions followed ASTM G99-04. The surface roughness of the disks, which were 120 mm in diameter, was 0.064, 0.1, 0.175, and 0.2 [micro]m (Ra), respectively. Round specimens were placed in contact with the disk at a position 37 mm from the center and rotated from 2,500 (0.58 km) to 30,000 revolutions (6.97 km) with a step increase of 2,500 revolutions at a constant speed of 70 rpm. The vertical load on the pin was 0.75 kg. The specimens were cleaned and dried after they were rotated. Then, the weight of each specimen was measured using an electronic weighing machine (Honeywell, UK). The weight loss Wl) was then calculated as follows:

[FIGURE 2 OMITTED]

Wl = [[W - W']/W] X 100%, (2)

where W is the weight before wearing and W' is the weight after wearing. To achieve high quality, low weight loss is required. The flexural (ASTM D638) and tension (ASTM D790) tests were conducted using a universal testing machine. The average strength was obtained from 10 specimens for every tested part.

RESULTS AND DISCUSSIONS

Filling Behavior of Microparts

The processing window of the microparts is shown in Fig. 2. The formability of a polymer compound is better when the processing window is larger. The quality of microparts is degraded when the processing parameters are outside the range of the window. Short-shot is produced in parts when processing exceeds the left-hand boundary of the window. Flash occurs in parts when processing exceeds the right-hand boundary of the window. In addition, the mechanical ability of the injection machine is exceeded when processing exceeds the top boundary of the window.

Photographs of polymer compound used in one mold to make four different microparts using the above processing window are shown in Fig. 3. It can be observed from the photographs that the microparts were well formed, with clear structural definition. The black microparts were compounded with 20 wt% added glass beads, while the white microparts were compounded with 25 wt% added glass beads. The differently colored parts were easily classified. The average part weights of the microgear, annular triangular plate, annular circular plate with a square hole, and annular circular plate with a donut hole were 2.5, 1.6, 2.0, and 1.9 mg, respectively.

The processing zone shifted to the right when the filler content increased. In other words, the injection pressure and mold temperature had to be increased as the filler content increased. Short-shot parts were formed when the forming pressure and mold temperature were low. The processing window of the polymer with 5 wt% added glass beads was close to that of pure POM.

Based on the processing window, in our experiments, the optimum parameters, which were near the center of the window, for using one mold to make four microparts were an injection pressure of 1,900 bars, a mold temperature of 90[degrees]C, and a melt temperature of 210[degrees]C.

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

Polymer Shrinkage With Varying Amounts of Added Reinforced Fillers

The shrinkage results for the micropolymer compound with various added glass bead contents are shown in Fig. 4. The shrinkage was significantly reduced when the weight content of the filler was increased. Polymer with 25 wt% content exhibited the least shrinkage. The shrinkage of polymer with 25 wt% glass bead content decreased by 1.5% compared to that of pure POM. This clearly shows that adding glass particles to the polymer effectively reduced shrinkage. When the polymer was in the melting stage, glass particles reduced the activity of the molecules, resulting in significantly lower volume expansion of the molecules. Then, crystallization of the molecules was retarded by particles during the frozen stage. Thus, shrinkage was reduced when glass particles were added to the polymer. The greater the particle content, the less shrinkage occurred.

However, large error deviations occurred in the polymer compound with 20 wt% and 25 wt% added glass bead content. A possible reason could be that the glass particles were not uniformly distributed in the polymer compound. In order to find the reason for this result, scanning electron microscopy (SEM; JSM-6500F, Japan) images of the polymer compound with 5 wt% and 20 wt% added glass bead content were taken and are shown in Fig. 5. In the case of POM with 5 wt% added glass bead content, the particles were uniformly distributed in the polymer compound, as shown in Fig. 5a. The error deviations among the 10 specimens were small because the glass particles were uniformly distributed in the polymer compound. However, the POM with 20 wt% added glass bead content revealed serious agglomeration in the polymer compound, as shown in Fig. 5b. Large error deviations among the 10 specimens thus occurred because the glass particles were not uniformly distributed in the polymer compound. These results indicate that uniform distribution of fillers in polymers is very important for the production of microreinforced composites. The same conditions occurred in the tensile and flexural tests, which will be discussed below.

Reproducibility in One Mold With Four Microparts

The reproducibility of microparts is normally evaluated as a standard of accuracy. As shown in Table 1, the total average deviation of the polymer compound with 20 wt% added glass beads was about 13 [micro]m. This result shows that the reproducibility of the microparts was outstanding because the achieved precision level was less than 20 [micro]m. That is, one mold with different microgeometries could be successfully applied in microinjection molding.

[FIGURE 5 OMITTED]

The average deviation between the designed and actual dimensions of the microgears, annular triangular plates, annular circular plates with square holes, and annular circular plates with donut holes was 18, 14, 9, and 11 [micro]m, respectively. The gear exhibited the largest deviation in precision, and the annular circular plate with a square hole exhibited the least deviation. That is, the reproducibility of the circular plate with a square hole was the best, while the reproducibility of the gear was the worst in this study. Here, the annular circular plate with a symmetric square hole could effectively resist distortion, resulting in less deviation. Plastic engineers favor using this shape to design parts, thus reducing the weight and increasing accuracy.

The average deviations of four different geometries for the first four molds were 17, 17, 21, and 14 [micro]m, respectively, higher than the total average deviation of 13 [micro]m. However, the average deviations were stayed between 6 and 12 [micro]m after the fifth mold was produced. In fact, the measured samples were chosen from the 11th to the 20th pieces because the previous pieces were low in quality. Fault parts were often produced at the beginning due to unstable mold and melt temperatures. That is, high quality parts could only be fabricated when the mold and melt temperatures were stable.

Wear Behavior of Polymer Compound With Added Fillers

The weight loss of the POM polymer with different weight contents of added glass beads is shown in Fig. 6. The surface roughness of the wear disk was 0.175 [micro]m. The weight loss decreased significantly when glass beads were added to the polymer compound. The greater the particle content, the lower the abrasion due to wear. The weight loss reached 0.09% for polymer compound with 25 wt% added glass particles after the specimen was rotated about 7 km. In addition, the weight loss of polymer with 25 wt% added glass bead content decreased by 0.53% compared to that of pure POM. This clearly shows that adding glass particles to the polymer effectively increased wear resistance. Figure 6 also shows that the wear loss increased steadily as the rotation distance increased for every tested specimen. The results strongly show that POM polymer with added glass beads is a suitable composite material for producing wear-resistant parts.

[FIGURE 6 OMITTED]

Meanwhile, the difference in weight loss between pure POM and POM with 5 wt% glass beads was large, compared with that between POM with 5 wt% and 20 wt% added glass beads. During wearing, the polymer in the compound was stripped, leaving only the glass beads, which were much harder than the polymer, to resist wear abrasion. After the glass beads were worn down, the next layer of polymer compound continued to wear. Thus, the polymer compound could significantly increase the achieved wear resistance although it had only 5 wt% added glass bead content. This clearly shows that only a small amount of filler needs to be added to the polymer compound to increase wear resistance significantly.

SEM images of pure POM and POM with 20 wt% added glass beads following wearing are shown in Fig. 7a and b. The worn surfaces of the pure POM were seriously damaged because there were no fillers to resist wear abrasion. On the other hand, the worn surfaces showed slight damage in the polymer with 20 wt% added glass particles. These results showed that the glass fillers effectively resisted wear abrasion.

POM polymer with 20 wt% added glass beads is widely used. Thus, POM polymer with 20 wt% added glass beads was selected for wear tests on disks with varying degrees of surface roughness. The weight loss of POM polymer with 20 wt% added glass beads following rotation on disks with varying degrees of roughness is shown in Fig. 8. The weight loss of the polymer compound decreased significantly when the surface roughness of the disk decreased. The lower the surface roughness, the less wear abrasion occurred. The weight loss of the polymer compound reached 0.24% after the specimen was rotated on a 0.2 [micro]m surface for about 7 km. However, the weight loss of the polymer compound was only 0.05% after the specimen was rotated on a 0.064 [micro]m surface for about 7 km.

In addition, 15W40 oil (Aral, Germany) was used to lubricate 0.2 and 0.064 [micro]m disk surfaces. Figure 8 reveals that the weight loss of the polymer compound was effectively decreased when lubrication oil was added to a 0.2 [micro]m rough surface. Furthermore, the weight loss of the polymer compound decreased slightly when lubrication oil was applied to a 0.064 [micro]m bright surface. Thus, lubrication oil was not needed when the polymer compound was sliding on a bright surface.

Mechanical Properties of Polymer Compound With Added Fillers

The mechanical strength of POM polymer with varying weight contents of added glass beads is shown in Figs. 9 and 10. The tensile and compressive strengths increased linearly when the glass filler content was increased in the polymer. The results agree with the rule of mixtures [21], which says that the mechanical properties of polymer compounds are proportional to the weight fractions of the polymer and filler. The tensile strengths of the polymer compound with 20 wt% and 25 wt% added glass beads were 17.3% and 24.3%, respectively, higher than that of pure POM. The compressive strength of the polymer compound with 20 wt% and 25 wt% added glass beads increased significantly by 57.6% and 73.4%, respectively, more than the increase that occurred in the case of pure POM. These results show that the mechanical properties of the POM polymer can be effectively improved by adding glass fillers.

[FIGURE 7 OMITTED]

On the other hand, large deviations in strength occurred in polymer with 20 wt% and 25 wt% added glass beads. The glass particles agglomerated in the polymer with 20 wt% and 25 wt% added fillers, as discussed in sec. 3.2, resulting in large error deviations. Distributing fillers uniformly in the polymer is the best way to decrease such a large deviation in strength.

[FIGURE 8 OMITTED]

CONCLUSION

Four different microparts in one mold could be successfully manufactured using the processing window for micro-injection molding. High pressure and adequate mold temperatures were needed to produce high quality microparts. Shrinkage was significantly reduced when the filler weight content was increased. The total average deviation between the designed and actual dimensions of the microparts was only 13 [micro]m. The results showed that the reproducibility of the microparts produced using microinjection molding was outstanding.

[FIGURE 9 OMITTED]

[FIGURE 10 OMITTED]

The weight loss decreased significantly when more glass beads were added to the polymer compound. The greater the particle content, the lower the wear abrasion. In addition, the weight loss of the polymer compound decreased significantly when a bright contact surface was used. Lubrication oil was not needed when the polymer compound was sliding on a bright surface.

The tensile and compressive strengths of the polymer compound increased linearly when the amount of glass filler added to the polymer was increased. Furthermore, the compressive strength of the polymer compound with 20 wt% and 25 wt% added glass beads increased significantly by 57.6% and 73.4%, respectively, more than the increase that occurred in the case of pure POM. Larger amounts of added glass beads agglomerated easily and were not uniformly dispersed in the polymer, resulting in large shrinkage and strength deviations. Thus, distributing glass particles uniformly in polymer is important for producing high quality parts.

REFERENCES

1. W. Michaeli, A. Spennemann, and R Gartner, Microsyst. Technol., 8, 55 (2002).

2. S. Hill, Micromolding--a small injection of technology, Mater. World, June, 24 (2001).

3. W. Michaeli, A. Rogalla, and C. Ziegmann, Kunststoffe, 89, 80 (1999).

4. C. Kukla, H. Loibl, and H. Detter, Kunststoffe Plast. Eur., 88, 1331 (1998).

5. O. Larsson, O. Ohman, A. Billman, L. Lundbladh, and C. Lindell, Proc. 1997 Int. Conf. Solid-State Sensors Actuators, Chicago, 1415 (1997).

6. R. Bourdon and W. Schneider, Kunststoffe Plast. Eur., 88, 1791 (1998).

7. M.S. Despa, K.W. Kelly, and J.R. Collier, Microsyst. Technol., 6, 660 (1999).

8. J. Fasset, Plast. Eng., 51, 35 (1995).

9. L. Merze, S. Rath, V. Piotter, R. Ruprecht, J. Ritzhaupt-Kleiss, and J. Hausselt, Microsyst. Technol., 8, 129 (2002).

10. T. Gietzelt, V. Piotter, O. Jacobi, R. Ruprecht, and J. Hausselt, Adv. Eng. Mater., 5(3), 139 (2003).

11. B. Su, T.W. Button, A. Schneider, L. Singleton, and P. Prewett, Microsyst. Technol., 8, 359 (2002).

12. A. Rota, T.V. Duong, and T. Hartwig, Microsyst. Technol., 8, 323 (2003).

13. V. Piotter, T. Gietzelt, and L. Merze, Sadhana , 28, 299 (2003).

14. W.C.J. Wei, R.Y. Wu, and S.J. Ho, J. Eur. Ceram. Soc., 20, 67 (2000).

15. M.C. Kuo, C.M. Tsai, J.C. Huang, and M. Chen, Mater. Chem, Phys., 90, 185 (2005).

16. J.S. Song and J.R.G. Evans, Ceram. Int., 21, 325 (1995).

17. P. Hagmann and W. Ehrfeld, Int. Polym. Process., 4, 3 (1989).

18. L. Yu, C.G. Koh, L.J. Lee, and K.W. Koelling, Polym. Eng. Sci., 8(5), 871 (2001).

19. S.Y. Yang, S.C. Nian, and I.C. Sun, Int. Polym, Process., 17, 355 (2002).

20. H. Eberle, Kunststoff Plast, Eur., 88, 1344 (1998).

21. B.D. Agarwal and L.J. Broutman, Analysis and Performance of Fiber Composites, 2nd. ed., Wiley Interscience (1990).

C.K. Huang, S.W. Chen, C.T. Yang

Department of Mechanical Engineering, Lunghwa University of Science and Technology, 300 Wan-Shou Rd., Sec. 1, Kueishan, Taoyuan, Taiwan, R.O.C.

Correspondence to: C.K. Huang; e-mail: [e-mail]

Contract grant sponsor: National Science Council of Taiwan, ROC; contract grant number: NSC 93-2622-E-262-004-CC3.

TABLE 1. Reproducibility of four microparts in one mold with POM and 20
wt% added glass beads.

                                      Annular           Circular plate
Geometry mold                         triangular plate  with square
no.                 Gear ([D.sub.1])  ([D.sub.2])       hole ([D.sub.3])

 1                   1.477 (-0.023)    1.482 (-0.018)    1.486 (-0.014)
 2                   1.476 (-0.024)    1.483 (-0.017)    1.487 (-0.013)
 3                   1.473 (-0.027)    1.477 (-0.023)    1.484 (-0.016)
 4                   1.478 (-0.022)    1.484 (-0.016)    1.497 (-0.003)
 5                   1.483 (-0.017)    1.484 (-0.016)    1.497 (-0.003)
 6                   1.486 (-0.014)    1.482 (-0.018)    1.497 (-0.003)
 7                   1.483 (-0.017)    1.496 (-0.004)    1.492 (-0.008)
 8                   1.486 (-0.014)    1.490 (-0.010)    1.491 (-0.009)
 9                   1.491 (-0.009)    1.491 (-0.009)    1.496 (-0.004)
10                   1.483 (-0.017)    1.491 (-0.009)    1.487 (-0.013)
Avg. deviation      -0.018            -0.014            -0.009
  (same cavity)
Total avg.                                              -0.0131
  deviation
Avg. shrinkage (b)   1.82%             1.59%             1.17%
Total avg.                                               1.48%
  shrinkage (b)

                    Circular plate    Avg. deviation (a)
Geometry mold       with donut        (different
no.                 hole ([D.sub.4])  cavities)

 1                   1.486 (-0.014)   -0.017
 2                   1.485 (-0.015)   -0.017
 3                   1.481 (-0.019)   -0.021
 4                   1.486 (-0.014)   -0.014
 5                   1.488 (-0.012)   -0.012
 6                   1.486 (-0.014)   -0.012
 7                   1.491 (-0.009)   -0.010
 8                   1.497 (-0.003)   -0.010
 9                   1.499 (-0.001)   -0.006
10                   1.487 (-0.013)   -0.011
Avg. deviation      -0.011
  (same cavity)
Total avg.
  deviation
Avg. shrinkage (b)   1.35%
Total avg.
  shrinkage (b)

Designed dimension: [D.sub.1] = [D.sub.2] = [D.sub.3] = [D.sub.4] = 1.5
mm.
(a) Deviation between the designed and actual dimensions of the
microparts, in millimeters.
(b) Shrinkage ([eta]) = [[d - d']/d] X 100%, where d and d' are the
outer diameter of the measured cavity and part, respectively.