Injection molding elastomeric parts with microstructured surfaces.
The low-volume manufacturing required in exploratory research at the early stages of product development has made casting the preferred method for the production of microstructures. This process is low cost and provides easy handling, but is time consuming and not optimal for scaling to high volume manufacturing. Injection molding, a highly-automated process with high production rates and low waste, is widely used in high rate manufacturing. The process also allows for greater material choices and more design flexibility. The key factors in injection molding of microstructured elastomeric surfaces are the machines, molds and the material behavior during processing.
Injection molding machines
Injection molding parts with microstructured surfaces usually requires faster injection rates, high pressures, tighter control of shot size (i.e., material forced into the mold) and better control systems. These improvements have been made in injection molding machines of all sizes to accommodate parts with thinner walls and smaller material volumes, as well as to reduce cycle times. With thinner-walled parts, injection velocities and pressures have increased to 1,500 mm/second (-60 inches/ second) and 3,400 bar (-50,000 psi), respectively. Response times of 2-50 ms have provided better control of shot size and more uniform flow of the polymer during injection. In microinjection molding machines, use of two-stage screw-plunger and plunger-plunger injection which separate plastication and injection, as well as shut offs at the end of the plastication and plunger units, allow even tighter control of shot size. As a result, injection molding is also one of the leading processes for the production of microstructures from thermoplastic materials.
Molding usually employs standard steel mold bases. The microstructured surfaces are not machined directly into the mold plates, but rather incorporated as inserts. In the mold design shown in figure 1, the insert is mounted in a steel cartridge which is bolted into the B-plate of the mold. This configuration permits easy changing of the microstructured insert and use of inserts made from nickel and other materials. The microstructured inserts, however, lack venting. Therefore, the molds are vacuum vented (i.e., a vacuum pump evacuates air from the cavity) between mold close and the start of injection. The vacuum port and o-ring (which provide a seal at the perimeter of the mold plates) are shown in figure la. As illustrated in figure 2, vacuum venting improves the definition (overall shape), but typically not the height or depth of the microfeatures molded from thermoplastics (ref. 17) and thermoplastic elastomers.
Tooling inserts with micro and nanoscale features are made from steel, nickel, and in some cases silicon. With steel and other metals, micromachined features as small as 50 [micro]m are available commercially; feature sizes as low as 10 [micro]m have been created with new machining techniques (ref. 18). Stainless steel and nickel alloys have been etched to create features as small as 0.02 [micro]m (ref. 19). With machined and etched metal surfaces, metal grain size and polishing determine the roughness of the tooling surface. Electroformed nickel tooling, created from lithography-based masters, is widely used and can achieve features with a resolution of 1 [micro]m and aspect ratio of 2.5. Silicon tooling with features fabricated using lithographic techniques has been used to achieve nanoscale features, but the tooling is fragile. Electroformed nickel tooling endures about 100,000 molding cycles, whereas silicon only endures about 3,000.
There are two different types of tooling inserts: positive tooling or negative tooling. As shown in figure 3a, the projections in positive tooling produce depressions or negative features in the molded surfaces. In contrast, the depressions in negative tooling create projections or positive features in molded surface parts (figure 3b). The former surfaces can be used for the production of microfluidic devices, whereas surfaces with positive features have potential as hydrophobic surfaces and adhesives. Achieving full feature depth is easier with positive tooling. For example, a copolyester and two polyurethane thermoplastic elastomers exhibited 83% replication of the feature depth with positive tooling, but only 65% replication with equivalent negative tooling. This difference is due to the flow patterns. With positive tooling, melt flows around the features, but with negative tooling, the molten material must flow into narrow channels where filling is hindered by greater pressure drops in the narrow features and high applied pressures which reduce the mobility of the polymer melt. As expected, replication is more difficult with smaller and deeper (higher aspect ratio) features. With thermoplastic materials, features with aspect ratios (i.e., ratio of feature height to feature width) of about 10:1 can be replicated when the tool is drafted appropriately (ref. 20). For elastomeric materials, draft angles are not as critical due to the flexibility of the material; however, the molding of features with aspect ratios higher than 2:1 has not been reported.
When molding parts with microstructured surfaces, the parts typically have a primary wall thickness of 0.5-3 mm (microparts can have thinner walls). The micro and nanoscale features are located on the surface of this primary wall (as shown in figure 4a). Due to the differences in thickness (and required fill pressures), mold filling is a two-stage process; melt fills the primary wall first and then flows into the micro and nanoscale features. This filling pattern is affected by gate location. Parallel flow is obtained when the gate is located perpendicular to the features (figure 4b). This flow pattern permits some filling of the features during injection of the polymer melt into the cavity. As a result, filling of the features increases with greater injection velocities, but usually is completed during packing of the part. Although melt flows directly into the microstructures with parallel flow, the gate vestige can interfere with the performance of the device. With impingement flow, however, the gate is located on the part edge or away from the features (as illustrated in figure 4c). This flow pattern causes filling of the primary wall during injection and filling of the features during the packing stage of the injection molding cycle. Hesitation of the melt at the entrance to micro features can produce a layer of partially-cooled melt which must be stretched during feature filling. In general, impingement flow is the preferred method because it provides a more uniform and better filling of microstructures. Regardless of gate location, the flow of melt across positive tooling features can influence the replication of those features. For example, the 5-50 [micro]m high projections that produce microfluidic channels can block or direct flow of the melt (ref. 21).
Issues that occur when injection molding microstructured surfaces are (1) undesirable adhesion between polymer melt and tooling insert and (2) high friction between the solidified polymer and the tooling insert. This adhesion hinders the smooth slipping (flowing) of the polymer melt into the microstructures and reduces replication of microstructures. High friction hinders part ejection and can cause stretching of elastomeric features and tearing of elastomeric surfaces (figures 5a5b). The use of lubricants and mold release agents, however, contaminates the tooling and part surfaces (figure 5c). Therefore, tooling inserts with microstructured surfaces usually are coated with an anti-stiction coating. Examples include monolayers of short-chain silanes, such as perfluorodecyltrichlorosilane, perfluorooctyl-trichlorosilane (refs. 22-24), and n-octadecyltri-chlorosilane (ref. 25), which adhere to the tool surface rather than the polymer. These coatings are prepared by a vapor self-assembled monolayer technique which provides relatively thin layers when compared to the microstructures. The antistiction coatings lower the surface energy of the tooling, reducing wettability of the polymer material with the tooling insert and friction between the feature walls and the polymer when ejected; this action improves filling and ejection of the microstructures (ref. 26).
Material behavior during processing
When injection molding parts with microstructured surfaces, thermoplastic elastomers do not behave in the same manner as thermoplastics. With thermoplastics, melt viscosity, solidification time and, depending on mold design, injection velocity or packing pressure affect replication of the microfeatures. Lower melt viscosities provide better replication, as indicated by the greater depth ratios (i.e., ratio of the vertical dimension of molded feature to the corresponding dimension in the tooling insert) and feature definition (shape) (ref. 26). Melt viscosity is controlled by selection of high flow resins and use of high melt temperatures during molding. Due to the small size of the features, retarding solidification of the melt is more critical than melt viscosity when molding microstructured surfaces. Enhanced replication is observed when the mold temperature is kept above the bulk glass transition temperature of amorphous and semicrystalline polymers; for semicrystalline plastics, however, the effect is less pronounced because of greater shrinkage. Replication can be improved with high mold temperatures, but these temperatures must be about 15[degrees]C below the material's glass transition temperature to prevent surface defects; i.e., depressed glass transition temperature at the mold wall causes smearing of melt along the microfeatures. Mounting insulating materials below rapidly cooling tooling inserts, such as electroformed nickel, retards cooling. Both high mold temperatures and insulating layers, however, produce unwanted increases in cycle time. The use of high mold temperatures during filling and packing and lower mold temperatures during cooling provide good replication and acceptable cycle times. Variable mold temperature is achieved with two sets of coolant lines in molds, heating of mold surfaces prior to injection, and mounting of heating inserts under the tooling insert. Overall, depth ratio has been shown to increase linearly with higher melt and mold temperatures (ref. 26).
Unlike thermoplastics, the replication of microstructured surfaces from both block copolymer thermoplastic elastomers and thermoplastic vulcanizates does not depend on melt viscosity and solidification time. Thermoplastic elastomers with large variations in melt viscosity have shown similar replication and similar optimum molding parameters (ref. 29). The lack of dependence on solidification time, however, is not unexpected because the softer phases of thermoplastic elastomers with thenlow transition temperatures remain mobile long after the part is ejected from the mold. In fact, the mobility of the soft and hard segments (phases) of the thermoplastic elastomers seems to control their processing conditions.
Increasing melt temperatures do not improve the replication of microstructured block copolymer thermoplastic elastomer and thermoplastic vulcanizate surfaces. Thermoplastic elastomers, including copolyester block copolymers (COPEs), polyetheramide block copolymers (PEBA), thermoplastic polyurethanes (TPUs) and thermoplastic vulcanizates (TPVs), also have narrow melt temperature windows in which microstructured parts can be molded. With too-low melt temperatures, feature replication is incomplete. When the melt temperatures are too high, the parts showed excessive shrinkage (which produces sink marks) and sometimes degradation of the melt. As with thermoplastics, melt temperatures above the thermoplastic elastomers' recommended melt temperatures were required to achieve replication of the features. This condition created less than 20[degrees]C-wide melt temperature windows for aromatic polyether-based thermoplastic polyurethanes (ref. 28) and polypropylene-based thermoplastic vulcanizates (ref. 23), whereas COPEs and PEBAs had wider temperature windows (ref. 27). Block copolymer thermoplastic elastomers with greater soft segment content, however, required melt temperatures that were 40-60[degrees]C lower than the temperature required for similar materials with greater hard segment content; these temperatures prevented expansion of features after part ejection (ref. 27).
Mold temperature selection requires balancing filling, stretching and shrinkage of the molded microfeatures (ref. 28). Excessively-low mold temperatures produce incomplete filling, whereas overly-high mold temperatures permit surface defects like tearing, sink marks in the parts and stretching of the features during part ejection. Stretching is caused by adhesion of the melt to tooling surface, and with thermoplastic elastomers can occur with an anti-stiction coating on the tooling insert. Within the mold temperature window that provides acceptable parts, increasing the mold temperature does not improve feature replication. This mold temperature range is also narrower than observed with thermoplastics, and typically greater than the manufacturer's recommended temperature range. Aromatic polyetherbased thermoplastic polyurethanes and PEBAs showed 1 (ECwide mold temperature windows, due to stretching of the TPU features (refs. 28 and 29) and tearing of the PEBA surfaces (ref. 25). COPEs and TPVs provided 20[degrees]C-wide mold temperature windows. Adhesion and stretching were also affected by feature size, with deeper features exhibiting greater stiction and stretching (ref. 28).
As discussed earlier, the direction of melt flow will dictate the significance of packing pressure and injection velocity. The injection velocity or packing pressure, however, must be sufficiently high to fill the features prior to solidification of the melt (ref. 24). High packing pressure improves replication with parallel flow, whereas for impingement flow, replication is improved with faster injection velocities. Even with parallel flow of thermoplastics, the high injection velocity must fill the primary wall rapidly enough to prevent cooling of the melt surface and poor replication of subsequent features.
When increasing injection velocity and packing pressure, thermoplastic elastomers show similar trends, but different magnitudes of these effects. Although direct filling of the TPU microfeatures exhibited a linear increase in depth ratio with increasing injection velocity, the increased depth ratio was not as pronounced as for thermoplastics (ref. 29). Packing pressures required for COPEs, PEBAs and TPVs were greater than those required for thermoplastics with similar melt viscosities (refs. 23 and 27). With polypropylene-based TPVs, higher pack pressures provided more consistent replication of the features (ref. 23). Excessively high pack pressures, however, cause softer thermoplastic elastomeric features to expand when the parts are removed from the mold (ref. 27).
When molding with constant mold temperatures, the cooling times for the block copolymer thermoplastic elastomers are longer than for thermoplastics. Cooling time also increases for softer grades due to their lower modulus. Thennoplastic vulcanizates, however, required twice the cooling time (e.g., 40 seconds instead of 20 seconds) of block copolymer thermoplastic elastomers; with shorter times, the microfeatures were not fully formed (ref. 23). The longer cooling times seemed to be related to relaxation of the rubber domains in the thermoplastic vulcanizates (ref. 23).
With the block copolymer, thermoplastic elastomers such as COPEs and PEBAs microfeature replication, specifically depth ratios, correlated with hard segment content. When molding 20-80 [micro]m wide microchannels from positive tooling, elastomers with moderate hard segment content (durometer D hardness of 40-45) permitted wider processing windows and better overall feature replication (ref. 27). COPEs and PEBAs with greater hard segment content (durometer D hardness of 60-65) exhibited some of the molding characteristics of thermoplastics, like the greater shrinkage associated with semi-crystalline thermoplastics, but required the high packing pressures of elastomers (ref. 27). As a result, they exhibited lower depth ratios and poorer feature definition. The surfaces also had more defects, with PEBAs having more surface tearing defects than COPEs (ref. 27). Harder TPUs also showed greater adhesion to the tooling and greater stretching of the features (ref. 28 and 29). Microfeatures molded from COPEs, PEBAs and TPUs with lower hard segment content (durometer D hardness ~ 30) were prone to feature deformation due to polymer relaxation after the parts were removed from the mold (refs. 27-29). In contrast, a COPE with a lower hard segment content provided the best replication, both in greater depth ratios and better feature definition (shape) for 0.10-50 [micro]m wide "pillars" molded from positive and negative tooling; this performance was attributed to the material's lower Vicat softening temperature which indicated higher flexibility and mobility of the material for longer times (ref. 30).
Thermoplastic vulcanizates (TPVs) behave differently than block copolymer thermoplastic elastomers. There was no correlation between feature filling and melt viscosity or material hardness (ref. 23). High temperatures, high pack pressures and longer cooling times were required for complete replication of 20 [micro]m diameter, 20 [micro]m deep pillars from negative tooling (ref. 23). This behavior suggested that the rubber domains in the TPVs, which are much larger than the domains in the block copolymer thermoplastic elastomers, were deformed during feature filling and relaxed to complete replication of the feature height. Feature diameters were constant, possibly due to the modulus of the polypropylene matrix. Increasing melt temperature and cooling time allowed for greater relaxation of the EPDM rubber domains (ref. 23). For a TPV with a greater rubber content and low molecular weight additives (durometer D hardness of 14), however, the polypropylene matrix could not always contain the relaxed rubber domains. With incomplete filling of features, the EPDM domains projected from the features, resulting in bumpy surfaces (figure 6a); when the features were completely filled, the surfaces were smooth (figure 6b). Similar relaxation and feature deformation has been reported for rubber-containing thermoplastics like impact modified polystyrene (HIPS) (ref. 31).
Machines and tooling developed for injection molding of microstructured surfaces from thermoplastics can be employed for injection molding thermoplastic elastomers into patterned surfaces with a wide range of applications. Filling microfeatures with thermoplastic elastomers requires higher than recommended melt and mold temperatures, high injection velocities and pack pressures, and longer cooling times. For block copolymer thennoplastic elastomers, replication of the microfeatures depends on the hard segment content of the materials, whereas no correlation has been observed between the rubber content of thennoplastic vulcanizates and feature replication. Deformation and relaxation of the elastomers, however, is a major factor governing their performance when injection molding parts with microstructured surfaces.
(1.) B. Bhushan and R.A. Sayer, "Surface characterization and friction of a bio-inspired reversible adhesive tape," Microsyst. Technol. 13 (2007), pp. 71-78.
(2.) S.E. Kim, E. Cheung and M. Sitti, "Wet self-cleaning biologically inspired elastomer mushroom shaped microfibrillar adhesives," Langmuir 25, 13(2009), pp. 7,196-7,199.
(3.) B. Aksak, M.P. Murphy and M. Sitti, "Adhesion of biologically inspired vertical and angled polymer microfiber arrays, " Langmuir 23 (2007), pp. 3,322-3,332.
(4.) B. Aksak, K. Sahin and M. Sitti, "The optimal shape of elastomer mushroom-like fibers for high and robust adhesion, " Beilstein J. Nanotechnol. 5 (2014), pp. 630-638.
(5.) B. Aksak, M.P. Murphy and M. Sitti, "Gecko inspired microfibrillar adhesives for wall climbing robots on micro/nanoscale rough surfaces," Robotics and Automation, Pasadena, CA, IEEE, 2008, pp. 3,058-3,063.
(6.) A.K. Geim, S. V. Dubonos, I. V. Grigorieva, K.S. Novoselov, A.A. Zhukov and S. Yu Shapoval, "Microfabricated adhesive mimicking gecko foot-hair, "Nature Materials 2 (2003), pp. 461463.
(7.) H.E. Jeong, J.-K. Lee, H.N. Kim, S.H. Moon and K.Y. Suh, "A nontransferring dry adhesive with hierarchical polymer nanohairs, "Proceedings of the National Academy of Sciences of the United States of America, 2009, pp. 5,639-5,644.
(8.) W.B. KhaledandD. Sameoto, "Fabrication and characterization of thermoplastic elastomer dry adhesives with high strength and low contamination, " Applied Materials & Interfaces, 6 (2014), pp. 6,806-6,815.
(9.) S. Kim and M. Sitti, "Biologically inspired polymer microfibers with spatulate tips as repeatable fibrillar adhesives, " Applied Physics Letters 89 (2006): 261911.
(10.) A. Charlambides and S. Bergbreiter, "A novel all-elastomer MEMS tactile sensor for high dynamic range shear and normal force sensing," J. Micromech. Microeng. 25 (2015): 095009.
(11.) J. Engel, J. Chen, N. Chen, S. Pandya and L. Chang, "Multiwalled carbon nanotube filled conductive elastomers: Materials and application to micro transducers, " Micro Electro Mechanical Systems (2006), pp. 246-249.
(12.) J. Harris, H. Lee, B. Vahidi, C. Tu, D. Cribbs, N.L. Jeon and C. Cotman, "Fabrication of a microfluidic device for the compartmentalization of neuron soma and axons, " Journal of Visualized Experiments 7 (2007): e261.
(13.) J.C. McDonald, J.C. and G.M. Whitesides. "Poly(dimethylsiloxane) as a material for fabricating microfluidic devices," Accounts of Chemical Research 35.7 (2002), pp. 491-499.
(14.) O.J.A. Schueller, X. Zhao, G.M. Whitesides, S.P. Smith and M. Prentiss, "Fabrication of liquid-core waveguides by soft lithography," Advanced Materials 11.1 (1999), pp. 37-41.
(15.) O.J.A. Schueller, D.C. Duffy', J.A. Roger, S.T. Brittain and G.M. Whitesides, "Reconfigurable diffraction gratings based on elastomeric microfluidic devices, " Sensors and Actuators 78 (1998), pp. 149-159.
(16.) D. Huh, Y. Torisawa, G.A. Hamilton, H.J. Kim and D.E. Ingber, "Microengineered physiological biomimicry: Organson-chips," Lab Chip 12 (2012), pp. 2,156-2,164.
(17.) S-H. Yoon, P. Padmanabha, N. Cha, J. Mead and C. Barry, "Evaluation of vacuum venting for micro-injection molding, " International Polymer Processing, 2011/04, pp. 346-353 (2011).
(18.) M. Koc and T. Ozel, Micro-Manufacturing: Design and Manufacturing of Micro-Products, Wiley (2011).
(19.) T.M. Doi, D. Ioan and S. Kurokawa, Advances in CMP/Polishing Technologies for the Manufacture of Electronic Devices--5.5.3 Micro-Machining Using Dry Etching, Elsevier (2012).
(20.) N.S. Ong, H. Zhang and W.H. Woo, Materials and Manufacturing Processes, 21, 824 (2006).
(21.) D.P. Dempsey, N. George, S.H. Yoon, E. Paszkowski, M. McGee, J.L. Mead and C.M.F. Barry, "Flow effects on replication of injection molded microchannels, " Society of Plastics Engineers Annual Technical Conference, 2011.
(22.) M. De Jesus Vega, S. Birkar, J. Mead and C. Barry, "Comparison of microstructuredsurfaces using injection molding and nanoimprint lithography, " ANTEC, Las Vegas, NV, 2014, pp. 1,971-1,975.
(23.) M. De Jesus Vega, S. Birkar, J. Mead and C. Barry, "Injection molding of microstructured surfaces from rubber-containing polymers, " Rubber Chemistry and Technology, Cleveland, OH, 2015.
(24.) S.-H. Yoon, N.-G. Cha, J. Lee, J.-G. Park, D. Carter, J. Mead and C. Barry, "Effect of processing parameters, antistiction coatings and polymer type when injection molding microfeatures, " Polymer Engineering and Science 50.2 (2010), pp. 411419.
(25.) B. Saha, S.B. Tor, L. Erjia, D.E. Hardt and J.H. Chun, "Hotembossing performance of silicon micromold coated with self-assembled n-octadecyltrichlorosilane, " Sensors and Actuators B: Chemical 160 (2011), pp. 207-214.
(26.) R. Thiruvenkataswamy, S.-H. Yoon, J.L. Mead and C.M.F. Barry, "Controlling dimensions when injection molding microfluidic devices, "Proceedings of Annual Technical Conference of Society of Plastics Engineers, 54 (2008).
(27.) Younghyo Kim, Sung-Hwan Yoon, Si-Hyeong Cho, Jin-Goo Park, Joey L. Mead and Carol M.F. Barry, "Effect of thermoplastic elastomer properties in molded microscale features," ANTEC, 2011.
(28.) M.W. Alabran, D. Carter, A.A. Busnaina, J.L. Mead and C.M.F. Barry, "Effects of tooling and injection molding conditions on nanoscale replication of thermoplastic elastomers," SPE ANTEC, 53 (2007), pp. 1,990-1,994.
(29.) S.H. Yoon, M.W. Alabran, J.S. Lee, D.J. Carter, J.L. Mead and C.M.F. Barry, "Micro-injection molding of high aspect ratio features with thermoplastic polyurethanes, " Annual Technical Conference Proceedings, Cincinnati, OH, 2007, pp. 1,9951,999.
(30.) Smita Birkar, Jin-Goo Park, Joey Mead and Carol Barry, "Injection molding of thermoplastic elastomers for microstructured substrates, "Rubber Chemistry and Technology, 87 (4), pp. 629-646 (2014).
(31.) J.W. Rodgers, M.E. Casey, S.S. Jedlicka and J.P Coulter, Journal of Micro and Nano-Manufacturing, 2, 011006-011013 (2014).
by Marisely De Jesus Vega, Joey Mead and Carol Barry, University of Massachusetts Lowell
Table 1--applications of microstructured surfaces from elastomers Application Material Feature shape Gecko adhesives, Polyvinylsilixane Pillars hydrophobic Polyurethane Mushroom surfaces shaped fibers Fibers Polyimide Hairs Polyurethane acrylate Nanohairs Styrene ethylene Mushroom shaped Butylene styrene pillars Tactile sensors PDMS with carbon Pillars material Micro transducers PDMS and Pillars polyurethane with MWCNTs Application Material Diameter (pm) Gecko adhesives, Polyvinylsilixane 50 hydrophobic Polyurethane 4.5 (tip: 9) surfaces 28-57 4 (tip: 8) 4-25 Polyimide 0.2-4 Polyurethane acrylate 0.35 (tip: 0.6) Styrene ethylene Tip: 40 Butylene styrene Tactile sensors PDMS with carbon -- material Micro transducers PDMS and -- polyurethane with MWCNTs Application Material Height (pm) Ref. Gecko adhesives, Polyvinylsilixane 70 1 hydrophobic Polyurethane 20 2 surfaces 114 3 20 4 -- 5 Polyimide 0.15-2 6 Polyurethane acrylate 3 7 Styrene ethylene 20 8 Butylene styrene Tactile sensors PDMS with carbon 300 10 material Micro transducers PDMS and -- 11 polyurethane with MWCNTs
|Printer friendly Cite/link Email Feedback|
|Author:||De Jesus Vega, Marisely; Mead, Joey; Barry, Carol|
|Date:||Dec 1, 2015|
|Previous Article:||Elastomer mixture mainly comprising a diene elastomer coupled by an aminoalkoxysilane group, rubber composition including the same and methods for...|
|Next Article:||Five things everyone should know about micro-molding.|