- Email: [email protected]
Experimental study of polymer microlens fabrication using partial-filling hot embossing technique
Microelectronic Engineering 162 (2016) 57–62
Contents lists available at ScienceDirect
Microelectronic Engineering journal homepage: www.elsevier.com/locate/mee
Experimental study of polymer microlens fabrication using partial-filling hot embossing technique Sean Moore a, Juan Gomez b, Devanda Lek b, Byoung Hee You a,b, Namwon Kim c, In-Hyouk Song a,b,⁎ a b c
Department of Engineering Technology, Texas State University, 601 University Dr., San Marcos, TX 78666-4684, United States Materials Science, Engineering, and Commercialization Program, Texas State University, 601 University Dr., San Marcos, TX 78666-4684, United States Ingram School of Engineering, Texas State University, 601 University Dr., San Marcos, TX 78666-4684, United States
a r t i c l e
i n f o
Article history: Received 18 November 2015 Received in revised form 11 April 2016 Accepted 10 May 2016 Available online 11 May 2016 Keywords: Hot embossing MEMS Partial-filling Polymer microlens Taguchi method
a b s t r a c t A method to fabricate microlens arrays in polymer substrates via hot embossing is presented in this paper. The partial-filling hot embossing technique of micro cavities on a mold insert through surface tension and capillary action is proven to be an effective means of limiting imperfections on the surface of microlens arrays. The effects of the investigated process parameters including temperature, embossing pressure, and holding time are analyzed via Taguchi method to identify effective processing conditions for microlens arrays of varying heights and diameters. Signal-to-noise (S/N) ratios are calculated for the focal length of the fabricated microlens arrays to identify key individual parameters and their interactions for a streamlined fabrication process. Experimental data indicates that the holding time in the embossing process has the most significant impact on lens focal length followed by embossing temperature and pressure. This study identifies a reliable means of microlens production and demonstrates the effects of varying process parameters in the partial-filling method of micro hot embossing for the production of lens arrays. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Microlenses and microlens arrays have been used in the biomedical, optical communication, and optoelectronics fields [1–5]. Used in fiber optic applications, the addition of microlens arrays at the objective and relay ends of a fiber bundle increases the fill factor of the bundle, as well as improves its coupling efficiency [2]. Microlenses have also been commonly employed in endoscopic applications for tissue and cellular imaging [1,5]. In response to growing demand, a variety of microlens manufacturing methods have been investigated. The most common methods include laser ablation, reflow techniques, ink jet molding, injection molding, and hot embossing [6–9]. While laser ablation techniques allow for later phase fabrication of micro-optics in established devices, they offer no versatility in lens height or diameter due to masking requirements [10]. Polymer reflow and ink jet molding techniques are effective for large aperture lenses. However, the processing parameters required are difficult to control in micro scale production [11]. The thermal and flow induced stresses introduced into small, thin profiles negate the mass production volumes that characterize injection molding processes.
⁎ Corresponding author at: Department of Engineering Technology, Texas State University, 601 University Dr., San Marcos, TX 78666-4684, United States. E-mail address: [email protected] (I.-H. Song).
http://dx.doi.org/10.1016/j.mee.2016.05.009 0167-9317/© 2016 Elsevier B.V. All rights reserved.
In contrast, a hot embossing technique offers advantages in regard to simplicity, operating cost, and replication accuracy for microstructures. The capability to fabricate microlens arrays with hot embossing has been demonstrated with two techniques: partial-cavity filling and full-cavity filling [12,13]. Full-cavity filling uses mold inserts with the desired microlens geometry that facilitate the complete filling of the mold cavity generating microlenses with limited variation. While this method does produce repeatable batch-to-batch outcomes, the polymer makes contact with the mold insert often resulting in surface defects that scatter incident light [14]. The partial-filling technique alleviates this issue by capitalizing on surface tension and capillary action to generate the lens structure thereby minimizing contact with mold surfaces and achieving optical surface characteristics [12]. Partial-filling is initiated with surface tension generated from the mold making contact with the polymer substrate under embossing pressure. The surface tension is a result of the elevated energy levels on the surface of the polymer opposed to its interior. In combination with polymer temperatures above glass transition temperature, this surface tension leads to a capillary action forcing the polymer into the recesses of the mold. The hemispherical convex lens formation of the polymer is a result of a reduction in surface tension as the center of the mold cavity allows the polymer to flow with less restriction [15]. Usable microlens fabrication requires an embossing temperature, an embossing pressure, and a holding time to be applied in the proper configuration. An effective combination of these parameters promotes the
58
S. Moore et al. / Microelectronic Engineering 162 (2016) 57–62
partial filling of the mold cavities in a hemispherical fashion without contacting the mold sidewalls. During embossing, the flow front of the polymer substrate flows into the mold cavity at a higher velocity in comparison to the flow at the sidewalls due to the pressure differential between the hole edge and its center [12]. Furthermore, alteration of process parameters facilitates the filling of the microholes at varying heights resulting in changes in focal length. Thus, microlens arrays of varying focal lengths can be achieved with the adjustment of processing parameters using only one mold insert. Despite these advantages, the partial-filling method is difficult to control with variations from lensto-lens and batch-to batch reproducibility being the primary concerns. In this study, polymer microlens arrays are demonstrated with polymethyl methacrylate (PMMA) substrates, commonly used in hot embossing to form a variety of microstructures [16–18]. To improve replication reliability, the effects of major processing parameters are systematically investigated in this paper. Three processing parameters including an embossing temperature, an embossing pressure, and holding time are considered with three levels arranged in the L9 orthogonal configuration of Taguchi method, a statistics based experimental design, which is beneficial in identifying significant impacts of individual processing parameters in addition to parameter interactions [19]. The focal length of the fabricated microlens arrays is measured and compared with the calculated value using the dimensions of the radius of curvature (ROC), sag height, and diameter to determine the effects of process parameters on the microlens array. 2. Design of microlens mold insert Hot embossing uses a patterned mold insert to transfer microstructures inherent to the mold onto polymer substrates. Effective fabrication of microstructures demands accurate mold patterns that are capable of generating high-resolution replications in polymer substrates. A variety of metallic materials have been implemented in the fabrication of mold inserts including nickel, stainless steel, and brass [20–22]. A brass alloy is selected for this study as it demonstrates the capability to withstand repetitions of high pressure and temperature variations present in hot embossing [14]. A HAAS MiniMill, a vertical CNC milling machine, is used to form microhole arrays in the brass. The design and fabrication of the mold insert is critical for defining the diameter of individual microlenses within an array, which determines the ROC and focal length of the microlenses. Fig. 1(a) shows the design and the dimension of the microhole array on the brass mold insert. A 10 × 10 array of 508 μm (0.02 in.) microholes is arranged in a rectangular configuration with a pitch of 711 μm (0.028 in.) and a 203 μm (0.008 in.) spacing between each hole. Fig. 1(b) is a cross-section view of A–A′ indicated on Fig. 1(a). Optical defects due to machining traces present on the mold insert are avoided using the partial-filling method. As the polymer substrate does not contact the sidewalls or bottom of the microholes during the embossing process, the length of the microholes is insignificant and set at 508 μm, a 1:1 depth to diameter ratio. Fig. 1(c) shows a schematic of the partial-cavity filling hot embossing technique. Fig. 2 illustrates the micro-milled mold insert on brass, (353 engravers brass, McMaster Carr) and a microscopic image of the microholes. Machining tolerances in conjunction with limitations inherent to end mill fabrication resulted in a deviation of hole diameter from mold design to fabrication. The measured diameter of fabricated microholes is 512 μm. The microhole array is placed in the center of the mold insert to mitigate thermal stresses created in the embossing process. 3. Fabrication of microlens array A 3 mm thick PMMA substrate is selected due to its appropriate optical transparency, and mechanical durability. A Carver Thermal Press (3893 4NE18, Carver Inc., Wabash, IN) is employed for the hot
Fig. 1. (a) A schematic representation of the microlens array design. (b) The cross-section view of A–A′ illustrated in panel (a) with a 1:1 depth to width ratio. (c) Illustration of partial-filling hot embossing process.
embossing process. The mold insert is mounted on the upper platen while a secondary brass stage is mounted to the lower platen. The lower stage, polished to mitigate surface roughness, is used to limit the transfer of surface artifacts to the polymer substrate. An oil-based cooling system is implemented to reduce thermal stress and its effects during cooling and demolding and to reduce process cycle time. The cooling system consists of an oil transfer pump, used to circulate a heat transfer fluid throughout the heating and cooling platens paired with a condenser that serves as a heatsink for the thermal fluid. In hot embossing, the polymer substrate is heated to an embossing temperature prior to molding. The glass transition temperature (Tg) of PMMA is 105 °C. The applied embossing pressure forces the heated polymer to flow into the microhole. The hot embossing process
S. Moore et al. / Microelectronic Engineering 162 (2016) 57–62
59
Fig. 3. A graphical representation of the hot embossing process with regard to embossing temperature and pressure is shown.
Fig. 2. Photo and microscopic image of the brass mold insert.
demonstrated in this experiment is divided into four steps including: pre-heating, embossing, demolding, and cooling. The process initiates by heating the upper and lower platens to the prescribed embossing temperature followed by the placement of the polymer substrate on the lower platen. Once the mold insert and the polymer substrate reach the prescribed embossing temperature, the press is closed and the embossing pressure is applied for a period defined as holding time [23]. During the holding time, the polymer substrate flows isotropically, constrained by the mold insert, into the mold cavity [24]. Throughout the embossing phase, the mold insert and the substrate plate continue to move closer to maintain constant embossing pressure. Temperature must remain constant during embossing, as any variations in temperature affect the fluidic resistance of the polymer substrate resulting in pattern defects [25]. Upon completion of the holding time, the molded pattern is separated from the mold insert by the relative displacement of the substrate plate, and cooled below 80 °C.
During the micro hot embossing experiments, three different levels (L1, L2, and L3) of each parameter are systematically applied in combinations dictated by the orthogonal array to identify significant interactions among them. Fig. 3 illustrates the process profile and the application of temperature and pressure as a function of time throughout the fabrication process. The temperature range for this research is set in accordance with the flow behavior of PMMA during embossing. At around 105 °C, PMMA begins to exhibit rubber elastic properties but thermal reflow is absent. Above 130 °C, PMMA is in a viscoelastic state typical of forming complete polymeric structures such as fluidic channels [27,28]. The purpose of this research is to avoid complete cavity filling. Thus, the maximum embossing temperature is set to 130 °C. The thermal press employed for this study exerts a minimum force of 1000 kg. While this force is sufficient to fabricate microlenses, this experiment uses forces of 1500 kg, 2000 kg, and 2500 kg exerting an embossing pressure of 285 psi, 380 psi, and 475 psi, respectively, to avoid any experimental replication issues associated with minimal force programming. Holding time is determined with preliminary parameter selection experiments whose results indicate that the holding time of 2 min under minimized embossing temperatures and pressures reliably produced lens structures. Therefore, symmetric holding times of 2 min, 3.5 min, and 5 min are employed. The process parameters are illustrated in Table 1. Upon completion of the prescribed holding time, the embossing pressure is immediately released and the substrate is allowed to cool under atmospheric conditions. In this manner, the parameter of holding time is isolated from embossing pressure allowing for independent study. The effects of process parameter configurations on the diameter and the height of the microlens arrays are measured and collected to determine the ROC and the focal length for each run of the experiment.
5. Results and discussion 4. Design of experiments A parameter design Taguchi methodology, which utilizes signal-tonoise ratios (S/N ratios) to establish relationships among parameters arranged in an orthogonal array is used in this study. Taguchi parameter and tolerance design not only identifies key parameter contributions in regard to stability, but also reduces the number of experimental runs required. Hot embossing experiments are performed at three symmetric intervals to study the effects of an embossing temperature, an embossing pressure, and a holding time in the partial cavity filling process. As process stability is the principle concern, the goal is to determine what functional relationship exists among factors, their levels, and the stability of the desired output. This research uses the L9 array with the fourth parameter omitted [26].
Experimental trials are conducted randomly with three repetitions. Fig. 4 shows a stereoscopic image of the microlens array fabricated with an embossing temperature of 120 °C, an embossing pressure of 285 psi, and a holding time of 2 min. Fig. 5 displays the 3-D profile image of the microlens scanned using KLA-Tencor P-7 stylus profiler. Table 1 The hot embossing parameters and their respective levels are shown. Process parameter
Level 1
Level 2
Level 3
(A) Embossing temp (°C) (B) Embossing pressure (psi) (C) Holding time (min)
120 285 2
125 380 3.5
130 475 5
60
S. Moore et al. / Microelectronic Engineering 162 (2016) 57–62
Angstroms Inc., Portsmouth, VA) as shown in Fig. 6(a). The ROC, R, of the analyzed lens is calculated by [29] 2
R¼
D2 þ 4h ; 8h
ð1Þ
where D is the diameter, and h is the height of the microlens. For a focal length measurement, the highest point in lens curvature is focused with a microscope (Nikon MM 800, Tokyo, Japan), denoted as position A of
Fig. 4. A stereoscopic image of a fabricated microlens array.
Fig. 5. 3-D profile of microlens measured using KLA-Tencor P-7 stylus profiler. Scanning area of 800 μm × 800 μm.
Ten microlenses are randomly selected and the sampling surfaces at the top of microlenses are scanned with a 20 μm scan length to measure surface roughness of the microlenses. Table 2 illustrates the surface roughness of the arithmetic average, Ra, and the root mean squared average, Rq. The average Ra and Rq are 6.98 ± 0.78 nm and 8.74 ± 0.97 nm, respectively. Each microlens is inspected and the average ROC and focal length are extracted using three individual lenses located along the bottom edge of the microlens array. A side profile photo of each of the measured lenses is captured with a surface tension analyzer (FTA 1000 B Class, First Ten
Table 2 The surface roughness of the fabricated microlenses. Scan #
Ra (nm)
Rq (nm)
1 2 3 4 5 6 7 8 9 10 Average Standard deviation
7.81 6.89 5.85 6.3 6.79 6.82 8.12 7.2 7.9 6.07 6.98 0.78
9.57 8.37 7.31 7.8 8.7 8.57 10.3 9.22 9.77 7.76 8.74 0.97
Fig. 6. (a) The methodology for obtaining lens height and diameter using scaling ratios is demonstrated. D and h are diameter and height of microlens, respectively, to extract the ROC of microlens and the focal length. (b) A schematic representation of lens anatomy and focal length measurement through a microscope. Here, R and f are ROC and focal length of microlens, respectively.
S. Moore et al. / Microelectronic Engineering 162 (2016) 57–62
61
Table 3 The calculated ROC and focal length data relevant to experiment level. Run
A
1 2 3 4 5 6 7 8 9
L1 L1 L1 L2 L2 L2 L3 L3 L3
B
L1 L2 L3 L1 L2 L3 L1 L2 L3
C
L1 L2 L3 L2 L3 L1 L3 L1 L2
ROC (μm)
Calculated focal length (μm)
Measured focal length (μm)
Y1
Y2
Y3
Ave.
Y1
Y2
Y3
Ave.
Y1
Y2
Y3
Ave.
1092 579 640 926 685 1099 675 800 506
989 668 584 948 671 1052 726 808 437
914 702 578 977 708 970 740 823 449
998 650 601 950 688 1040 714 810 464
2229 1181 1306 1889 1399 2242 1378 1633 1032
2019 1364 1191 1935 1347 2146 1482 1650 870
1866 1432 1180 2039 1445 1980 1509 1679 899
2038 1326 1226 1954 1397 2123 1456 1654 934
2353 1304 1336 1946 1465 2557 1428 1737 1119
1997 1431 1227 2029 1372 2292 1498 1706 958
2191 1450 1229 2279 1523 2174 1538 1741 996
2180 1395 1264 2085 1453 2341 1488 1728 1024
and calculated values and exhibits a correlation between the two methods. A “smaller is better” configuration of the S/N ratio is applied to minimize the influence of changing multiple parameters within a run thus highlighting individual factor effects on experimental variation. The focal length is chosen for analysis as it is dependent upon discrete ROCs, sag heights, and diameters. Each S/N ratio is calculated by
S=N ¼ −10 log
Fig. 7. Measured and calculated focal length of the fabricated microlenses.
Fig. 6(b). While the lens remains fixed on the microscope stage, the objective lens of the microscope moves vertically upward to focal point B, where incident light rays converge. The travel distance of the objective lens along the z-axis indicates the focal length of the individual microlens. Using the ROC of the lens, R, and the refractive index, n, of the lens media, the focal length, f, is given by [29] f ¼
R : n−1
ð2Þ
In the case of PMMA, the refractive index is 1.49 for the wavelength of 600 nm. Using the measured heights and diameters, ROC and focal length for each run are calculated and shown in Table 3, using Eq. (1) and Eq. (2). The table reflects the collective average Y1, Y2, and Y3 of three lenses prepared from individual sample arrays generated by the replications of each run. Three sample lenses from each replication are selected, measured, and averaged to obtain each run's respective Y values for ROC, calculated focal length, and measured focal length. These values are collectively averaged to achieve an overall run average as shown for the nine experimental levels. The calculated values are compared with measured focal length values to ensure efficacy of the measurement method. Fig. 7 illustrates the deviation between average measured
! n 1X Y 2i ; m i¼1
ð3Þ
where Yi values are experimental data from Table 3 for each run and m is the number of measurements per run. The S/N ratio range is calculated for each parameter to determine the main effect. A parameter with the highest ranking reflects the greatest effect throughout the experiment. Table 4 shows the parameter S/N ratios and their respective main effect for focal length, calculated using Eq. (3). The calculated main effects of temperature, pressure, and holding time for the microlens focal length are 2.73 dB, 2.51 dB, and 3.13 dB while the measured focal length has main effects of 2.88 dB, 2.32 dB, and 3.40 dB, respectively. Fig. 8 graphically illustrates the main effect of each parameter for the measured focal length of the microlenses. Results indicate the most critical aspect of the hot embossing process for microlens focal length is the holding time. An increase in holding time is directly related to an increase in energy applied to the substrate. The increase in energy facilitates the flow of the heated polymer into the only available expansion region, the microholes. The convex hemispherical shape is formed by capillary action and surface tensions resultant from constraining the polymer flow with the mold insert. Similarly, experimental data indicates that the embossing temperature has an effect on ROC and focal length of the microlens array. The increase in the embossing temperature promoted the flow of the polymer substrate into the mold cavity by relaxing the molecular bonds in the substrate increasing the polymer elasticity and thereby decreasing fluidic resistance. Embossing temperature determines the physical phase of the PMMA substrate to promote cavity filling. An increase in temperature corresponds to an increase in the ability of the substrate to flow into the recesses of the mold insert. Embossing pressure, while still a requirement for the formation of microlens structures, is the least contributing factor to the fabrication process.
Table 4 S/N ratio (dB) and main effect for the focal length of microlenses. Parameter
Embossing temp Embossing pressure Holding time
Calculated focal length
Measured focal length
Level 1
Level 2
Level 3
Main effect
Level 1
Level 2
Level 3
Main effect
−63.49 −65.10 −65.71
−65.09 −63.25 −62.58
−62.36 −62.59 −62.65
2.73 2.51 3.13
−63.91 −65.55 −66.32
−65.69 −63.64 −63.18
−62.81 −63.23 −62.92
2.88 2.32 3.40
62
S. Moore et al. / Microelectronic Engineering 162 (2016) 57–62
Fig. 8. The main effect plot of the S/N ratios for measured focal length of the microlens array.
6. Conclusion The effects of embossing temperature, pressure, and holding time are investigated for the fabrication of microlens arrays in PMMA substrates using the partial-filling hot embossing technique. The partial-filling technique replicated microlens arrays of varying focal lengths and using one mold insert. Although surface imperfections are mitigated due to the reduced contact with the mold walls, variations between the microlens arrays is of principle concern. This research characterizes the effects of process parameters including embossing temperatures, embossing pressures, and holding times on microlens fabrication. A three level, three parameter, L9 orthogonal Taguchi design study shows parameter interactions. Additionally, the most critical parameter in relation to lens ROC and focal length is determined. Results indicate that holding time is the most influential parameter with resonating impact throughout each experimental level. This study provides in-depth information in regard to the appropriate processing conditions for the fabrication of polymer microlenses using the partial-filling technique. References [1] V.K. Shinoj, V. Murukeshan, S.B. Tor, N.H. Loh, S.W. Lye, Design fabrication and characterization of thermoplastic microlenses for fiber-optic probe imaging, Appl. Opt. 53 (2014) 1083–1088, http://dx.doi.org/10.1364/AO.53.001083. [2] X.-T. Yan, J. Yang, X. Bin, X. Ma, F. Li, Y. Zhao, F. Bu, Design of the microlens arrays coupling with imaging fiber bundle, Opt. Lett. 9 (2013) 169–172, http://dx.doi. org/10.1007/s11801-013-3016-4. [3] R. Ramakrishnan, N. Saran, R. Petcavich, Ink jet printing for selective conductor processes for displays and flexible circuits, J. Disp. Technol. 8 (2011) 337–344, http:// dx.doi.org/10.1889/1.3499831. [4] K. Li, C. Feng, H. Choi, Analysis of micro-lens integrated flip-chip INGaN light-emitting diodes by confocal microscopy, Appl. Phys. Lett. 104 (2014) 051107, http://dx. doi.org/10.1063/1.4863925. [5] A. Hassanfiroozi, Y.P. Huang, B. Javidi, H.P. Shieh, Hexagonal liquid crystal lens array for 3D endoscopy, Opt. Express 23 (2015) 971–981, http://dx.doi.org/10.1364/OE. 23.000971. [6] A. Zukauska, M. Malinauskas, C. Reinhardt, B. Chichkov, R. Gadonas, Closely packed hexagonal conical microlens array fabricated by direct laser photopolymerization, Appl. Opt. 51 (2012) 4995–5003, http://dx.doi.org/10.1364/AO.51.004995.
[7] H. Zhang, L. Li, D. McCray, D. Yao, A. Yi, A microlens array on curved substrates by 3D microprojection and reflow process, Sensors Actuat. A- Phys. 179 (2012) 242–250, http://dx.doi.org/10.1016/j.sna.2012.03.002. [8] D. Xie, H. Zhang, X. Shu, J. Xiao, Fabrication of polymer micro-lens array with pneumatically diaphragm-driven drop on demand ink-jet technology, Opt. Express. 20 (2012) 15186–15195, http://dx.doi.org/10.1364/OE.20.015186. [9] P. Xie, P. He, Y.C. Yen, K.J. Kwak, D. Gallego-Perez, L. Chang, W.C. Liao, A. Yi, J. Lee, Rapid hot embossing of polymer microstructures using carbide-bonded graphene coating on silicon stampers, Surface Coat. Tech. 258 (2014) 174–180, http://dx. doi.org/10.1016/j.surfcoat.2014.09.034. [10] K. Naessens, H. Ottevaere, R. Baets, P.V. Daele, H. Thienpont, Direct writing of microlenses in polycarbonate with excimer laser ablation, Appl. Opt. 42 (2003) 6349–6359, http://dx.doi.org/10.1364/AO.42.006349. [11] J. Stringer, B. Derby, Limits to feature size and resolution in ink jet printing, J. Europ. Ceramic Soci. 29 (2009) 913–918, http://dx.doi.org/10.1016/j.jeurceramsoc.2008. 07.016. [12] C.Y. Chang, S.Y. Yang, S.L. Huang, K.H. Hsieh, Fabrication of polymer microlens arrays using capillary forming with a soft mold of micro-holes array and UVcurable polymer, Opt. Express 14 (2006) 6253–6258, http://dx.doi.org/10. 1364/OE.20.015186. [13] K. Li, C. Feng, H.W. Choi, X.C. Shan, R. Maeda, Y. Murakoshi, Micro hot embossing for replication of microstructures, Jpn. J. Appl. Phys. 42 (2003) 3859–3862, http://dx. doi.org/10.1143/JJAP.42.3859. [14] D. Yao, N. Pratapkumar, L. Lei, A. Yi, A two-station embossing process for rapid fabrication of surface microstructures on thermoplastic polymers, Polym. Eng. Sci. 47 (2007) 530–539, http://dx.doi.org/10.1002/pen.20721. [15] X. Chang, D. Xie, X. Ge, H. Li, Noncontact microembossing technology for fabricating thermoplastic optical polymer microlens array sheets, Scientific World J. 2014 (2014) 1–8, http://dx.doi.org/10.1155/2014/736562. [16] N. Hashim, S.K. Kamarudin, W.R.W. Daud, Design fabrication and testing of a PMMAbased passive single-cell and a multi-cell stack micro-DMFC, Int. J. Hydrogen Energ. 34 (2009) 8263–8269, http://dx.doi.org/10.1016/j.ijhydene.2009.07.043. [17] J. Liu, L. Wang, W. Ouyang, W. Wang, J. Qin, Z. Xu, S. Xu, D. Ge, L. Wang, C. Liu, L. Wang, Fabrication of PMMA nanofluidic electrochemical chips with integrated microelectrodes, Biosens. Bioelectron. 72 (2015) 288–293, http://dx.doi.org/10.1016/ j.bios.2015.05.031. [18] N. Shiraishi, M. Kimura, Y. Ando, Development of PMMA-based gas sensor and its evaluation using a VOC dilution flow system, Microelectron. Eng. 119 (2014) 115–121, http://dx.doi.org/10.1016/j.mee.2014.03.021. [19] R. Ranjit, A Primer on the Taguchi Method, Society of Manufacturing Engineers, Dearborn, Michigan, 1990 8–10, http://dx.doi.org/10.1016/j.ijhydene.2009.07. 043. [20] J.T. Wu, Y.T. Chu, S.Y. Yang, C.C. Li, Low temperature embossing technique for fabrication of large-area polymeric microlens arrays with supercritical carbon dioxide, Microelectron. Eng. 87 (2010) 2620–2624, http://dx.doi.org/10.1016/j.mee.2010. 07.028. [21] B.K. Lee, J.H. Kim, D.S. Kim, S.S. Chang, T.H. Kwon, Microfabrication of a nickel mold insert by a modified deep X-ray lithography process and its application to hot embossing, Microelectron. Eng. 87 (2010) 2449–2455, http://dx.doi.org/10.1016/j. mee.2010.04.023. [22] C.Y. Chang, S.Y. Yang, M.H. Chu, Rapid fabrication of ultraviolet-cured polymer microlens arrays by soft roller stamping process, Microelectron. Eng. 84 (2007) 355–361, http://dx.doi.org/10.1016/j.mee.2006.11.004. [23] L. Heyderman, H. Schift, C. David, J. Gobrecht, T. Schweizer, Flow behaviour of thin polymer films used for hot embossing lithography, Microelectron. Eng. 54 (2000) 229–245, http://dx.doi.org/10.1016/S0167-9317(00)00414-7. [24] J. Gomez, G. Conner, D.H. Chun, N. Kim, I.-H. Song, B. You, Mold filling analysis of an alignment structure in micro hot embossing, Fibers Polym. 15 (2014) 1197–1201, http://dx.doi.org/10.1007/s12221-014-1197-5. [25] M. Heckele, W. Bacher, K. Muller, Hot embossing - the molding technique for plastic microstructures, Microsyst. Technol. 4 (1998) 122–124, http://dx.doi.org/10.1007/ s005420050112. [26] R. Ranjit, A Primer on the Taguchi Method, Society of Manufacturing Engineers, Dearborn, Michigan, 1990 193–194. [27] X. Zhue, T.W. Simon, T. Cui, Hot embossing at viscous state to enhance filling process for complex polymer microstructures, Microsyst. Technol. 18 (2012) 257–265, http://dx.doi.org/10.1007/s0054201113895. [28] W.B.D. Forfang, T.G. Conner, B. You, T. Park, I.-H. Song, Fabrication and characterization of polymer microprisms, Microsyst. Technol. 20 (2014) 2071–2077, http://dx. doi.org/10.1007/s00542-014-2067-1. [29] F. Trager, Springer Handbook of Lasers and Optics, 449Springer-Verlag Berlin, Heidelberg, Berlin, Germany, 2012.
Contents lists available at ScienceDirect
Microelectronic Engineering journal homepage: www.elsevier.com/locate/mee
Experimental study of polymer microlens fabrication using partial-filling hot embossing technique Sean Moore a, Juan Gomez b, Devanda Lek b, Byoung Hee You a,b, Namwon Kim c, In-Hyouk Song a,b,⁎ a b c
Department of Engineering Technology, Texas State University, 601 University Dr., San Marcos, TX 78666-4684, United States Materials Science, Engineering, and Commercialization Program, Texas State University, 601 University Dr., San Marcos, TX 78666-4684, United States Ingram School of Engineering, Texas State University, 601 University Dr., San Marcos, TX 78666-4684, United States
a r t i c l e
i n f o
Article history: Received 18 November 2015 Received in revised form 11 April 2016 Accepted 10 May 2016 Available online 11 May 2016 Keywords: Hot embossing MEMS Partial-filling Polymer microlens Taguchi method
a b s t r a c t A method to fabricate microlens arrays in polymer substrates via hot embossing is presented in this paper. The partial-filling hot embossing technique of micro cavities on a mold insert through surface tension and capillary action is proven to be an effective means of limiting imperfections on the surface of microlens arrays. The effects of the investigated process parameters including temperature, embossing pressure, and holding time are analyzed via Taguchi method to identify effective processing conditions for microlens arrays of varying heights and diameters. Signal-to-noise (S/N) ratios are calculated for the focal length of the fabricated microlens arrays to identify key individual parameters and their interactions for a streamlined fabrication process. Experimental data indicates that the holding time in the embossing process has the most significant impact on lens focal length followed by embossing temperature and pressure. This study identifies a reliable means of microlens production and demonstrates the effects of varying process parameters in the partial-filling method of micro hot embossing for the production of lens arrays. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Microlenses and microlens arrays have been used in the biomedical, optical communication, and optoelectronics fields [1–5]. Used in fiber optic applications, the addition of microlens arrays at the objective and relay ends of a fiber bundle increases the fill factor of the bundle, as well as improves its coupling efficiency [2]. Microlenses have also been commonly employed in endoscopic applications for tissue and cellular imaging [1,5]. In response to growing demand, a variety of microlens manufacturing methods have been investigated. The most common methods include laser ablation, reflow techniques, ink jet molding, injection molding, and hot embossing [6–9]. While laser ablation techniques allow for later phase fabrication of micro-optics in established devices, they offer no versatility in lens height or diameter due to masking requirements [10]. Polymer reflow and ink jet molding techniques are effective for large aperture lenses. However, the processing parameters required are difficult to control in micro scale production [11]. The thermal and flow induced stresses introduced into small, thin profiles negate the mass production volumes that characterize injection molding processes.
⁎ Corresponding author at: Department of Engineering Technology, Texas State University, 601 University Dr., San Marcos, TX 78666-4684, United States. E-mail address: [email protected] (I.-H. Song).
http://dx.doi.org/10.1016/j.mee.2016.05.009 0167-9317/© 2016 Elsevier B.V. All rights reserved.
In contrast, a hot embossing technique offers advantages in regard to simplicity, operating cost, and replication accuracy for microstructures. The capability to fabricate microlens arrays with hot embossing has been demonstrated with two techniques: partial-cavity filling and full-cavity filling [12,13]. Full-cavity filling uses mold inserts with the desired microlens geometry that facilitate the complete filling of the mold cavity generating microlenses with limited variation. While this method does produce repeatable batch-to-batch outcomes, the polymer makes contact with the mold insert often resulting in surface defects that scatter incident light [14]. The partial-filling technique alleviates this issue by capitalizing on surface tension and capillary action to generate the lens structure thereby minimizing contact with mold surfaces and achieving optical surface characteristics [12]. Partial-filling is initiated with surface tension generated from the mold making contact with the polymer substrate under embossing pressure. The surface tension is a result of the elevated energy levels on the surface of the polymer opposed to its interior. In combination with polymer temperatures above glass transition temperature, this surface tension leads to a capillary action forcing the polymer into the recesses of the mold. The hemispherical convex lens formation of the polymer is a result of a reduction in surface tension as the center of the mold cavity allows the polymer to flow with less restriction [15]. Usable microlens fabrication requires an embossing temperature, an embossing pressure, and a holding time to be applied in the proper configuration. An effective combination of these parameters promotes the
58
S. Moore et al. / Microelectronic Engineering 162 (2016) 57–62
partial filling of the mold cavities in a hemispherical fashion without contacting the mold sidewalls. During embossing, the flow front of the polymer substrate flows into the mold cavity at a higher velocity in comparison to the flow at the sidewalls due to the pressure differential between the hole edge and its center [12]. Furthermore, alteration of process parameters facilitates the filling of the microholes at varying heights resulting in changes in focal length. Thus, microlens arrays of varying focal lengths can be achieved with the adjustment of processing parameters using only one mold insert. Despite these advantages, the partial-filling method is difficult to control with variations from lensto-lens and batch-to batch reproducibility being the primary concerns. In this study, polymer microlens arrays are demonstrated with polymethyl methacrylate (PMMA) substrates, commonly used in hot embossing to form a variety of microstructures [16–18]. To improve replication reliability, the effects of major processing parameters are systematically investigated in this paper. Three processing parameters including an embossing temperature, an embossing pressure, and holding time are considered with three levels arranged in the L9 orthogonal configuration of Taguchi method, a statistics based experimental design, which is beneficial in identifying significant impacts of individual processing parameters in addition to parameter interactions [19]. The focal length of the fabricated microlens arrays is measured and compared with the calculated value using the dimensions of the radius of curvature (ROC), sag height, and diameter to determine the effects of process parameters on the microlens array. 2. Design of microlens mold insert Hot embossing uses a patterned mold insert to transfer microstructures inherent to the mold onto polymer substrates. Effective fabrication of microstructures demands accurate mold patterns that are capable of generating high-resolution replications in polymer substrates. A variety of metallic materials have been implemented in the fabrication of mold inserts including nickel, stainless steel, and brass [20–22]. A brass alloy is selected for this study as it demonstrates the capability to withstand repetitions of high pressure and temperature variations present in hot embossing [14]. A HAAS MiniMill, a vertical CNC milling machine, is used to form microhole arrays in the brass. The design and fabrication of the mold insert is critical for defining the diameter of individual microlenses within an array, which determines the ROC and focal length of the microlenses. Fig. 1(a) shows the design and the dimension of the microhole array on the brass mold insert. A 10 × 10 array of 508 μm (0.02 in.) microholes is arranged in a rectangular configuration with a pitch of 711 μm (0.028 in.) and a 203 μm (0.008 in.) spacing between each hole. Fig. 1(b) is a cross-section view of A–A′ indicated on Fig. 1(a). Optical defects due to machining traces present on the mold insert are avoided using the partial-filling method. As the polymer substrate does not contact the sidewalls or bottom of the microholes during the embossing process, the length of the microholes is insignificant and set at 508 μm, a 1:1 depth to diameter ratio. Fig. 1(c) shows a schematic of the partial-cavity filling hot embossing technique. Fig. 2 illustrates the micro-milled mold insert on brass, (353 engravers brass, McMaster Carr) and a microscopic image of the microholes. Machining tolerances in conjunction with limitations inherent to end mill fabrication resulted in a deviation of hole diameter from mold design to fabrication. The measured diameter of fabricated microholes is 512 μm. The microhole array is placed in the center of the mold insert to mitigate thermal stresses created in the embossing process. 3. Fabrication of microlens array A 3 mm thick PMMA substrate is selected due to its appropriate optical transparency, and mechanical durability. A Carver Thermal Press (3893 4NE18, Carver Inc., Wabash, IN) is employed for the hot
Fig. 1. (a) A schematic representation of the microlens array design. (b) The cross-section view of A–A′ illustrated in panel (a) with a 1:1 depth to width ratio. (c) Illustration of partial-filling hot embossing process.
embossing process. The mold insert is mounted on the upper platen while a secondary brass stage is mounted to the lower platen. The lower stage, polished to mitigate surface roughness, is used to limit the transfer of surface artifacts to the polymer substrate. An oil-based cooling system is implemented to reduce thermal stress and its effects during cooling and demolding and to reduce process cycle time. The cooling system consists of an oil transfer pump, used to circulate a heat transfer fluid throughout the heating and cooling platens paired with a condenser that serves as a heatsink for the thermal fluid. In hot embossing, the polymer substrate is heated to an embossing temperature prior to molding. The glass transition temperature (Tg) of PMMA is 105 °C. The applied embossing pressure forces the heated polymer to flow into the microhole. The hot embossing process
S. Moore et al. / Microelectronic Engineering 162 (2016) 57–62
59
Fig. 3. A graphical representation of the hot embossing process with regard to embossing temperature and pressure is shown.
Fig. 2. Photo and microscopic image of the brass mold insert.
demonstrated in this experiment is divided into four steps including: pre-heating, embossing, demolding, and cooling. The process initiates by heating the upper and lower platens to the prescribed embossing temperature followed by the placement of the polymer substrate on the lower platen. Once the mold insert and the polymer substrate reach the prescribed embossing temperature, the press is closed and the embossing pressure is applied for a period defined as holding time [23]. During the holding time, the polymer substrate flows isotropically, constrained by the mold insert, into the mold cavity [24]. Throughout the embossing phase, the mold insert and the substrate plate continue to move closer to maintain constant embossing pressure. Temperature must remain constant during embossing, as any variations in temperature affect the fluidic resistance of the polymer substrate resulting in pattern defects [25]. Upon completion of the holding time, the molded pattern is separated from the mold insert by the relative displacement of the substrate plate, and cooled below 80 °C.
During the micro hot embossing experiments, three different levels (L1, L2, and L3) of each parameter are systematically applied in combinations dictated by the orthogonal array to identify significant interactions among them. Fig. 3 illustrates the process profile and the application of temperature and pressure as a function of time throughout the fabrication process. The temperature range for this research is set in accordance with the flow behavior of PMMA during embossing. At around 105 °C, PMMA begins to exhibit rubber elastic properties but thermal reflow is absent. Above 130 °C, PMMA is in a viscoelastic state typical of forming complete polymeric structures such as fluidic channels [27,28]. The purpose of this research is to avoid complete cavity filling. Thus, the maximum embossing temperature is set to 130 °C. The thermal press employed for this study exerts a minimum force of 1000 kg. While this force is sufficient to fabricate microlenses, this experiment uses forces of 1500 kg, 2000 kg, and 2500 kg exerting an embossing pressure of 285 psi, 380 psi, and 475 psi, respectively, to avoid any experimental replication issues associated with minimal force programming. Holding time is determined with preliminary parameter selection experiments whose results indicate that the holding time of 2 min under minimized embossing temperatures and pressures reliably produced lens structures. Therefore, symmetric holding times of 2 min, 3.5 min, and 5 min are employed. The process parameters are illustrated in Table 1. Upon completion of the prescribed holding time, the embossing pressure is immediately released and the substrate is allowed to cool under atmospheric conditions. In this manner, the parameter of holding time is isolated from embossing pressure allowing for independent study. The effects of process parameter configurations on the diameter and the height of the microlens arrays are measured and collected to determine the ROC and the focal length for each run of the experiment.
5. Results and discussion 4. Design of experiments A parameter design Taguchi methodology, which utilizes signal-tonoise ratios (S/N ratios) to establish relationships among parameters arranged in an orthogonal array is used in this study. Taguchi parameter and tolerance design not only identifies key parameter contributions in regard to stability, but also reduces the number of experimental runs required. Hot embossing experiments are performed at three symmetric intervals to study the effects of an embossing temperature, an embossing pressure, and a holding time in the partial cavity filling process. As process stability is the principle concern, the goal is to determine what functional relationship exists among factors, their levels, and the stability of the desired output. This research uses the L9 array with the fourth parameter omitted [26].
Experimental trials are conducted randomly with three repetitions. Fig. 4 shows a stereoscopic image of the microlens array fabricated with an embossing temperature of 120 °C, an embossing pressure of 285 psi, and a holding time of 2 min. Fig. 5 displays the 3-D profile image of the microlens scanned using KLA-Tencor P-7 stylus profiler. Table 1 The hot embossing parameters and their respective levels are shown. Process parameter
Level 1
Level 2
Level 3
(A) Embossing temp (°C) (B) Embossing pressure (psi) (C) Holding time (min)
120 285 2
125 380 3.5
130 475 5
60
S. Moore et al. / Microelectronic Engineering 162 (2016) 57–62
Angstroms Inc., Portsmouth, VA) as shown in Fig. 6(a). The ROC, R, of the analyzed lens is calculated by [29] 2
R¼
D2 þ 4h ; 8h
ð1Þ
where D is the diameter, and h is the height of the microlens. For a focal length measurement, the highest point in lens curvature is focused with a microscope (Nikon MM 800, Tokyo, Japan), denoted as position A of
Fig. 4. A stereoscopic image of a fabricated microlens array.
Fig. 5. 3-D profile of microlens measured using KLA-Tencor P-7 stylus profiler. Scanning area of 800 μm × 800 μm.
Ten microlenses are randomly selected and the sampling surfaces at the top of microlenses are scanned with a 20 μm scan length to measure surface roughness of the microlenses. Table 2 illustrates the surface roughness of the arithmetic average, Ra, and the root mean squared average, Rq. The average Ra and Rq are 6.98 ± 0.78 nm and 8.74 ± 0.97 nm, respectively. Each microlens is inspected and the average ROC and focal length are extracted using three individual lenses located along the bottom edge of the microlens array. A side profile photo of each of the measured lenses is captured with a surface tension analyzer (FTA 1000 B Class, First Ten
Table 2 The surface roughness of the fabricated microlenses. Scan #
Ra (nm)
Rq (nm)
1 2 3 4 5 6 7 8 9 10 Average Standard deviation
7.81 6.89 5.85 6.3 6.79 6.82 8.12 7.2 7.9 6.07 6.98 0.78
9.57 8.37 7.31 7.8 8.7 8.57 10.3 9.22 9.77 7.76 8.74 0.97
Fig. 6. (a) The methodology for obtaining lens height and diameter using scaling ratios is demonstrated. D and h are diameter and height of microlens, respectively, to extract the ROC of microlens and the focal length. (b) A schematic representation of lens anatomy and focal length measurement through a microscope. Here, R and f are ROC and focal length of microlens, respectively.
S. Moore et al. / Microelectronic Engineering 162 (2016) 57–62
61
Table 3 The calculated ROC and focal length data relevant to experiment level. Run
A
1 2 3 4 5 6 7 8 9
L1 L1 L1 L2 L2 L2 L3 L3 L3
B
L1 L2 L3 L1 L2 L3 L1 L2 L3
C
L1 L2 L3 L2 L3 L1 L3 L1 L2
ROC (μm)
Calculated focal length (μm)
Measured focal length (μm)
Y1
Y2
Y3
Ave.
Y1
Y2
Y3
Ave.
Y1
Y2
Y3
Ave.
1092 579 640 926 685 1099 675 800 506
989 668 584 948 671 1052 726 808 437
914 702 578 977 708 970 740 823 449
998 650 601 950 688 1040 714 810 464
2229 1181 1306 1889 1399 2242 1378 1633 1032
2019 1364 1191 1935 1347 2146 1482 1650 870
1866 1432 1180 2039 1445 1980 1509 1679 899
2038 1326 1226 1954 1397 2123 1456 1654 934
2353 1304 1336 1946 1465 2557 1428 1737 1119
1997 1431 1227 2029 1372 2292 1498 1706 958
2191 1450 1229 2279 1523 2174 1538 1741 996
2180 1395 1264 2085 1453 2341 1488 1728 1024
and calculated values and exhibits a correlation between the two methods. A “smaller is better” configuration of the S/N ratio is applied to minimize the influence of changing multiple parameters within a run thus highlighting individual factor effects on experimental variation. The focal length is chosen for analysis as it is dependent upon discrete ROCs, sag heights, and diameters. Each S/N ratio is calculated by
S=N ¼ −10 log
Fig. 7. Measured and calculated focal length of the fabricated microlenses.
Fig. 6(b). While the lens remains fixed on the microscope stage, the objective lens of the microscope moves vertically upward to focal point B, where incident light rays converge. The travel distance of the objective lens along the z-axis indicates the focal length of the individual microlens. Using the ROC of the lens, R, and the refractive index, n, of the lens media, the focal length, f, is given by [29] f ¼
R : n−1
ð2Þ
In the case of PMMA, the refractive index is 1.49 for the wavelength of 600 nm. Using the measured heights and diameters, ROC and focal length for each run are calculated and shown in Table 3, using Eq. (1) and Eq. (2). The table reflects the collective average Y1, Y2, and Y3 of three lenses prepared from individual sample arrays generated by the replications of each run. Three sample lenses from each replication are selected, measured, and averaged to obtain each run's respective Y values for ROC, calculated focal length, and measured focal length. These values are collectively averaged to achieve an overall run average as shown for the nine experimental levels. The calculated values are compared with measured focal length values to ensure efficacy of the measurement method. Fig. 7 illustrates the deviation between average measured
! n 1X Y 2i ; m i¼1
ð3Þ
where Yi values are experimental data from Table 3 for each run and m is the number of measurements per run. The S/N ratio range is calculated for each parameter to determine the main effect. A parameter with the highest ranking reflects the greatest effect throughout the experiment. Table 4 shows the parameter S/N ratios and their respective main effect for focal length, calculated using Eq. (3). The calculated main effects of temperature, pressure, and holding time for the microlens focal length are 2.73 dB, 2.51 dB, and 3.13 dB while the measured focal length has main effects of 2.88 dB, 2.32 dB, and 3.40 dB, respectively. Fig. 8 graphically illustrates the main effect of each parameter for the measured focal length of the microlenses. Results indicate the most critical aspect of the hot embossing process for microlens focal length is the holding time. An increase in holding time is directly related to an increase in energy applied to the substrate. The increase in energy facilitates the flow of the heated polymer into the only available expansion region, the microholes. The convex hemispherical shape is formed by capillary action and surface tensions resultant from constraining the polymer flow with the mold insert. Similarly, experimental data indicates that the embossing temperature has an effect on ROC and focal length of the microlens array. The increase in the embossing temperature promoted the flow of the polymer substrate into the mold cavity by relaxing the molecular bonds in the substrate increasing the polymer elasticity and thereby decreasing fluidic resistance. Embossing temperature determines the physical phase of the PMMA substrate to promote cavity filling. An increase in temperature corresponds to an increase in the ability of the substrate to flow into the recesses of the mold insert. Embossing pressure, while still a requirement for the formation of microlens structures, is the least contributing factor to the fabrication process.
Table 4 S/N ratio (dB) and main effect for the focal length of microlenses. Parameter
Embossing temp Embossing pressure Holding time
Calculated focal length
Measured focal length
Level 1
Level 2
Level 3
Main effect
Level 1
Level 2
Level 3
Main effect
−63.49 −65.10 −65.71
−65.09 −63.25 −62.58
−62.36 −62.59 −62.65
2.73 2.51 3.13
−63.91 −65.55 −66.32
−65.69 −63.64 −63.18
−62.81 −63.23 −62.92
2.88 2.32 3.40
62
S. Moore et al. / Microelectronic Engineering 162 (2016) 57–62
Fig. 8. The main effect plot of the S/N ratios for measured focal length of the microlens array.
6. Conclusion The effects of embossing temperature, pressure, and holding time are investigated for the fabrication of microlens arrays in PMMA substrates using the partial-filling hot embossing technique. The partial-filling technique replicated microlens arrays of varying focal lengths and using one mold insert. Although surface imperfections are mitigated due to the reduced contact with the mold walls, variations between the microlens arrays is of principle concern. This research characterizes the effects of process parameters including embossing temperatures, embossing pressures, and holding times on microlens fabrication. A three level, three parameter, L9 orthogonal Taguchi design study shows parameter interactions. Additionally, the most critical parameter in relation to lens ROC and focal length is determined. Results indicate that holding time is the most influential parameter with resonating impact throughout each experimental level. This study provides in-depth information in regard to the appropriate processing conditions for the fabrication of polymer microlenses using the partial-filling technique. References [1] V.K. Shinoj, V. Murukeshan, S.B. Tor, N.H. Loh, S.W. Lye, Design fabrication and characterization of thermoplastic microlenses for fiber-optic probe imaging, Appl. Opt. 53 (2014) 1083–1088, http://dx.doi.org/10.1364/AO.53.001083. [2] X.-T. Yan, J. Yang, X. Bin, X. Ma, F. Li, Y. Zhao, F. Bu, Design of the microlens arrays coupling with imaging fiber bundle, Opt. Lett. 9 (2013) 169–172, http://dx.doi. org/10.1007/s11801-013-3016-4. [3] R. Ramakrishnan, N. Saran, R. Petcavich, Ink jet printing for selective conductor processes for displays and flexible circuits, J. Disp. Technol. 8 (2011) 337–344, http:// dx.doi.org/10.1889/1.3499831. [4] K. Li, C. Feng, H. Choi, Analysis of micro-lens integrated flip-chip INGaN light-emitting diodes by confocal microscopy, Appl. Phys. Lett. 104 (2014) 051107, http://dx. doi.org/10.1063/1.4863925. [5] A. Hassanfiroozi, Y.P. Huang, B. Javidi, H.P. Shieh, Hexagonal liquid crystal lens array for 3D endoscopy, Opt. Express 23 (2015) 971–981, http://dx.doi.org/10.1364/OE. 23.000971. [6] A. Zukauska, M. Malinauskas, C. Reinhardt, B. Chichkov, R. Gadonas, Closely packed hexagonal conical microlens array fabricated by direct laser photopolymerization, Appl. Opt. 51 (2012) 4995–5003, http://dx.doi.org/10.1364/AO.51.004995.
[7] H. Zhang, L. Li, D. McCray, D. Yao, A. Yi, A microlens array on curved substrates by 3D microprojection and reflow process, Sensors Actuat. A- Phys. 179 (2012) 242–250, http://dx.doi.org/10.1016/j.sna.2012.03.002. [8] D. Xie, H. Zhang, X. Shu, J. Xiao, Fabrication of polymer micro-lens array with pneumatically diaphragm-driven drop on demand ink-jet technology, Opt. Express. 20 (2012) 15186–15195, http://dx.doi.org/10.1364/OE.20.015186. [9] P. Xie, P. He, Y.C. Yen, K.J. Kwak, D. Gallego-Perez, L. Chang, W.C. Liao, A. Yi, J. Lee, Rapid hot embossing of polymer microstructures using carbide-bonded graphene coating on silicon stampers, Surface Coat. Tech. 258 (2014) 174–180, http://dx. doi.org/10.1016/j.surfcoat.2014.09.034. [10] K. Naessens, H. Ottevaere, R. Baets, P.V. Daele, H. Thienpont, Direct writing of microlenses in polycarbonate with excimer laser ablation, Appl. Opt. 42 (2003) 6349–6359, http://dx.doi.org/10.1364/AO.42.006349. [11] J. Stringer, B. Derby, Limits to feature size and resolution in ink jet printing, J. Europ. Ceramic Soci. 29 (2009) 913–918, http://dx.doi.org/10.1016/j.jeurceramsoc.2008. 07.016. [12] C.Y. Chang, S.Y. Yang, S.L. Huang, K.H. Hsieh, Fabrication of polymer microlens arrays using capillary forming with a soft mold of micro-holes array and UVcurable polymer, Opt. Express 14 (2006) 6253–6258, http://dx.doi.org/10. 1364/OE.20.015186. [13] K. Li, C. Feng, H.W. Choi, X.C. Shan, R. Maeda, Y. Murakoshi, Micro hot embossing for replication of microstructures, Jpn. J. Appl. Phys. 42 (2003) 3859–3862, http://dx. doi.org/10.1143/JJAP.42.3859. [14] D. Yao, N. Pratapkumar, L. Lei, A. Yi, A two-station embossing process for rapid fabrication of surface microstructures on thermoplastic polymers, Polym. Eng. Sci. 47 (2007) 530–539, http://dx.doi.org/10.1002/pen.20721. [15] X. Chang, D. Xie, X. Ge, H. Li, Noncontact microembossing technology for fabricating thermoplastic optical polymer microlens array sheets, Scientific World J. 2014 (2014) 1–8, http://dx.doi.org/10.1155/2014/736562. [16] N. Hashim, S.K. Kamarudin, W.R.W. Daud, Design fabrication and testing of a PMMAbased passive single-cell and a multi-cell stack micro-DMFC, Int. J. Hydrogen Energ. 34 (2009) 8263–8269, http://dx.doi.org/10.1016/j.ijhydene.2009.07.043. [17] J. Liu, L. Wang, W. Ouyang, W. Wang, J. Qin, Z. Xu, S. Xu, D. Ge, L. Wang, C. Liu, L. Wang, Fabrication of PMMA nanofluidic electrochemical chips with integrated microelectrodes, Biosens. Bioelectron. 72 (2015) 288–293, http://dx.doi.org/10.1016/ j.bios.2015.05.031. [18] N. Shiraishi, M. Kimura, Y. Ando, Development of PMMA-based gas sensor and its evaluation using a VOC dilution flow system, Microelectron. Eng. 119 (2014) 115–121, http://dx.doi.org/10.1016/j.mee.2014.03.021. [19] R. Ranjit, A Primer on the Taguchi Method, Society of Manufacturing Engineers, Dearborn, Michigan, 1990 8–10, http://dx.doi.org/10.1016/j.ijhydene.2009.07. 043. [20] J.T. Wu, Y.T. Chu, S.Y. Yang, C.C. Li, Low temperature embossing technique for fabrication of large-area polymeric microlens arrays with supercritical carbon dioxide, Microelectron. Eng. 87 (2010) 2620–2624, http://dx.doi.org/10.1016/j.mee.2010. 07.028. [21] B.K. Lee, J.H. Kim, D.S. Kim, S.S. Chang, T.H. Kwon, Microfabrication of a nickel mold insert by a modified deep X-ray lithography process and its application to hot embossing, Microelectron. Eng. 87 (2010) 2449–2455, http://dx.doi.org/10.1016/j. mee.2010.04.023. [22] C.Y. Chang, S.Y. Yang, M.H. Chu, Rapid fabrication of ultraviolet-cured polymer microlens arrays by soft roller stamping process, Microelectron. Eng. 84 (2007) 355–361, http://dx.doi.org/10.1016/j.mee.2006.11.004. [23] L. Heyderman, H. Schift, C. David, J. Gobrecht, T. Schweizer, Flow behaviour of thin polymer films used for hot embossing lithography, Microelectron. Eng. 54 (2000) 229–245, http://dx.doi.org/10.1016/S0167-9317(00)00414-7. [24] J. Gomez, G. Conner, D.H. Chun, N. Kim, I.-H. Song, B. You, Mold filling analysis of an alignment structure in micro hot embossing, Fibers Polym. 15 (2014) 1197–1201, http://dx.doi.org/10.1007/s12221-014-1197-5. [25] M. Heckele, W. Bacher, K. Muller, Hot embossing - the molding technique for plastic microstructures, Microsyst. Technol. 4 (1998) 122–124, http://dx.doi.org/10.1007/ s005420050112. [26] R. Ranjit, A Primer on the Taguchi Method, Society of Manufacturing Engineers, Dearborn, Michigan, 1990 193–194. [27] X. Zhue, T.W. Simon, T. Cui, Hot embossing at viscous state to enhance filling process for complex polymer microstructures, Microsyst. Technol. 18 (2012) 257–265, http://dx.doi.org/10.1007/s0054201113895. [28] W.B.D. Forfang, T.G. Conner, B. You, T. Park, I.-H. Song, Fabrication and characterization of polymer microprisms, Microsyst. Technol. 20 (2014) 2071–2077, http://dx. doi.org/10.1007/s00542-014-2067-1. [29] F. Trager, Springer Handbook of Lasers and Optics, 449Springer-Verlag Berlin, Heidelberg, Berlin, Germany, 2012.