Sensors and Actuators A: Physical

Transcription

1 Sensors and Actuators A 159 (010) Contents lists available at ScienceDirect Sensors and Actuators A: Physical journal homepage: Fabrication of various dimensions of high fill-factor micro-lens arrays for OLED package K.H. Liu a, M.F. Chen b, C.T. Pan b,, M.Y. Chang c, W.Y. Huang c a Department of Mechanical Engineering, R.O.C. Military Academy, Kaohsiung, Taiwan, ROC b Department of Mechanical and Electro-Mechanical Engineering, Center for Nanoscience and Nanotechnology, National Sun-Yat-Sen University, 70 Lien-hai Rd., Kaohsiung 804, Taiwan, ROC c Department of Photonics, National Sun-Yat-Sen University, Kaohsiung 804, Taiwan, ROC article info abstract Article history: Received May 009 Received in revised form 1 January 010 Accepted 19 February 010 Available online 1 March 010 Keywords: OLED Gapless micro-lens array NiCo electroplating Fill-factor Organic light-emitting diode (OLED) has been the subject of much interest in the lighting and display module. This study focuses on how to enhance brightness and scatter through a micro-lens array (MLA) with high fill-factor. A LIGA-like (Lithographie Galvanoformung Abformung, LIGA) process was applied to the fabrication process because it has good replication for microstructures. Seven kinds of micro-lens arrays with different dimensions were designed. These dimensions with different aspect ratio were divided into two groups, i.e. gapless high fill-factor and low fill-factor with gap. They were used to compare their optical properties of brightness with each other. In addition, a controllable method was also developed to design the contour of a micro-lens. Then, a metallic mold with MLA was fabricated by nickel cobalt (NiCo) electroplating process. A highly accurate and strong mold can be obtained. Finally, a UV (ultraviolet) curable polymer was used as the material of an optical film in this replication process of a MLA. After these optical films with MLAs were obtained, an optical microscope (OM) and a photometer were used to measure and characterize the optical results. The experimental results of the MLA were compared with those of the design. 010 Elsevier B.V. All rights reserved. 1. Introduction Micro-lens arrays (MLAs) have been applied more and more in the optical and lighting systems in recent years, such as in optical fibers, image systems and illumination. There was also much literature associated with the function of a MLA presented [1 3]. An organic light-emitting display (OLED) has been developed quickly, and the brightness and luminance efficiency become an important issue. An OLED with a MLA has been researched to enhance light extraction [4,5]. Forrest and coworkers [6] proposed the relationship of a curvature of a micro-lens and out-coupling efficiency in an OLED. A simulation and an investigation about the fill-factor of a MLA were presented and its luminance intensity was examined by Wei et al. [7]. It revealed that the intensity increased with a fill-factor. An optical thin film with a microstructure of a MLA has been developed and investigated about its various fabrication processes. Varied fabrication methods of MLAs have been described, such as the reflow process [8 10], laser-aided fabrication [11,1], lithography process with a gray-scale mask [13], a photolytic technique Corresponding author. Tel.: x439; fax: address: panct@mail.nsysu.edu.tw (C.T. Pan). [14], molding [15,16], etching [17], assembled by surface properties [18,19], liquid crystal type [0], closed-packed colloidal monolayer [1,], and proton beam writing [3]. Among these processes, molding has more advantages such as fast, high precision and mass product. The LIGA-like (Lithographie Galvanoformung Abformung, LIGA) [4] technique was applied to fabricate the mold for a molding process, which is usually a metallic mold. The LIGA process was developed to fabricate micro-lens technology in 1980s from the nuclear energy research in Germany. It combines with the X-ray lithography, the micro-electroplating, and the microinjection or the micro-embossing [5]. The difference between the LIGA and the LIGA-like technique is that ultraviolet light replaces X-ray. For a MLA, fill-factor is an important factor to light efficiency which is defined as the percentage of lens area to the full area. To collect the maximum amount of light, the lens area should be as close to 100% as possible [3]. However, most literatures only discussed circular micro-lens in an array. The maximum fill-factor of a micro-lens with circular apertures in different layout is 78.5% for rectangular layout and 90.6% for hexagonal layout [6]. To obtain 100% fill-factor of a MLA, a non-circular geometrical lens has been presented [6 3]. Among these literatures, there were hexagonal [7], triangular [8,9] and dual-curvature MLAs [3] presented in our laboratory. But, in these polygonal MLA studies, only a /$ see front matter 010 Elsevier B.V. All rights reserved. doi: /j.sna

2 K.H. Liu et al. / Sensors and Actuators A 159 (010) Table 1 The designed variables of seven micro-lens arrays (unit: m) (aspect ratio = H/D). No. Diameter (D) Thickness (H) Vertical (V) Period (T) Hypotenuse Layout Aspect ratio Gapless Gapless Gapless Gapless Not gapless Not gapless Not gapless 3.33 single dimension was designed, and the MLAs usually were regularpolygon. A controllable dimension for a micro-lens was developed, and a non-regular MLA was obtained in this study. Regular MLA is defined as a micro-lens in the array with a regular hexagonal shape; while non-regular MLA is defined as a micro-lens in the array with a non-regular hexagon. Seven kinds of dimensions of hexagonal MLAs with different aspect ratios (the ratio of height to diameter) and fill-factor were designed and fabricated to investigate the relationship of the luminance intensity of gap and gapless MLA. The fabrication process of these hexagonal MLAs consisted of a UV (ultraviolet)-lithography process, reflow, NiCo (nickel cobalt) electroplating and a replication method. The NiCo electroplating was used to obtain a metallic mold of a polygonal MLA. The UV-cured process is a kind of the replication process. It has excellent forming ability and fast fabrication. For the measurement, a green OLED was used as the light source in this study. The result shows that the distribution of luminance is improved when the MLAs have both high fill-factor and high aspect ratio. The distribution of luminance is the physical quantity of uniformity of a film. Therefore, a better uniformity can be obtained.. The design principle A key point to fabricate a MLA with high fill-factor is the layout of patterns on a mask. First, fill-factor is defined as the percentage of lens area to the whole area, thus the high fill-factor means more MLAs in this area. How to increase the fill-factor is an important topic of fabricating a MLA. In this study, a hexagonal MLA is a solution of the topic. A single hexagonal micro-lens was hard to manufacture. But a specific configuration of patterns on a mask can make a desired hexagonal MLA. And various dimensions of the layouts on mask can form different MLAs and optical effects. Thus, the design of a mask is the current study subject. There were total seven different dimensions designed in this study. Table 1 shows the detailed sizes of the seven patterns. Fig. 1 shows the schematic diagram of the parametric definition listed in Table 1. The three smaller circles (A C, see Fig. 1) are the patterns on a mask. Three bigger circles which are drawn in dotted line are used to describe the micro-lenses how to interwork. Finally, the interworked microlenses become hexagonal micro-lenses. Diameter (D) is the original diameter of a pattern on a mask, and the values are 40, 30, 5 and 15 m. Vertical (V) is the distance of OC, and the value is 100 m; period (T) is the distance of AB, and it is also the distance of centers of two circles which have different values of 50, 40, 35 and 5 m (all the symbols are shown in Fig. 1). The reason why the variations of the dimensions are chosen is that the dimensional variations in optical film for the LED or OLED package from 10 to 50 m. But, when the dimensions are less than 10 m, the yield rate would be low due to diffraction effect using traditional I-line and G-line exposure system. When the dimensions are larger than 50 m, it would be too large for the LED or OLED package. The period is chosen according to the dimensions of diameters. The relationship between diameters and periods is described below: ( D1 T + D ) or T< ( D1 + D ) There are three cases; if D 1 = D = D, T = D, it means that the circles are tangential. If T > D, they are separate from each other, called gap group. If T < D, they are intersectional, called gapless group. V is a constant value, but T is a variable. D is less than T, thus, the gaps between any two patterns are kept at 10 m after lithography. In addition, thicknesses (H) of photoresist in the lithography process are kept at 1.5 and 5 m. Thus, the aspect ratio of Samples 3 and 6 is higher than that of the others. The influence of aspect ratio on optical properties is discussed. A micro-lens becomes a regular hexagon when the Hypotenuse (i.e. AC and BC) is the same to Period. It means that the positions of three patterns are like a regular triangle. Finally, there are three gap-mlas used to compare the optical properties with other gapless MLAs. 3. Fabrication processes of gapless hexagonal micro-lens arrays The fabrication of optical films of these micro-lens arrays is introduced in this section. The first step was to fabricate masks based on the dimension listed in Table 1. A sketch of a circle array on a mask is shown in Fig.. There are three parameters used to Fig. 1. The three smaller circles (A, B and C) are the patterns of a mask; three bigger circles which are drawn in dotted line are used to describe the micro-lenses how to interwork; finally, the interworked micro-lenses become hexagonal micro-lenses. Diameter (D) is the original diameter of a pattern on a mask; vertical (V) is a distance of; period (T) is a distance of AB, and it is also a distance of centers of two circles.

3 18 K.H. Liu et al. / Sensors and Actuators A 159 (010) Fig. 4. A replication of UV-cured process: (a) spin coating the polymer on the secondary mold, and exposed to UV light; (b) after several seconds, the polymer was cured. Fig.. A sketch of a circle array on a mask with three parameters. Fig. 3. Schematic lithography process: (a) spin coating the photoresist AZ460; (b) a cylindrical array was obtained after development; (c) reflow the microstructures at 140 C; (d) sputtering a Ni-film as a seed layer; (e) NiCo electroplating was used to wrap the photoresist micro-lenses to form a gapless mold; (f) a passivation treatment with thermal method was applied on the surface of NiCo alloy; (g) a secondary electroplating was performed, and CMP process was used to flat the surface; the substrate was removed.

4 K.H. Liu et al. / Sensors and Actuators A 159 (010) Fig. 5. Experimental pictures of micro-lens arrays, including gapless and gap-mlas. (a) Gapless; (b) gapless; (c) gapless; (d) gapless; (e) not gapless; (f) not gapless; (g) not gapless.

5 130 K.H. Liu et al. / Sensors and Actuators A 159 (010) Table The measured data of seven different micro-lens arrays and the error of measured and designed value. Positive errors mean the micro-lenses are bigger than designed one, and negative errors mean they are smaller (unit: m). No. Dimension a Dimension b Dimension c Hypotenuse Error (c) Error (hypotenuse) define the dimensions of a micro-lens array, which are D, T and V. A silicon substrate was prepared and sent to spin coating photoresist AZ460, and the thickness of photoresist had two values of 1.5 and 5 m(fig. 3(a)). Then, that was sent to a mask aligner to expose and develop after soft-baking process at 90 C. A cylindrical array on the silicon substrate was defined (Fig. 3(b)). The cylindrical array on the silicon substrate was heated to a temperature above the glass transit temperature (T g ) of the photoresist. The photoresist microstructures were melted and reflowed, and then the profiles were changed into a hemispherical shape due to the surface tension effect. It was known as reflow process (Fig. 3(c)). After obtaining the hemispherical micro-lens array, NiCo alloy electroplating technique was applied to make a metal mold. The process includes two steps: first, Ni thin film was sputtered on the substrate and hemispherical surface served as a seed layer (Fig. 3(d)). Second, NiCo alloy deposited by electroplating was performed to form a metal mold. In the second step, every micro-lens grew gradually until they touched each other. The deposition rate of NiCo alloy was uniformly controlled by a cathode-rotated mechanism to cover whole conductive-template. Finally, it was wrapped up by NiCo alloy to form a metallic hexagonal MLA, also called a primary master mold (Fig. 3(e)). The residual stress and hardness of the NiCo mold were measured by a spiral contractometer and a Vickers microhardness tester. A spiral contractometer, purchased from Yamamoto Corporation in Japan, was applied to measure the internal stress of electrolyte. The testing steps for the hardness measurement: first, 1-mm-thick Ni Co alloy was electroplated on a substrate with 10 mm 10 mm in area. Then, the Ni Co alloy was mounted in epoxy resin, followed by grinding and polishing processes. Next, the hardness of Ni Co alloy was measured. The Vickers microhardness tester, TECH FM-100e Vickers (brand name in Taiwan), was used to measure the hardness. A load of 100 g per 15 s was compressed on ten different points of the sample. The hardness was an average data of ten different measurements. The hardness of the NiCo alloy was over Hv 650 and its residual stress after electroplating process was below 1.5 kg/mm. Then, a passivation treatment with thermal method was applied on the surface of NiCo alloy (Fig. 3(f)). Then, the NiCo alloy was electroplated again on the passivation surface. After the second electroplating process, a multilayer structure was formed as shown in Fig. 3(g). CMP (chemical mechanic polishing) process was performed to obtain a flatness surface before next step. The brittle property of the passivation layer makes the secondary master mold easily separate, as shown in Fig. 3(h). The purpose is to obtain an independent metallic mold, which is an inverse mold of the primary master mold. The mold may wear out or deform during the manufacturing process. If this mold was deformed or broken down, a new secondary master mold could be fabricated immediately from this primary master mold. UV-cured process [1] is a kind of replication. It has advantages of fast molding and simplicity. In this study, it was used to replicate the micro-lens array. The UV-cured process included two steps; one was spin coating and the other was exposure by UV light. A commercial UV curable polymer (Sil-more Industrial Company, Ltd., Taiwan) was used in this experiment. The refractive index is 1.50, and the viscosity is 500 cps at 5 C. The spinning speed was set at 000 rpm for 0 s, as shown in Fig. 4(a). Then, it was exposed by UV light to cure, as shown in Fig. 4(b). When the UV curable polymer was cured, it became a solidified film with the MLAs, as listed in Table 1. The films that were attached to an OLED can enhance the extractive rate of an OLED display and the brightness. 4. Results and discussions Fig. 5 shows the seven experimental photos of hexagonal microlens arrays whose unit is in micro-meter ( m). Samples 1,, 3 and 4 belong to the gapless group, and Samples 5, 6, and 7 belong to the group with gap. These shapes are controlled by the designed dimensions listed in Table 1. When T is shorter, the shape becomes longer. Sample 1 is a regular hexagon, and Sample 4 is a long hexagon. When V is longer, the hexagonal MLA has gaps. The measured data are listed in Table to compare with Table 1. The dimension of measurement is indicated in Fig. 6. The symbol, Dimension a, is the length of the boundary of two micro-lens; Dimension b, is the diagonal length of a hexagonal micro-lens; and Dimension c, is the width of a micro-lens. Compared with Fig. 1, the relationship of measured lengths and designed values is given by: T = c (1) Fig. 6. The definition of three measured distance of a micro-lens. The three values can be used to calculate the length of Hypotenuse as shown in Fig. 1. This figure was drawn by computer software, Solidworks. V = a + b Thus, ()

6 K.H. Liu et al. / Sensors and Actuators A 159 (010) Fig. 7. The SEM pictures of the MLAs: (a) Sample 3 (aspect ratio = 0.5); (b) Sample 4 (aspect ratio = 0.33); (c) Sample 5 (aspect ratio = 0.5). ( Hypotenuse = a ) + b ( + c ) As shown in Table, dimensions of the MAL are similar to the designed variables. The errors of Hypotenuse and the Dimension c are also shown in Table. The SEM (scanning electron microscope) pictures in Fig. 7 show that the difference between the MLAs is quite obvious. Fig. 7(a) is a Sample 3 picture which has an aspect ratio of 0.5, and the ratio is higher than that of Sample 4(Fig. 7(b)). Fig. 7(c) shows a SEM picture of Sample 5 which is a MLA with gap. The MLAs were fabricated on a polymer film, called an optical film, and were tested by SpectraScan Colorimeter PR-650 (Photo Research, Inc.). The measured set-up was controlled by a computer, including a test sample and a power supply. It can measure the luminance of a light source. A clamping apparatus is mounted on X Y Z table which can move at specific distance. A power supply (400 Source Meter, Keithley) can provide DC (direct current) to an OLED as a light source in this measurement. When an OLED is supplied with a larger voltage, it can product more luminance. The purpose of the test is to realize that when light passes through an optical film with hexagonal MLA, it can change the distribution of its intensity. The changes are measured and analyzed by PR-650. An OLED without an optical film was measured first, then other measurements of optical films were operated one by one. The first measurement was used as base data to compare with the others. Fig. 8 is a schematic diagram of the measurement of an OLED with an optical film. One optical film was measured at nine Fig. 8. Nine points of measurement of an optical film, and an OLED as light source under the optical film (a region of a gray rectangle). points which are shown as nine dark circles, and the optical film was attached to an OLED which is shown as a gray rectangle in Fig. 8. The OLED luminesces green light, and its area is mm 1.5 mm in area. The dimension of an optical film is 3 mm 3 mm. There are six points (1, 3, 4, 6, 7 and 9) measured for their luminance at large-angle. Fig. 9 shows the results of nine points, as shown in Fig. 8. Fig. 9 is the relationship between the luminance (cd/m ) and the input voltage of a power supply. Fig. 10 shows that the relationship between the luminance and voltage is a linear relationship. The results of points, 5 and 8 are much larger than those of the other points. Fig. 10 shows that the luminance of the three points is linearly proportional to the applied voltage. The data of every point include one base and seven optical films. They all have a trend toward more luminance with an increase in the voltage. Among the results, points 1, 3, 4, 6, 7 and 9 show that the luminance of optical films is greater than that of the base. The results are similar to those presented in literature [5]. But, points, 5 and 8 (in the center of the OLED) show the results of decreases. These results can be divided into two groups: gapless and gap group. The definition of gapless is defined as the boundaries of each micro-lens without interspace, but the gap group has interspace between micro-lenses. In Fig. 9, these points (i.e. Fig. 9(a) indicates point 1, (b) indicates point, and (c) point 3, etc.) indicate the position on a film in the measurement. Each chart in Fig. 9 has seven measured values at one specific point from seven films. The Samples 1 4 belong to the gapless group and the others belong to gap group. In most of measured points, the luminance of gapless group is more than that of{ok} gap group. In addition, in the gapless group, Sample 3 has more luminance because of its higher aspect ratio (about 0.5) than the others (about ). And, in the gap group, Sample 6 also has similar trend, but its luminance is less than that of{ok} the Sample 3. In the measurement, the light source was an OLED cell which was made in our lab. The OLED luminance was not uniform. Thus the luminance, in the center, of the points, 5 and 8 shown in Fig. 8 was brighter than that of the other regions. But, using this non-uniform OLED can help us understand the function of our films. When the light from the source goes through the films, the luminance distribution can be changed and formed uniformly, which is the purpose of this study. In Fig. 9, the luminance of the points 1, 3, 4, 5, 7 and 9 was increased except for Sample 7. Although the luminance of the film had little decrease in the points, 5 and 8, the whole uniformity was increased, especially for Sample 3.

7 13 K.H. Liu et al. / Sensors and Actuators A 159 (010) Fig. 9. The measured luminance of one base and seven films at nine points at three different voltages. Not only Sample 3 is a gapless MLA, but also its aspect ratio is the highest of the design. A luminance comparison of a base and Sample 3 is especially discussed in Fig. 11. Fig. 11(a) shows the comparison of points 1, 3, 4, 6, 7 and 9 of a base and Sample 3. The increments are 10.59%, %, 313.4%, 50.86%, 51.35% and 599.8%. Fig. 11(b) shows the comparison of points, 5 and 8 of a base and Sample 3. The decrements are 6.88%, 7.74% and 1.85%. Thus, the optical film can raise the luminance of the large-angle, and the uniform of the luminance of a light source can be improved. Therefore, a gapless micro-lens array provides a 100% fill-factor to improve the effect of a light source. And, a MLA with higher aspect ratio can obtain an extra gain. A layout of a micro-lens array is a key factor to improve luminance.

8 K.H. Liu et al. / Sensors and Actuators A 159 (010) can raise the luminance at the large-angle, and the uniform of the luminance of a light source can also be improved. Acknowledgements The authors would like to thank National Science Council (NSC) for their financial supports to the project (granted numbers: 95-1-E MY, and NSC96-6-E CC3). Also, the authors would like to thank C.W. Haung and the Center for Micro/Nano Technology Research, National Cheng Kung University, Tainan, Taiwan, for equipment access and technical support. Fig. 10. The linear relationship between the luminance and voltage. Fig. 11. The luminance comparison of a base and an optic film, Sample 3, (a) the comparison of points 1, 3, 4, 6, 7 and 9 of a base and Sample 3, and the raising percentages are 10.59%, %, 313.4%, 50.86%, 51.35% and 599.8%, respectively; (b) the comparison of points, 5 and 8 of a base and Sample 3, and the decreasing percentages are 6.88%, 7.74% and 1.85%, respectively. 5. Conclusion In this study, the improvement of brightness and scatter was investigated. Seven dimensions of hexagonal MLAs were designed and fabricated to investigate the relationship of the luminance intensity. They were divided into two groups, i.e. gapless high fill-factor and low fill-factor with gap. In addition, a controllable method has been developed to design the shape and aspect ratio of a micro-lens. A LIGA-like process was applied in this study, because it has good replication for microstructures. A metallic mold with micro-lens array was fabricated by NiCo electroplating process. Finally, a UV (ultraviolet) curable polymer was the material of a micro-lens array in the replication process. From the results, the configuration of the micro-lens array was compared with the design. The photometer was also used to test the brightness of the micro-lens arrays. In comparison to points 1, 3, 4, 6, 7 and 9 of a base and Sample 3, the increments are 10.59%, %, 313.4%, 50.86%, 51.35% and 599.8%, respectively. And, in comparison to points, 5 and 8 of a base and Sample 3, the decrements are 6.88%, 7.74% and 1.85%, respectively. The effect of an optical film with gapless and high aspect ratio micro-lens array can show more obvious results than that of the other designs. A layout of a microlens array is a key factor to improve luminance. The optical film References [1] L.C. Ling, Y.I. Yen, F.C. Ho, An Efficient Illumination System for Single-pane LCD Projector, SID 01 DIGEST, 001, pp [] C. Ke, X. Yi, Z. Xu, J. Lai, Monolithic integration technology between microlens arrays and infrared charge coupled devices, Opt. Laser Technol. 37 (005) [3] N.F. Borrelli, Efficiency of microlens array for projection LCD, in: Proceedings of the 44th Electronic Components and Technology Conference, Washington, DC, May 1 4, 1994, pp [4] F. Li, X. Li, J. Zhang, B. Yang, Enhanced light extraction from organic lightemitting devices by using microcontact printed silica colloidal crystals, Org. Electron. 8 (007) [5] S. Möller, S.R. Forrest, Improved light out-coupling in organic light emitting diodes employing ordered microlens arrays, J. Appl. Phys. 91 (008) [6] Y. Sun, S.R. Forrest, Organic light emitting devices with enhanced outcoupling via microlenses fabricated by imprint lithography, J. Appl. Phys. 100 (006), [7] M. Wei, J. Lee, H. Lin, Y. Ho, K. Chen, C. Lin, C. Wu, H. Lin, J. Tsai, T. Wu, Efficiency improvement and spectral shift of an organic light-emitting device by attaching a hexagon-based microlens array, J. Opt. A: Pure Appl. Opt. 10 (008) [8] M.C. Hutley, Optical techniques for the generation of microlens arrays, J. Mod. Opt. 37 () (1990) [9] C.H. Chien, C.T. Pan, C.C. Hsieh, C.M. Yang, K.L. Sher, A study of the geometry of microball lens arrays using the novel batch-fabrication technique, Sens. Actuators A: Phys. 1 (005) [10] C.K. Chung, Y.Z. Hong, Fabrication and analysis of the reflowed microlens arrays using JSR THB-130N photoresist with different heat treatments, Microsyst. Technol. 13 (007) [11] J.R. Ho, T.K. Shin, J.W. Cheng, C.K. Sung, C.F. Chen, A novel method for fabrication of self-aligned double microlens arrays, Sens. Actuators A 135 (007) [1] Y.C. Lee, C.Y. Wu, Excimer laser micromachining of aspheric microlenses with precise surface profile control and optimal focusing capability, Opt. Lasers Eng. 45 (007) [13] W. Däschner, P. Long, R. Stein, C. Wu, S.H. Lee, General aspheric refractive microoptics fabricated by optical lithography using a high energy beam sensitive glass gray-level mask, J. Vacuum Sci. Technol. B: Microelectron. Nanometer Struct. 14 (005) [14] N.F. Borrelli, D.L. Morse, R.H. Bellman, Photolytic technique for producing microlenses in photosensitive glass, Appl. Opt. 4 (16) (1985) [15] 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 () (007) [16] H. Yang, C.K. Chao, M.K. Wei, C.P. Lin, High fill-factor microlens array mold insert fabrication using a thermal reflow process, J. Micromech. Microeng. 14 (8) (004) [17] P. Zhang, G. Londe, J. Sung, E. Johnso, M. Lee, H.J. Cho, Microlens fabrication using an etched glass master, Microsyst. Technol. 13 (007) [18] W.K. Huang, C.J. Ko, F.C. Chen, Organic selective-area patterning method for microlens array fabrication, Microelectron. Eng. 83 (006) [19] C. Tsou, C. Lin, A new method for microlens fabrication by a heating encapsulated air process, IEEE Photon. Technol. Lett. 18 (3) (006) 490. [0] H. Ren, Y.H. Lin, S.T. Wu, Flat polymeric microlens array, Opt. Commun. 61 (006) [1] H.J. Nam, D.Y. Jung, G.R. Yi, H. Choi, Close-packed hemispherical microlens array from two-dimensional ordered polymeric microspheres, Langmuir (006) [] F. Li, X. Li, J. Zhang, B. Yang, Enhanced light extraction from organic lightemitting devices by using microcontact printed silica colloidal crystals, Org. Electron. 8 (5) (007) [3] R.K. Dutta, J.A. van Kan, A.A. Bettiol, F. Watt, Polymer microlens replication by nanoimprint lithography using proton beam fabricated Ni stamp, Nucl. Instrum. Methods Phys. Res. B 60 (007) [4] Ehrfeld, W.F. Gotz, D. Munchmeyer, W. Schelb, D. Schmidt, K.K. GmbH, LIGA process: sensor construction techniques via X-ray lithography, in: Proceedings of the IEEE Technical Digest: Solid-State Sensor and Actuator Workshop, 1988, pp [5] H. Guckel, T. Christenson, K. Skrobis, Metal micromechanisms via deep X-ray lithography, electroplating and assembly, J. Micromech. Microeng. (199) 5 8.

9 134 K.H. Liu et al. / Sensors and Actuators A 159 (010) [6] A.Y. Smuk, N.M. Lawandy, Direct laser fabrication of dense microlens arrays in semiconductor-doped glass, J. Appl. Phys. 87 (8) (00) [7] M.C. Chou, C.T. Pan, S.C. Shen, M.F. Chen, K.L. Lin, S.-T. Wu, A novel method to fabricate gapless hexagonal micro-lens array, Sens. Actuators A 118 (005) [8] C.T. Pan, C.H. Su, Fabrication of gapless triangular micro-lens array, Sens. Actuators A 134 (007) [9] C.T. Pan, C.H. Su, Fabrication of high fill factor optical film using two-layer photoresists, J. Mod. Opt. 55 (1) (008) [30] C.P. Lin, H. Yang, C.K. Chao, Hexagonal microlens array modeling and fabrication using a thermal reflow process, J. Micromech. Microeng. 13 (003) [31] C.P. Lin, H. Yang, C.K. Chao, A new microlens array fabrication method using UV proximity printing, J. Micromech. Microeng. 13 (003) [3] C.T. Pan, M.F. Chen, P.J. Cheng, Y.M. Hwang, S.D. Tseng, J.C. Huang, Fabrication of gapless dual-curvature microlens as a diffuser for a LED package, Sens. Actuators A 150 (009) Biographies K.H. Liu was born in Nantou, Taiwan, Republic of China, in He received his Masters degree in 1997, from Graduate Department of Industrial Technology Education, National Kaoshiung Normal University in Kaoshiung, Taiwan. He was a researcher in the field of mechanical and manufacturing Engineering and MEMS Division. He joined R.O.C. Military Academy, Kaohsiung, Taiwan, Republic of China, as a lecturer in Dr. M.F. Chen was born in Tainan, Taiwan, Republic of China, in He received his Doctorate in 009, from Department of Mechanical and Electro-Mechanical Engineering of National Sun-Yat-Sen University in Kaohsiung, Taiwan. He was a researcher in the field of MEMS, LIGA process and optical units. He joined National Sun Yat-Sen University, Kaohsiung, Taiwan, Republic of China, as a post-doctoral fellowship in 009. Prof. C.T. Pan was born in Nauto, Taiwan, Republic of China, in He received his Masters Engineering degree and Doctorate in 1993 and 1998, respectively, from Power Mechanical Engineering Department of National Tsing Hua University in Hsinchu, Taiwan. He was a researcher in the field of laser machining polymer in the TU Berlin (IWF) in Germany from 1997 to 1998 and a researcher of MEMS Division in the MIRL/ITRI, Hsinchu in Taiwan from 1998 to 003. He joined National Sun Yat-Sen University, Kaohsiung, Taiwan, Republic of China, as an assistant professor in 003. His current research interests focus on MEMS, NEMS, and LIGA process.