Open Access
Add to BibSonomyAbstract
In this study, we presented a light-emitting diode-based (LED) miniaturized optical pattern imager for slim mobile phone application. To meet volume constraints, we designed a miniaturized compact illuminating and imaging optical component. The objective was to minimize optical loss using several nano- and micro-fabrication methods. After integration into a single optical body, the prototype imager–with dimensions of 6.8 × 2.2 × 2.5 mm and a weight of 0.4 g–demonstrated clear feasibility in measuring 2D micropatterns with widths of 50 and 10 μm.
©2010 Optical Society of America
1. Introduction
Light-emitting diodes (LEDs) are effective lighting sources for optical illumination systems in portable multimedia devices such as mobile phones and navigators. LEDs are not only lightweight, but consume less power. Such advantages have convinced manufacturers to adopt LEDs as the preferable light source for commercialized optical devices, e.g., fingerprint detectors, miniaturized projectors, as well as the optical mouse and scanners. Thus, the design and manufacture of LED-based optical devices have become one of the most heavily researched topics relating to optics in recent years [1–4]. In particular, manufacturers have integrated such devices into a single platform within mobile phones to offer multifunctional operation while still maintaining small and slim features. The same has been done for the optical mouse, mobile projectors, and illumination displays. Further, chip scale package (CSP) LEDs play an important role in such devices, offering optical design flexibility due to extremely minute features [5]. Since manifesting many common parameters in optical design, the projector and imager–among such an array of devices–have attracted our interest. As a result of miniaturization, such LED-based optical devices face many technological issues, e.g., beam collimation, beam homogenization, étendue limitation, and minimum optical loss [6–9].
where θ is the half-angle of angular beam distribution.Figure 1 shows the similarities between an LED-based projector and imager within optical systems. In miniaturized projectors or optical imaging devices, imagers require a small illuminated area and restricted angular input, both of which result in relatively small conserved optics, i.e., étendue, as described in Eq. (1). An efficient optical illumination component collects large angular light coverage, converting it to smaller angles; the light engine–the guide for illumination optics–shapes such light spatially to fit the imaging area while conserving étendue. In mobile phones, geometric dimensions for optical imaging systems have decreased as demands for image quality have increased. Unfortunately, as étendue of the imager optics reaches to maximum limit, total optic volume grows. Although several studies have been conducted on miniaturized LED-based optical systems–particularly fingerprint scanners–mechanical volume with thicknesses more than several tens of millimeter render slim mobile phones inadaptable [10,11]. Furthermore, image quality, based on image-detection of scattered light from fingerprints, is not much higher than frustrated total internal reflection light. Focusing on such volume constraints, i.e., étendue limit, we aimed to design and fabricate various types of étendue-limited compact illumination and imaging optics that are viable in a miniaturized LED-based projector and imager. Integration of such optical components–manufactured by various types of nano- and micro-machining and fabrication processes into a single optical body–can minimize volume consumption and assembling tolerance.
Fig. 1 Schematic of a LED-based étendue-limited optical system: miniaturized pocket projector (left) and imager (right).
In this paper, we proposed a miniaturized 2D optical imager for application in a slim mobile phone. We examined the proposed imager and experimentally verified corresponding optical performance. Through such a challenging study, we were able to show that the design methodology of high-quality LED-based compact illumination optics and imaging lenses can provide small étendue with minimum loss. To meet volume constraints for slim mobile phones, we employed various types of nano- and micro-fabrication methods to realize a thin specialized optical profile. After assembly of several optical components for jig measurement integration, the fabricated miniaturized 2D imager clearly demonstrated feasibility when measuring 10 μm of width in micropatterns with extremely small features.
2. Optical design of the miniaturized 2D optical imager
Figure 2 illustrates the schematic and ray path of the proposed 2D optical imager. Light generated from a LED chip was extracted from the CSP profile. The LED beam-shaping lens (BSL) reduced large illumation angles and quasi-uniformly redistributed light. Further, the integrated custom microprism effectively steered shaped LED beams. Although half of the light flux initially reflected off of the aluminum-coated surface–refracting a certain incident angle, to the imaging area (Aimager)–the other half was refracted only within the microprism, resulting in lost rays. Although lost ray luminous flux was as high as 50%, the microprism consumed extremely small mechanical volume, steering light and providing enough brightness on the imaging area via optical path length reduction of several millimeters. Imaging objects absorbed or partially scattered some light; still, other light was reflected due to total internal reflection (TIR). Thus, only micropattern input or fingerprint by frustrated TIR fingerprints–as well as incidents related to array Fresnel imaging lens diffraction–were reflected on the TIR surface. The image was finally formed in the image sensor by means of a microscope. Subsequently, a charge-coupled device (CCD) sensor was able to detect not only bright regions where air gaps were located, but darker regions where skin or other materials with an appropriate refractive index contacted polymethyl methacrylate (PMMA). In addition, the microscope could have been replaced with a complementary metal-oxide-semiconductor (CMOS) image sensor in the actual unit.
An imaging optical system requires high sensor luminance with a high contrast ratio. The efficiency of each optical component should contribute to final sensor luminance. Image brightness (Isensor) of the image plane can be defined as follows:
For a given imaging area (Aimager) and sensor size (Asensor), image sensor brightness can be enhanced by either increasing luminance of the light source (LLED) or by improving illumination efficiency for optical components (ηprism and ηBSL), as shown in Eq. (2). Although increased imaging area–accomplished by increasing systematic étendue limits–directly leads to overall increments of imager height, such increments are unacceptable for volume-constrained optical systems. Therefore, controlling numerical aperture (NA) of the imaging lens is the only variable option for increasing optical efficiency.
Image contrast ratios rely on light absorption and scattered imaging object characteristics. Optical gain (Gimager) describes imaging object capacity of light reflection in particular directions vis-a-vis perfectly absorbed diffusing mediums, i.e., fingerprints. Accordingly, extremely thin blue CSP LEDs were adopted due to high absorption in relation to human skin. Given such parameters, image quality perceived by sensors was fully estimated during approximation of uniformly illuminated object areas.
2.1 Design for illumination optical components
The miniaturized 2D optical imager consisted of five optical components: a CSP LED, a LED BSL, an integrated custom microprism with LED illumination steering, a light guide, and an array diffractive Fresnel imaging lens. The optical design of such an imager was highly challenging because optical component incorporation–from the light source to the imaging lens–had to complement each other within extremely small mechanical volume. Optical illumination software was used because ultra slim configuration rendered the system highly dependent on light source radiation patterns. Such details were addressed within component design methodology.
Since light distribution patterns immediately above the LED surface were not able to be approximated with either the extended surface light or point source, miniaturized optical component performance was likewise not predictable without concrete LED optical characteristic analysis [12,13]. In this region, CSP LED is ideally considered as a volume light source due to the generation of light from multiple quantum wells (MQWs). To minimize light source volume, the thinnest packaged CSP LED, currently made by the Kingbright co. ltd. (Model no. APG1608QBC/D; λ = 466 ± 20 nm; 0.495 lumen, 3.1V, 20 mA) and consuming only 1.6 ´ 0.8 ´ 0.25 mm of volume, was chosen. The physical dimensions of the chosen CSP LED are shown in Fig. 3(a) .
Fig. 3 (a) Physical dimensions of the LED and the precisely modeled LED, (b) molded LED and measurement coordinates, (c) comparison of light distributions between molded LED and actual LED in polar coordinates, tracing of a hundred million rays.
As the distance between light source and BSL was five times less than source diameter, i.e., “near-field” in illumination optics, the LED was designed according to volume light source and optical performance, as well as chip packaging [12–15]. We previously demonstrated a LED fabrication method by Monte Carlo ray tracing that functions almost identically as commercial LEDs [16]. Figure 3(b) shows the molded CSP LED including MQWs, a chip, electrodes, and wires with measurement coordinates.
For LED model verification, angular luminous intensity distribution was measured experimentally, as demonstrated in Fig. 3(c). To determine accurately the similarity between simulated light distribution patterns and those in the measured set-up, normalized cross correlation (NCC) was applied [16–18]. As high as a 97% rate of correlation existed between measured and modeled illumination profiles, as previously reported [16].
Based on such results, the modeled source in this study adequately complemented the actual LED. Thus, modeled LEDs sufficiently described the optical characteristics of gallium nitride-based (GaN) blue CPS LEDs within an acceptable variation range.
Due to the specialized minute features of the imager, as well as low luminous CSP LED flux, luminous intensity was amplified within the acceptance angle (θ), while Lambertian-like light distribution was homogenized by a beam shaping lens with extremely small features [Fig. 2]. We previously demonstrated a near-field beam shaping lens (NBSL) for GaN-based blue CSP LEDs [16,19]. The NBSL displays effective homogeneity, as well as narrow illumination angles, within the CSP LED. For miniaturized 2D optical imager application, we adopted similar design methodology for NBSL, having to modify it to accommodate a much thinner CSP LED than those used in the previous study. Since CSP LED height was only two-thirds of the previous one, we focused on top CSP LED emission as the dominant light emission, thus reducing central aspheric lens focal length from 0.5 to 0.4 mm.
The design resulted in illumination angle reduction from 120° to 14° in full width at half maximum (FWHM). For given étendue of the source, the NBSL was theoretically able to collect 74.6% of light, however collecting only 60% from the CPS LED during simulation. An extra loss of 14.6% was caused either by Fresnel loss at entrance and exit interfaces or distribution mismatch. Hence, the NBSL amplified usable light flux within θ and helped to integrate the custom microprism during homogeneous illuminance distribution within the imaging area. The ratio of center versus edge brightness at 3.4 millimeters, i.e., optical path length from the rear NBSL surface to the imaging area, was about 79.4% during simulation.
The maximum invariant in an optical system is limited by the optical component with the smallest étendue. Thus, restricted étendue of the imager directly affects light guide design, as well as imaging lenses. To meet étendue requirements for illuminated planes, i.e., imaging area, the light guide was set not only to provide sufficient luminous flux, but restrain the shaped beam within a smaller luminous area (Aimager) within the incident angle (ν). The geometric design requirements of the light guide addressed height minimization (H), as well as appropriate ranges of the imaging area. An integrated microprism is ideally able to steer shaped LED illumination in incident angles. Silicon-based bulk micromachining facilitated such a design requirement by means of a particular etched angle and single-sided aluminum coating [20].
Equation (3) shows the light-steering characteristics of the microprism. The incident angle (ν) on the imaging area depended on the NBSL acceptance angle (θ). Equation (3) demonstrates the imaging (Aimager) and input areas of the imaging lens (Alens), also depending on the NBSL acceptance angle (θ) for a given imager height. Thus, the incident angle (ν), as well as limited imager height, possibly restricted imaging area size (Aimager), i.e., the shorter system height, the smaller imaging area. Therefore, efficient light guide design represents an important consideration in minimizing imager height within limited mechanical volume, as the following formula demonstrates:
Due to the NBSL and an integrated custom microprism with LED illumination steering, as well as optimized light guide design, average luminous intensity within the acceptance angle was much higher than such intensity from the original light source. Thus, image brightness from the image plane was able to be enhanced for a given imaging area.
In such a case, we chose a height of 1.5 mm for the light guide, while the maximum imaging area of the imager stood at about 2.8 mm2. In Eq. (4), the input diameter of the imaging lens was about 0.9 mm, with an image compression ratio (Rimage) of about 0.32, suggesting that an effective imaging area can enlarge despite original image distortion. However, such images can be easily restored when image processing technology is integrated into the imaging system.
2.2 Design of the imaging lens
Restricted étendue of an imaging area requires higher NA from the imaging lens, as shown in Eq. (2). Thus, the key design parameter of the imaging lens is represented by short focal length and increased lens aperture. If the radius curvature of the lens enlarges, i.e., short focal length, a single lens will be unable to cover large imaging areas. Limitations on lens diameter, as well as thin lens profile fabrication difficulty, are overcome by adopting an array diffractive Fresnel imaging lens. Figure 4(a) illustrates rays originating from an imaging area during frustrated TIR incidence of an individual Fresnel imaging lens, being formed on the imaging sensor with several incident angles. As shown in Fig. 4(a), the thickness of each Fresnel imaging lens (d, wavelength over refractive index difference from the PMMA to the air) is calculated using a simple diffraction lens equation, as well as imaging lens diameter, i.e., the number of lens arrays being directly calculated [21]. The array diffractive Fresnel imaging lens in the study had a focal length of 500 μm, a diameter of 250 μm, and a thickness of 0.94 μm. The 3 × 3 array diffractive Fresnel imaging lens, with 330 μm of duration, was chosen for this study.
Fig. 4 (a) Schematic of a diffractive Fresnel imaging lens design, (b) MTF performance of the designed imaging lens.
The modulation transfer function (MTF) value was representative of optical performance for the diffractive Fresnel imaging lens, as shown in Fig. 4(b). Although the graph was plotted according to a plane wave assumption, effective image quality–which simply means a higher MTF value–was discernible during calculation. For example, we detected values greater than 0.7 MTF for images with 500 μm of focal length by using a Fresnel imaging lens with 50 μm patterns; for 10 μm patterns, we detected 0.3 MTF values near the 9.5° range. Using a thin lens equation in air, image plane location was calculated, as shown in Eq. (5) [21]. In this case, we revealed 695 μm when the curvature radius was set as follows: C1 = 0, C2 = 4.
2.3 Ray tracing simulation for a LED-based miniaturized 2D optical imager
LightTools optical simulation software (Optical Research Associates, CA, USA), capable of analyzing various types of illumination systems, was used to estimate designed miniaturized 2D optical imager performance. Figure 5(a) shows overall system configuration in LightTools after integration of each designed optical component. Figure 5(b), illustrating the simulation results, clearly shows that corresponding illumination distribution of the binary 20 μm pattern on the imaging area was well-imaged in the receiving detector. Absorption-induced Lambertian light-scattering of the micropattern stood at about 8%, similar to a fingerprint.
Fig. 5 (a) Designed miniaturized 2D optical imager in the optical software, (b) image formed on the detector, (c) illuminance distribution on the imaging area without imaging object, (d) total luminous flux variation on the imaging area based on assembly errors. (a hundred-million rays being traced during simulation)
Light distribution uniformity within the imaging area represents another critical imager requirement. Figure 5(c) shows imaging area illuminance via working rays. A lateral cross section of brightness distribution revealed uniformity at 73% and longitudinal direction at 63%. Most of the incidents of lost rays with less than critical angles, i.e., about 42°, on the imaging area involved refraction with the imaging area, thus not being accounted for in actual imaging uniformity.
Such results verified designed imager validity. Over total luminous flux of the CSP LED during simulation, ηBSL and ηprism stood at about 60% and 40%, respectively. Although such illumination efficiencies seem to be relatively large, imaging area brightness was still large enough for clear imaging due to steered LED beam illumination with optical path lengths of 3.4 mm in the small imaging area (2.8 mm2). Meanwhile, 30% of CSP LED luminous flux–along with some lost rays–was transferred onto the imaging area. Inevitable assembly error between the NBSL and CSP LED resulted in decreased illumination in the imaging area, as shown in Fig. 5(d). Assembly tolerance is ideally controlled at less than 0.1 mm in order to contain luminous flux loss at about 25%.
3. Manufacturing of the mold
To achieve required optical performance with respect to volume constraint, various nano- and micro-fabrication methods, i.e., electrical discharge machining (EDM), silicon-based electron beam lithography, and ultra-precision machining, were combined.
We have previously demonstrated the manufacture and fabrication of a diffractive Fresnel imaging lens, as well as an integrated custom microprism [22,23]. Using the identical method, an array diffractive Fresnel imaging lens and an integrated custom microprism were fabricated separately on a silicon substrate, afterward being bonded onto a metal core with high temperature-durable thermal curable resin [Fig. 6(a) ]. The bottom core, machined by EDM [Fig. 6(b)], was polished several times to ensure surface quality (Rq = 10 nm) for use with TIR [Fig. 6(c)]. Total Integrated Scattering (TIS) on the imaging area, based on Eq. (6), stood at less than 2% [24]. Thus, 98% of lights with incident angles (v) were accurately reflected due to TIR; the majority of the lost rays smaller than the critical angle were refracted out onto the imaging area.
Fig. 6 (a) Assembled fabricated optical component on silicon wafer in the top core, (b) surface quality of the bottom core after EDM, (c) enhanced surface quality after mechanical polishing.
3.1. Manufacturing of the LED beam shaping lens
NBSL geometry is very complex, consisting of a series of specialized prismatic facets and a central aspheric lens. For practical application of such complex geometric designs, we adopted Single Point Diamond Turning (SPDT), addressing extremely miniaturized features in unison with machined NBSL. Machined NBSL provided a profile error of less than 1 μm of deviation, surface quality for TIR being sufficient on the outer facets [Fig. 7(a) ]. Further, the CSP LED illumination angle of the machined lens was reduced from 120° to 16° in FWHM [Fig. 7(b)]. LED–compared to original CSP LED–luminous intensity within the acceptance angle was increased ten-fold [25]. Measured efficiency of total NBSL luminous flux stood at about 58.5%.
Fig. 7 (a) Photography of machined NBSL and microscope image of facets ( × 50), (b) optical performance of the CSP LED with/without NBSL in logarithmic.
3.2 Fabrication of the integrated optical system
Although fabrication techniques such as silicon-based electron beam lithography for imaging lens and machined NBSL with SPDT may not be scalable for large volume production, such techniques can be used to make individual mold templates in thermal nano-imprint lithography. Systematic and repeated use of such templates possibly facilitates practical device manufacture [26,27].
We previously examined fabrication of integrated custom microprisms and array diffractive Fresnel imaging lenses for mold-based thermal silicon nano-imprints [22,23]. The method was expanded in the present study to include metallic mold bonding of fabricated optical components onto silicon wafers. The mold of the miniaturized 2D optical imager was placed within thermal nano-imprint equipment, controlling the temperature and pressure inside both top and bottom cores [Fig. 8(a) ]. Inside-core temperature was monitored by a thermal coupler, any gas being released by the gas release pass during mudding. The overall mold area–excluding optical components–was coated with polytetrafluoroethylene (PTFE) in order to obtain not only the highest operating temperature, but an extremely low friction coefficient. Such coating prevented any adhesion-driven surface damage during fabrication. The integrated custom microprism and imaging lens were well-fabricated with very fine nano-and micro-scale patterns [Fig. 8(b)].
Fig. 8 (a) Assembled mold set with hot embossing equipment, (b) fabricated optical system by thermal nano-imprint (4 Mpa, 160° C).
After fabrication of the integrated optical imager body, reflective mirror coating was executed on one side of the microprism by thermal evaporation via an electron beam, providing directionality for the evaporated aluminum [Fig. 9(a) ]. For uniform and selective aluminum PMMA coating, the fabricated sample was aligned precisely at a specific tilted angle (δ) of 50.89°. Figure 9(b) shows the schematic for tilted angle calculation as determined by Eq. (7). Coated aluminum thickness of the reflection area stood at about 800 nm with a 80Å/s- deposition rate, as follows:
4. Measurement of micro pattern images by the miniaturized 2D optical imager
Due to extremely miniaturized volume, assembly of optical components onto a single module became an issue in determining imager optical performance. To manufacture assembly parts, rapid prototyping was adopted during assembly jig fabrication, manifested as complex geometry for the prototype. The assembled miniaturized 2D optical imager with two assembling jigs is shown in Fig. 10(a) . The NBSL and CSP LED was positioned precisely within the bottom jig. Integrated optics assembled by the top jig was aligned precisely at the assembling guide, being assembled with the bottom jig by means of hooks. Since the assembled imager had a tendency to flip over, it was secured by the measurement frame. Figure 10(b) shows all assembled parts prepared for measurement, as well as an enlarged image of all assembled component and jigs in side view.
Fig. 10 (a) Assembled and fabricated optical components with measurement jigs, (b) image tester setup of an LED-based micro pattern recognizer and side view of all assembled optics and jigs.
The formed image using the imager was verified in relation to the original image taken by microscope of the same sample. The well-imaged extruded grid sample, made by copper, was sized at 50 × 50 μm and placed on a copper plate, as shown in Fig. 11(a) . As only extruded grids absorbed light and in-between valleys, the pattern reflected the light well. The acquired image was darker at the grid and brighter at the valleys, as shown in Fig. 11(b). The central area of the images was clearer, both in simulation and measurement, than the edge because relatively small MTF values were obtained as light reached the sensor margins.
Fig. 11 (a) Physical dimension of the target specimen, (b) formed images on image plane for target specimen through microscopes.
To verify minimum recognizable line width, a 10 μm pattern width with a 100 μm duration [Fig. 12(a) ] was imaged using the imager. As shown in Fig. 12(b), the total number of periods measurable in a single lens was approximately seven. Thus, total computed imaging width of the three rows of imaging lens was about 1.8 mm. Although absorption and scattering, as well as fabrication or assembly errors between simulation and measurement, possibly existed, acquired images from the proposed imager supported theoretical analysis and simulation results.
Fig. 12 (a) Physical dimension of the micropattern, (b) formed images on an image plane as the target specimen via a microscope.
The amount of light scattered from the object and reaching the imaging lens was dependent on the reflecting characteristics of not only the target object, but corresponding wavelength. In addition, absorbed fingerprints less likely scattered blue light, possibly enhancing the contrast ratio of the acquired fingerprint image. Thus, the miniaturized 2D optical imager was applicable not only for pattern imaging, but also fingerprint detection for devices with extremely small features.
5. Conclusion
In this study, we proposed a LED-based miniaturized 2D optical imager for application in slim mobile phones. To meet volume constraints, high-quality LED-based compact illumination optics, as well as an imaging lens, were designed and fabricated using a specialized thin profile to minimize optical loss and maintain optical performance. To achieve required optical performance with volume constraints, i.e., étendue limitation, different nano-and micro-fabrication methods, such as EDM, electron beam lithography and ultraprecision machining, were adopted. By means of thermal nano-imprinting, nano- and micro patterns were effectively fabricated. Machined NBSL enhanced usable flux within the acceptance angle; the light guide–with an integrated custom microprism including LED illumination steering–provided homogeneous luminance distribution with restricted angular input in the imaging area. By assembling the optical components with measurement jigs, we successfully produced a prototype of a miniaturized fingerprint imager with dimensions of 6.8 × 2.2 × 2.5 mm and a weight of 0.4 g. LED-based miniaturized 2D optical imager feasibility for slim mobile phones was finally demonstrated by imaging 50 μm and 10 μm micropatterns. With the high image quality of the imager and high absorption rate of fingerprint for the blue light, the imager was applicable not only for pattern imaging, but also fingerprint detection for slim mobile phone application with extremely small features.
Acknowledgement
This study was supported by the Korea Science and Engineering Foundation–National Research Laboratory Program under grant no. 20100018458, as well as the Core Technology Development Program for Next-generation Solar Cells in the Research Institute for Solar and Sustainable Energies (RISE).
References and links
1. I. Fujieda, K. Tai, E. Matsuyama, and M. Kurita, “Dual-LED imaging system for secure fingerprint identification,” Proc. of SPIE 6049, 60490B–1- 60490B–9 (2005).
2. http://www.crucialtec.com/.
3. X. J. Yu, Y. L. Ho, L. Tan, H. C. Huang, and H. S. Kwok, “LED Based Projection Systems,” J. Disp. Technol. 3(3), 295–303 (2007). [CrossRef]
4. R. Iijima, M. Fujimoto, Y. Maruo, M. Nakatsuka, T. Inui, S. Abe, and Y. Unuma, “Document scanner using polymer waveguides with a microlens array,” Opt. Eng. 41(11), 2743–2748 (2002). [CrossRef]
5. L. Nguyen, “Wafer-Level chip-scale packaging,” in Proceedings of Professional Development Course 55th Electronic Components & Technology Conference, (Orlando, Florida, 2005), pp. 4–19.
6. P. Schreiber, S. Kudaev, P. Dannberg, and A. Gebhardt, “Microoptics for homogeneous LED-illumination,” Proc. SPIE 6196, 61960–1 - 61960–9 (2006).
7. Y. Ding, X. Liu, Z. R. Zheng, and P. F. Gu, “Freeform LED lens for uniform illumination,” Opt. Express 16(17), 12958–12966 (2008). [CrossRef] [PubMed]
8. F. Fournier and J. Rolland, “Design Methodology for High Brightness Projectors,” J. Disp. Technol. 4(1), 86–91 (2008). [CrossRef]
9. X. Zhao, Z. L. Fang, J. C. Cui, X. Zhang, and G. G. Mu, “Illumination system using LED sources for pocket-size projectors,” Appl. Opt. 46(4), 522–526 (2007). [CrossRef] [PubMed]
10. T. Imamura, K. Kakutani, and I. Fujieda, “Imaging simulated smeared fingers with a sensor based on scattered-light detection,” Opt. Rev. 15(3), 143–147 (2008). [CrossRef]
11. K. Tai, E. Matsuyama, M. Kurita, and I. Fujieda, “Dual-LED imaging for finger liveliness detection and its evaluation with replicas,” Appl. Opt. 45(24), 6263–6269 (2006). [CrossRef] [PubMed]
12. C. C. Sun, W. T. Chien, I. Moreno, C. C. Hsieh, and Y. C. Lo, “Analysis of the far-field region of LEDs,” Opt. Express 17(16), 13918–13927 (2009). [CrossRef] [PubMed]
13. I. Ashdown, and M. Salsbury, “A Near-field Goniospectroradiometer for LED Measurements,” Proc. SPIE 6342, 634215–1 - 634215–11 (2006).
14. F. Hu, K. Y. Qian, and Y. Luo, “Far-field pattern simulation of flip-chip bonded power light-emitting diodes by a Monte Carlo photon-tracing method,” Appl. Opt. 44(14), 2768–2771 (2005). [CrossRef] [PubMed]
15. I. Moreno and C. C. Sun, “LED array: where does far-field begin?” Proc. SPIE 7058, 70580R (2008). [CrossRef]
16. J. Y. Joo, C. S. Kang, S. S. Park, and S. K. Lee, “LED beam shaping lens based on the near-field illumination,” Opt. Express 17(26), 23449–23458 (2009). [CrossRef]
17. C. C. Sun, T. X. Lee, S. H. Ma, Y. L. Lee, and S. M. Huang, “Precise optical modeling for LED lighting verified by cross correlation in the midfield region,” Opt. Lett. 31(14), 2193–2195 (2006). [CrossRef] [PubMed]
18. I. Moreno and C. C. Sun, “Modeling the radiation pattern of LEDs,” Opt. Express 16(3), 1808–1819 (2008). [CrossRef] [PubMed]
19. J. Y. Joo and S. K. Lee, “Miniaturized TIR Fresnel lens for miniature optical LED applications,” Int. J. Precis. Eng. Manuf. 10(2), 137–140 (2009). [CrossRef]
20. N. Maluf, and K. Williams, An Introduction to microelectromechanical systems engineering (Artech house Inc., Norwood, 2004).
21. F. L. Pedrotti, and S. Leno, Introduction to optics (Prentice-Hall, New Jersey, 1987).
22. D. K. Woo, K. Hane, and S. K. Lee, “Fabrication of a multi-level lens using independent-exposure lithography and FAB plasma etching,” J. Opt. A, Pure Appl. Opt. 10(4), 044001 (2008). [CrossRef]
23. D. K. Woo, K. Hane, and S. K. Lee, “High order diffraction grating using v-shaped groove with refractive and reflective surfaces,” Opt. Express 16(25), 21004–21011 (2008). [CrossRef] [PubMed]
24. J. S. Fender and J. E. Harvey, “Specifying surface finish and scattering tolerances of conical optical elements,” Opt. Eng. 21, 983–986 (1982).
25. J. Y. Joo, C. S. Kang, S. S. Park, and S. K. Lee, “Fabrication of a beam shaping lens for chip scale packaged LEDs,” International Conference on Precision Engineering 2010 (to be published).
26. S. H. Ng and Z. F. Wang, “Hot embossing on polymethyl methacrylate,” Singapore Inst. Manuf. Technol. Rep. 8(3), 174–177 (2007).
27. M. T. Gale, “Replication techniques for diffractive optical elements,” Microelectron. Eng. 34(3-4), 321–339 (1997). [CrossRef]
