A 25-pixel illumination system composed of a 5 × 5 dielectric liquid-lens (DLL) zoom module array, 25 light-emission diodes (LEDs), and a secondary optical lens demonstrates 3D light field manipulation. LEDs function as 2D illumination pixels while the DLL module array performs longitudinal illuminance adjustability by zooming each illumination pixel. A test on the similarity of two illuminance patterns between experiments and simulations shows a normalized cross correlation (NCC) higher than 0.8, indicating the feasibility of the system design. Also, the illumination system is further applied to correct a distorted light pattern on a 45° tilt screen as well as to perform light compensation on distance-differential objects.
© 2011 OSA
Light has been extensively used in our daily lives since the invention of the incandescent light bulb by Thomas Edison. The illuminance of a point-of-light source is inversely proportional to the travel distance of light, squared. Such an illumination phenomenon induces a non-uniform illumination on distance-differential objects (the objects with different distances from the light source) so that the object closer to the source appears brighter. Highly controllable illumination is still desired to improve quality of life and comfort via room atmosphere, decoration, captured image quality, public security, and so forth. A typical solution to the problem of increasing light controllability is to introduce a lens module to focus, magnify, or zoom the projected beam of light, providing longitudinal adjustability of illuminance (i.e., one-dimensional light field controllability) [1,2]. Still, non-uniform light distribution could be improved to some extent, but it remains an issue. To further enhance planar light field controllability, the pixel concept of illumination, usually achieved by an LED array, was introduced to permit individual adjustment of the outgoing light intensity based on each pixel . To really achieve spatial light manipulation, our approach is to implement zoom ability on each pixel of an LED array.
In the study, we demonstrate the integration of dielectric liquid zoom lenses with 2D illumination LED pixels to form a 3D illumination system where the design is primarily based on paraxial ray approximation using OSLO and TracePro simulation software. To equip focus or zoom ability, methods that could be employed include mechanically tuning the spacing between solid lens modules [1,2], redistributing the refractive index distribution of liquid crystals [4–7], and deforming liquid-lens profiles [8–17]. Among them, liquid lenses driven by electrowetting or dielectric forces is one promising solution because of low power consumption, no moving parts, high transparency, and special miniaturization potential of large pixel systems [11–17]. A small form factor along with high power efficiency could lead to portable illumination systems in the future. The illumination system proposed demonstrated light controllability, including spatial light pattern manipulation, correction of distorted light patterns, and a 3D flashlight for the camera. The experimental data such as illuminance patterns were captured and compared with simulation predictions via normalized cross correlation (NCC) to estimate the similarity of light patterns and to show the feasibility of the illumination system design .
2. The 3D illumination system
The 3D illumination system consists of a 5 × 5 dielectric liquid-lens zoom module array, 25 LEDs, and a secondary optical lens shown in Fig. 1(a) . LED illumination pixels assembled in a 2D array in combination with zoom liquid lenses provide light controllability over space for each single module. Figure 1(b) depicts the composition of a single module that contains one LED, two face-to-face dielectric liquid lenses (P1 and P2), and a spacer in between the two liquid lenses. The face-to-face assembly permits larger beam convergence and more light transmission. The dielectric liquid lenses, the core components of the system, realize the zoom ability for each pixel as the focal length of each liquid lens is adjusted via applied voltages. The voltages biased at the liquid lenses are controlled by a microprocessor PICI18F4550 (Microchip Technology Inc., USA) that converts digital signals to voltages we want via pulse width modulation and interface of an inter-integrated circuit (I2C). LED power is adjusted using a digital I/O controller, manipulating the light intensity of each pixel. A secondary optical lens is employed to reduce excessive overlapping of the light beams so that the light pattern can be distributed with higher uniformity.
The system for dynamic light projection onto treasure, jewelry, and antiques with 3D appearance is designed to fully control light patterns in the range of 30 to 60 cm from the illumination system (i.e., within 2X distance variation). The design is primarily based on paraxial ray approximation using OSLO and TracePro simulation software. The LED (Everlight Electronics, 59-146UWD/TR8) has a spot diameter of 2 mm and a viewing angle of ± 25.0°. Each single module has a divergence angle varying from ± 14.0° to ± 7° under a dielectric liquid-lens (DLL) voltage of 40 Vrms, achieving angular magnification of two times and illuminance variation of four times. The spacing between the LED and P1 and that between P1 and P2 are 0.72 mm and 5.22 mm, respectively. The module design is symmetric along the x and y axes. The pitch of adjacent liquid lenses varies a bit to satisfy the secondary lens. For row assembly, the separation gap between the first and second row/column and between the fourth and fifth row/column (close to the outer edges of the secondary lens) is a distance of 14.63 mm. The separation gap between the second and third row/column and third and fourth row/column is a distance of 15.93 mm. The secondary optical lens made of BK7 is designed to have a radius of −294.12 mm, a lens thickness of 15 mm, and a diameter of 50 mm.
The system assembly started from each single module in which LEDs were integrated with two liquid lenses and one spacer. Two liquid lenses were axially aligned face-to-face with the insertion of a spacer in between them using a laser aligner, minimizing tilt and de-center aspects. The lens set was then individually mounted onto a LED-embedded PCB as shown in Fig. 2(a) , forming a single module. Each single module was plugged into the control circuit board. The secondary optical lens was positioned in front of the 5 × 5 module array. After assembly, the 3D illumination system is shown in Fig. 2(b).
3. Experimental results
3.1 Performance of dielectric liquid lenses
Dielectric liquid lenses consist of two iso-density immiscible liquids that have a refractive index difference of 0.11 and a dielectric constant difference of 43 [15–17]. Mixed alcohol and silicone oil are used as the sealing liquid and the liquid droplet, respectively. The droplet profile along with the refractive index difference determines the optical powers of the liquid lenses. Each liquid lens has a volume of 5 mm3, corresponding to the droplet diameter of 4.7 mm at the rest state. As DLL voltage is applied, the dielectric force produced on the interface of the two liquids squeezes the droplet, resulting in the change of the effective focal length (EFL). When the applied DLL voltage increased from zero bias to 40 Vrms, the corresponding EFL varied from 53.8 mm to 19.5 mm for the liquid lenses (see Fig. 3 .). The response time of the DLL was measured to be 400 ms for the full-range tuning. Each single module containing two face-to-face liquid lenses had a divergence angle of ± 14.0° and ± 7.0° when both of the two liquid lenses were set at zero bias and 40 Vrms, respectively, obtaining 2X angular magnification (14/7 = 2) and 4X illuminance variation.
3.2 Light pattern similarity test
The illumination system assembled was first investigated by a similarity test to verify the spatial light controllability. The similarity of two illuminance patterns can be estimated by normalized cross correlation (NCC) . The test was conducted by projecting the letter H onto the screens at 30 cm and at 60 cm from the system, as shown in Fig. 4 . To form the letter H at 30 cm away, all LEDs were turned on except the ones in the gray region (see Fig. 4). The DLL voltages at the corner pixels 1, 5, 21, and 25 were set to be 25 Vrms to converge the light at the corners, and the DLL voltages of other pixels at 6, 10, 11–15,16, and 20 were set at zero bias. For the screen located at 60 cm, the DLL voltages of all LED-on pixels were changed to 40 Vrms to increase the illuminance. The H patterns at the two different distances were measured to have nearly the same geometrical dimensions––20 cm in height and 20 cm in width. MATLAB software was used to convert captured light patterns at 30 cm and 60 cm to illuminance patterns by eliminating hue and saturation information . The illuminance pattern was calibrated with respect to the central illuminance of the light pattern measured by the lux meter (LUTRON, LX-1108). Illuminance at the center of the pattern H was directly measured to be 91 and 84 lux at 30 cm and 60 cm, respectively. The resultant illuminance difference was about 7.7%, much smaller than the illuminance decay of 75%, due to the 2X distance variation in nature. The NCC between the two experimental illuminance patterns was calculated to be about 0.9, showing the system can project similar light patterns from 30 to 60 cm. As compared with simulation predictions, the experimental illuminance patterns at 30 or 60 cm have an NCC higher than 0.8 shown in Table 1 , implying the feasibility of the illumination system design.
3.3 Correction of light pattern distortion on 45° tilted screen
The illumination system proposed was further applied to restore a distorted illumination pattern on the tilted screen (see Fig. 5 ). The DLL bias setting of each pixel initially remained unchanged like the one at 30 cm away, as stated in the previous session. As the screen was rotated by 45°, the light pattern projected was inevitably distorted, particularly in the region away from the rotation axis. Compared with the reference pattern at 30 cm (see Fig. 4), the distorted illuminance pattern had a low NCC value of 0.14. By increasing the DLL voltage to 40 Vrms for pixels 5, 10, 13–15, 20, and 25, the distortion reduced; the corrected illuminance pattern with the improved NCC value of 0.64 had a central illuminance of 85 lux and an illuminance difference of about 6.6%. This case implies the potential of zoom ability for each pixel to enable the system to correct the light pattern on a tilted screen. As compared with simulation predictions, the experimental illuminance patterns before or after modification had an NCC higher than 0.65 as shown in Table 2 .
3.4 Lighting compensation demonstration
For lighting applications, the illumination system was used to compensate for unwanted light distribution for a better image capture. Desired illumination is sometimes difficult to obtain with commercial flashlights. For instance, illumination merely on one side of objects (e.g., due to sunset or sunrise or in an unmanageable environment) might cause some pictures to be overexposed or underexposed. A lighting test was performed on a tiger doll positioned at 30 cm away. The test-bed setup is shown in Fig. 6 . Initially, it was illuminated by a commercial flashlight on the right-hand side, as shown in Fig. 6 (a). The doll image that was captured appeared to be dimmer at the left-hand side due to insufficient illumination. To increase illuminance properly, LEDs at pixels 1, 2, 6, 7, 11, 12, and 16-25 were turned on with the DLL voltage set at zero bias. With this setting, the features of the doll on the left-hand side displayed clearly, as shown in Fig. 6(b). The other plan was to investigate the capability of simultaneous lighting compensation on distance-differential objects (see Fig. 7 ). Two doll images, a Mickey Mouse and a princess, were positioned at 30 cm and 60 cm away from the illumination system, respectively. Four LEDs at pixels 9, 12, 14, and 17 were turned on with the DLL voltage set at zero bias. Illuminance decay by the inverse-square law made the princess doll dimmer by 75% as compared to the Mickey Mouse doll. To have equal illuminance, the voltage of liquid lenses at pixels 9 and 14 were increased to 40 Vrms to converge the light beams onto the princess doll. As a result, the princess doll appeared to be equally bright with the Mickey Mouse doll, each one having benefitted from the zoom ability of each pixel.
Obtaining 3D light controllability is greatly desired in order to enhance convenience and comfort in our daily life. We integrate dielectric liquid zoom lenses with 2D illumination LED pixels to form a 25-pixel, 3D illumination system. The illumination system was used to execute a light pattern similarity test, distorted light pattern correction, and lighting compensation demonstration. From the investigation of NCC between experimental data and simulation predictions, the system has an NCC higher than 0.8 in the spatial light pattern test and an NCC close to 0.7 in the image correction test, showing the feasibility of our illumination system design.
References and links
1. J. R. Richardson, “Motor-controlled lens system,” US Patent 5,029,992 (July1991).
2. A. Uke and S. Wright, “Multi-lens zoom system and method for flashlights,” US Patent App. 11/668,605 (February 2007).
3. P. Schreiber, S. Kudaev, P. Dannberg, and U. D. Zeitner, “Homogeneous LED-illumination using microlens arrays,” Proc. SPIE 5942, 59420K (2005). [CrossRef]
4. T. Nose, S. Masuda, and S. Sato, “A liquid crystal microlens with hole-patterned electrodes on both substrates,” Jpn. J. Appl. Phys. 31(Part 1, No. 5B), 1643–1646 (1992). [CrossRef]
7. C. C. Cheng, C. A. Chang, C. H. Liu, and J. A. Yeh, “A tunable liquid-crystal microlens with hybrid alignment,” J. Opt. A, Pure Appl. Opt. 8(7), S365–S369 (2006). [CrossRef]
8. B. Berge and J. Peseux, “Variable focal lens controlled by an external voltage––an application of electrowetting,” Eur. Phys. J. E 3(2), 159–163 (2000). [CrossRef]
9. T. Krupenkin, S. Yang, and P. Mach, “Tunable liquid microlens,” Appl. Phys. Lett. 82(3), 316–318 (2003). [CrossRef]
10. S. Kuiper and B. H. W. Hendriks, “Variable-focus liquid lens for miniature cameras,” Appl. Phys. Lett. 85(7), 1128–1130 (2004). [CrossRef]
15. C. G. Tsai, C. N. Chen, L. S. Cheng, C. C. Cheng, J. T. Yang, and J. A. Yeh, “Planar liquid confinement for optical centering of dielectric liquid lenses,” IEEE Photon. Technol. Lett. 21(19), 1396–1398 (2009). [CrossRef]
17. C. C. Yang, C. G. Tsai, and J. A. Yeh, “Miniaturization of dielectric liquid microlens in package,” Biomicrofluidics 4(4), 43006 (2010). [CrossRef]
18. 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]
19. S. C. Chapra, Applied Numerical Methods with MATLAB for Engineers and Scientists (McGraw-Hill Higher Education, 2008).