Demonstration of a simplified optical mouse lighting
module by integrating the non-Lambertian
LED chip and the free-form surface
Jui-Wen Pan
1,2,3,* and Sheng-Han Tu
41Institute of Photonic System, National Chiao Tung University, Tainan 71150, Taiwan
2Biomedical Electronics Translational Research Center, National Chiao Tung University, Hsin-Chu, 30010 Taiwan 3Department of Medical Research, Chi Mei Medical Center, Tainan 71004, Taiwan
4Genesis Photonics Inc., No. 5 Dali 3rd Road, Tainan Science-Based Industrial Park, 741, Taiwan *Corresponding author: juiwenpan@gmail.com
Received 5 December 2011; revised 28 March 2012; accepted 30 March 2012; posted 2 April 2012 (Doc. ID 159429); published 15 May 2012
A cost-effective, high-throughput, and high-yield method for the efficiency enhancement of an optical mouse lighting module is proposed. We integrated imprinting technology and free-form surface design to obtain a lighting module with high illumination efficiency and uniform intensity distribution. The imprinting technique can increase the light extraction efficiency and modulate the intensity distribution of light-emitting diodes. A modulated light source was utilized to add a compact free-form surface ele-ment to create a lighting module with 95% uniformity and 80% optical efficiency. © 2012 Optical Society of America
OCIS codes: 350.3950, 230.3670, 220.4000, 220.2945.
1. Introduction
Recently, GaN light-emitting diodes (LEDs) have at-tracted plenty of interest due to their favorable char-acteristics, such as high flux efficiency, reliability, low power consumption, long life, and environmental friendliness [1,2]. These LEDs have already been ex-tensively adopted in several practical applications, for example as backlight light sources for flat panels, in mobile projectors, for street lighting, traffic sig-nals, automotive lighting, and general lighting [1,2]. However, the ratio of refractive index between GaN (n 2.5) and air (n 1.0) is 2.5, which causes con-siderable photons to be trapped inside the GaN chip due to the total internal reflection (TIR), where they are then absorbed and converted into heat. Fortu-nately, the external quantum efficiency of GaN-based
LEDs can easily be improved by enhancing the es-cape probability of the trapper photons. Methods that can increase the external quantum efficiency include reducing the surface roughness [3], utilizing photonic crystals [4], patterning the sapphire sub-strate [5], and imprinting techniques [6,7]. These methods introduce a relative roughness interface be-tween the air and the LED chip so that the opportu-nity for photons to escape from inside the GaN chip increases.
Agilent was the first company to develop an optical mouse with an LED light source. The conventional LED mouse optical module included two subparts. One was the image subpart, which consisted of an imaging lens and a complementary metal-oxide semiconductor (CMOS) sensor, and the other the il-lumination path. The imaging lens constructed a con-jugation relationship between the CMOS sensor and the active area. Following the illumination path, the illumination lens would collect the light emitted from
1559-128X/12/152834-08$15.00/0 © 2012 Optical Society of America
the LED chip, simultaneously homogenizing the dis-tribution of the illumination over the active area. The illumination lens also converted the incident angle of the emitting light so that it impinged on the illumi-nated area from 17° to 22° for the production of a high-contrast image [8]. In a conventional illumina-tion lens, the double TIR and three exit facets were adopted to produce a homogenized illumination dis-tribution throughout the active area [9]. However, there is decay in the reflection efficiency when the double TIR surface is contaminated by dust or parti-cles. Moreover, the double TIR mechanism limits the angle of ray propagation and reduces the total optical efficiency of the illumination lens to less than the manufacturer’s recommended tolerance [10]. An-other characteristic of the conventional illumination lens is the un-uniformity on the active area. In a con-ventional illumination lens, the double TIR condition is utilized to produce a virtual image. The design of the three exit facets ensures that the virtual images are superimposed on the illuminated surface [11]. However, the uniformity produced by the illumina-tion lens is only about 65%; moreover, the three exit facets also produce three hot spots that reduce the illumination uniformity. There is an additional loss of optical efficiency owing to the two hot spots outside the conjugation area of the image sensor. A reduction in the efficiency loss and increase in the uniformity both lead to improved performance of the illumina-tion lens for adopillumina-tion in the optical mouse.
In this study, we demonstrate a novel method that satisfies both demands for an illumination lens for the optical mouse. The imprinting technique intro-duced after the chip fabrication process helps to in-crease the light extraction efficiency of the LED by 14% and modulates the near-field pattern at the chip level to match the peak intensity of 22° needed for optical mouse applications. Furthermore, a corre-sponding optical element designed for a free-form surface is also integrated with the imprinted LED chip to optimize the illumination of the lens. The uni-formity on the illuminated surface was raised to 95% after removal of TIR loss and three hot spots. Most important of all, both the imprinting technique and the free-form illumination lens are cost-effective, having the high throughput, compact size, and high yield needed for mass production. Finally, the novel design for free-form surface illumination was fabri-cated and tested. All optical performance results were consistent with the optical simulation results. Clearly, the novel illumination system offered im-proved illumination efficiency and image contrast. 2. Imprinting and Chip Process
The intensity pattern of the planar LED chip has the characteristics of a Lambertian light source: a peak intensity located in a direction normal to the lighting surface, and a half intensity at60° relative to the normal axis [12]. The structure imprinted on the pla-nar LED chip surface cause a shift in the Lambertian properties.
The chip fabrication process included mesa defini-tion and etching, the addidefini-tion of a transparent con-ductive layer (TCL) and depositioning of the Cr-Au electrode pad [13]. After the chip fabrication process, a Si mold and spin-on-glass (SOG) were used to fab-ricate a one-dimensional (1D) grating structure on the top of the TCL. The period of this grating struc-ture was 6 μm with a filling factor of 0.5, and the height was about 385 nm. Figures1(a)and1(b)show the optical microscopic (OM) and atomic force micro-scopic images of the LED chip with the imprinted structure. The imprinted structure, as depicted in Fig. 1(c), has a serrated profile with an asymmetri-cally blazed structure. The blazing structure has
Fig. 1. (Color online) Microscopic images and geometric depiction of the imprinted structures: (a)50× optical microscope picture of the imprinted LED chip; (b) AFM picture of the 1D blazed grating structure; (c) 1D blazed grating geometric scheme.
two different angles, 10° and 50°, between each-sidewall and the bottom surface.
3. Electrical and Optical Performances
The results of the electric performance of the im-printed and planar LED chips are shown in Fig. 2. Figure2(a)shows the applied voltage versus current (I-V) curve measured in the probe stage. The forward
voltage of the imprinted LED was 3.4 at a current of 20 mA, similar to that of the planar LED. Although the imprinting process provided a high temperature and pressure environment, the electrical properties of the LED chip were not affected by the imprinting process. Even with an additional pressing and an-nealing processes, the forward voltage remained at 3.4 V. Figure2(b)shows the injection current versus output power (L-I) curve measured by an integral sphere. Figure2(b)indicates an enhancement in the output power of 14% for the imprinted LED, com-pared with that of the planar LED at an input current of 20 mA. The enhancement in the output power was a result of the insertion of medium index material and surface roughness. The indexes were gradually changed effectively by the insertion of an SOG surface texture layer, because the refractive index of the SOG was 1.5. As a consequence, there was a reduction in light trapping from Fresnel loss. The other reason for the light output enhancement was due to the enhancement of escape probability of trapped photons by the imprinted surface texture. Near-field measurement of the LED chip is used to obtain the near-field intensity distribution [14]. Figure 2(c) shows the intensity distribution results from 1D blazed grating LED chips with scanning an-gles of 180°,−180°, and 90° to the imprinted grating period. The imprinted LED shows a non-Lambertian light with a shift in the peak intensity in both the 180° and −180° scanning direction. Consequently, when the scanning direction is 90°, a Lambertian-like pattern, like that of a planar LED chip, is presented. 4. Optical Design and Simulation
There are many different technologies used in this field for moving sensing applications, especially to meet the specifications for an LED optical mouse [9]. The image contrast is affected by the incident an-gle between the light source and the illuminated sur-face. Generally, the best incident angle is around 17° to 22° [8]. A conventional illumination lens has three functions: first, collection of the emitting light; second, homogenization of the light distribution; and third, control of the incident angle in the active area. In this study, the incident angle is controlled by the blazing grating fabricated on the LED chip surface. The illumination lens is designed to enhance the illumination homogenization and the collection efficiency.
In order to attain these two goals, we use a free-form surface (FFS) with compact size. The Ad-vanced System Analysis Program (ASAP) blazer curves are adapted to construct the free-form surface, which can be described as anm by n tensor productPu; v, as shown in Eq.1 [15].
Pm
i0Pnj0PijWijBmi uBnjv
Pm
i0Pnj0WijBmi uBnjv ; 1
wherePij,Wijis a set ofm 1n 1 control points and weight factors; Bmi u 1−ui!m−i!m−1uim! is the
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0 10 20 30 40 50 Current (mA) Voltage (V) Planar LED 1D blazed grating 0 20 40 60 80 100 0 1 2 3 4 5 6 7 Power (mW) Current (mA)
Planar LED chip 1D blazed grating
(a)
(b)
(c)
Fig. 2. (Color online) The optical performance of imprinted and planar LEDs: (a) applied voltage versus current curve; (b) injection current versus luminous curve; (c) intensity distributions of imprinted LEDs for 180°,−180°, and 90° measurement.
Bernstein polynomial; and each control point has three values,x, y, z, in the Cartesian coordinate sys-tem. The m 1n 1 control points and weight factors are used to control the surface normal vector of each local unit surface of the free-form surface. Then we can modulate the local ray impinging on the local unit surface of the free-form surface to the active area on the illumination surface.
In order to obtain the two aforementioned goals, two kinds of FFSs are used to homogenize the illumi-nation distribution and collect the emitted light [16,17]. They are shown in Fig.3. The radiation pat-tern from the LED chip with the imprinted micro-structure can be divided into two regions by the optical axis. In our system, the optical axis is set as thez global axis. One region is under the optical axis with the peak intensity, and the other is above the optical axis. The four FFSs can also be divided into two parts. One part has two green surfaces for
collecting the intensity distribution underneath the optical axis and homogenizing the illumination dis-tribution on the target area. The two green surfaces are symmetrical with thez-x plane owing to the 1D blazing grating. The green surface consists of7 × 13 control points and weight factors for m 6 and n 12. The other part is marked by two red surfaces for collecting the light above the optical axis and homogenizing the illumination distribution on the il-luminated area. The two red surfaces are also sym-metrical with thez-x plane owing to the 1D blazing grating. The red surface consists of 7 × 13 control points and weight factors for m 6 and n 12. The green and red FFSs with control points for the two regions are shown in Figs.3(a) and3(b).
Owing to the division of the illumination system into two FFSs, there are also two optical paths. Along the first optical path, the emitting ray passes through the green surface and then impinges on the active area. The rays coming from the first optical path comprise the main intensity controlled by blazed grating. First, the merit function is used to calculate7 × 13 control points and 7 × 13 weight fac-tors for the green surface with uniformity and the peak intensity angle. The green surface is the com-mon surface for the two optical paths; thus, we de-signed this surface first. After finishing the green surface, the other surface was designed. The other free-form surface controls only the second optical path, which is comprised of emitted rays collected by the red surface, then passing through green sur-face, until finally impinging on the active area. In the second step, the 7 × 13 control points and 7 × 13 weight factors of the red surface are also calculated. Two sets of control points and weight factors are used in the merit function to modulate the green and red FFSs to control the emitted light. The merit func-tions adopted in this paper are the uniformity and the peak intensity at 22° on the illuminated surface (1 mm × 1 mm active area with black rectangle) as follows: For the peak intensity in the active area:
Tmaxθ ∼ maximum; (2)
when θ 22°. For uniformity in the active area (1 mm2 area):
UniformityIcen Iavg
×100%: 3
For the total optical efficiency: E
P
p;qIp;q×Ap;q
Wtotal
×100%: 4
For the angle space at the active area, we obtained the intensity distribution function of Tθ, so as to find the peak intensity Tmaxθ. The merit function was used to control the two sets of control points and the weight factors to locate the peak intensity at 22° in the active area. In other words, the peak
Fig. 3. (Color online) Free-form surface with control points: (a) green surface; (b) red surface.
intensity was 68° with the surface normal of the ac-tive area. To find the spatial intensity distribution over the active area, we divided the active area into ap by q mesh with both p and q being integers. Each mesh had local irradianceIp;q. TheIcen is the irradi-ance of the mesh at the center of the active area; the Iaveis the average irradiance at the active area.Wtotal
is the total power emitted from the LED chip. The optimization process is divided into three main steps. In the first step, we try to simplify the LED source by using point source. The angle bution of the point source is apodizated by the distri-bution function shown in Fig. 2(c). By using this simple light source, we can calculate the start point and end point for each local unit surface of green sur-face. After the first step, we calculate only the start point and end point for the free-form surface. The
control point is set at the middle of the start point and the end point for each local unit surface. The weight factor is 1 for all local unit surfaces. With the optimization condition at this step, we only use Eq. (2) to get the correct angle distribution and the peak intensity angle.
In the second step, we try to add eight point sources with apodization at the x-y plane by three times. The first time, we add the two points at (0.1, 0) and −0.1; 0, then do the optimization pro-cess with only the start point and end point for each local unit surface. The second time, we add the two points at (0, 0.1) and0; −0.1 then do the optimiza-tion process with only the start point and end point for each local unit surface. The final time, we add the four points at (0.1, 0.1),−0.1; 0.1, −0.1; −0.1, and 0.1; −0.1 and then do the optimization process with
only start point and end point for each local unit surface. With the optimization condition, we use Eqs. (2–4) to get the correct angle distribution and uniformity distribution with high efficiency.
In the third step, we try to use the extended light source (0.2 mm × 0.2 mm) with apodization to re-place the night point sources. We still optimize the start point and end point once again. Then we use the control point and weight factor to do the optimi-zation and freeze all parameters of the start point and end point of all local unit surfaces. Finally all parameters, including start point, end point, control point, and weight factor, were used to do the optimi-zation with feedback function for uniformity. The feedback times are eight in optimization code [18].
The final optimization results for the four FFSs are shown in Fig4(a). A representation of the ray tracing of the four FFSs is shown in Figs.4(b),4(c), and4(d). The energy distribution and radiant intensity distri-bution at the illuminated surface are shown in Fig5. The incident angle at the peak intensity is 68° with the normal vector of the illuminated surface. A high-contrast image is obtained with the image lens with this incident angle [8]. The uniformity at the active area (1 mm × 1 mm active area with back rectangle) is 95%. The optical efficiency is 80%. The results of a comparison of our design with the conventional de-sign are shown in Table1. An examination of the ta-ble shows that we get better performance in terms of optical efficiency, uniformity, and compact size with our device. The tolerance is the key manufacturing issue for optical systems. We also compared the two systems for LED misalignment issues for mass
Table 1. Comparison of the Traditional and Novel Illumination System
Illumination system
Conventional illumination system [9] (light pipe
and three facets at exit port)
Four FFS system (four FFSs) Optical Efficiency 40% 80% Uniformity 65% 95% Alignment method Horizontal Horizontal Size (mm3) 16 mm3 4 mm3 Total internal
reflection
2 0
Fig. 5. (Color online) Radiant intensity distribution and energy distribution of the illumination distribution with the active area.
Table 2. Comparison of Given Alignment Tolerance for the Two Illumination Systems
Misalignment of LED
Conventional illumination system [9] Four FFS system Optical efficiency Uniformity Optical efficiency Uniformity Decenter alongx-axis* 0.05 mm 36% 63% 75% 89%
−0.05 mm 35% 64% 73% 90%
Decenter alongy-axis 0.05 mm 36% 64% 73% 89%
−0.05 mm 37% 63% 74% 90%
Decenter alongz-axis 0.05 mm 38% 64% 78% 91%
−0.05 mm 40% 65% 77% 90%
Rotation aroundx-axis** 2° 34% 61% 73% 88%
−2° 35% 60% 72% 87%
Rotation aroundy-axis 2° 33% 62% 71% 89%
production. The results are shown in Table2. The op-tical efficiency and uniformity of our design was bet-ter than that of the traditional design. We proposed a highly efficient and compact volume optical motion sensor. Even under the worst alignment conditions, the optical performance of our device was better than that of the conventional system. Finally, a working sample of the novel free-form surface lens was fin-ished, as shown in Fig. 6(a). An optical mouse was assembled as well. A CMOS camera was used to cap-ture the illumination distribution. The illumination distribution is shown in Fig.6(b). The simulation re-sults for the illumination distribution are consistent with the CMOS picture. The CMOS sensor intensity showed the intensity to be enhanced 1.9 times over that of the conventional illumination system. The result is also similar to the tolerance analysis. In a word, the novel free surface system is better than the traditional illumination system.
5. Conclusions
In this study we applied a cost-effective, high-throughput, and high-yield imprinting technique for fabrication of an LED chip. The imprinted layer not only increased the light extraction efficiency of the LED chip by 14%, but also modulated the far-field pattern on the chip level. A tilted intensity dis-tribution was achieved by using this imprinted layer. Furthermore, the corresponding free-form surfaces were designed for an optical mouse application. The uniformity at the illuminated surface was 95%. The optical efficiency was 80%. The characteristics of the four free-form surfaces included the compact size, low fabrication cost, and the total internal reflection loss was excluded in our lighting module. The im-printed LED and optical element design are inte-grated to obtain an illumination system with a higher optical efficiency, more uniform intensity dis-tribution, and more compact size. Finally, a sample of the novel free-form surface illumination system was finished and tested. The test results show a perfor-mance very similar to the simulation results.
This study was supported in part by the National Science Council, project number NSC100-2220-E-009-023 and NSC101-2314-B-384-001, and in part by the “Aim for the Top University Plan” of the National Chiao Tung University and Ministry of Education, Taiwan, Republic of China.
References
1. E. F. Schubert and J. K. Kim,“Solid-state light sources getting smart,” Science 308, 1274–1278 (2005).
2. M. S. Zukauskas, M. S. Schur, and R. Gaska, Introduction to Solid State Lighting (Wiley-Interscience 2002).
3. T. Fuji,“Increase in the extraction efficiency of GaN-based light-emitting diodes via surface roughening,” Appl. Phys. Lett.84, 855 (2004).
4. K. Orita,“High-extraction-efficiency blue light-emitting diode suing extended-pitch photonic crystal,” Jpn. J. Appl. Phys. 43, 5809–5813 (2004).
5. M. Khizar, Z. Y. Fan, K. H. Kim, J. Y. Lin, and H. X. Jiang, “Nitride deep-ultraviolet light-emitting diodes with microlens array,” Appl. Phys. Lett. 86, 173504 (2005).
6. Z. M. Wang, X. J. Lio, S. Wang, C. X. Lio, M. H. Sun, K. Bao, B. Zhang, G. Y. Zhang, Y. G. Wang, Y. Chen, H. Ji, and Q. Ouyang, “Light output enhancement of a GaN-based light emitting diode by polymer film imprinting,” Semicond. Sci. Tech. 22, 279 (2007).
7. K. Bao, X. N. Kang, B. Zhang, T. Dai, C. Xiong, H. Ji, G. Y. Zhang, and Yong Chen,“Improvement of light extraction from patterned polymer encapsulated GaN-based flip-chip light emitting diodes by imprinting,” IEEE Photon. Tech. Lett. 19, 1840–1842 (2007).
8. B. J. Tullis,“Method and device for achieving high contrast sur-face illumination,” U.S. patent 5686720, (11 November, 1997). 9. Avago (Agilent),“HDNS-2100 solid-state optical mouse lens data sheet,”http://www.avagotech.com.tw/docs/AV02‑1278EN. 10. R. E. Fisher and B. Tadic-Galeb, Optical System Design
(McGraw-Hill, 2000).
11. G. E. Smith,“Illumination optics and method,” U.S. patent 6476970, (5 November, 2003).
12. E. F. Schubert,“The Lambertian emission pattern,” in Light-Emitting Diodes (Cambridge University, 2006), pp. 94–95. 13. S. H. Tu, S. Y. Wu, Y. C. Lee, J. W. Pan, C. H. Kuo, C. M. Wang,
and J. Y. Chang,“Functional imprinting structures on GaN-based light-emitting diodes for light pattern modulation Fig. 6. (a) FFS lens sample; (b) illumination patent captured by
and light extraction efficiency enhancement,” Opt. Rev. 17, 379–384 (2010).
14. Radiant Zemax, “SIG-400TM,” http://www.radiantimaging .com/index.php?q=products/sig400.
15. “Advanced System Analysis Program technical guide,” Edges in ASAP,http://www.breault.com/k‑base.php?kbaseID= 29&catID=44&page=2.
16. J. W. Pan, S. H. Tu, C. M. Wang, and J. Y. Chang,“High effi-ciency pocket-size projector with a projecting lens and light
emitting diode-based light source system,” Appl. Opt. 19, 3406–3414 (2008).
17. J. W. Pan, S. H. Tu, W. S. Sun, C. M. Wang, and J. Y. Chang, “Integration of non-Lambertian LED and reflective optical element as efficient street lamp,” Opt. Express 18, A221– A230 (2010).
18. Y. Luo, Z. Fang, Y. Han, and H. Li,“Design of compact and smooth free-form optical system with uniform luminance for LED light source,” Opt. Express 18, 9055–9063 (2010).
