Open Access
Add to BibSonomyAbstract
GaN LED array in packaging is important for the demand of high optical flux. In order to tighten the whole packaging size, the spacing among LED chips becomes an important factor in the packaging design. This study is to investigate the change of the light extraction when a GaN LED chip array packaging is applied. The shielding effect with various spacing for the GaN LED array with or without silicon encapsulation is obtained. We apply the Monte Carlo ray-tracing method in the simulations to analyze the optical behavior of the two major types of the GaN LED array. The shielding effect is more dominant for bare chip packaging. When a silicone thin dispensing layer is applied, the shielding effect is not obvious because of more light extraction, the neighbor dies play an important role in photon recycling.
© 2015 Optical Society of America
1. Introduction
Recently light emitting diodes (LEDs) have been regarded as the most promising solid-state light source in next-generation lighting owing to its advantages in more energy saving, more design freedom, better color performance, higher reliability, and more environmental benefits [1–10].One of the obstacles to spread LEDs into practical applications is the flux density. It arises from the lumen per watt which performed by a high-power white LEDs is lower than some traditional light sources. To solve the problem, increasing optical throughput with advanced optics is a common approach to an illumination system. If a heavy injection current cannot afford the flux required by an optical system, the only solution is to enlarge the area of the LED chip. This approach, however, also enlarges the etendue of the light source, and the optical efficiency of the optical system could drop. Therefore, how to tighten the LED array is one of the important topics. Numerous studies have provided thermal analytical results to determine the optimum pitch on the design of the LED array [11–20]. The luminous efficacy LED arrays are affected by the heat dissipation levels, available spacing, and lens encapsulation [21]. Generally, larger spacing between LED chips yields higher luminous efficacy due to better heat dissipation. The shortage is that the etendue is also enlarged. In contrast, small spacing in the LED array could perform small light extraction owing to possible shielding effects, which describes the effect of the emitting light re-entering the neighboring chips. Therefore, a study to determine optimal spacing for the LED array packaging to achieve maximum light extraction is demanded. This study primarily performs simulations and experimental analysis on GaN LEDs to determine their light extraction efficiency (LEE) with tightened array spacing. The objective is to investigate the LEE changes when the spacing of the GaN LED chips is changed. All simulations and the corresponding experiments are presented and demonstrated.
2. Simulation of LED chip array
2.1 Wire-bonding array
The simulation of the LEE is based on the Monte Carlo ray-tracing [22–32]. Figure 1 shows the geometry of the blue GaN chip in the simulation. The lateral dimensions of the selected LED chip are 1185 × 1185 μm. To simplify the simulation conditions, the electrode reflectivity is set at 40%, the reflectivity of the bottom reflector is set at 90%, and the refractive index of the encapsulated silicone is 1.5.
Fig. 1 The wire bond chip structure: (a) image by scanning electron microscopy (SEM); (b) the simulation structure by the advanced systems analysis program (ASAP)
To enhance the light extraction for a wire-bonding LED, patterned sapphire substrates (so called PSS) with a microstructure array are always introduced as shown in Fig. 2. The PSS is composed of pyramid array with a fill factor of 46%. The filling factor represents the ratio of the cyclic microstructure area, which is adjusted using microstructure spacing. The side view of the PSS are shown in Fig. 2(c). The key parameters of the LED chip are shown in Fig. 2(d), where indium tin oxide (ITO) is used as the transparent and electricity-conducting layer. In the simulations, because the blue-ray nearly completely penetrates the ITO, the blue-ray absorption of the ITO is not considered. Regarding the epitaxial layer, because the actual absorption coefficient of the chip active layer is unknown (causing an undetermined absorption effect magnitude), this study refers to the literature to obtain the actual chip reference values. In the simulations, the absorption coefficient is set at 2000 cm−1 and 1000 cm−1 for the active layer and the p-GaN layer, respectively.
Fig. 2 Structure of the wire-bonding GaN LED: (a) geometry of epitaxy layers; (b) top view of the PSS; (c) side view of the PSS as in (b);(d) the simulation parameters used for wire bond and flip chip LEDs.
2.2. Thin-GaN array
The Cree thin-film EZ-1000 blue-chip LED is used as the experimental sample. The LEDs chip is developed to create a thin-GaN LED with a texture surface. The selected dimensions are based on the size of the EZ-1000 blue-chip LED, which are 980 × 980 μm. In the simulation, the electrode shielding effect on the LEE is considered, and the electrode reflectivity is set at 40%. Figure 3 shows the LED structure. To simplify the simulation conditions, the complex electrode scattering behavior is disregarded, and the reflective mirror surface with 90% of reflectivity is used as the reflective layer at the bottom. Because the chip design is also focused primarily on light extraction, a microstructure is developed on the thin-GaN LEDs chip to increase the LEE, as shown in Fig. 3. This study not only refers to relevant literature but also determines the epitaxial parameters and active layer absorption coefficient that are added to the simulation. In the simulation, the active layer absorption coefficient and the encapsulated epoxy refractive index are set at 200 cm−1 and 1.5, respectively.
Fig. 3 (a) A photo ofthe GaN chip of EZ-1000. (b) The simulation parameters of the thin-GaN with surface texture.
Figure 4(a) shows the surface structure image of the Cree EZ-1000 chip. A chemical etching process is used to produce surface microstructures. The size, direction, and the inclination angle of the surface structures are dissimilar and randomly distributed. The inclination angles are estimated to range between 58° and 63° based on the obtained images. In the simulations, the inclination angle of the pyramids is set at 60°. The pyramids comprise randomly arranged hexagonal pyramids with sizes ranging between 0.1 and 3 μm. To use the randomly distributed and various-sized microstructure pyramids in the simulations, this study divides the irregular surface structures into three microstructure groups, as shown in Fig. 4(b). Because of the hexagonal pyramidal shapes, pyramids that differ in size produce a gap (space), and triangular pyramids are used to fill these gaps. To enable random distribution of the pyramids, the three dissimilarly sized pyramids are used randomly to fill a “unit area.” Then the unit areas are used to fill the die surface until the fill factor reaches 100%. Figure 4(c) shows the ratio of each pyramid type in the microstructures (in percentages). The simulated LEE with (without) microstructures is 68.53% (19.56%).
Fig. 4 (a) A photo of surface texture structure of the Cree EZ-1000 chip. (b) Top view of the texture structure. (c) The parameters of different microstructures in the simulation. The cover area indicates the filling factor of each structure in the whole surface.
3. Simulation and the corresponding experiment
This study presents simulation as well as experiment to determine the effects on the LEE caused by changing LED chip spacing with or without silicon encapsulation.
In the first case, we study the wire-bonding structure. Two 12-inch integrating spheres are used to measure and confirm the radiant flux of the central chip in the six packaging samples, which are shown in Fig. 5, where the average LED spacings in each chip are 45, 95, 137, 193, 292, and 384 μm, separately. The geometry of the array packaging is shown in Fig. 6, where Fig. 6(b) shows the thin silicone dispensing layer covering the LED array to extract the light. The thin silicone is aimed to change the refractive index of the spacing. If the silicone is a dome shape, a different effect will happen in lighting extraction [21]. In order to describe the shielding effect, we define a flux ratio between the single chip packaging and the multiple-chip packaging (simplified FRSM). The FRSM is the ratio between the flux by a LED chip located at the center of a chip array and the flux by a single LED chip. In the experiment, only the central chip in the 9-in-1 array is turned on. The FRSM is smaller than 100% owing to some lights going into the neighbor chip and resulting in additional loss in comparison with a single chip. The measured FRSM of the six samples and the corresponding simulation are shown in Fig. 7. In the corresponding simulation, the reflectivity of the bottom surface (R) in the simulation varies from 50% to 100%. The experimental measurement fits the condition with reflectivity of around 85%, which coincides with the practical experience. From the measurement, we find that in the case without lens encapsulation, the shielding effect is not obvious when the spacing reaches 200 μm and above. When the lens encapsulation is applied to the packaging, the shielding effect is not obvious. The reason is that silicone encapsulation extracts more lights from the top surface rather than in the sidewalls.
Fig. 5 Images of the wire bond array chips with varying LEDs spacing, (a) without silicone dispensing, and (b) with silicone dispensing.
Fig. 6 Schematic diagrams of the multi-chip packaging, (a) without silicone dispensing, and (b) with silicone dispensing.
Fig. 7 The experimental measurement and the corresponding simulation in the case of wire-bonding with different reflectivity of the bottom surface (R) of the flux ratio between the single chip packaging and multiple-chip packaging, (a) without silicone dispensing, and (b) with silicone dispensing.
The second case is for the thin-GaN arrays, and the average LED spacings in the seven samples in experiment are 1, 15, 24, 46, 56, 67, and 93 μm, respectively, as shown in Fig. 8. The measured flux ratios of the seven samples and the corresponding simulation are shown in Fig. 9, where reflectivity of the bottom surface in the simulation varies from 70% to 90%. From the simulation and the corresponding experimental measurement in the both cases with and without lens encapsulation, the shielding effect is not obvious until the spacing is reduced to 10 μm or less. The reason for less shielding effect is that the effective emitting surface of a GaN chip in a thin-GaN structure only exists on the top surface, because the thin-GaN LED does not contain sapphire substrate. Thus the probability of photons entering the neighbor chips is low and the change in LED spacing has no significant effect on the overall luminous efficacy.
Fig. 8 Images of the seven thin-GaN arrays with different LED chip spacing, (a) without silicone dispensing, (b) with silicone dispensing.
Fig. 9 The experimental measurement and the corresponding simulation in the case of thin-GaN with different reflectivity (R) of the flux ratio between the single chip packaging and multiple-chip packaging, (a) without silicone dispensing, and (b) with silicone dispensing.
4. Discussion
The study in both simulation and experimental measurement shows that the spacing effect is an important factor for light extraction for multi-chip packaging. However, the effect is different in different LED structures. In the case of bare chip array, around 10% enhancement of FRSM is observed when the spacing is large enough. In the case with silicone dispensing, the enhancement is reduced to 5% or even less. The reason is that the silicone dispensing changes the surrounding refractive index and the escaping light cone in the chip is enlarged. However, the silicone dispensing layer could form a cavity to enhance the interaction among the neighbor chips. An advanced analysis of light extraction for the two packaging structures is shown in Fig. 10, where the blue bar is for the LEE by the light directly emitted from the central chip or reflected by the bottom reflector. The shielding effect is shown with red and green bars. In the wire-bonding case, the side walls of the sapphire form effective light escape window in the chip so that the side-emitting lights cannot be neglected. Some of the lights emit from the side walls, entering the neighbor chip, and finally escaping from the chip from the side wall or the top surface. The second effect is caused by the thin coating silicone layer, which may form a narrow escaping cone and reflects parts of the light to the neighbor chips as shown in Fig. 11. If the silicone coating is a dome shape, the reflected light will become less, and it induces less effect on the light extraction by the neighbor chips. From Fig. 10, we can find that more lights are directly from the emitting chip when the spacing increases, and meanwhile, the lights from the side wall of the neighbor chips decrease while the lights from the top surface almost keep constant. In the case of the thin-GaN, the spacing effect is not obvious. Both the lights from the emitting chip or the neighbor chips are almost unchanged. However, there are around 30% of escaping lights from the top surfaces in the neighbor chips. Thus we can find that 20% to 30% of LEE is affected by the neighbor chips in both packaging structures. If the absorption is heavy or other optical effects cannot be neglected, the shielding effect could play a more important role in multi-chip packaging. In such a condition, photon recycling [33,34] in packaging design should be paid attention.
Fig. 10 The detailed analysis for light extraction with silicone dispensing in the case of (a) wire-bonding and (b) thin-GaN. The dash line indicates the LEE by the emitter with single chip packaging.
The shielding effect could be different in the chip size and the location. In the size issue, we simulate four sizes, including 237 μm, 593 μm, 1185 μm, and 1422 μm, while the other packaging parameters are kept the same. The size effect is not obvious in the case of thin-GaN. In the case of wire-bonding, the smaller size will have larger FRSM and smaller shielding effect, but the difference is within 10% in FRSM.
The location of the chip could have different shielding effects, which could become less in the edge chip since the neighbor chip is less. The simulation shows that the location effect is not obvious in both wire-bonding and thin-GaN cases when the spacing is larger than 50μm. The difference of FRSM in the central chip and the edge chip is around 1% to 2% in the wire-bonding case, and can reach up to 1% when the spacing is shorter than 50 μm in the thin-GaN case. When the chip is near the silicone border of the package, more backward light from the interface between the silicone and the air will cause less FRSM.
5. Conclusion
This paper is to present a study of the shielding effect when multiple-chip packaging is applied, which is highly demanded in high-power luminaire with high optical flux and less etendue. The GaN LED chips are classified into wire-bonding and thin-GaN, and the main difference between two types of LED chips are the side wall windows. The studied packaging is a 3 × 3 array with or without thin silicone dispensing layer. In the simulation and corresponding experiment, the analysis is done and measures the light extraction in comparison with the packaging with a single chip.
In order to be close to the practical application, the wire-bonding chip is with the pattern sapphire substrate and the thin-GaN is with surface texture. Both structures are precisely constructed in the simulation. In the case of wire-bonding, the side wall plays a more obvious role in the shielding effect rather than in the case of thin-GaN. The experimental measurement shows that 9% and 5% of FRSM drops when the spacing is decreased from 400 μm to 45μm without/with silicone dispensing. In the case of thin-GaN, the shielding effect is not obvious except that the spacing is as short as 1μm or less in both cases of with and without silicone dispensing.
In the advanced analysis of LEE with silicone dispensing, the neighbor chips always have effects on the LEE even the spacing is large. There are many lights falling into the side walls or top surface of the neighbor chips. Therefore, the absorption of other optical effects in the neighbor chips could affect the light extraction, and photon recycling is important and needs more attention.
Acknowledgment
This study was supported in part by the National Central University’s “Plan to Develop First-class Universities and Top-level Research Centers” (Grant numbers 995939 and 100G-903-2), the Ministry of Science and Technology of the Republic of China (contract numbers 99-2623-E-008-002-ET, 100-2221-E-008-088-MY3, and NSC101-2221-E-008-108 and 103-2221-E-008-063-MY3). The author would like to thank BRO for providing ASAP simulation tool.
References and links
1. C. P. Kuo, R. M. Fletcher, T. D. Osentowski, M. C. Lardizabal, M. G. Craford, and V. M. Robbins, “Highperformance AlInGaP visible light emitting diodes,” Appl. Phys. Lett. 57(27), 2937–2939 (1990). [CrossRef]
2. H. Sugawara, M. Ishikawa, and G. Hatakoshi, “High-efficiency InAlGaP/GaAs visible light-emitting diodes,” Appl. Phys. Lett. 58(10), 1010–1012 (1991). [CrossRef]
3. A. Žukauskas, M. S. Shur, and R. Gaska, Introduction to Solid-State Lighting (Wiley, 2002).
4. T. Fujii, Y. Gao, R. Sharma, E. L. Hu, S. P. DenBaars, and S. Nakamura, “Increase in the extraction efficiency of GaN-base light emitting diodes via surface roughening,” Appl. Phys. Lett. 84(6), 855–857 (2004). [CrossRef]
5. M. G. Craford, “LEDs for solid state lighting and other emerging applications: status, trends, and challenges,” Proc. SPIE 5941, 594101 (2005). [CrossRef]
6. E. F. Schubert and J. K. Kim, “Solid-state light sources getting smart,” Science 308(5726), 1274–1278 (2005). [CrossRef] [PubMed]
7. J. Q. Xi, H. Luo, A. J. Pasquale, J. K. Kim, and E. F. Schubert, “Enhanced light extraction in GaInN light-emitting diode with pyramid reflector,” IEEE Photon. Technol. Lett. 18(22), 2347–2349 (2006). [CrossRef]
8. K. Bao, B. Zhang, X. N. Kang, T. Dai, C. Xiong, G. Y. Zhang, and Y. Chen, “Improvement of light extraction from micro-pattern encapsulated GaN-based LED by imprinting,” Proc. SPIE 6910(69100N), 1–8 (2008).
9. G. He and L. Zheng, “White-light LED clusters with high color rendering,” Opt. Lett. 35(17), 2955–2957 (2010). [CrossRef] [PubMed]
10. C. C. Sun, Y. Y. Chang, T. H. Yang, T. Y. Chung, C. C. Chen, T. X. Lee, D. R. Li, C. Y. Lu, Z. Y. Ting, B. Glorieux, Y. C. Chen, K. Y. Lai, and C. Y. Liu, “Packaging efficiency in phosphor-converted white LEDs and its impact to the limit of luminous efficacy,” J. Solid State Lighting 1(1), 19 (2014). [CrossRef]
11. Cree, Inc., http://www.cree.com/.
12. L. Kim, J. H. Choi, S. H. Jang, and M. W. Shin, “Thermal analysis of LED array system with heat pipe,” Thermochim. Acta 455(1-2), 21–25 (2007). [CrossRef]
13. I. Moreno, M. Avendaño-Alejo, and R. I. Tzonchev, “Designing light-emitting diode arrays for uniform near-field irradiance,” Appl. Opt. 45(10), 2265–2272 (2006). [CrossRef] [PubMed]
14. L. Yang, J. Hu, L. Kim, and M. W. Shin, “Thermal analysis of GaN-based light emitting diodes with different chip sizes,” IEEE Trans. Device Mater. Reliab. 8(3), 571–575 (2008). [CrossRef]
15. A. Christensen and S. Graham, “Thermal effects in packaging high power light emitting diode arrays,” Appl. Therm. Eng. 29(2-3), 364–371 (2009). [CrossRef]
16. D. Wu, K. Wang, X. Luo, S. Liu, “Enhancement of light extraction of multi-chips light-emitting diode (LED) modules with various micro-structure arrays,” in 11th International Conference on Electronic Packaging Technology and High Density Packaging IEEE, 2010), pp 1398-1400.
17. D. Ramane and A. Shaligram, “Optimization of multi-element LED source for uniform illumination of plane surface,” Opt. Express 19(S4Suppl 4), A639–A648 (2011). [CrossRef] [PubMed]
18. H. H. Wu, K. H. Lin, and S. T. Lin, “A study on the heat dissipation of high power multi-chip COB LEDs,” Microelectron. J. 43(4), 280–287 (2012). [CrossRef]
19. R. H. Horng, H. L. Hu, R. C. Lin, L. S. Tang, C. P. Hsu, and S. L. Ou, “Cup-shaped copper heat spreader in multi-chip high-power LEDs application,” Opt. Express 20(S5Suppl 5), A597–A605 (2012). [CrossRef] [PubMed]
20. C. J. M. Lasance and A. Poppe, Thermal Management for LED Applications (Solid State Lighting Technology and Application Series) (Springer, 2014).
21. I. Moreno, D. Bermúdez, and M. Avendaño-Alejo, “Light-emitting diode spherical packages: an equation for the light transmission efficiency,” Appl. Opt. 49(1), 12–20 (2010). [CrossRef] [PubMed]
22. S. J. Lee, “Analysis of light-emitting diodes by Monte Carlo photon simulation,” Appl. Opt. 40(9), 1427–1437 (2001). [CrossRef] [PubMed]
23. N. Linder, S. Kugler, P. Stauss, K. P. Streubel, R. Wirth, and H. Zull, “High-Brightness Light-Emitting Diodes Using Surface Texture,” Proc. SPIE 4278, 19–25 (2001). [CrossRef]
24. A. Badano and J. Kanicki, “Monte Carlo analysis of the spectral photon emission and extraction efficiency of organic light-emitting devices,” J. Appl. Phys. 90(4), 1827–1830 (2001). [CrossRef]
25. C. C. Sun, C. Y. Lin, T. X. Lee, and T. H. Yang, “Enhancement of light extraction of GaN-based LED with introducing micro-structure array,” Opt. Eng. 43(8), 1700–1701 (2004). [CrossRef]
26. T. X. Lee, C. Y. Lin, S. H. Ma, and C. C. Sun, “Analysis of position-dependent light extraction of GaN-based LEDs,” Opt. Express 13(11), 4175–4179 (2005). [CrossRef] [PubMed]
27. isC.-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 (2006).
28. D. S. Han, J. Y. Kim, S. I. Na, S. H. Kim, K. D. Lee, B. Kim, and S. J. Park, “Improvement of LEE of flip-chip light-emitting diode by texturing the bottom side surface of sapphire substrate,” IEEE Photon. Technol. Lett. 18, 1406–1408 (2006). [CrossRef]
29. T. X. Lee, K. F. Gao, W. T. Chien, and C. C. Sun, “Light extraction analysis of GaN-based light-emitting diodes with surface texture and/or patterned substrate,” Opt. Express 15(11), 6670–6676 (2007). [CrossRef] [PubMed]
30. Y. Qin, X. Guo, W. J. Jiang, R. Fang, and G. D. Shen, “Light Extraction Analysis of AlGaInP based LED with Surface Texture,” Proc. SPIE 7635(763505), 1–7 (2009).
31. Z. Liu, K. Wang, X. Luo, and S. Liu, “Precise optical modeling of blue light-emitting diodes by Monte Carlo ray-tracing,” Opt. Express 18(9), 9398–9412 (2010). [PubMed]
32. J. W. Pan and C. S. Wang, “Light extraction efficiency of GaN-based LED with pyramid texture by using ray path analysis,” Opt. Express 20(S5Suppl 5), A630–A640 (2012). [CrossRef] [PubMed]
33. C. C. Sun, W. T. Chien, I. Moreno, C. T. Hsieh, M. C. Lin, S. L. Hsiao, and X. H. Lee, “Calculating Model of Light Transmission Efficiency of Diffusers Attached to a Lighting Cavity,” Opt. Express 18(6), 6137–6148 (2010). [CrossRef] [PubMed]
34. C. C. Sun, T. X. Lee, Y. C. Lo, C. C. Chen, and S. Y. Tsai, “Light extraction enhancement of GaN-based LEDs through passive/active photon recycling,” Opt. Commun. 284(20), 4862–4868 (2011). [CrossRef]
