We apply a microstructure array (MSA) film to improve the angular color uniformity (ACU) of a correlated-color-temperature-tunable LED (CCT-tunable LED) with tunable CCT ranging from 2700 to 6500 K. The effects of the MSA film area and the height between the film and LED are investigated and optimized. The resulting ACU is greatly improved for all CCT ranges with little luminous flux loss. For a typical CCT range of 3000-4000 K, with a full-covering MSA film and height H = 5 mm, the CCT deviation is significantly reduced from 1090 K to 218 K, with only 1.8% luminous flux loss.
© 2016 Optical Society of America
The light-emitting diode (LED) is considered the next-generation light source, and it has gained widespread popularity in its applications to medical apparatus, backlit displays, spot lamps, and residential illumination while offering the advantages of energy saving, environment sustainability, and long lifetimes [1,2]. A large number of studies have focused on angular color uniformity (ACU) as a crucial indicator of the optical properties of high-quality LEDs. One efficient method to improve the LED ACU involves optimization of the structure of the phosphor layers in the LED. Zheng et al.  presented a novel package structure in which the thickness of the phosphor layer reduced from the center to the edge, which enhanced the light ratio of yellow to blue over the angular range from 0° to 180°. Further, Zheng et al.  tailored the phosphor concentration distribution of three superimposed layers to reduce the variation in the power ratio of yellow light to blue light between the center and the periphery, which significantly ameliorated the ACU. Another valid method is the application of an optical functional film to LED systems. In this regard, Lin et al.  applied distributed Bragg reflector (DBR) with various reflection characteristics for different light wavelengths on the top of the device package. They observed that yellow photons were mostly “spread” through the film, while blue photons were reflected efficiently backward, which improved the excitation of phosphor. Further, the proportion between yellow and blue rays in normal direction got improved and so it was with performance in the correlated color temperature (CCT). Ding et al.  applied a micro-patterned array (MPA) optical film to remote phosphor LED lamps, which array could scatter blue light from the normal direction over a large angle. The deviation of the CCT distribution decreased while affording acceptable output power. However, all these studies on the ACU have primarily focused on LEDs with a monochromatic CCT; few studies have reported on CCT-tunable LEDs.
With the development of LED technology, the CCT of LEDs is becoming tunable, as human preference to CCT varies according to the time of day or mood. Inappropriate lighting conditions have been demonstrated to upset body chemistry and lead to deleterious health effects, even leading to cancer . Thus, a suitable CCT-tunable LED that can adjust to the human body’s physiological rhythm and moods and activate thinking could be of great benefit . A CCT-tunable LED can be obtained in two major configurations. One option involves combining and controlling R/G/B dies individually [9,10], while the other involves the assembling of cool-white and warm-white LEDs with separate luminous intensities [11,12]. The R/G/B combination approach suffers from limited luminous efficacy owing to the low external quantum efficiency (EQE) of green-emitting semiconductors (~25%) , an unsatisfactory color rendering index (CRI), and the need for complex controlling electrical circuits [14,15]. The approach of assembling cool-white and warm-white phosphor-coated LEDs appears more feasible and superior for affording higher luminous efficacy and a more satisfactory CRI along with simplicity of the manufacturing process . Several well-known LED manufacturers have adopted the second scheme to make available a variety of products. These include the Tiger Zenigata CCT-tunable chip-on-board (COB) series by SHARP Corp., with rectangular cool and warm color temperature regions coated alternately, affording a tunable CCT range of 2700-5700 K  and the CHI CCT-tunable COB series by EVERLIGHT, with band-like cool and warm-white regions alternately displayed, affording a seamless CCT change over the range of 2700-5700 K .
Most recent studies on CCT-tunable LEDs have mainly focused on improving the extraction efficiency and color control of LED systems . Nevertheless, the mismatch between the lambertian emission pattern of blue LED chip and the omni-directional emission of phosphor in phosphor-converted LED would cause the yellow ring effect. As for the COB LED, non-uniformity of the CCT distribution inevitably exists, manifesting as the “blue ring” and “yellow halo” in the CCT-tunable LEDs . Further, when a CCT-tunable LED is combined with reflectors in practical lamps, there is an induced difference of maximum emitting angles between the cool-white LEDs and warm-white LEDs. Due to this emitting angle difference and non-uniform light mixing, the non-uniformity problem of the CCT-tunable LED is further compounded. The method to improve the ACU performance of LED by optimizing the structure of phosphor coating is mainly applied in the LED with sink , Which is not suitable for COB LED with multi-chips. Meanwhile, it is expensive to improve the color uniformity of LED by applying different types of lens with different designed molds . Although a roughened diffusing film is capable to ameliorate the problem, it would significantly lower the luminous efficiency of the whole system, leading to energy loss . Therefore, the CCT-tunable LEDs are still facing the great challenging: how to improve the CCT uniformity.
In this context, we applied a microstructure array (MSA) film with excellent scattering characteristics to enhance the ACU of a CCT-tunable LED system. We first constructed an LED system affixed with MSA film. We studied the effects of an MSA film (fully covering the LED) on light mixing, and we analyzed the mechanisms of the microstructure working as regards the incident light waves through the finite-difference time-domain (FDTD) method. Next, parameters including the area of the optical film and the height between the film and LED were further optimized experimentally. Finally, the optical performance of the film operating over different CCT ranges from 2700 K to 6500 K was systematically analyzed.
2. Description of CCT-tunable LED
Figure 1(a) shows the experimental design of the CCT-tunable LED used in the study. The size of the board is 18mm*24mm. The LED system consists of a central circular (R = 3.9 mm) warm-white region (WWR) surrounded by a ring-shaped (Rinner = 4.3 mm, Router = 7.5 mm) cool-white region (CWR). A high-concentration phosphor colloid was dispersed in the WWR to achieve a CCT of 2700-3000 K, while the CWR consisted of a low-concentration colloidal dispersion to achieve a CCT of 6000-6500 K. Both the WWR and the CWR were packaged on the same lead frame substrate and separated by a white silicone compound ring. The luminous intensities of the WWR and CWR were controlled individually, and the CCT could be tuned intelligently via regulating the luminous intensity proportion of the two regions. When the intensity ratio between the CWR and WWR emissions decreased, the CCT-tunable LED displayed warm-white spot, and when the intensity ratio increased, the LED displayed cool-white spot. Thus, the CCT of the LED system was made tunable over the range from 2700 K to 6500 K adapt to the occasions.
Figure 1(b) shows the schematic of the CCT-tunable LED. The CCT-tunable LED is enclosed by a reflector of a certain height in a practical lamp system. The maximum emitting angle of the WWR () is smaller than that of the CWR (), and the peripheral region is only illuminated by CWR emission. The CCT of the periphery is significantly higher in comparison with that of the central area, where the light from the two regions is “mixed” thoroughly. This gives rise to an obvious “blue ring,” as shown in Fig. 1(b). In addition, as the two regions are separated by the silicone compound ring, the light waves close to the boundary areas do not mix effectively, thus leading to ACU variation as well. The factors of the maximum emitting angle difference and ineffective light mixing result in CCT deviation in the CCT-tunable LED system.
In order to solve the abovementioned non-uniformity problem of the CCT-tunable LED, an MSA film was affixed on the surface of the quartz glass covering the LED device. Figure 1(c) shows the scanning electron microscopy (SEM) images of the MSA, with the micro-cone structure having a base diameter of 2.8 μm and height of 1.5 μm. As shown in Fig. 1(d), the “ameliorative” CCT-tunable LED system consisted of four parts: a CCT-tunable LED, reflector, quartz glass, and a section of MSA film affixed to the glass. Parameter H represents the height between LED and the MSA film, and the area of the MSA optical film is related to its diameter d and the film shape. Both these parameters can be adjusted to enhance the ACU performance of the CCT-tunable LED system. The films applied in the system were fabricated by means of a low-cost and efficient spin-coating process.
3. Experiment setup
The CCT distribution, luminous intensity, spectrum over the zenith angle range from −90° to 90°(in the plane XOZ), and the luminous flux of the CCT-tunable system were measured experimentally. The instrumentation used for these measurements is shown in Fig. 2. The system comprises a rotating-axis arm surrounding the testing bench, an optical fiber probe to detect emitted light, and a spectrometer (Ocean USB2000 + ). A KESSLY® adjustable DC source was used to provide the driving current for the LED. The total current in the two regions was maintained at 300 mA, and the CCT of the LED system could be adjusted by changing the input power proportion of the two regions. We mention here that the CCT was set in the typical range of 3000-4000K during optimization. During the testing progress, the probe revolved around the LED system in the vertical plane to measure the CCT, luminous intensity, and spectrum. The luminous flux of the system was measured with the use of an integrating sphere system. The CCT-tunable LED system was mounted on aluminum fins to minimize the effect of the increased junction temperature, as shown in the inset in Fig. 2.
4. Results and discussion
To investigate the effectiveness of the LED fully covered by the MSA film (hereafter “full-covering film”) on the light mixing of the CCT-tunable LED system, the emission spectrum of the illumination spot was measured and recorded over the zenith angle range in steps of 5°. The spectrum consisted of two components: blue light emitted from the chips and the re-emitted light from the phosphor. For the CWR, the blue light from the chips was dominant, while re-emitted reddish light was dominant in the WWR. We figure out the color coordinates from the emission spectrum and then calculate the angle-dependent CCT with the Mc Camy’s formula. The CCT distributions of the systems with or without film are shown in Fig. 3(a). The parameter ΔCCT is defined as the deviation between the maximum and the minimum CCT values. We note that for the system without the MSA film, the CCT distribution fluctuates greatly. As the zenith angle increases from 0° to 90°, the CCT exhibits a valley value of 3513K at 30°, a peak value of 4487K at 55°, and the deviation ΔCCT reaches 974K. As the CCT of the system is decided by the luminous intensity ratio of the WWR to the CWR. We can infer that the intensity of the WWR is relatively stronger at 30°, while it is weaker comparing to the CWR intensity at 55° according to the figure. It results from the fact that in the CCT-tunable LED system, rays emitted from the WWR at small angles travel directly toward the glass in straight lines. However, only the rays emitted from the inner side of the CWR can propagate in the same manner, and the rays from the CWR periphery are reflected off the walls of the reflector. Thus, the WWR luminous intensity distribution in the central zone is more uniform and concentrated, while that of the CWR decreases as the angle increases. The intensity ratio of the WWR to the CWR increases and the CCT reach the minimum at 30°. On the other hand, rays from the WWR emitted at large angles that cannot escape directly, would suffer more times of reflections inside the reflector than the CWR one. The presence of the reflector also causes that the maximum emitting angle of the WWR is smaller than that of the CWR. The WWR intensity distribution reduces drastically when the angle is larger than 30°. The intensity ratio decrease as the angle increase, and the CCT reach the peak at 55°, where is the maximum emitting angle of the WWR. The actual spot illumination of the system is shown in the inset of Fig. 3(a). The spot can be classified into three illumination regions according to the angular deviation from the center to the perimeter: neutral white, warm white, and an obvious cool-white ring. Further, a “blue ring” and “yellow halo” are also simultaneously observed. After MSA film fixture to the system, we observe that the angle-dependent CCT fluctuation is greatly moderated. The ΔCCT is only 294K, reduced by 69.8% of that of the system without the film covering. This result indicates that the intensity distribution difference of the two regions is very small in the full-covering film case. The inset of Fig. 3(a) shows the illumination spot obtained with the film covering; the spot is uniform and the yellow and blue aberrations observed at 30° and 55° are nearly eliminated.
The spectra of the systems with and without the film at 30° and 55° are shown in Fig. 3(b). The measured peak wavelengths of the CWR and the WWR are 609 nm (reddish-light) and 460 nm (blue-light), respectively. The intensity of the two peak wavelengths in the spectrum is closely linked to the CCT after rays mixing. We note that the intensity changes of the blue-light remains the same as the filmless system; however, the reddish-light intensity reduces at 30° and increases at 55° after the application of the MSA film. This corresponds to decrease in the peak CCT and increase in the valley CCT in system with full-covering film. This result indicates that the MSA film “scattered” the centralized incident power of the WWR, dominated by reddish light, from the center to the periphery. The intensity proportion variation for different wavelengths from 0° to 90° diminished, and the ACU performance of the CCT-tunable LED was enhanced.
To obtain a better understanding of the mechanism of the MSA film to incident light waves from the LED, FDTD algorithm was introduced in the study. The FDTD analysis setups have been detailed in our previous study . Since the scattering characteristics of the microstructure could be effectively characterized with the input normal plane wave . We utilize the similar method to explore the performance of the MSA film for simplification. The FDTD analysis model and results are shown in Figs. 4(a) and (b), respectively. When a plane wave propagates from the film structure surface to the air, light waves are scattered at the interface (Fig. 4(b)). The concentrated vertical plane wave is dispersed to propagate along the entire length of the interface between air and the MSA. Most waves are transmitted symmetrically, while the rest are reflected backward. This result indicates that the microstructure exhibits a large scattering ability in dispersing incident light waves to larger angles. There is better mixing of the rays from the two regions in the angular range of 0° to ± 90° with enhancement in the ACU when the MSA film is applied to the system.
The relevant parameters of the MSA optical film in practical application including the film area and height between the MSA and LED (H) were subsequently optimized experimentally.
First, the area of the MSA film was optimized. Three types of films were fabricated and affixed to the surface of the glass considering the contour of the two regions: half-covering film with radius R1 = 3.9 mm, ring film with Rinner = 4.3 mm and Router = 7.5 mm, and full-covering film with radius R2 = 7.5 mm. The height between the film and the LED was set to H = 7.5 mm. Figure 5(a)-(d) plots the angle-dependent CCT distributions of the systems optimized by the MSA films with different degrees of covering. The emission spot was characterized into the central zone (0° to ± 30°), middle zone ( ± 30° to ± 60°), and the peripheral zone ( ± 60° to ± 90°).
For the system without film (Fig. 5(a)), ΔCCT was equal to 1200 K. As the zenith angle varied from 0° to ± 90°, the CCT reduced from 4000 K at 0° to 3430 K at ± 35°, and subsequently, it increased rapidly to 4620 K at ± 60°. This large CCT deviation ascribed the luminous intensity distribution variation between the two regions, as shown in Fig. 5(e). From the figure, we note that the luminous intensity of the CWR emission is dispersive from the center to the perimeter, while the distribution of the WWR emission is mostly confined to the central zone, and it drops drastically to nearly zero in the peripheral zone. This is because the rays of the CWR suffer more reflections of the reflector in central zone, while the maximum emitting angle is larger than the CWR. The ratio of the WWR to CWR intensities is greater or equal to 1 in the angular range from 0° to ± 65°, and it drops steeply to 0 at larger angles. The CCT is low or high corresponding to large or small values of this ratio, respectively, when the rays from the two regions mix together. The luminous intensity distribution divergence of different regions in the same COB LED mainly owing to the reflection and absorption of the reflector, distinguishes the CCT-tunable LED from the ordinary LED. The abovementioned difference leads to the observed poor ACU performance of the system.
When the half-covering film is affixed to the surface of the glass, ΔCCT slightly increases to 1400 K. The corresponding CCT distribution tends to be high in the center and low in the periphery as shown in Fig. 5(b). The CCT at 0° increases to 4600 K, but decreases to 3200 K at ± 35° and 4200 K at ± 55°. These changes are due to the fact that the half-covering film scatters the rays of the CWR and WWR when they propagate through it. Both the luminous intensities of the WWR and CWR in the center would be “removed” from the center to peripheral zone by the MSA film. The WWR intensity distribution is concentrated and uniform in the central zone, while the CWR distribution is relatively dispersive. Therefore, when the half-covering MSA film was applied in the system, the WWR intensity reduction in the central zone is more than that of the CWR, as shown in Fig. 5(f). The CCT in the central zone consequently decreases. Meanwhile, the CCT decreases in the peripheral zone as larger amount of warm-white rays are scattered to large angles.
When the ring film is affixed to the surface of the glass, ΔCCT decreases to 800 K. As shown in the plot in Fig. 5(c), the CCT distribution tends to increase gradually as the angle increases, and the angle corresponding to the peak CCT is shifted by 15° toward the center with respect to the no-film-covering case. The CCT drops to 3200 K at 0°, and the peak CCT decreases to around 4000 K. When the mixed rays propagate through the glass, they are scattered lateral to the peripheral and central zones by the ring film. As in the original system, the WWR half-value intensity angle is larger than that of the CWR, and thus, stronger power of WWR emission is scattered to the center by the ring film, while the CWR emission is mainly dispersed to the middle zone, as shown in Fig. 5(g). As a result, the power of the WWR tends to be more concentrated in ± 15°, while the power of the CWR is concentrated in the middle zone after the application of the ring film. CCT in the central zone decreases while that in the middle increases, and the peak CCT shifts toward the center.
When the full-covering film is affixed to the surface of glass, ΔCCT drastically reduces to 300 K as shown in Fig. 5(d). The ACU performance of the system is excellent, with the exhibition of a uniform CCT distribution. The WWR luminous intensity distribution trend is nearly identical to that of the CWR. The maximum WWR emitting angle increases to 80°, and the intensity variation between the two regions at large angles reduces, as shown in Fig. 5(h). This result indicates that the application of the full-covering MSA film combines the benefits offered by the half-covering film and the ring film. The full-covering film scatters the power of the WWR in central zone to large angle, overcoming the angle constraint of the reflector with regard to the peripheral zone, while simultaneously scattering CWR emission to the middle zone. Light rays from the two regions mix evenly throughout the illumination space, and the angle-dependent CCT variation between the perimeter and the middle is minimized. The full-covering MSA film was chosen for further optimization for practical application in the next stage of the study.
The height (H) between the film and the LED forms another crucial parameter that requires optimization. In our study, H was set to six values as follows: 1.5 mm, 2.5 mm, 5 mm, 7.5 mm, 10 mm, and 12.5 mm. The CCT distributions of the LED system with these different reflector heights are shown in Fig. 6. For the system without the film, as the height decreases, the angles corresponding to both the valley and peak CCT values increase, and ΔCCT increases from 1200 K (H = 12.5 mm) to 1600 K (H = 1.5 mm) gradually. This is because the maximum emitting angle of the two regions increases, and the intensity at large angles becomes significant such that a slight intensity variation leads to a drastic change in the CCT. After the application of the full-covering MSA film, ΔCCT decreases, and this effect subsequently varies with the height; parameter ΔCCT decreases and then increases as H decreases from 12.5 mm to 1.5 mm. In the height range of 12.5-5 mm, the CCT distribution of the system is large in the central and middle zones but small in the periphery. With decrease in height, the CCT at large angles increases while that in the center remains unchanged; ΔCCT gradually decreases from 500 K to 300 K. With decrease in the height, the emitting angle increases, and the amount of CWR rays reflected by the reflector decreases. A larger amount of cool-white light propagates through the film and is scattered to the periphery. Thus, the CCT in periphery increases gradually. In the case when H ≤ 5 mm, the CCT at large angles increases to a level that exceeds that of the central zone. As a relatively larger number of cool-white rays impinge the periphery, the scattering ability of the MSA is limited to dispersion of the WWR rays to such large angles, and thus, ΔCCT again deteriorates. In summary, the ACU of the system with the MSA optical film is optimal for H = 2.5-5 mm.
In terms of practical application, the CCT of the system could be tuned by varying the CWR/WWR intensity ratio. When this ratio increased, the CCT was high, and when it reduced, the CCT was low. Our analysis indicates that the film mainly affects the scattering of the WWR rays over large angles, while the film’s effect on the CWR intensity distribution is relatively milder. The lack of WWR intensity in the periphery was compensated, and the ACU performance was enhanced. Further, the MSA film performance varied over different CCT ranges owing to the intensity ratio differences. The tunable CCT range was divided into five sub-ranges as 2700-3000 K, 3000-4000 K, 4000-5000 K, 5000-6000 K, and 6000-6500 K according to the overall CCT.
Table 1 lists the ΔCCT values and the luminous flux of the LED system with and without optimal MSA film over the five ranges in experiment. The results indicate that the MSA film significantly enhances the ACU of the system over all ranges. Parameter ΔCCT reached the optimum value in the range of 2700-3000 K, and it increased with the CCT. The increasing CCT deviation mainly results from increase in the CWR intensity and decrease in the WWR scattered intensity at large angles. The intensity of the CWR increased at high CCT values along with increase in the corresponding CWR intensity in the periphery, while WWR intensity decreased. In our system, when light rays propagate through the film, although a portion of the centralized incident power of the WWR is scattered over large angles, the scattered WWR intensity is too weak to compensate for the CWR intensity at large angles; thus, the CCT deviation in the peripheral zone again deteriorates. Furthermore, we also evaluated the luminous flux. When compared with the system without film, the largest luminous flux loss was just 3.3% in the range of 6000-6500 K, and the smallest loss was as low as 1.8% in the typical CCT range of 3000-4000 K. In summary, our findings indicate that the use of an optimally configured MSA film can enhance the ACU performance in the range from 2700 K to 6500 K with very little luminous flux loss, while yielding a superior ACU performance even for low CCT values.
In this work, we applied a microstructure array (MSA) film to a CCT-tunable LED (with CCT range of 2700-6500 K) consisting of a cool-white region and a warm-white region assembled on the same lead frame substrate in order to improve the angular color uniformity (ACU). The outstanding scattering ability of the film aided in dispersing the stronger warm-white region rays over large angles. The spatial distribution variation in the luminous intensity of the two regions was effectively improved, and the rays “mixed” evenly, which led to improved ACU. Parameters including the film area and the height between the film and LED were optimized experimentally in the CCT range of 3000-4000 K. When a full-covering MSA optical film was affixed to the surface of the glass and the reflector height was set to H = 5 mm, ΔCCT improved from 1090 K to 218 K with a flux loss of only 1.8% when compared with that of the system without the MSA film. Furthermore, the effects of the film over the CCT ranges of 2000-3000 K, 3000-4000 K, 4000-5000 K, 5000-6000 K, and 6000-6500 K were all measured. Our results indicate that the use of an MSA film can indeed improve the ACU performance with an acceptable output luminous flux loss. We believe that the application of the MSA optical film can significantly contribute to the development of high-quality CCT-tunable LEDs.
This work is financially supported by the National Natural Science Foundation of China (NSFC) (Grant Nos. 51405161 and 51375177), the Postdoctoral Science Foundation of China (No. 2014M560659), the Natural Science Foundation of Guangdong Province (No. 2014A030312017), and Fundamental Research Funds for the Central Universities.
References and links
1. L. Yang, Y. Liu, Y.-L. Zhong, X.-X. Jiang, B. Song, X.-Y. Ji, Y.-Y. Su, L.-S. Liao, and Y. He, “Fluorescent silicon nanoparticles utilized as stable color converters for white light-emitting diodes,” Appl. Phys. Lett. 106(17), 173109 (2015). [CrossRef]
2. M. Kim, S. M. Lee, and K. C. Choi, “Optical tuning of phosphors by plasmonic gold nanoparticles for phosphor-converted white light emitting diodes,” Appl. Phys. Lett. 105(14), 141119 (2014). [CrossRef]
3. H. Zheng, X. Fu, R. Hu, S. Liu, and X. Luo, “Angular color uniformity improvement for phosphor-converted white light-emitting diodes by optimizing remote coating phosphor geometry,” presented at the 13th International Conference on Electronic Packaging Technology and High Density Packaging (ICEPT-HDP), 2012, 1483–1486. [CrossRef]
4. M. Zheng, W. Ding, F. Yun, Z. Tian, Q. Li, Y. Li, M. Guo, Y. Zhang, and X. Yang, “ACU enhancement of WLEDs realized by multilayer phosphor with convex shape,” IEEE Photonics Technol. Lett. 28(2), 111–114 (2016). [CrossRef]
5. H. Y. Lin, K. J. Chen, S. W. Wang, C. C. Lin, K. Y. Wang, J. R. Li, P. T. Lee, M. H. Shih, X. Li, H. M. Chen, and H. C. Kuo, “Improvement of light quality by DBR structure in white LED,” Opt. Express 23(3), A27–A33 (2015). [CrossRef] [PubMed]
6. X. Ding, J. Li, Q. Chen, Y. Tang, Z. Li, and B. Yu, “Improving LED CCT uniformity using micropatterned films optimized by combining ray tracing and FDTD methods,” Opt. Express 23(3), A180–A191 (2015). [CrossRef] [PubMed]
7. D. E. Blask, R. T. Dauchy, L. A. Sauer, J. A. Krause, and G. C. Brainard, “Growth and fatty acid metabolism of human breast cancer (MCF-7) xenografts in nude rats: impact of constant light-induced nocturnal melatonin suppression,” Breast Cancer Res. Treat. 79(3), 313–320 (2003). [CrossRef] [PubMed]
10. K. Sakuma, K. Omichi, N. Kimura, M. Ohashi, D. Tanaka, N. Hirosaki, Y. Yamamoto, R. J. Xie, and T. Suehiro, “Warm-white light-emitting diode with yellowish orange SiALON ceramic phosphor,” Opt. Lett. 29(17), 2001–2003 (2004). [CrossRef] [PubMed]
11. A. T. L. Lee, J. K. O. Sin, and P. C. H. Chan, “Scalability of quasi-hysteretic FSM-based digitally controlled single-inductor dual-string buck LED driver to multiple strings,” IEEE Trans. Power Electron. 29(1), 501–513 (2014). [CrossRef]
12. A. Lee, H. Chen, S. C. Tan, and S. Y. R. Hui, “Precise dimming and color control of light-emitting diode systems based on color mixing,” IEEE Trans. Power Electron. 31(1), 65–80 (2016). [CrossRef]
14. H. Y. Lin, S. W. Wang, C. C. Lin, K. J. Chen, H. V. Han, Z. Y. Tu, H. H. Tu, T. M. Chen, M. H. Shih, and P. T. Lee, “Excellent color quality of white light-emitting diodes by embedding quantum dots in polymers material,” IEEE J. Sel. Top. Quantum Electron.22(1), 2000107 (2015).
15. H. C. Chen, K. J. Chen, C. C. Lin, C. H. Wang, C. C. Yeh, H. H. Tsai, M. H. Shih, and H. C. Kuo, “Improvement of lumen efficiency in white light-emitting diodes with air-gap embedded package,” Microelectron. Reliab. 52(5), 933–936 (2012). [CrossRef]
19. C. Y. Liu, K. J. Chen, D. W. Lin, C. Y. Lee, C. C. Lin, S. H. Chien, M. H. Shih, G. C. Chi, C. Y. Chang, and H. C. Kuo, “Improvement of emission uniformity by using micro-cone patterned PDMS film,” Opt. Express 22(4), 4516–4522 (2014). [CrossRef] [PubMed]
20. L. Zongyuan, L. Sheng, W. Kai, and L. Xiaobing, “Optical Analysis of Phosphor’s Location for High-Power Light-Emitting Diodes,” IEEE Trans. Device Mater. Reliab. 9(1), 65–73 (2009). [CrossRef]
22. S. Guo, S. Zhou, H. Li, and B. You, “Light diffusing films fabricated by strawberry-like PMMA/SiO2 composite microspheres for LED application,” J. Colloid Interface Sci. 448, 123–129 (2015). [CrossRef] [PubMed]
23. Y. Kawaguchi, K. Nishizono, J.-S. Lee, and H. Katsuda, “Light extraction simulation of surface-textured light-emitting diodes by finite-difference time-domain method and ray-tracing method,” Jpn. J. Appl. Phys. 46, 31 (2007).