Open Access
Add to BibSonomyAbstract
The chromaticity coordinates and spectra of phosphor-converted LEDs are demonstrated to be well controlled in this study. Through the feedback coating method of stacked yellow, green, and red phosphor layers, the color rendering index (CRI), correlated color temperature (CCT), and spectra are determined to match precisely the desired target. In addition, the reabsorption effect is strongly influenced by the order of stacked phosphor layers and the selected excitation wavelength of phosphors. The degree of reabsorption will modify the original spectra and cause a mismatch between the experimental measurement and the simulation based on the linear superposition of blue light and phosphor-emitted light. This feedback coating method offers an easy approach towards optimized spectra, which can offer the highest luminous efficacy of radiation with excellent color-rendering properties.
© 2015 Optical Society of America
1. Introduction
Several countries have recently issued a regulation for prohibiting incandescent lamps of 40, 60, 75, and 100 watts. With the phase-out of light sources with low efficiency and high energy consumption, the penetration rate of white light-emitting-diodes (LEDs) into the lighting market has annually increased. Given their high efficiency, long life time, no absence of mercury, and dimming capacity, LEDs are already used in numerous applications, such as backlight displays and intelligent light control systems [1].
The requirements of LED characteristics in various application fields are much in demand to date. The correlated color temperature (CCT) must be restricted within the small variation, and the color rendering index (CRI) should be sufficiently high. The maximum value of the CRI is 100 for sunlight, which indicates the distortion-free true colors of an object. The luminous efficacy of radiation (LER) of a light source is closely related to the sensitive of human eye. The highest LER is equal to the eye sensitivity function multiplied by 683 lm/W for a monochromatic emission at 555 nm [2]. However, the CRI of a monochromatic emission at 555 nm is very poor because the spectrum of sunlight is continuous. Consequently, a trade-off between CRI and LER is inevitable. The specific applications will determine which of the two will need to be maximized.
Three approaches are available for producing white light. The first one uses blue LEDs to excite the yellow phosphor to generate white light [3–5]. This method has benefits of easy process and low cost. Therefore it is widely adopted for lighting industry. However, this spectrum lack of green and red region, the CRI is quite low. The second approach is exciting red, green, and blue phosphors from UV LEDs [6,7]. The emission spectrum includes a wide region and has a high CRI. However, the light efficiency is not sufficiently high because of the very low efficiency of UV LEDs. The third method is mixing trichromatic red, green, and blue LEDs [2]. Although these LEDs have better light efficiency and acceptable CRI, the different driving voltage and light time of each LED increases the difficult of operation and the cost.
In this study, a simple and precise feedback control technology is proposed to approach the target specifications of white light LEDs. The LER, CCT, CRI, and the chromaticity coordinates are easily adjusted by coating layer-by-layer multiple phosphors. Compared with the narrow emission bandwidth of multiple single-color LEDs, the phosphor-converted LEDs (pc-LEDs) provide board emission spectra with great color stability [8,9]. Through the measurement of feedback spectra of pc-LEDs, the thickness of the phosphor coated onto the LEDs is determined for each specific phosphor concentration. The well-controlled feedback of the LED light source is extremely flexible for lighting applications.
2. Experiments
In this study, the multi-layered phosphors are conformal coated on blue LEDs by the pulsed spray coating method shown in Fig. 1. It provides the best angular color distribution comparing to remote or dispense deposition methods [10,11]. Phosphor suspension slurry is formed by mixing phosphor powder, silicone binder, and an alkyl-based solvent. Then spraying phosphor slurry on the substrates and baking at 150 °C. The weight-percentage of phosphor slurry is fixed to 50 wt. %, and the total thickness of stacked phosphor layers is determined by the feedback spectra to approach the target color points. The density of the phosphor is approximately 1 mg/cm2 and the thickness of each coating run is approximately 10 μm [12]. The process of feedback spectrum control involves:
- • Coating the first layer.
- • Optical properties measurement.
- • Checking the consistency between the experimental result and the target result.
- • Feedback discrepancy for determining the next coating parameter.
The feedback spectra are measured by an integrating sphere with the USB 2000 + spectrometer from Ocean Optics Company. The peak emission wavelength of blue LEDs are approximately 450 nm and the driving current is 350 mA. The blue LEDs are used to excite the phosphor and convert the blue light into yellow, yellow-green, and red light. To achieve a high degree of freedom to tune the color chromaticity, several phosphors are chosen. The peak emission wavelength of yellow phosphor (Y) is 558 nm, that of the yellow-green (YG) phosphor is 520 nm, and those of the red phosphor are 630 nm (R630 or R) and 660 nm (R660 or R’) at an excitation wavelength of 460 nm. The LER, CCT, CRI, and the color chromaticity of the stacked multi-phosphor layers are strongly influenced by the layer order, phosphor thickness, phosphor power distribution, and the reabsorption effect [12–16].
3. Discussion and results
The commercial white light LEDs are generated by pumping blue LEDs into a single yellow phosphor, but it typically only has a CRI of approximately 70 and a CCT above 4000 K [5–7]. To achieve superior optical quality for warm white illumination, the two-phosphor combination of a yellow or yellow-green phosphor with a red phosphor is a suitable approach.
In this study, the CCT region can be adjusted from 2500 K to 8000 K by simply mixing different ratios of Y or YG with R phosphors in a fixed blue LED intensity. The CRI is upgraded from 70 to 80–95. The trend chart of color points on the CIE x, y chromaticity diagram does not simply follow the straight line between the blue light source point and the phosphor point. The introduction of a secondary phosphor involved the parabolic curve based on its chromaticity point. Figure 2 shows the results of Y and R630 multi-phosphor layers. The Y phosphor is selected as the front coating at 6, 8, 10, and 12 mg, followed by 1, 2, 3, and 4 mg of the R630 phosphor in succession. In the beginning, the color point followed the straight line between blue source point and Y phosphor emission point. Subsequently, when the R630 phosphor is involved, the color point starts to turn right. How far this point will reach depends on the ratio of the Y and R630 phosphors. This feedback control implies that with the appropriate yellow and red phosphor, the curve of color point can be moved along the black-body radiation line.
Fig. 2 Color points on the CIE x, y chromaticity chart of blue (450 nm) LED-pumped Y phosphors with 4, 6, 8, 10, and 12 mg, followed by R630 phosphor with 1, 2, 3, and 4 mg.
Figure 3 shows a reverse coating order as compared with Fig. 2. With thicknesses of 1, 2, 3, and 4 mg, R630 is used as the front phosphor layer; a portion of blue light will preferentially be converted to red light. In the following layers, the color points x, y will accordingly move to the upper right with the increasing thickness of the Y phosphor. However, this approach can only obtain one cross point through the black-body radiation line. Therefore, if the desired point is at a lower CCT, the feedback process must be repeated.
Fig. 3 Color points on the CIE x, y chromaticity chart of blue (450 nm) LED -pumped R630 phosphor with 1, 2, 3, and 4 mg followed by Y phosphors with 2, 4, 6, 8, 10, and 12 mg.
The transmitted blue photons will lose more energy within the red phosphor than within the same amount of yellow phosphor because of the Stokes shift. In addition, Fig. 3 shows that if the R630 thickness is greater than 3 mg, the color points will move and exceed the center of the straight light between the blue LED and red phosphor points. Regardless of how the subsequent yellow phosphor is added, the color points still can’t fall into the ANSI C78.377 quadrangles [17]. The ANSI set up has 8 CCT points (6500 K, 5700 K, 5000 K, 4500 K, 4000 K, 3500 K, 3000 K, and 2700 K), which are described by the quadrangles on the chromaticity chart. Given this consideration, only the first red phosphor of less than or equal to 2 mg is analyzed.
A smart strategy for approaching the target x, y points is suggested in this study. Although the same amount of total phosphor shown in Figs. 2 and 3, the chromaticity values are significantly different. However, for the evaluation of good optical quality, the correct color points are not enough to represent it. Only the correct color points with matched spectra can achieve superior light quality. The evolution of emission spectra based on different order of two Y/R630 phosphor layers is analyzed in Fig. 4. The spectra are normalized to the peak wavelength of the white broadband. The “Y6” in Fig. 4 means that the blue LED light source is converted to board-band yellow emission light through thickness of 6 mg Y phosphor (peak wavelength of 550 nm). Y6R2 and Y6R4 are added with 2 and 4 mg of R630 phosphor layer above the 6 mg Y6 layer. The peak wavelength shifts from 550 nm to 588 nm for Y6R2 and 601 nm for Y6R4. Except for the conversion from blue to yellow light, the emitted yellow light (480–580 nm) will reabsorbed by the red phosphor and will convert yellow light to longer wavelength energy.
Fig. 4 Emission spectra of blue (450 nm) LED-pumped Y and R630 phosphors with different thicknesses and orders.
To compare the influence of the phosphor coating order on optical characteristics, R1Y10 is selected in the vicinity of Y6R2, as shown in the inset of Fig. 4. The spectra are almost the same, but the light efficiency dropped from the 85.2 lm/W of Y6R2 to the 74.5 lm/W of R1Y10. Therefore, a thicker phosphor layer will consume the light efficiency even with the same color points and spectra. The unit lm/W mentioned in this study refers to the light efficiency before the lens package.
The R2 is expressed as the blue LED light source is converted to the broadband red emission light through thickness of 2 mg of the R630 phosphor (peak wavelength of 630 nm). R2Y6 and R2Y12 have the additional 6 mg and 12 mg of Y phosphor layer above the R630 phosphor layer. The peak wavelengths shown in Fig. 4 have almost no shift within the 608 nm of R2Y6 and of R2Y12. The spectra of the board-band region are almost the same, except the blue light region continuously decays with the increasing Y phosphor thickness. This trend indicates that the red emitted light from the first R630 phosphor will not be absorbed by the subsequent Y phosphor but the blue light source will be absorbed by Y phosphor. Although more amount of blue light is absorbed by the overlying Y phosphor, the probability of light transmission into air will drop off extremely because of the over-thick phosphor.
The accuracy feedback method of chromaticity coordinates and spectra is demonstrated above. For high quality of LED lighting, the crucial issue is that the green gap must be overcome [18–21]. Given the lack of high-efficiency green LEDs and the need to lower costs, phosphor-converted LED have recently become a more attractive option than color-mixed LEDs. However, the inevitable Stokes shift loss, which results from the wavelength conversion, still exists. In this study, the Y phosphor is replaced by a YG phosphor to complete the green gap of the excitation spectra. The YG phosphor is combined with R630 phosphor to raise CRI. The feedback processes are repeated according to Figs. 2 and 3. Given that 6 mg of the Y phosphor added with the R630 phosphor moved along the black-body line, the same thickness as that of the 6 mg YG phosphor is selected for further discussion.
The locus of chromaticity x, y points according to different orders of combining the two YG/R630 phosphor layers is analyzed in Fig. 5. The overlap region between the spectra and the eye sensitivity function is the major influence of its efficiency. Although the CRI can be easily boosted by changing the Y phosphor to a YG phosphor (the green gap makes up the deficiency), the light efficiency will be sacrificed. The CRI of Y6R2 is 71 and that of YG6R2 is increased to 87; the light efficiency of Y6R2 is 85.2 lm/W and that of YG6R2 is decreased to 69.79 lm/W. The measured optical parameters are listed in Table 1. The light efficiency of YG6R2 is 69.79 lm/W and that of R2YG6 is 45.18 lm/W. From Fig. 5, if the red phosphor as the first excitation layer, the subsequently added YG phosphor must be thicker to reach the black-body line. With a thicker phosphor layer, a lower light efficiency is obtained because of the phosphor absorption.
Fig. 5 Color points on the CIE x, y chromaticity chart of blue (450 nm) LED-pumped YG/R630 phosphors with different thicknesses and orders.
Table 1. Chromaticity coordinates, CRI, CCT, and lm/W of blue (450 nm) LED-pumped YG/R630 phosphors with different thicknesses and orders. The chromaticity coordinates of the Y, YG, and R630 phosphors are measured as reference values for the analysis.
The emission spectra of blue LED-pumped YG and R630 phosphors with different thicknesses and orders are shown in Fig. 6. The spectra are normalized to the peak wavelength of the white broadband. The lower efficiency of R2YG6 is ascribed to the small overlap region between the spectra of R2YG6 and the eye sensitivity function. Although more of the YG phosphor layer was added above the red phosphor layer to make the color points approach the black-body line, the spectra in the green region of R2YG6 and R2YG10 did not show much increase. Consequently, the light efficiency will not increase as much as expected. The light efficiency of R2YG6 is 45.18 lm/W and that of R2YG10 is 49.16 lm/W.
Fig. 6 Emission spectra of blue (450 nm) LED-pumped YG and R630 phosphors with different thicknesses and orders.
By mixing the green phosphor with the R630 phosphor, the CRI can easily exceed 85 because of green gap filling. However, the long wavelength region attributed from R630 is not sufficient to reach higher CRI. R630 must be replaced by R660 to enlarge the region of longer wavelength. Figure 7 shows the color points of the YG phosphor mixed with R630 and R660 phosphor. The Y phosphor mixed with the R660 phosphor is used as a reference. The R9/CRI can increase from the 40/87 of YG6R2 to the 96/95 of YG6R2’. The R9/CRI of Y6R2 is 21/71 and that of Y6R’2 is 33/81.
Fig. 7 Color points on the CIE x, y chromaticity chart of blue (450 nm) LED-pumped Y/R660 or YG/ R660 phosphors with different thicknesses and orders.
Two issues must be addressed when changing longer excitation wavelength of red phosphor. First, the light efficiency will be sacrificed when the CRI is increased. Second, the linear deviation will be severe when the excitation wavelength of the red phosphor is deeper. By comparing the total phosphor (Y6R2/Y6R’2 and YG6R2/YG6R’2) of the same thickness, R660 will force the color points to move along the large curvature of the parabolic line. To approach the black-body line, a certain thickness of the first yellow or green phosphor must be added. As shown in Fig. 7, when the YG thickness is increased from 6 mg to 7 mg, the YG7R’2 will be closer to the black-body line.
In addition, Fig. 7 demonstrates the 2 mg red R660 phosphor as the first deposed layer. Although the subsequent YG phosphor is more than twice this value at 14 mg, the color points still cannot easily reach the black-body line. Table 2 shows the characteristics of the YGR2, R2YG6, YGR’2, and R’2YG6 phosphors, which all have a total phosphor thickness of 8 mg. The CRI of YGR’2 is 95, which is higher than the 87 of YGR2, but the light efficiency is deceased from 69.79 lm/W to 63.78 lm/W. The CRI and light efficiency are trade-offs based on all the results.
Table 2. Chromaticity coordinates, CRI, CCT, and lm/W of blue (450 nm) LED pumped Y/R660 or YG/R660 phosphor with different thicknesses and orders. The chromaticity coordinates of R660 phosphor are measured as a reference for the analysis.
The emission spectra of blue LED-pumped YG and R660 phosphors with different thicknesses and orders are shown in Fig. 8. The spectra are normalized to the peak wavelength of white broadband. The blue light is first absorbed by the YG phosphor then converted to yellow-green light. The short emission wavelength of the YG phosphor as the first excitation layer provides a benefit to the white LED efficiency because of the lesser Stokes loss. After the addition of the 2 mg red phosphor, the peak wavelength shifts from 533 nm to 625 nm. When the long emission R660 phosphor is the first excitation layer, the peak wavelength of R’2, R’2YG6, and R’2YG14 shows almost no shift. The spectra between 500 and 600 nm did not significantly increase from R’2YG6 to R’2YG14. Therefore, the light efficiency does not increase too much.
Fig. 8 Emission spectra of blue (450 nm) LED-pumped YG and R660 phosphors with different thicknesses and orders.
To obtain a broadband spectrum with high CRI, the yellow or yellow-green phosphor is stacked with a red phosphor in this study. To analyze the reabsorption phenomenon within the multi-phosphor layers, the amount of phosphor is maintained at 8 mg. The excitation and emission spectra of the Y, YG, R630, and R660 phosphors are shown in Fig. 9. When a fixed-wavelength BLED is used as the pumped source, the phosphor combination of YG/R will induce more reabsorption than Y/R. To distinguish the spectral distortion resulting from the reabsorption effect, the same amount of Y/R630 and YG/R660 with the 6 mg/2 mg combination is investigated in Fig. 10.
Fig. 9 Excitation and emission spectra of the Y, YG, R630, and R660 phosphors. The BLED is the pumped source.
Fig. 10 Emission spectra of blue (450 nm) LED pumped (a) Y/ R630 and (b) YG/R660 phosphors with total amount of 8 mg.
Based on the linear superposition of weighted blue and phosphor emission spectra, the resultant spectra of Y6R2 and R2Y6 did not significantly deviate from the experimental results. For Y6R2 in Fig. 10(a), the first yellow phosphor layer will absorb the blue light and be transferred into yellow light. The transmitted blue and yellow light into the second red phosphor layer will be transferred into red light. The emission spectral of yellow phosphor is in the excitation spectral region of red phosphor and will cause the reabsorption of yellow light and result in spectrum deviation. For R2Y6 in Fig. 10(a), the emission light from first red phosphor layer will not excite the subsequent yellow phosphor layer with a smaller spectrum deviation than Y6R2. Changing the yellow phosphor to the yellow-green phosphor aggravated the reabsorption. A large deviation between the spectra from the linear superposition and the experimental results is demonstrated in Fig. 10(b), especially for YG6R2. The ratio of reabsorption cannot easily be precisely quantified in experiments because of the different phosphor order, thickness, concentration, excitation wavelength, and emission wavelength. The concept of reabsorbed short wavelength transfer for emitting longer wavelengths may be a good technique to save on the amount of expensive red phosphor used.
4. Conclusion
A feedback phosphor coating method for precisely controlling the chromaticity and spectra is demonstrated in this study. By using a short emission wavelength of yellow or green as the first stacked phosphor layer the black-body line is more easily reached with a thinner total phosphor thickness. In addition, the selection of a phosphor with a short emission wavelength as the first layer will cause reabsorption by subsequent phosphor layer with a long emission wavelength. Although the reabsorption effect is complicated to analyze, the appropriate stacked order has the benefit of reducing the amount of red phosphor used. The luminous efficacy with acceptable color-rendering index can be maximized by optimizing the spectra via the proposed feedback coating method.
Acknowledgment
This research was supported by the Bureau of Energy, Ministry of Economic Affairs in Taiwan.
References and links
1. DOE SSL R&D Multi-Year Program Plan, http://www1.eere.energy.gov/buildings/ssl/
2. E. F. Schubert, Light-Emitting Diodes (Cambridge, (2003).
3. S. Nakamura, “Present performance of InGaN based blue/green/yellow LEDs,” Proc. SPIE 3002, 26–35 (1997). [CrossRef]
4. R. Mueller-Mach, G. O. Mueller, and M. R. Krames, “Phosphor materials and combination for illumination grade white pcLED,” Proc. SPIE 5187, 115–122 (2004). [CrossRef]
5. M. R. Krames, O. B. Shchekin, R. Mueller-Mach, G. O. Mueller, L. Zhou, G. Harbers, and M. G. Craford, “Status and future of high-power light-emitting diodes for solid-state lighting,” J. Disp. Techno. 3(2), 160–175 (2007). [CrossRef]
6. T. Nishida, T. Ban, and N. Kobayashi, “High-color-rendering light sources consisting of a 350 nm ultraviolet light-emitting diode and three-basal-color phosphors,” Appl. Phys. Lett. 82(22), 3817–3819 (2003). [CrossRef]
7. T. Fukui, K. Kamon, J. Takeshita, H. Hayashi, T. Miyachi, Y. Uchida, S. Kurai, and T. Taquchi, “Superior illuminant characteristics of color rendering and luminous efficacy in multilayered phosphor conversion white light sources excited by near-ultraviolet light-emitting diodes,” Jpn. J. Appl. Phys. 48(11), 112101 (2009). [CrossRef]
8. R. Mueller-Mach, G. Mueller, M. R. Krames, H. A. Hoppe, F. Stadler, W. Schnick, T. Juestel, and P. Schmidt, “Highly efficient all-nitride phosphor-converted white light emitting diodes,” Phys. Status Solidi A 202(9), 1727–1732 (2005). [CrossRef]
9. R. Mirhosseini, M. F. Schubert, S. Chhajed, J. Cho, J. K. Kim, and E. F. Schubert, “Improved color rendering and luminous efficacy in phosphor-converted white light-emitting diodes by use of dual-blue emitting active regions,” Opt. Express 17(13), 10806–10813 (2009). [CrossRef] [PubMed]
10. Z. Liu, S. Liu, K. Wang, and X. Luo, “Optical analysis of color distribution in white LEDs with various packaging method,” IEEE Photon. Technol. Lett. 20(24), 2027–2029 (2008). [CrossRef]
11. H. T. Huang, C. C. Tsai, and Y. P. Huang, “Conformal phosphor coating using pulsed spray to reduce color deviation of white LEDs,” Opt. Express 18(S2), A201–A206 (2010). [CrossRef] [PubMed]
12. B. C. Liu, J. K. Huang, K. J. Chen, S. H. Chiu, Z. Y. Tu, C. C. Lin, P. T. Lee, M. H. Shih, M. T. Wang, J. M. Hwang, and H. C. Kuo, “Luninous efficiency enhancement of white light-emitting diodes by using a hybrid phosphor structure,” J. Photonics Energy 5(1), 057603 (2015). [CrossRef]
13. Y. Zhu and N. Narendran, “Investigation of remote-phosphor white light-emitting diodes with multi-phosphor layers,” Jpn. J. Appl. Phys. 49(10), 100203 (2010). [CrossRef]
14. J. P. You, N. T. Tran, and F. G. Shi, “Light extraction enhanced white light-emitting diodes with multi-layered phosphor configuration,” Opt. Express 18(5), 5055–5060 (2010). [CrossRef] [PubMed]
15. R. Hu, B. Cao, Y. Zou, Y. Zhu, S. Liu, and X. Luo, “Modeling the light extraction efficiency of bi-layer phosphors in white LEDs,” IEEE Photon. Technol. Lett. 25(12), 1141–1144 (2013). [CrossRef]
16. R. Hu, Y. Wang, Y. Zou, X. Chen, S. Liu, and X. Luo, “Study on phosphor sedimentation effect in white-emitting diodes package by modeling multi-layer phosphors with the modified Kubelka-Munk theory,” J. Appl. Phys. 113(6), 063108 (2013). [CrossRef]
18. J. M. Phillips, M. E. Coltrin, M. H. Crawford, A. J. Fischer, M. R. Krames, R. Mueller-Mach, G. O. Mueller, Y. Ohno, L. E. S. Rohwer, J. A. Simmons, and J. Y. Tsao, “Research challenges to ultra-efficient inorganic solid-state lighting,” Laser Photon. Rev. 1(4), 307–333 (2007). [CrossRef]
19. D. Schiavon, M. Binder, A. Loeffler, and M. Peter, “Optically pumped GaInN/GaN multiple quantum wells for the realization of efficient green light-emitting devices,” Appl. Phys. Lett. 102(11), 113509 (2013). [CrossRef]
20. W. Schnick, “Shine a light with nitrides,” Phys. Status Solidi RRL 3(7–8), A113–A114 (2009). [CrossRef]
