Red-blue-green solid state light sources using a narrow line-width green phosphor A. Liu,* A. Khanna, P.S. Dutta, and M. Shur Smart Lighting Engineering Research Center, Rensselaer Polytechnic Institute, Troy, NY, USA *[email protected]
Abstract: We demonstrate that using a narrow line-width green phosphor with the peak wavelength closely aligned with the peak in the human eye sensitivity significantly improves the Luminous Efficacy of Radiation (LER) for Red-Green-Blue (RGB) emitters. Compared to the traditional RGB sources, the improvement in LER of 20 lm/W can be achieved. Combining the narrow band green phosphor with conventional wide band red and blue phosphors allows for trading off these improvements against the deviation from the Planckian locus for even higher LER. The light sources with the narrow line green phosphor are particularly promising for high energy efficiency and high intensity illumination, where somewhat compromises can be made in the color quality such as in automotive, outdoor spaces, industrial ware-houses, public places (train stations, airports) etc.. © 2015 Optical Society of America OCIS codes: (160.5690) Rare-earth-doped materials; (230.3670) Light-emitting diodes; (330.1715) Color, rendering and metamerism.
References and links 1. 2. 3. 4. 5. 6. 7. 8. 9.
10. 11. 12. 13. 14.
A. Žukauskas, M. Shur, and R. Gaska, Introduction to Solid-State Lighting (Wiley, 2002). E. F. Schubert, Lighting-Emitting Diodes (Cambridge University, 2003). M. H. Crawford, “LEDs for Solid-State Lighting: Performance Challenges and Recent Advances,” IEEE J. Sel. Top. Quantum Electron. 15(4), 1028–1040 (2009). I. Akasaki and C. Wetzel, “Future challenges and directions for nitride materials and light emitters,” Proc. IEEE 85(11), 1750–1751 (1997). S. Nakamura, M. Senoh, N. Iwasa, and S. Nagahama, “High-brightness InGaN blue, green and yellow lightemitting diodes with quantum well structures,” Jpn. J. Appl. Phys. 34(Part 2, No. 7A), L797–L799 (1995). C. Wetzel, T. Salagaj, T. Detchprohm, P. Li, and J. S. Nelson, “GaInN/GaN growth optimization for high-power green light-emitting diodes,” Appl. Phys. Lett. 85(6), 866–868 (2004). Lightimes News, “Osram Achieves 147 lm/W with Green LED,” http://www.sslighting.net/documents/articles/news/124526.html Wyszecki, Günter and Stiles, Color Science - Concepts and Methods, quantitative Data and Formulae (2nd ed.) (Wiley-Interscience, 2000) American National Standards Institute, “Energy Star Program Requirements for Integral LED Lamps No. C78.377A,” http://www.energystar.gov/ia/partners/product_specs/program_reqs/archive/ENERGY_STAR_Integral_LED_La mp_Specification_V1.4_FINAL.pdf A. Khanna and P. S. Dutta, “Narrow spectral emission CaMoO4: Eu3+, Dy3+, Tb3+ phosphor crystals for white light emitting diodes,” J. Solid State Chem. 198, 93–100 (2013). A. Khanna and P. S. Dutta, “CaWO4:Eu3+, Dy3+, Tb3+ Phosphor Crystals for Solid-State Lighting Application,” ECS Trans. 41(37), 39–48 (2012). A. Žukauskas, R. Vaicekauskas, A. Tuzikas, A. Petrulis, R. Stanikūnas, A. Švegžda, P. Eidikas, and P. Vitta, “Firelight LED source: Toward a balanced approach to the performance of solid-state lighting for outdoor environments,” IEEE Photonics J. 6(3), 1–16 (2014). P. Vitta, R. Stanikūnas, A. Tuzikas, I. Reklaitis, A. Stonkus, A. Petrulis, H. Vaitkevičius, and A. Žukauskas, “Energy-saving approaches to solid state street lighting,” in Proceedings of IEEE Eleventh International Conference on Solid State Lighting 81231H (2011). S. Muthu, F. J. P. Schuurmans, and M. D. Pashley, “Red, green, and blue LEDs for white light illumination,” IEEE J. Sel. Top. Quantum Electron. 8(2), 333–338 (2002).
#233251 - $15.00 USD (C) 2015 OSA
Received 26 Jan 2015; revised 28 Feb 2015; accepted 2 Mar 2015; published 9 Mar 2015 6 Apr 2015 | Vol. 23, No. 7 | DOI:10.1364/OE.23.00A309 | OPTICS EXPRESS A309
15. A. Žukauskas, R. Vaicekauskas, P. Vitta, A. Zabiliūtė, A. Petrulis, and M. Shur, “Color rendition engineering of phosphor-converted light-emitting diodes,” Opt. Express 21(22), 26642–26656 (2013). 16. L. Zhang, X. Guo, T. Liang, X. Gu, Q. Lin, and G. Shen, “Color rendering and luminous efficacy of trichromatic and tetrachromatic LED-based white LEDs,” Microelectron. J. 38(1), 1–6 (2006). 17. 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 solidstate lighting,” Laser and Photonics Reviews 1(4), 307–333 (2007). 18. A. Liu, M. Sandipan, and M. Shur, “LED illuminant on the Ambient Light”, Proc. SPIE 9190, Thirteenth International Conference on Solid State Lighting, 919004 (2014). 19. M. Sivak, “Mercury-free HID headlamps: glare and color rendering,” http://deepblue.lib.umich.edu/handle/2027.42/55201.
1. Introduction One of the key challenges in solid-state lighting is to improve the radiant efficiency of green emitters [1–6]. Figure 1 showing the typical radiant efficiencies of colored LEDs versus the wavelength clearly illustrates this problem. One of the approaches to improve the radiant efficiency of the green solid-state sources is to use the phosphor converted green LED source. An example is a green-emitting LED prototype from Osram demonstrating a record efficacy of 147 lumens per watt (lm/W) at a 530 nanometers (nm) wavelength with a spectral width of 35 nm [7].
Fig. 1. Radiant efficiency of color LEDs. Source: Philips Rebel Colors series and CREE XP-E color series. Dashed line is the guide for an eye.
In this paper, we demonstrate the significance of a new approach using narrow line-width green phosphor. Figure 2 schematically shows a typical wide band green phosphor and narrow line green phosphor (used in this study) superimposed on the photopic eye sensitivity function [8]. Because of a very strong dependence of the eye sensitivity on the wavelength near 555 nm, the conversion of light using a narrow line phosphor with the peak wavelength closely aligned with the peak in the human eye sensitivity might be much more efficient for increasing LER.
#233251 - $15.00 USD (C) 2015 OSA
Received 26 Jan 2015; revised 28 Feb 2015; accepted 2 Mar 2015; published 9 Mar 2015 6 Apr 2015 | Vol. 23, No. 7 | DOI:10.1364/OE.23.00A309 | OPTICS EXPRESS A310
Fig. 2. SPD of wide band and narrow line green phosphors. The data for the wide band phosphor is from CREE XP-E Green LED. The narrow line phosphor is our data. The SPD is measured at a drive current of 350mA at a temperature of 25°C.
2. Experimental The wide band red, green and blue LEDs used in this work are commercial LEDs from Philips Rebel Color LED series. For the narrow line green phosphor, we used calcium molybdate activated byTb3+ ions (CaMoO4: Tb3+).The concentration of the Tb3+ ions in the lattice was fixed at 5 mole percent as it gives rise to the highest possible luminescence intensity from Tb3+ ions [10]. Sodium ions (Na+) were also introduced in the calcium molybdate lattice to compensate for the extra positive charge accompanying trivalent lanthanide dopants (Tb3+). Charge compensation with monovalent Group (I) ions like Na+ has been shown to be very effective in improving the luminescence intensity of narrow line-width phosphors [10, 11]. The intense phosphor emission at 550 nm in Fig. 2 results from the f-f transition (5D4→7 F5) of Tb3+ ion. This transition is unaffected by the host lattice crystal field and lattice vibrations giving rise to the narrow emission line with full width at half maxima of ~5-10 nm. The Ca0.9MoO4: Tb0.053+/ Na0.053+ phosphor was synthesized by flux growth method described in references 10 and 11. The following precursor materials purchased from Alfa Aesar were used for the flux crystal growth: CaO (99.95% purity), MoO3 (99.8% purity), NaCl (99% purity) and TbCl3.6H2O (99.9% purity). Molybdenum (VI) oxide (MoO3) with a melting point of 795 °C acted as the flux material for dissolution of other precursors during the phosphor crystal growth. Prior to the synthesis of the phosphor compound, the aforementioned chemicals were mixed in the appropriate stoichiometric ratio in the Retsch PM 100 Planetary Ball Mill for 30 minutes at 600 rpm. The ball milled powders were reacted in alumina crucibles at 1150 °C for 15-24 hours to form a homogeneous liquid phase. Then the temperature of the reaction mixture was lowered at 2-3 °C per hour (hr) to 650 °C for crystal growth using a self nucleation process. Thereafter, the crystals were cooled to room temperature at a rate of 50-100 °C per hour. The crystals were extracted by dissolving the unreacted flux in hot water or in mild HCl-water mixture (pH ~2). For characterization of the photoluminescence properties of the phosphors, the flux grown crystals were ground into fine powders in the Retsch PM 100 Planetary Ball Mill and dispersed in silicone (RTV 6126 GE Silicones) on a fixed area of 2.25 cm2 of clear glass slides. The area density of the phosphors dispersed in silicone was kept constant at 10 mg/cm2. The spectral power distribution of the phosphor samples on glass slides were measured and analyzed using Ocean Optics Spectrometer (Jaz Spectroscopy Suite) coupled with an integrating sphere and Spectral Suite software. A near-UV LED with a peak emission wavelength of 380 nm was used as the excitation source for the phosphor samples. The near#233251 - $15.00 USD (C) 2015 OSA
Received 26 Jan 2015; revised 28 Feb 2015; accepted 2 Mar 2015; published 9 Mar 2015 6 Apr 2015 | Vol. 23, No. 7 | DOI:10.1364/OE.23.00A309 | OPTICS EXPRESS A311
UV LEDs were driven at 350 mA and 3.5 V. The radiant efficiency of near-UV LEDs is rapidly approaching 50% and in the future, it is anticipated to yield efficiency values close to that of the blue LEDs. To demonstrate the improvement achieved by using the narrow line phosphor, we compared two pair of RGB emitters (see Fig. 3). First, the conventional wide band RGB light source (Fig. 3(a)) is compared with the RGB source with a wide band green phosphor augmented with the narrow line green phosphor (Fig. 3(b)). Second, to explore the potential of narrow line phosphor in spectral design, we compared RGB emitters with optimized SPDs. The reference source is a wide band RGB emitter with optimized peak wavelengths and channel intensities for optimal color rendering and high efficacy (Fig. 3(c)). The reference source is also compared to the high efficacy RGB source with wide band blue and red phosphors along with a narrow line green phosphor (Fig. 3(d)). The efficacy is quantified by the Luminous Efficacy of Radiation (LER), while the color rendering is represented by Color Rendering Index (CRI).
Fig. 3. Spectral power distribution of four RGB emitters.
In this work, we alter the SPD of various emitters by independently tuning the intensities of various color channels. For the wide band RGB channels, the SPDs are linearly scaled to various intensities. For the narrow line green channel, we used the Gaussian function to approximate the phosphor emission: S (λ , λc ) = I ⋅ e
− ( λ − λc )2
σ2
(1)
Here S is the relative spectral intensity, I is the intensity scaling factor ( I ≥ 0 ), λ is the wavelengths, λc is the peak wavelength, set to be 550 nm in this study and σ 2 = 18nm 2 . This approximation gives Gaussian of a Full Width at Half Maximum (FWHM) of about 7 nm. The peak wavelength and the full width at half maximum (FWHM) are listed in Table 1. Table 1. Peak wavelength and FWHM of LEDs LED type Red Wide green Blue Narrow green
Peak wavelength (nm) 625 520 477 550
FWHM (nm) 16 31 22 7
LER (lm/W) 242.7 481.5 109.8 678.4
For the wide band RGB emitter (Fig. 3(a)), the intensities of three channels were set to achieve the desired Correlated Color Temperature (CCT) on the blackbody curve. The selected CCTs ranged from 3000 K to 8000 K, with a step of 100 K. For the narrow green channel augmented RGB emitter (Fig. 3(b)), the narrow green channel was gradually added to the original wide band RGB channels. As a result, the color coordinate of the combined light is deviated from the blackbody curve, moving towards the coordinate of the narrow green phosphor. The color deviation in this work was measured by Duv, which is the distance between the color coordinates of the combined light and the corresponding blackbody of the #233251 - $15.00 USD (C) 2015 OSA
Received 26 Jan 2015; revised 28 Feb 2015; accepted 2 Mar 2015; published 9 Mar 2015 6 Apr 2015 | Vol. 23, No. 7 | DOI:10.1364/OE.23.00A309 | OPTICS EXPRESS A312
same CCT, measured in CIE 1960 color space. We also used the RGB emitters with optimized SPDs (Fig. 2(c) and Fig. 2(d)) for comparison. In this case, the peak wavelengths of wide band phosphors were allowed to vary. The corresponding phosphor spectra were modeled using the analytical approximation method described by Žukauskas et al. [15]. The reference source was a high fidelity emitter with the optimized efficacy. The corresponding multi-objective optimization method used was described in [18]. The target points were on the blackbody curve (Duv = 0) with the range of CCTs from 3000 K to 7000 K. The experimental source used the narrow green phosphor instead. The efficacy was optimized with CRI being constraint to 85, color point being constraint to various CCT with Duv = 0. To trade CRI for higher efficacy, narrow green phosphor component was gradually added to the experimental source. As a result, the color point deviated from the blackbody curve. 3. Result Figure 4 shows the LER result of the augmented RGB emitter (Fig. 3(b) compare to Fig. 3(a)). The white circles show the LER of the original RGB emitters with the color coordinates on blackbody curve. The filled squares show the LER of augmented RGB emitter. As more intense narrow green channel is added to the RGB emitter, the combined light becomes more and more greenish deviating from the blackbody curve as expected. Figure 4 shows the LERs with Duv from 0 to 0.02. White squares and triangles marked the LERs of Duv at 0.007 and 0.02 respectively.
Fig. 4. Efficacy of wide band RGB emitter augmented by narrow line-width green LED. Black square, LER of the greenish light. White circle, square and triangle showed LER at specified Duv. The inset shows the SPD of the improved source for CCT = 5051 K and Duv = 0.007.
The results in Fig. 4 demonstrate the improvement of efficacy after the narrow green channel is added. Although the combined light becomes greenish, the value of Duv is less than 0.007 which is considered as tolerable for most applications [9]. Before reaching this threshold, 10 ~20 lm/W higher efficacy could be achieved, depending on the CCT. For applications, where the intense illumination is the main requirement such as street or automotive headlight lighting [12–14, 19], reducing energy consumption is more important than a higher color quality. Hence adding a narrow green light to the wide band RGB emitter could be an effective solution. Figure 4 shows the results up to a Duv = 0.02. At this value, 35~50 lm/W efficacy improvement can be expected.
#233251 - $15.00 USD (C) 2015 OSA
Received 26 Jan 2015; revised 28 Feb 2015; accepted 2 Mar 2015; published 9 Mar 2015 6 Apr 2015 | Vol. 23, No. 7 | DOI:10.1364/OE.23.00A309 | OPTICS EXPRESS A313
Fig. 5. Efficacy and color quality of optimized reference RGB cluster and narrow green phosphor source. (a) LER (b) CRI. The inset shows the SPD of the improved source for CCT = 5000 K and Duv = 0. The black square shows LER and CRI of the optimized conventional RGB source. Colored shapes show the LER and CRI of the RGB emitter with narrow green phosphor. Duv at 0, 0.007, 0.014 and 0.020 are demonstrated respectively.
Figure 5 shows the CRI and LER result of the RGB emitter with replaced green channel (see SPD in Fig. 3(d)). The result is compared with reference sources (black squares in Fig. 5). The reference sources are a series of high fidelity wide band RGB emitters with modeled SPD (Fig. 3(c)) to achieve optimal LER. For the narrow green source, as extra green component being added to the emitter, the color points are deviated from the blackbody curve. Figure 5 plotted the LER and CRI scores for Duv at 0, 0.007, 0.014 and 0.020 respectively (magenta circle, green square blue triangle and red inverted triangles respectively). Comparing the narrow green source at Duv = 0 (see magenta squares in Fig. 5) with the reference source, a small fraction of the color quality (CRI = 5~10) is sacrificed for as much as ~20 lm/W efficacy. As seen in the figure, this trade-off is more beneficial at lower CCTs such as 3000 K. The drop of CRI can be understood by the barely covered wavelengths in the green range, except for the vicinity of 555 nm. It must be noted that the established SPD and the efficacy of the reference source can be enhanced by an additional 20 lm/W by adding the narrow green phosphor, thus demonstrating the significance of high efficiency narrow green phosphor. Moreover, with extra green component added to the narrow green source, even higher efficacy can be expected. In these cases, both color quality (CRI) and color appearance (Duv) can be achieved simultaneously. According to the ENERGY STAR standard for general lighting applications [9], a Duv of 0.007 can be tolerated. At the 0.007 Duv, up to 340 lm/W efficacy as well as more than 82 CRI can be achieved (see green squares in Fig. 5). A comparison of related studies by Zhang et al [16] demonstrated trichromatic white light source with Duv = 0.021 and CCT = 4400 K. The achieved CRI and LER are 82.4 and 316
#233251 - $15.00 USD (C) 2015 OSA
Received 26 Jan 2015; revised 28 Feb 2015; accepted 2 Mar 2015; published 9 Mar 2015 6 Apr 2015 | Vol. 23, No. 7 | DOI:10.1364/OE.23.00A309 | OPTICS EXPRESS A314
lm/W. In comparison, at the same CCT, the narrow green phosphor provides a CRI of 83 and LER of 330 lm/W, with a much smaller Duv (0.014 or less). At the 0.020 Duv, LER of the narrow green phosphor can be as high as 357 lm/W, while CRI is sacrificed by 10. Comparing to the conventional high efficacy RGB emitters presented in Fig. 5, narrow green phosphor provides ~40 lm/W extra efficacy over a wide range of CCTs. In general, narrow line-width LED sources are expected to provide high efficacy. In [17], Philips et al. theoretically simulated white light using 1 nm width Red, Yellow, Green and Blue (RYGB) laser light sources and observed a LER of 408 lm/W and 90 CRI for a CCT of 3000 K. 4. Conclusions Our results show that using a narrow line green phosphor with the peak wavelength closely aligned with the peak of the human eye sensitivity significantly improves the Luminous Efficacy of Radiation (LER) for Red-Green-Blue (RGB) emitters. For practical solutions, the improvement in LER increases with the deviation from the Planckian locus reaching 35 to 50 lm/W (depending on the color temperature) for tolerable Duv = 0.02. For optimized high efficacy conventional RGB emitters, an improvement up to 20 lm/W for efficacies has been demonstrated at an expense of 5 to 10 lower CRI values. Acknowledgments This work was supported by the National Science Foundation under NSF Cooperative Agreement No. EEC-0812056 and in part by New York State under NYSTAR contract C090145.
#233251 - $15.00 USD (C) 2015 OSA
Received 26 Jan 2015; revised 28 Feb 2015; accepted 2 Mar 2015; published 9 Mar 2015 6 Apr 2015 | Vol. 23, No. 7 | DOI:10.1364/OE.23.00A309 | OPTICS EXPRESS A315
Abstract: We demonstrate that using a narrow line-width green phosphor with the peak wavelength closely aligned with the peak in the human eye sensitivity significantly improves the Luminous Efficacy of Radiation (LER) for Red-Green-Blue (RGB) emitters. Compared to the traditional RGB sources, the improvement in LER of 20 lm/W can be achieved. Combining the narrow band green phosphor with conventional wide band red and blue phosphors allows for trading off these improvements against the deviation from the Planckian locus for even higher LER. The light sources with the narrow line green phosphor are particularly promising for high energy efficiency and high intensity illumination, where somewhat compromises can be made in the color quality such as in automotive, outdoor spaces, industrial ware-houses, public places (train stations, airports) etc.. © 2015 Optical Society of America OCIS codes: (160.5690) Rare-earth-doped materials; (230.3670) Light-emitting diodes; (330.1715) Color, rendering and metamerism.
References and links 1. 2. 3. 4. 5. 6. 7. 8. 9.
10. 11. 12. 13. 14.
A. Žukauskas, M. Shur, and R. Gaska, Introduction to Solid-State Lighting (Wiley, 2002). E. F. Schubert, Lighting-Emitting Diodes (Cambridge University, 2003). M. H. Crawford, “LEDs for Solid-State Lighting: Performance Challenges and Recent Advances,” IEEE J. Sel. Top. Quantum Electron. 15(4), 1028–1040 (2009). I. Akasaki and C. Wetzel, “Future challenges and directions for nitride materials and light emitters,” Proc. IEEE 85(11), 1750–1751 (1997). S. Nakamura, M. Senoh, N. Iwasa, and S. Nagahama, “High-brightness InGaN blue, green and yellow lightemitting diodes with quantum well structures,” Jpn. J. Appl. Phys. 34(Part 2, No. 7A), L797–L799 (1995). C. Wetzel, T. Salagaj, T. Detchprohm, P. Li, and J. S. Nelson, “GaInN/GaN growth optimization for high-power green light-emitting diodes,” Appl. Phys. Lett. 85(6), 866–868 (2004). Lightimes News, “Osram Achieves 147 lm/W with Green LED,” http://www.sslighting.net/documents/articles/news/124526.html Wyszecki, Günter and Stiles, Color Science - Concepts and Methods, quantitative Data and Formulae (2nd ed.) (Wiley-Interscience, 2000) American National Standards Institute, “Energy Star Program Requirements for Integral LED Lamps No. C78.377A,” http://www.energystar.gov/ia/partners/product_specs/program_reqs/archive/ENERGY_STAR_Integral_LED_La mp_Specification_V1.4_FINAL.pdf A. Khanna and P. S. Dutta, “Narrow spectral emission CaMoO4: Eu3+, Dy3+, Tb3+ phosphor crystals for white light emitting diodes,” J. Solid State Chem. 198, 93–100 (2013). A. Khanna and P. S. Dutta, “CaWO4:Eu3+, Dy3+, Tb3+ Phosphor Crystals for Solid-State Lighting Application,” ECS Trans. 41(37), 39–48 (2012). A. Žukauskas, R. Vaicekauskas, A. Tuzikas, A. Petrulis, R. Stanikūnas, A. Švegžda, P. Eidikas, and P. Vitta, “Firelight LED source: Toward a balanced approach to the performance of solid-state lighting for outdoor environments,” IEEE Photonics J. 6(3), 1–16 (2014). P. Vitta, R. Stanikūnas, A. Tuzikas, I. Reklaitis, A. Stonkus, A. Petrulis, H. Vaitkevičius, and A. Žukauskas, “Energy-saving approaches to solid state street lighting,” in Proceedings of IEEE Eleventh International Conference on Solid State Lighting 81231H (2011). S. Muthu, F. J. P. Schuurmans, and M. D. Pashley, “Red, green, and blue LEDs for white light illumination,” IEEE J. Sel. Top. Quantum Electron. 8(2), 333–338 (2002).
#233251 - $15.00 USD (C) 2015 OSA
Received 26 Jan 2015; revised 28 Feb 2015; accepted 2 Mar 2015; published 9 Mar 2015 6 Apr 2015 | Vol. 23, No. 7 | DOI:10.1364/OE.23.00A309 | OPTICS EXPRESS A309
15. A. Žukauskas, R. Vaicekauskas, P. Vitta, A. Zabiliūtė, A. Petrulis, and M. Shur, “Color rendition engineering of phosphor-converted light-emitting diodes,” Opt. Express 21(22), 26642–26656 (2013). 16. L. Zhang, X. Guo, T. Liang, X. Gu, Q. Lin, and G. Shen, “Color rendering and luminous efficacy of trichromatic and tetrachromatic LED-based white LEDs,” Microelectron. J. 38(1), 1–6 (2006). 17. 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 solidstate lighting,” Laser and Photonics Reviews 1(4), 307–333 (2007). 18. A. Liu, M. Sandipan, and M. Shur, “LED illuminant on the Ambient Light”, Proc. SPIE 9190, Thirteenth International Conference on Solid State Lighting, 919004 (2014). 19. M. Sivak, “Mercury-free HID headlamps: glare and color rendering,” http://deepblue.lib.umich.edu/handle/2027.42/55201.
1. Introduction One of the key challenges in solid-state lighting is to improve the radiant efficiency of green emitters [1–6]. Figure 1 showing the typical radiant efficiencies of colored LEDs versus the wavelength clearly illustrates this problem. One of the approaches to improve the radiant efficiency of the green solid-state sources is to use the phosphor converted green LED source. An example is a green-emitting LED prototype from Osram demonstrating a record efficacy of 147 lumens per watt (lm/W) at a 530 nanometers (nm) wavelength with a spectral width of 35 nm [7].
Fig. 1. Radiant efficiency of color LEDs. Source: Philips Rebel Colors series and CREE XP-E color series. Dashed line is the guide for an eye.
In this paper, we demonstrate the significance of a new approach using narrow line-width green phosphor. Figure 2 schematically shows a typical wide band green phosphor and narrow line green phosphor (used in this study) superimposed on the photopic eye sensitivity function [8]. Because of a very strong dependence of the eye sensitivity on the wavelength near 555 nm, the conversion of light using a narrow line phosphor with the peak wavelength closely aligned with the peak in the human eye sensitivity might be much more efficient for increasing LER.
#233251 - $15.00 USD (C) 2015 OSA
Received 26 Jan 2015; revised 28 Feb 2015; accepted 2 Mar 2015; published 9 Mar 2015 6 Apr 2015 | Vol. 23, No. 7 | DOI:10.1364/OE.23.00A309 | OPTICS EXPRESS A310
Fig. 2. SPD of wide band and narrow line green phosphors. The data for the wide band phosphor is from CREE XP-E Green LED. The narrow line phosphor is our data. The SPD is measured at a drive current of 350mA at a temperature of 25°C.
2. Experimental The wide band red, green and blue LEDs used in this work are commercial LEDs from Philips Rebel Color LED series. For the narrow line green phosphor, we used calcium molybdate activated byTb3+ ions (CaMoO4: Tb3+).The concentration of the Tb3+ ions in the lattice was fixed at 5 mole percent as it gives rise to the highest possible luminescence intensity from Tb3+ ions [10]. Sodium ions (Na+) were also introduced in the calcium molybdate lattice to compensate for the extra positive charge accompanying trivalent lanthanide dopants (Tb3+). Charge compensation with monovalent Group (I) ions like Na+ has been shown to be very effective in improving the luminescence intensity of narrow line-width phosphors [10, 11]. The intense phosphor emission at 550 nm in Fig. 2 results from the f-f transition (5D4→7 F5) of Tb3+ ion. This transition is unaffected by the host lattice crystal field and lattice vibrations giving rise to the narrow emission line with full width at half maxima of ~5-10 nm. The Ca0.9MoO4: Tb0.053+/ Na0.053+ phosphor was synthesized by flux growth method described in references 10 and 11. The following precursor materials purchased from Alfa Aesar were used for the flux crystal growth: CaO (99.95% purity), MoO3 (99.8% purity), NaCl (99% purity) and TbCl3.6H2O (99.9% purity). Molybdenum (VI) oxide (MoO3) with a melting point of 795 °C acted as the flux material for dissolution of other precursors during the phosphor crystal growth. Prior to the synthesis of the phosphor compound, the aforementioned chemicals were mixed in the appropriate stoichiometric ratio in the Retsch PM 100 Planetary Ball Mill for 30 minutes at 600 rpm. The ball milled powders were reacted in alumina crucibles at 1150 °C for 15-24 hours to form a homogeneous liquid phase. Then the temperature of the reaction mixture was lowered at 2-3 °C per hour (hr) to 650 °C for crystal growth using a self nucleation process. Thereafter, the crystals were cooled to room temperature at a rate of 50-100 °C per hour. The crystals were extracted by dissolving the unreacted flux in hot water or in mild HCl-water mixture (pH ~2). For characterization of the photoluminescence properties of the phosphors, the flux grown crystals were ground into fine powders in the Retsch PM 100 Planetary Ball Mill and dispersed in silicone (RTV 6126 GE Silicones) on a fixed area of 2.25 cm2 of clear glass slides. The area density of the phosphors dispersed in silicone was kept constant at 10 mg/cm2. The spectral power distribution of the phosphor samples on glass slides were measured and analyzed using Ocean Optics Spectrometer (Jaz Spectroscopy Suite) coupled with an integrating sphere and Spectral Suite software. A near-UV LED with a peak emission wavelength of 380 nm was used as the excitation source for the phosphor samples. The near#233251 - $15.00 USD (C) 2015 OSA
Received 26 Jan 2015; revised 28 Feb 2015; accepted 2 Mar 2015; published 9 Mar 2015 6 Apr 2015 | Vol. 23, No. 7 | DOI:10.1364/OE.23.00A309 | OPTICS EXPRESS A311
UV LEDs were driven at 350 mA and 3.5 V. The radiant efficiency of near-UV LEDs is rapidly approaching 50% and in the future, it is anticipated to yield efficiency values close to that of the blue LEDs. To demonstrate the improvement achieved by using the narrow line phosphor, we compared two pair of RGB emitters (see Fig. 3). First, the conventional wide band RGB light source (Fig. 3(a)) is compared with the RGB source with a wide band green phosphor augmented with the narrow line green phosphor (Fig. 3(b)). Second, to explore the potential of narrow line phosphor in spectral design, we compared RGB emitters with optimized SPDs. The reference source is a wide band RGB emitter with optimized peak wavelengths and channel intensities for optimal color rendering and high efficacy (Fig. 3(c)). The reference source is also compared to the high efficacy RGB source with wide band blue and red phosphors along with a narrow line green phosphor (Fig. 3(d)). The efficacy is quantified by the Luminous Efficacy of Radiation (LER), while the color rendering is represented by Color Rendering Index (CRI).
Fig. 3. Spectral power distribution of four RGB emitters.
In this work, we alter the SPD of various emitters by independently tuning the intensities of various color channels. For the wide band RGB channels, the SPDs are linearly scaled to various intensities. For the narrow line green channel, we used the Gaussian function to approximate the phosphor emission: S (λ , λc ) = I ⋅ e
− ( λ − λc )2
σ2
(1)
Here S is the relative spectral intensity, I is the intensity scaling factor ( I ≥ 0 ), λ is the wavelengths, λc is the peak wavelength, set to be 550 nm in this study and σ 2 = 18nm 2 . This approximation gives Gaussian of a Full Width at Half Maximum (FWHM) of about 7 nm. The peak wavelength and the full width at half maximum (FWHM) are listed in Table 1. Table 1. Peak wavelength and FWHM of LEDs LED type Red Wide green Blue Narrow green
Peak wavelength (nm) 625 520 477 550
FWHM (nm) 16 31 22 7
LER (lm/W) 242.7 481.5 109.8 678.4
For the wide band RGB emitter (Fig. 3(a)), the intensities of three channels were set to achieve the desired Correlated Color Temperature (CCT) on the blackbody curve. The selected CCTs ranged from 3000 K to 8000 K, with a step of 100 K. For the narrow green channel augmented RGB emitter (Fig. 3(b)), the narrow green channel was gradually added to the original wide band RGB channels. As a result, the color coordinate of the combined light is deviated from the blackbody curve, moving towards the coordinate of the narrow green phosphor. The color deviation in this work was measured by Duv, which is the distance between the color coordinates of the combined light and the corresponding blackbody of the #233251 - $15.00 USD (C) 2015 OSA
Received 26 Jan 2015; revised 28 Feb 2015; accepted 2 Mar 2015; published 9 Mar 2015 6 Apr 2015 | Vol. 23, No. 7 | DOI:10.1364/OE.23.00A309 | OPTICS EXPRESS A312
same CCT, measured in CIE 1960 color space. We also used the RGB emitters with optimized SPDs (Fig. 2(c) and Fig. 2(d)) for comparison. In this case, the peak wavelengths of wide band phosphors were allowed to vary. The corresponding phosphor spectra were modeled using the analytical approximation method described by Žukauskas et al. [15]. The reference source was a high fidelity emitter with the optimized efficacy. The corresponding multi-objective optimization method used was described in [18]. The target points were on the blackbody curve (Duv = 0) with the range of CCTs from 3000 K to 7000 K. The experimental source used the narrow green phosphor instead. The efficacy was optimized with CRI being constraint to 85, color point being constraint to various CCT with Duv = 0. To trade CRI for higher efficacy, narrow green phosphor component was gradually added to the experimental source. As a result, the color point deviated from the blackbody curve. 3. Result Figure 4 shows the LER result of the augmented RGB emitter (Fig. 3(b) compare to Fig. 3(a)). The white circles show the LER of the original RGB emitters with the color coordinates on blackbody curve. The filled squares show the LER of augmented RGB emitter. As more intense narrow green channel is added to the RGB emitter, the combined light becomes more and more greenish deviating from the blackbody curve as expected. Figure 4 shows the LERs with Duv from 0 to 0.02. White squares and triangles marked the LERs of Duv at 0.007 and 0.02 respectively.
Fig. 4. Efficacy of wide band RGB emitter augmented by narrow line-width green LED. Black square, LER of the greenish light. White circle, square and triangle showed LER at specified Duv. The inset shows the SPD of the improved source for CCT = 5051 K and Duv = 0.007.
The results in Fig. 4 demonstrate the improvement of efficacy after the narrow green channel is added. Although the combined light becomes greenish, the value of Duv is less than 0.007 which is considered as tolerable for most applications [9]. Before reaching this threshold, 10 ~20 lm/W higher efficacy could be achieved, depending on the CCT. For applications, where the intense illumination is the main requirement such as street or automotive headlight lighting [12–14, 19], reducing energy consumption is more important than a higher color quality. Hence adding a narrow green light to the wide band RGB emitter could be an effective solution. Figure 4 shows the results up to a Duv = 0.02. At this value, 35~50 lm/W efficacy improvement can be expected.
#233251 - $15.00 USD (C) 2015 OSA
Received 26 Jan 2015; revised 28 Feb 2015; accepted 2 Mar 2015; published 9 Mar 2015 6 Apr 2015 | Vol. 23, No. 7 | DOI:10.1364/OE.23.00A309 | OPTICS EXPRESS A313
Fig. 5. Efficacy and color quality of optimized reference RGB cluster and narrow green phosphor source. (a) LER (b) CRI. The inset shows the SPD of the improved source for CCT = 5000 K and Duv = 0. The black square shows LER and CRI of the optimized conventional RGB source. Colored shapes show the LER and CRI of the RGB emitter with narrow green phosphor. Duv at 0, 0.007, 0.014 and 0.020 are demonstrated respectively.
Figure 5 shows the CRI and LER result of the RGB emitter with replaced green channel (see SPD in Fig. 3(d)). The result is compared with reference sources (black squares in Fig. 5). The reference sources are a series of high fidelity wide band RGB emitters with modeled SPD (Fig. 3(c)) to achieve optimal LER. For the narrow green source, as extra green component being added to the emitter, the color points are deviated from the blackbody curve. Figure 5 plotted the LER and CRI scores for Duv at 0, 0.007, 0.014 and 0.020 respectively (magenta circle, green square blue triangle and red inverted triangles respectively). Comparing the narrow green source at Duv = 0 (see magenta squares in Fig. 5) with the reference source, a small fraction of the color quality (CRI = 5~10) is sacrificed for as much as ~20 lm/W efficacy. As seen in the figure, this trade-off is more beneficial at lower CCTs such as 3000 K. The drop of CRI can be understood by the barely covered wavelengths in the green range, except for the vicinity of 555 nm. It must be noted that the established SPD and the efficacy of the reference source can be enhanced by an additional 20 lm/W by adding the narrow green phosphor, thus demonstrating the significance of high efficiency narrow green phosphor. Moreover, with extra green component added to the narrow green source, even higher efficacy can be expected. In these cases, both color quality (CRI) and color appearance (Duv) can be achieved simultaneously. According to the ENERGY STAR standard for general lighting applications [9], a Duv of 0.007 can be tolerated. At the 0.007 Duv, up to 340 lm/W efficacy as well as more than 82 CRI can be achieved (see green squares in Fig. 5). A comparison of related studies by Zhang et al [16] demonstrated trichromatic white light source with Duv = 0.021 and CCT = 4400 K. The achieved CRI and LER are 82.4 and 316
#233251 - $15.00 USD (C) 2015 OSA
Received 26 Jan 2015; revised 28 Feb 2015; accepted 2 Mar 2015; published 9 Mar 2015 6 Apr 2015 | Vol. 23, No. 7 | DOI:10.1364/OE.23.00A309 | OPTICS EXPRESS A314
lm/W. In comparison, at the same CCT, the narrow green phosphor provides a CRI of 83 and LER of 330 lm/W, with a much smaller Duv (0.014 or less). At the 0.020 Duv, LER of the narrow green phosphor can be as high as 357 lm/W, while CRI is sacrificed by 10. Comparing to the conventional high efficacy RGB emitters presented in Fig. 5, narrow green phosphor provides ~40 lm/W extra efficacy over a wide range of CCTs. In general, narrow line-width LED sources are expected to provide high efficacy. In [17], Philips et al. theoretically simulated white light using 1 nm width Red, Yellow, Green and Blue (RYGB) laser light sources and observed a LER of 408 lm/W and 90 CRI for a CCT of 3000 K. 4. Conclusions Our results show that using a narrow line green phosphor with the peak wavelength closely aligned with the peak of the human eye sensitivity significantly improves the Luminous Efficacy of Radiation (LER) for Red-Green-Blue (RGB) emitters. For practical solutions, the improvement in LER increases with the deviation from the Planckian locus reaching 35 to 50 lm/W (depending on the color temperature) for tolerable Duv = 0.02. For optimized high efficacy conventional RGB emitters, an improvement up to 20 lm/W for efficacies has been demonstrated at an expense of 5 to 10 lower CRI values. Acknowledgments This work was supported by the National Science Foundation under NSF Cooperative Agreement No. EEC-0812056 and in part by New York State under NYSTAR contract C090145.
#233251 - $15.00 USD (C) 2015 OSA
Received 26 Jan 2015; revised 28 Feb 2015; accepted 2 Mar 2015; published 9 Mar 2015 6 Apr 2015 | Vol. 23, No. 7 | DOI:10.1364/OE.23.00A309 | OPTICS EXPRESS A315
