Open Access
Add to BibSonomyAbstract
A solution for multi-functional indoor light sources is proposed to achieve the new concept of healthy lighting. A remotely controllable light source that embodies a quadruple-chip light-emitting diode and driven by pulse-width-modulation currents is designed. Therefore, spectral power distributions (SPDs) of the light source can be readily controlled. An algorithm, namely the optical power ratio algorithm, is developed to select all suitable SPDs adapted for various applications. Principles of selection are based on those traditional visual indices, as well as on some non-visual parameters such as circadian action factor, circadian efficacy of radiation and circadian illuminance. We investigate in detail the correlation among these parameters and provide SPDs with both decent visual and non-visual performances for three typical cases. The study suggests some fundamental principles for designing healthy light sources, and can be regarded as a guide for designing indoor light sources of the next generation.
© 2016 Optical Society of America
1. Introduction
In the year 2002, the intrinsically photoreceptive retinal ganglion cell was discovered [1]. Shortly thereafter, this cell was found to be associated with the secretion of melatonin, which took charge of the circadian rhythm of human being. Components of the spectral power distribution (SPD) in the blue region of a light source play a key role in reducing the nocturnal melatonin level and thus exert impacts on circadian physiology, alertness and cognitive performance levels [2]. A more detailed study of suppression on nocturnal melatonin has shown that a modeling utilizing spectral weighting functions is urgently called for. Hitherto, Branard et al. [3], Tharpan et al. [4], Gall et al. [5], and Rea et al. [6] have proposed their own functions, respectively. Among these studies, Gall’s model suggests that not only can visible electromagnetic radiation affect the human visual system, but also it can alter the circadian clock that lies behind all biological functions following the daily rhythm [7]. The existence of such non-visual effects lead to a new concept namely ‘healthy lighting’, which requires an light source to take into account physiological and psychological impacts on human beings, in addition to those tradition characteristics. Although popular key parameters of various types concern the quality of a light source, such as chromaticity coordinates (CIE x, CIE y), color rendering index (CRI), correlated color temperature (CCT), luminous efficacy of radiation (LER), the luminous flux and the illuminance (VIL) for vision performance, they fail to gauge non-visual effects [8–10]. Therefore, circadian action factor (CAF), circadian efficacy of radiation (CER) and circadian illuminance (CIL) have been introduced as measures of non-visual effects of light sources [11], and are briefly described below. These parameters, like those traditional ones, highly depend on SPDs of light sources. Consequently, light-emitting diodes (LEDs), whose SPDs can be readily controlled and whose circadian parameters can be optimized, qualify as health-friendly light sources [12,13].
As illumination sources for indoor and outdoor applications, LEDs have been widely utilized in general lighting and, specifically, more advanced – intelligent lighting [14,15]. Hitherto, at least four methods that are capable of attaining a white light source exist. First, although blue chips combined with yellow phosphor can provide high CRI (> 80), their fixed CCT values cannot be suitably controlled for intelligent lighting. Second, a wide range of CCT can be achieved by mixing lights from warm-white and cold-white LEDs, but the chromaticity coordinate of overall emission will drift away from the blackbody locus into the non-white region at certain CCTs. Third, the RGB-tricolor LEDs can emit the pure white light, of which coordinates are located in the vicinity of blackbody locus, however, their CRI cannot be always kept in a high level when the CCT is changed. Forth, Adding into a white LED to form the red-green-blue-white (RGBW) LED will significantly improve the performance [16], for the wide spectral span of white chip will extend the hue, saturation and brightness values available in the color system. These merits make RGBW LED the best candidate for the solution of intelligent lighting. To obtain correct color, we need to accurately control the (mean) light intensity of each component of RGBW LED. A popular solution involves driving individual chips in RGBW LED independently by pulse-width-modulation (PWM) current signal. The mean brightness of each color can be precisely controlled by adjusting the duty cycle.
In this article, a solution is introduced to design the quadruple-color-LED-based indoor light source, which has taken into account both non-visual effects and traditional properties of visible light. At its core lies an algorithm named optical power ratio algorithm (OPRA) that functions as a liaison between desired SPDs and PWM duty cycles. With this solution, we can obtain proper light emissions with finely tailored SPDs that comfort both eyes and bodies.
2. Some relative concepts
2.1 Non-visual parameters
In this paper, we employ the circadian spectral sensitivity function in the Gall’s model [5], shown in Fig. 1(a), to measure the circadian flux of optical sources. In the same figure we also present photopic and scotopic sensitivity functions, which gauge spectral resolved sensitivities of human eyes under bright and dark cases respectively [9].
Fig. 1 (a) Normalized spectra of the circadian sensitivity curve C(λ) [5], the photopic sensitivity curve V(λ) and the scotopic sensitivity curve V’(λ) [9]; (b) The relative melatonin suppression value versus CIL of a 470 nm blue LED (B470 LED). Data are from Ref [17].
Definitions of key parameters associated with the vision and circadian performance are summarized in Table 1. Resembling LER, the CER is defined as the ratio of the circadian luminous flux to the radian flux [18]. The CAF, namely the ratio of CER to LER [18], is used for optimizing circadian performance because of its positive correlation with melatonin suppression. Generally, higher values of CAF indicate more light energy concentrating on the blue spectral region, which is associated with the melatonin suppression. In workplaces, high CAFs can enhance workers’ level of vigilance and excitation, while low CAFs which are required in bedrooms can help people relax. The CIL, defined as the product of VIL and CAF, is considered as a threshold for the relative melatonin suppression regardless of the type of optical sources [11,17]. Therefore, in Fig. 1(b), we can use the relative melatonin suppression value (MSV [19]) curve of a 470 nm blue LED as a standard to analyze the circadian performance [17]. Such a standard curve shows that melatonin suppression rarely occurs below 47 blx (MSV < 10%) whereas 80% of melatonin will be effectively suppressed when the CIL exceeds 250 blx. Briefly speaking, SPDs have to be adapted according to functionalities of rooms.
Table 1. Key parameters for the vision and circadian performance
2.2 Metamerism
As a phenomenon that should be critically considered in SPD designs, the metamerism indicates that two different SPDs exhibit the same color to human’s eyes as they share the identical chromaticity coordinates (thus identical CCTs). Since two SPDs do not actually match, their CRIs and non-visual parameters appear different from each other. For example, two simulated SPDs showed in Fig. 2 are generated by the same RGBW LED with different PWM currents, and their key parameters are presented in Table 2. Results indicate that values of CRI and CAF enormously differ although two combinations share the same chromaticity. Typically, the value of CRI is recommended to surpass 80 for general lighting [20] and to surpass 90 for advanced lighting, such as product-showcase lighting that requires accurate color rendering [21]. Therefore, the PWM current of RGBW LED should be designed carefully to ensure that not only the chromaticity coordinate but also visual and non-visual parameters can meet fundamental requirements.
Fig. 2 SPDs of metameric white lights at 4000K, generated by the same RGBW LED with different PWM currents. (a) The overall SPD of which CRI equals 32. (b) The overall SPD of which CRI equals 82.
Table 2. Parameters of the metameric white light at 4000K
2.3 Principles of choosing visual and non-visual parameters
Principles of fine visual and non-visual parameters diversify with applications of spaces. Although these cases all prefer high CRI for correct color rendering, their requirements for CCT and CAF differ. The product-showcase lighting requires maximum CRI but no specific CAF; the office-at-night lighting needs maximum CAF and CRI ≥ 80; and the bedroom-at-night lighting demands minimum CAF and CRI ≥ 80.
3. Methodology and algorithm
In this section, we present a methodology for obtaining SPDs with high CRIs as well as proper CAFs. For a RGBW LED, the overall SPD consists of SPDs of four individual components (chips). It is worth noting that, under the PWM driving current, the normalized SPD of each chip, which can be measured in advance, is maintained constant at different pulse widths. Therefore, being the linear combination of SPDs of four components, the overall SPD depends on optical powers of each chip, indicated explicitly as
where P and E(λ) denote the optical power and relative SPD; subscripts “o”, “r”, “g”, “b” and “w” denote overall, red, green, blue, and white, respectively. With all SPDs normalized to the area under curves, we obtainIn Eq. (2), Rr, Rg, Rb and Rw stand for the optical power ratio of each LED subject to . Equation (1) can be rewritten as
In the OPRA presented in this article, by changing optical power ratios, we can simulate overall SPDs to achieve optimized parameters of both visual and non-visual effects (Fig. 3).
The CRI, CAF, CCT, and chromaticity coordinate all depend on the relative overall SPD, which further depends on optical power ratios of four chips. Therefore, by simulating all relative overall SPDs, we can select combinations possessing decent CRIs as well as proper non-visual parameters, such as CAF. Subsequently, we can divide the specific optimization process into five steps listed below.
Step I: Optimization goals should be set according to different applications in advance. Three typical applications are discussed in this article and respective goals are summarized in Table 3.
Table 3. Optimization goals for three different situations
Step II: Measure emission spectra of red, green, blue, and white LEDs respectively. Each color is driven by PWM current signals with different duty cycles and constant amplitude at 350 mA. Additionally, identify the correlation between optical powers and duty cycles of each LED chip.
Step III: Change the optical power ratio of each RGBW chip to enumerate all available relative overall SPDs. Because each set of Rr, Rg, Rb and Rw correlates to a unique relative overall SPD, all possible relative overall SPDs can be obtained after all processes shown in Fig. 4 are finished.
Fig. 4 The flow chart of enumerating all available relative SPDs by changing the optical power ratio of RGBW LED.
Step IV: Calculate chromaticity coordinates for each overall SPD. If the chromaticity coordinate is accepted, then we calculate all other parameters concerning visual and circadian performances, such as CRI, LER, VIL, CAF, CER, and CIL. Afterwards, select relative overall SPDs that suits for a specific application most. Note that, under the amplitude-350-mA PWM driving current, optical powers of those absolute overall SPDs with the same relative overall SPD encounter a limitation, at which the largest duty cycle of individual chips equals 100%. With the set of optical power of RGBW chips at a certain CCT, the set of four duty cycles can be calculated via functions between optical powers and duty cycles. Store all sets of duty cycles in the required CCT range.
Step V: Drive the RGBW LED with PWM duty cycle corresponding to a certain application, and adjust the luminous intensity by tuning four duty cycles in the same proportion.
4. Simulation and experimental results
4.1 Hardware
The OPRA can be applied universally to any multiple-chip light sources. In this section, we employ the Cree XLamp XM-L Color LED as an example for evaluating the accuracy of the OPRA. Each chip is first driven by the 350 mA forward current, and the temperature of the heat sink on which the RGBW LED is fixed is maintained at 75°C. Measured by an integrating sphere and a spectrometer (Spectro 320, Instrument Systems, Germany), spectra and parameters of four LED chips are shown in Fig. 5 and Table 4.
Table 4. Parameters of the RGBW LED
Then, we use PWM currents to drive each chip. Plots of optical powers versus duty cycles of each chip respectively are illustrated in Fig. 6. Duty cycles of individual chips are denoted by Dr, Dg, Db and Dw. All R-squares reach 0.999, indicating an obvious linearity which greatly facilitates the control of optical powers.
The control system is accomplished in the manner as follows. Schematics of the current-drive circuit for RGBW LED is shown in Fig. 7. As a system-on-chip Bluetooth module, the CC2540 combines an excellent RF transceiver with an industrial-standard-enhanced 8051 microcontroller unit, and serves for receiving control signals sent by cell phones. The MBI6661, a step-down DC/DC converter, can be regarded as a current source for LED chips, and is controlled by the CC2540.
4.2 Applications
In this section, we discuss three typical applications which possess diverse requirements for visual and non-visual qualities of the indoor lighting.
(1) Product showcase
Large CRI values, at least 90 [21], are desired for product-showcase lighting, in order to own the best color restoration, whereas no particular circadian performance is needed. We change the optical power ratio of red, green, blue and white LEDs to simulate all possible relative SPDs and choose the combinations that possess the maximum CRI with the CCT from 3000K to 6500K. Then, we use corresponding PWM currents of each combination to drive individual LEDs and measure parameters (Table 5), exhibiting satisfactory agreements between measured and set values, with less than 3.5% relative errors, and demonstrating high accuracies and practicalities of the OPRA.
Table 5. Experimental results of RGBW LED for product-showcase lighting
The considerable similarity between set and measured values of CRI and CAF is also showed in Fig. 8(a). Additionally, a positive correlation between CAF and CCT is proved, which indicates that bluish light are more effective than the reddish on the circadian system of human being. As shown in Fig. 8(b), Dw is always kept in high levels because the white LED contains the spectral region corresponding to color amber which can efficiently improve the CRI of the overall SPD. In summary, because CRI remains larger than 93, the Cree XLamp XM-L Color LED can completely fulfil product-showcase lighting requirements with CCT from 3000K to 6500K under PWM driving currents that are generated by the OPRA.
Fig. 8 Product-showcase lighting: (a) Measured and set values of CRI and CAF; (b) PWM duty cycles versus CCT.
(2) Office at night and bedroom at night
Due to applications of offices and bedrooms, optimization goals are to maximize CAF for better work incentive at office and minimize CAF for relax at home, in addition, to make sure CRIs keep higher than 80. According to the OPRA, we survey all available SPDs and select two for each CCT that meet requirements for both cases respectively. Table 6 and Table 7 list experimental results of the office and the bedroom at night, respectively. It is apparent that experimental results and simulation values consist with each other.
Table 6. Experimental results of RGBW LED for office-at-night lighting
Table 7. Experimental results of RGBW LED for bedroom-at-night lighting
Using the data in Table 6 and Table 7, we compare measured and set values of CRI and CAF and then draw plots of duty cycle versus CCT for both office-at-night and bedroom-at-night applications respectively in Fig. 9 and Fig. 10. It is presented in Fig. 9(a) and Fig. 10(a) that the measured CRI and CAF can match set values quite well. Additionally, similar with the product-showcase lighting, Dw of both office-at-night and bedroom-at-night lighting are always kept in high level and Db of both applications are all increasing gradually when the CCT is raising.
Fig. 9 Office-at-night lighting: (a) Measured and set values of CRI and CAF; (b) PWM duty cycles versus CCT.
Fig. 10 Bedroom-at-night lighting: (a) Measured and set values of CRI and CAF; (b) PWM duty cycles versus CCT.
We also compare trends of CRI and CAF with respect to CCT of three different lighting applications in Fig. 11. Subscripts “PS”, “ON”, and “BN” denote product-showcase, office-at-night, and bedroom-at-night, respectively. Since CRI plays a key role for product-showcase lighting, we maximize CRIPS so that values of CRIPS are kept higher than others in Fig. 11(a). When it comes to the office-at-night lighting and the bedroom-at-night lighting, both CRI and CAF should be carefully considered. Form experimental results in Fig. 11(b), as expected, it is clear that for all CCTs, the curve of CAFON is located on the top while that of CAFBN lies at the bottom.
Fig. 11 (a) CRI versus CCT of three different applications. (b) CAF versus CCT of three different applications.
As mentioned before [Fig. 1(b)], in order to create a working environment with incentive, the threshold value of CIL has to surpass 250 blx to ensure MSV can exceed 80%. According to the measured CAF in Table 6, we can calculate VILthON (minimum threshold of VILON) in the required CCT range for office-at-night lighting (Fig. 12). It is obvious that to achieve the same MSV, the warm-white light (3000K) needs more than twice VIL comparing with the cool-white light (6500K). If the overall SPD of the RGBW LED is not appropriately optimized, its CAF cannot reach the maximum value, resulting in an increase of VILthON, and leading to much more energy consumption.
Fig. 12 The threshold value of VILON for better work activity in office and the threshold value of VILBN for a relaxing environment in bedroom.
Similarly, according to the threshold that CIL < 47 blx (MSV < 10%) in bedroom at night, the VILthBN (maximum threshold value of VILBN) is also calculated and presented in Fig. 12. If the actual VILBN exceeds VILthBN at certain a CCT, it will be difficult for people to feel relax. Moreover, in the warm-white region, much more VIL margins exist for people to adjust the lighting intensity comparing with that in the cool-white region. If CAF is not optimized to the minimum value, the curve of VILthBN will drop as a whole, and that may cause the insufficient illuminance especially in high CCT.
5. Conclusions
In this article, a solution for multi-function healthy indoor lighting has been investigated. We have accomplished a two-step mission. The first step involves accomplishing a quadruple-LED based light source of which SPDs are allowed to be readily tailored by tuning duty cycles of PWM driving currents. In the second step, we propose the OPRA for driving the light source to meet requirements of healthy lighting. Unlike previous studies that only investigate the luminous efficiency and color rendering, we consider CAF as the core index of non-visual effects on circadian performance. For adapting the light source to various indoor-lighting applications, we control SPDs to adjust both visual and non-visual parameters. In addition, experimental results demonstrate high accuracies and practicalities of the light source which can be remotely controlled by cellphones. We believe that these light sources concerning non-visual effects will prosper in the near future.
Appendix
For convenient reading, all acronyms or abbreviations are listed in Table 8 below.
Table 8. Acronyms of technical terms presented in this article
Funding
International Science and Technology Cooperation Program of China (2015DFG62190); Major Science and Technology Project between University–Industry Cooperation in Fujian Province (2013H6024); NNSF of China (61504112); Fundamental Research Funds for the Central Universities (20720150026); NSF of Fujian Province (2016R0091).
References and links
1. D. M. Berson, F. A. Dunn, and M. Takao, “Phototransduction by retinal ganglion cells that set the circadian clock,” Science 295(5557), 1070–1073 (2002). [CrossRef] [PubMed]
2. S. M. Pauley, “Lighting for the human circadian clock: recent research indicates that lighting has become a public health issue,” Med. Hypotheses 63(4), 588–596 (2004). [CrossRef] [PubMed]
3. G. C. Brainard, J. P. Hanifin, J. M. Greeson, B. Byrne, G. Glickman, E. Gerner, and M. D. Rollag, “Action spectrum for melatonin regulation in humans: evidence for a novel circadian photoreceptor,” J. Neurosci. 21(16), 6405–6412 (2001). [PubMed]
4. K. Thapan, J. Arendt, and D. J. Skene, “An action spectrum for melatonin suppression: evidence for a novel non-rod, non-cone photoreceptor system in humans,” J. Physiol. 535(1), 261–267 (2001). [CrossRef] [PubMed]
5. D. Gall and K. Bieske, “Definition and measurement of circadian radiometric quantities,” in Proceedings of the CIE Symposium 2004 on Light and Health: Non-visual Effects, (Vienna, Austria, 2004), pp. 129–132.
6. M. S. Rea, M. G. Figueiro, and J. D. Bullough, “Circadian photobiology: an emerging framework for lighting practice and research,” Light. Res. Technol. 34(3), 177–190 (2002). [CrossRef]
7. M. S. Rea, M. G. Figueiro, A. Bierman, and J. D. Bullough, “Circadian light,” J. Circadian Rhythms 8(1), 2 (2010). [CrossRef] [PubMed]
8. Y. Ohno, “Spectral design considerations for white LED color rendering,” Opt. Eng. 44(11), 111302 (2005). [CrossRef]
9. Z. Q. Guo, T. M. Shih, Y. J. Lu, Y. L. Gao, L. H. Zhu, G. L. Chen, J. H. Zhang, S. Q. Lin, and Z. Chen, “Studies of scotopic/photopic ratios for color-tunable white light-emitting diodes,” IEEE Photonics J. 5(4), 8200409 (2013). [CrossRef]
10. Y. Lin, Z. Deng, Z. Guo, Z. Liu, H. Lan, Y. Lu, and Y. Cao, “Study on the correlations between color rendering indices and the spectral power distribution,” Opt. Express 22(S4Suppl 4), A1029–A1039 (2014). [CrossRef] [PubMed]
11. J. H. Oh, S. J. Yang, and Y. R. Do, “Healthy, natural, efficient and tunable lighting: four-package white LED for optimizing the circadian effect, color quality and vision performance,” Light Sci. Appl. 3(2), 1–9 (2014).
12. Z. Luo, H. Chen, Y. Liu, S. Xu, and S. T. Wu, “Color-tunable light emitting diodes based on quantum dot suspension,” Appl. Opt. 54(10), 2845–2850 (2015). [CrossRef] [PubMed]
13. L. Zan, D. Lin, P. Zhong, and G. He, “Optimal spectra of white LED integrated with quantum dots for mesopic vision,” Opt. Express 24(7), 7643–7653 (2016). [CrossRef] [PubMed]
14. E. F. Schubert and J. K. Kim, “Solid-state light sources getting smart,” Science 308(5726), 1274–1278 (2005). [CrossRef] [PubMed]
15. M. H. Crawford, “LEDs for solid state lighting: Performance challenges and recent advances,” IEEE J. Sel. Top. Quantum Electron. 15(4), 1028–1040 (2009). [CrossRef]
16. G. He and L. Zheng, “Color temperature tunable white-light light-emitting diode clusters with high color rendering index,” Appl. Opt. 49(24), 4670–4676 (2010). [CrossRef] [PubMed]
17. J. H. Oh, H. Yoo, H. K. Park, and Y. R. Do, “Analysis of circadian properties and healthy levels of blue light from smartphones at night,” Sci. Rep. 5(28), 11325 (2015). [CrossRef] [PubMed]
18. A. Žukauskas, R. Vaicekauskas, and P. Vitta, “Optimization of solid-state lamps for photobiologically friendly mesopic lighting,” Appl. Opt. 51(35), 8423–8432 (2012). [CrossRef] [PubMed]
19. M. Aubé, J. Roby, and M. Kocifaj, “Evaluating potential spectral impacts of various artificial lights on melatonin suppression, photosynthesis, and star visibility,” PLoS One 8(7), e67798 (2013). [CrossRef] [PubMed]
20. H. Sasabe and J. Kido, “Development of high performance OLEDs for general lighting,” J. Mater. Chem. C Mater. Opt. Electron. Devices 1(9), 1699–1707 (2013). [CrossRef]
21. Q. Dai, L. Hao, Y. Lin, and Z. Cui, “Spectral optimization simulation of white light based on the photopic eye-sensitivity curve,” J. Appl. Phys. 119(5), 053103 (2016). [CrossRef]
