Spectral Design Considerations for White LED Color Rendering
Total Page:16
File Type:pdf, Size:1020Kb
Optical Engineering 44͑11͒, 111302 ͑November 2005͒ Spectral design considerations for white LED color rendering Yoshi Ohno Abstract. White LED spectra for general lighting should be designed for National Institute of Standards and Technology high luminous efficacy as well as good color rendering. Multichip and 100 Bureau Drive phosphor-type white LED models were analyzed by simulation of their Gaithersburg, Maryland 20899 color characteristics and luminous efficacy of radiation, compared with E-mail: [email protected] those of conventional light sources for general lighting. Color rendering characteristics were evaluated based on the CIE Color Rendering Index ͑CRI͒, examining not only Ra but also the special color rendering indices ⌬ * Ri, as well as on the CIELAB color difference Eab for the 14 color samples defined in CIE 13.3. Several models of three-chip and four-chip white LEDs as well as phosphor-type LEDs are optimized for various parameters, and some guidance is given for designing these white LEDs. The simulation analysis also demonstrated several problems with the current CRI, and the need for improvements is discussed. © 2005 Society of Photo-Optical Instrumentation Engineers. ͓DOI: 10.1117/1.2130694͔ Subject terms: color rendering; colorimetry; CRI; lighting; luminous efficacy; solid- state lighting; LED; white LED. Paper SS040576 received Aug. 24, 2004; accepted for publication Jun. 13, 2005; published online Nov. 30, 2005. This paper is a revision of a paper presented at the SPIE conference on Solid State Lighting, Aug. 2004, Denver, Colorado. The paper presented there appears ͑unrefereed͒ in SPIE Proceedings Vol. 5530. 1 Introduction The main driving force for solid-state lighting is the po- tential of huge energy savings on the national or global One of the most important characteristics of light sources scale.6 Thus, when considering spectra of light sources for for general lighting is color rendering. Color rendering is a general illumination, another important aspect to consider property of a light source that tells how natural the colors is luminous efficacy ͑lumens per watt͒. The term luminous of objects look under the given illumination. If color ren- efficacy is normally used for the conversion efficiency from dering is poor, the light source will not be useful for gen- ͑ ͒ 1 the input electrical power watts to the output luminous eral lighting. The U.S. Energy Policy Act of 1992 specifies flux ͑lumens͒. The luminous efficacy of a source is deter- ͑ minimum requirements for both the luminous efficacy lu- mined by two factors: the conversion efficiency from elec- ͒ ͑ ͒2 mens per watt and the Color Rendering Index CRI for trical power to optical power ͑called radiant efficiency or several common types of lamp products sold in the USA. external quantum efficiency7͒ and the conversion factor This is an important aspect to be considered for white from optical power ͑watts͒ to luminous flux ͑lumens͒. The LEDs being developed for general lighting. latter is called the luminous efficacy of radiation ͑LER͒. White light from LEDs is realized by mixture of multi- Since the LER and color rendering are determined solely color LEDs or by combinations of phosphors excited by by the spectrum of the source, white LED spectra should be blue or UV LED emission, and thus they have greater free- optimized for both of these aspects. dom in spectral design than conventional sources. Ques- The difficulty is that color rendering and the LER are tions arise on how the spectra of white LEDs should be generally in a trade-off. Based on the CRI, color rendering designed for good color-rendering performance, e.g., is best achieved by broadband spectra distributed through- whether RGB white LEDs can satisfy the need, or a four- out the visible region, while luminous efficacy is highest color mixture is needed, or whether much broader, continu- with monochromatic radiation at 555 nm. This trade-off is ous spectra are required. To evaluate the color-rendering 2 evident in many existing lamps. By studying the CRI, some performance of light sources, the CRI, recommended by people are led to believe that white LED spectra should ͑ ͒ the Commission Internationale de l’Éclairage CIE ,is mimic the spectrum of the sun or a blackbody. While such available and widely used, but it is known to have 3,4 spectra would give high CRI values, they would suffer sig- deficiencies, especially when used for sources having nificantly from low LER. The challenge in creating LEDs narrowband spectra. A poor correlation between visual 5 for use as illumination sources is to provide the highest evaluation of RGB white LEDs and the CRI is reported. possible energy efficiency while achieving best color ren- The color-rendering problems of white LEDs are being in- dering possible. For this purpose, an accurate metric of vestigated by the CIE Technical Committee 1-62, with a color rendering is of importance. If the metric is incorrect, future plan to develop a new metric. energy will be wasted. To analyze the possible performance of white LEDs and 0091-3286/2005/$22.00 © 2005 SPIE also the problems of the CRI, a simulation program has Optical Engineering 111302-1 November 2005/Vol. 44͑11͒ Downloaded from SPIE Digital Library on 12 Feb 2010 to 129.6.147.88. Terms of Use: http://spiedl.org/terms Ohno: Spectral design considerations… been developed. Various white LED spectra, of multichip type and phosphor type, were modeled and analyzed in comparison with conventional lamps. The results of the simulation are presented, and the problems and necessary improvements of the CRI are discussed. 2 Color-Rendering Index The CRI is currently the only internationally agreed-on metric for color rendering evaluation. The procedure for its ⌬ calculation is, first, to calculate the color differences Ei ͑in the 1964 W*U*V* uniform color space—now obsolete͒ of 14 selected Munsell samples when illuminated by a ref- erence illuminant and when illuminated by the given illu- minant. The first eight samples are medium-saturated col- ors, and the last six are highly saturated colors ͑red, yellow, ͑͒ Fig. 1 LED model SLED at 464 nm compared with the SPD of a green, and blue͒, complexion, and leaf green. The reference typical real blue LED. illuminant is the Planckian radiation for test sources having a correlated color temperature ͑CCT͒ Ͻ5000 K, or a phase of daylight† for test sources having CCT ജ5000 K. The process incorporates the von Kries chromatic adaptation ͵ ͑͒ ͑͒ Km V S d transformation. The Special Color Rendering Indices Ri for ͑ ͒ each color sample are obtained by K = , where Km = 683 lm/W. 4 ͵ S͑͒ d ⌬ ͑ … ͒ ͑ ͒ Ri = 100 − 4.6 Ei i =1, ,14 . 1 Here Km is the maximum LER, and its value, 683 lm/W This gives the evaluation of color rendering for each par- ͑for monochromatic radiation at 555 nm͒, is defined in the ͑ ticular color. The maximum value of Ri zero color differ- international definition of the candela. While various other ence͒ is 100, and the values can be negative if color differ- terms are used in the LED industry, the terms introduced ences are very large. The General Color Rendering Index here are the ones officially recommended internationally.7 Ra is given as the average of the first eight color samples: 4 White LED Simulation Program 8 R i ͑ ͒ Mathematical models have been developed for multichip Ra = ͚ . 2 LEDs and phosphor-type LEDs in order to analyze numer- i 8 =1 ous spectral designs of white LEDs. To simulate multichip ͑ LEDs, the following mathematical model for LED spectra The score for perfect color rendering zero color differ- ͑ ͒ ͒ has been developed. The spectral power distribution SPD ences is 100. Note that “CRI” is often used to refer to Ra, ͑͒ ͑ of a model LED, SLED , for a peak wavelength 0 and but the CRI actually consists of 15 numbers: Ra and Ri i ⌬ =1 to 14͒. half spectral width 0.5, is given by g͑, ,⌬ ͒ +2g5͑, ,⌬ ͒ S ͑, ,⌬ ͒ = 0 0.5 0 0.5 , ͑5͒ 3 Luminous Efficacy of Radiation LED 0 0.5 3 ͑ ⌬ ͒ ͕ ͓͑ ͒ ⌬ ͔2͖ The energy efficiency of a light source is evaluated as its where g , 0 , 0.5 =exp − − 0 / 0.5 . The unit of luminous efficacy v, which is the ratio of the luminous flux wavelength is the nanometer. Figure 1 shows an example of ͑lumens͒ emitted by the source to the input electrical power this LED model compared with the SPD of a typical real ͑watts͒. It is determined by two factors: blue LED spectrum ͓measured at NIST with a relative ex- panded uncertainty ͑k=2͒ less than 5%, depending on the ͔ = K, ͑3͒ wavelength . v e Using the LED model described, spectra of a three-chip ͑ where e is the radiant efficiency of the source ratio of ͑RGB͒ white LED and four-chip white LEDs with various output radiant flux to input electrical power; “external combinations of peak wavelengths and spectral widths can quantum efficiency” is often used with the same meaning͒, be created. For these white LED spectra, the simulation ͑ and K is the luminous efficacy of radiation ratio of lumi- program calculates the general CRI, Ra, and special CRIs, ͒ ⌬ * nous flux to radiant flux, abbreviated as LER in this paper , R1 to R14, as well as color differences Eab in the CIELAB and is determined by the spectral distribution S͑͒ of the color space8 and the LER K. In addition, a broadband source phosphor-type white LED model has been developed, based on Planckian radiation in a limited spectral range †One of daylight spectra at varied correlated color temperatures.