Lighting designers and engineers use different terms for lighting equipment than are used in the vernacular. In this report, we will be using the engineering terms. A luminaire is the combination of light fixture hardware, a ballast or driver if applicable, and a light source, commonly called a lamp (i.e., a light bulb). Thus, the term lamp can refer to an incandescent bulb, a compact fluorescent light (CFL) bulb, or a light-emitting diode (LED) replacement “bulb.” This report will use the term lamp. A luminaire consists of, minimally, an integrated light source or lamp holder, commonly called a socket, and the way to connect to the electrical supply. Most fixtures also contain optical elements that distribute the light as desired, such as a reflector, lens, shade, or globe. When needed, fixtures and luminaires contain a ballast or a driver. A ballast is an electronic device that converts incoming electricity to the proper voltage and current required to start and maintain the operation of a lamp. The term driver refers to the corresponding device used in an SSL luminaire. Luminaire examples include chandeliers, downlights, table lamps, wall sconces, recessed or pendant mounted luminaires, and exterior streetlights. When equipped with lamps, they are called luminaires. The types of lamps typically encountered are discussed below in the section, “Lamps.”
METRICS FOR MEASURING LIGHT OUTPUT
The portion of the electromagnetic spectrum that can be perceived by the human visual system is called the visible spectrum. The amount of light, weighted by the sensitivity of the visual system, emitted by a source per unit time is its luminous flux (Figure C.1) and is measured in lumens (lm). This makes lumens one of the appropriate pieces of information for lamp packaging to help consumers choose the appropriate replacement lamps. Lumens provide a description most closely related to brightness and should be referred to when choosing replacement lamps. A proliferation of fact sheets and labels has accompanied the recent introduction of new lighting technologies, leaving some consumers confused about the relationship between watts (W) and lumens. That relationship is determined by the energy efficiency of the product. Watts describe the amount of electrical power consumed by the product, and lumens describe the rate at which it emits light. For example, most 60 W incandescent lamps emit approximately 850 lumens. Similarly, many 13 W CFLs emit 850 lumens.
Luminous intensity (Figure C.2) is the luminous flux per unit solid angle, evaluated in terms of a standardized visual response and expressed in candela. The magnitude of luminous intensity results from luminous flux being redirected
by a reflector or magnified by a lens.1 This measurement is used primarily to describe the specific light intensity and distribution of a luminaire. Illuminance is the concentration of luminous flux incident on a surface (Figure C.3). The unit of illuminance is lux (lx), and it indicates the number of lumens per square meter. Lumens per square foot are called footcandles (fc). Whereas luminous flux relates to the total output of a lamp or lighting product, illuminance relates to the amount of light striking a surface or point. Illuminance depends on the luminous flux of the light sources and their distances from the illuminated surface.
Luminance is a measure used for self-luminous or reflective surfaces (Figure C.4). It expresses the amount of light, weighted by the sensitivity of the visual system, per unit area of the surface that is traveling in a given direction and is expressed as candelas per square meter (cd/m2). When referring to illuminated surfaces, luminance is determined by the incident light (illuminance) and the reflectance characteristics of the surface. For instance, light and dark colored walls will have different luminance values when they have the same illuminance. Luminance is a metric used for internally illuminated variable-sized flat light sources forms, such as sheets or tapes (Figure C.4), since the total luminous flux will depend on the surface area of the product.
The luminous efficacy of a lighting product is the ratio of the luminous flux to the total electrical power consumed and has units of lumens per watt (lm/W). A perfect light source—that is, one that converts all the electricity into visible light—would have an efficacy of 408 lm/W for an assumed color rendering index (CRI; a measure of color quality, discussed below) of 90 (Phillips et al., 2007).2 The luminous efficacy of a traditional 60 W incandescent lamp (luminous flux of 850 lumens) is such that only 14.2 lumens are emitted per watt of power drawn by the light bulb. As efficacies increase, more of the power is used to generate visible light, and this leads to a more efficient product. High color quality LEDs currently are being manufactured with efficacies in the range of 60 to 188 lm/W.
It is important to note that efficacy is different from efficiency. The efficiency of a lighting system is the ratio between the obtained efficacy and the theoretical maximum efficacy of a light source (408 lm/W for a CRI of 90) and is always expressed as a percentage. Thus, it accounts for the ballast efficiency (if there is one), the light source efficacy, and the luminaire efficiency (see Figure C.5) in one lumped parameter. Thus, incandescent lamps with system efficacies ranging from 4 to 18 lm/W (depending largely on the wattage of the bulb) will have system efficiencies of only about 0.2 to 2.6 percent. Efficiency does not, however, account for the perceived quality of the light. Using the theoretical maximum of 408 lm/W and the ranges of efficacies for different lighting technologies leads to the ranges of system efficiencies shown in Figures C.5 and C.6.
1 The concept of solid angle has a strict geometric definition but can be thought of as a way to describe the focusing and redirecting of a light source by the lenses and reflectors in the luminaire.
2 A different choice of CRI = 80 would lead to a maximum efficacy of 423 lm/W, and so forth.
VISIBLE SPECTRUM AND QUALITY OF LIGHT
The human eye can generally detect light with wavelengths between 380 nm (corresponding to blue/violet light) and 780 nm (corresponding to red light). The spectral power distribution (SPD) determines several important properties of a light source. The SPD describes the relative amount of light per wavelength emitted by a light source and is often graphically represented, as shown in Figure C.7. This figure shows the SPDs of a halogen lamp, an RGB LED (which produces white light by combining red, green, and blue component LEDs), an OLED, and a combination of four colored lasers.
The color of emitted light as perceived by people, called chromaticity, is regulated by the spectral composition. The human visual system does not process light on a wavelength-by-wavelength basis. Instead, the brain receives signals from only three input channels, the different cone photopigments found in the eye. Because of this, countless different SPDs can produce light identical in chromaticity. To illustrate this,
the four widely varying SPDs shown in Figure C.7 would all produce light that would appear indistinguishable.
On the correlated color temperature (CCT) scale, all four spectra in Figure C.7 are approximately 3,000 K. CCT is used to describe nominally white light sources and refers to the temperature of a blackbody radiator that produces a light perceived to be most similar in chromaticity to the white light source. A typical incandescent lamp has a CCT of 2,500 to 3,000 K, whereas office and school lighting is often 4,000 to 5,000 K. Lower CCTs include more light nearer the red end of the visible spectrum and are perceived to be “warmer,” while higher CCTs tend toward the blue end and are perceived to be “cool.” In somewhat of a misnomer, the labeling is indicative of the feelings they evoke rather than their actual temperatures. Though the color of daylight changes throughout the day and with location on Earth, it is commonly described as having a CCT of 6,500 K. Though CCT is widely used among lighting manufacturers and designers, it only describes one dimension of light source chromaticity, in the blue-yellow direction. It does not consider pink-green shifts in white light color, though Duv is a measure increasingly used for that information.
The most common system for specifying and communicating the precise chromaticity of light sources uses CIE 1931 (x, y) chromaticity coordinates (CIE, 2004). The CIE 1931 (x, y) chromaticity diagram is shown in Figure C.8. The curved edge of the outer horseshoe shape on the diagram is the spectrum locus and consists of the colors of monochromatic (only one wavelength) radiation. The straight edge line is the purple line, and the colors are always a combination of red and blue (not monochromatic).
Chromaticity does not provide all of the color information of interest for general illumination applications. The color of the light itself does not predict the appearance of colored objects illuminated by the source, a property referred to as color rendering. Though color rendering is determined by the spectral output of a light source, it cannot be predicted by a cursory inspection of the shape of the spectral power distribution and subtle differences in SPD can produce marked differences in the chromaticity of illuminated objects (Ohno, 2005). The SPD also determines the LER (i.e., the luminous efficacy of radiation) of a light source. In technical terms, LER is the ratio of luminous flux to radiant flux.3 In simple terms, the LER is luminous efficacy that could be achieved if the light source was able to convert electricity to light perfectly, with no losses. The final luminous efficacy of a light source is determined from both the LER and the efficiency with which the technology converts electricity to light. The sensitivity of the human visual system differs for the various wavelengths in the visible range. The relationship between wavelength and the relative sensitivity of the
3 Radiant flux is the amount of electromagnetic energy emitted per unit time at all wavelengths including visible light and other spectral bands. As such it will exceed the luminous flux.
human visual system is described by the Spectral Luminous Efficiency Function (Vl) (CIE, 1926), which is shown by the dashed curves in Figure C.9. This function peaks at 555 nm. Light of this wavelength has a LER of 683 lm/W, setting the upper bound for luminous efficacy, as illustrated by the 555 nm laser in panel a. It is important to note that white light cannot achieve 683 lm/W, only light at 555 nm can. Visual sensitivity is markedly lower for light in the short and long wavelength regions of the visible spectrum. The other three panels of Figure C.9 show different SPDs and their corresponding LER. Panel b shows an RGB white LED, panel c shows a different type of white LED (called a phosphor LED), and panel d shows the SPD of a typical incandescent lamp. As shown, the effect of spectral power distribution on luminous efficacy can be substantial. The incandescent SPD has a relatively low LER because it has a lot of energy in the very long visible and infrared wavelengths, to which the visual system is either minimally or completely insensitive.
Though the wavelengths of light to which the eye is most sensitive lie in the middle of the spectrum, a light source composed of light only in the middle of the visible spectrum would not be useful for general illumination. To achieve desirable color characteristics, light of other wavelengths must be present. There is generally a trade-off between luminous efficacy and color quality (Ohno, 2005). Depend-
ing on the application and goals of a lighting product or lit environment, a luminaire manufacturer or lighting designer may choose to prioritize one trait over the other. For example, in a parking garage with lights on 24 hours a day, a specifier may require excellent efficacy and accept subpar color quality. On the other hand, a museum may require superior color and be willing to sacrifice efficacy.
Good color rendering can be achieved with such discontinuous light spectra because of the properties of the other two elements in the process of perceiving object colors: the reflectance of the objects and the absorption of the cone photopigments in the human visual system. All objects, natural or man-made, reflect as a function of wavelength in a very broad and continuous manner. The reflectance factors of these objects (the proportion of light reflected as a function of wavelength) do not show sudden spikes or isolated dips in reflectivity across the visible spectrum. Because of this, the general shape of the reflectance factor can be interpolated with fairly coarse wavelength sampling. The three cone photopigments responsible for color vision have absorption functions that are very broad, continuous, and overlapping in wavelength sensitivity. Each cone type responds to many wavelengths, though sensitivity is different depending on the wavelength. The outputs of these photoreceptors do not signal the wavelength composition of the stimulus to the brain. For instance, a certain level of activity from one cone type could result from a small amount of energy at every wavelength it is sensitive to or a lot of energy at only one wavelength it is sensitive to. The visual system makes absolutely no distinction between these two situations (Ruston, 1972). The perception of color arises from combining and comparing the activity among the three cone types. Therefore, countless combinations of input wavelengths can lead to the exact same perception of color. These circumstances, in which objects reflect in a fairly predictable manner and the visual system interprets incoming light in terms of three broadly sensitive channels, allow a great deal of flexibility for the spectral content of light sources. A recent study demonstrated an extreme case of this in which light sources were developed composed of only four lasers (i.e., sources with extremely narrow emission spectra) with high color rendering quality (Neumann et al., 2011).
A light source need not emit energy at every visible wavelength in order to achieve high color quality (Figure C.9). An understanding of the spectral power distribution’s effects on luminous efficacy and the color properties of a light source will enable SSL developers to optimize energy efficiency while maintaining good color quality.
There are many different kinds of lamps. Most of the lamps used in residential applications are omnidirectional (emit light in all directions) incandescent lamps, typically with a medium screw base (Figure C.10) that fits into most residential luminaires. In addition, there are candelabra and intermediate base lamps that are commonly used in residential applications, especially in chandeliers and wall sconces. Incandescent lamps produce light by heating a tungsten filament to a temperature of approximately 2,500 to 3,000 K where the filament glows or incandesces.
Halogen lamps are incandescent lamps in which the tungsten filament has been enclosed in a capsule containing a halogen gas, typically bromine, which allows the filament to operate at a slightly higher temperature without reducing the rated life and resulting in a somewhat higher light output than the standard incandescent lamp. Halogen lamps are available that emit light omnidirectionally, as well as directional varieties, often known as reflector lamps. Reflector lamps are designated by the properties of their reflectors, such as PAR (parabolic aluminized reflector (Figure C.11), or MR (multifaceted mirror reflector), and are most commonly either standard incandescent or halogen. The low voltage MR-16 lamp (Figure C.12) commonly used in accent, task, and display lighting uses halogen technology.
Fluorescent lamps are available in a range of shapes and sizes. Linear fluorescent lamps are frequently used in commercial spaces (offices, stores) and are typically 4-foot-long
tubes. They are often installed in recessed luminaires in the ceiling or are pendant mounted from the ceiling. All fluorescent lamps require a ballast. Compact fluorescent lamps (CFLs) are available with screw bases and an integral ballast (Figure C.13) for use as replacements for incandescent lamps or with pin bases for use with a separate ballast (Figure C.14). Both CFLs and linear fluorescent lamps produce light by exciting phosphors, which then fluoresce, with ultraviolet energy. A small amount of mercury is added to the lamp to emit UV radiation at a suitable wavelength for exciting the phosphor.
High-intensity discharge (HID) lamps are electric lamps with tubes filled with gas and metal salts. The gas initiates an arc, which evaporates the metal salts, forming a plasma. This results in an efficient and high-intensity light source. These lamps are suitable for both indoor and outdoor applications and are generally used to light large spaces or roadways. All
HID lamps require a ballast. Mercury vapor, metal halide (Figure C.15), and high-pressure sodium lamps are examples of specific types of HID lamps. HID lamps require a warmup period to reach stable output as well as a cool-down period before restarting.
Azevedo, I.L., M.G. Morgan, and F. Morgan. 2009. The transition to solidstate lighting. Proceedings of the IEEE 97:481-510.
CIE (Commission Internationale de l’Eclairage). 1926. Commission Internationale De L’Eclairage Proceedings, 1924. Cambridge, U.K.: Cambridge University Press.
CIE. 2004. Colorimetry. Third edition. Technical Report. Vienna, Austria: CIE Central Bureau.
Neumann, A., J.J. Wierer, W. Davis, Y. Ohno, S.R.J. Brueck, and J.Y. Tsao. 2011. Four-color laser white illuminant demonstrating high color-rendering quality. Optics Express 19(14):A982-A990.
Ohno, Y. 2005. Spectral design considerations for white led color rendering. Optical Engineering 44(11).
Phillips, J.M., M.E. Coltrin, M.H. Crawford, A.J. Fischer, M.R. Krames, R. Mueller-Mach, G.O. Mueller, et al. 2007. Research challenges to ultra-efficient inorganic solid-state lighting. Laser and Photonics Reviews 1(4):307-333.